Characterization of gp120 Hydrolysis by IgA Antibodies from Humans without HIV Infection

AIDS RESEARCH AND HUMAN RETROVIRUSESVolume 23, Number 12, 2007, pp. 1541–1553© Mary Ann Liebert, Inc.DOI: 10.1089/aid.2007.0081

Characterization of gp120 Hydrolysis by IgA Antibodies from Humans without HIV Infection

STEPHANIE PLANQUE,1 YUKIE MITSUDA,1 HIROAKI TAGUCHI,1 MARIA SALAS,2

MARY-KATE MORRIS,2 YASUHIRO NISHIYAMA,1 ROBERT KYLE,3 PABLO OKHUYSEN,1

MIGUEL ESCOBAR,1 ROBERT HUNTER,1 HAYNES W. SHEPPARD,2 CARL HANSON,2

and SUDHIR PAUL1

ABSTRACT

Antibody hydrolysis of the superantigenic gp120 site and HIV-1 neutralization was studied as a potential anti-HIV mechanism in uninfected humans. gp120 hydrolysis by purified serum and salivary antibodies was de-termined by electrophoresis and peptide sequencing, the proteolytic mechanism was analyzed using elec-trophilic peptide analogs, and viral neutralization was studied using peripheral blood mononuclear cells ashosts. Polyclonal and monoclonal IgA but not IgG preparations selectively catalyzed the cleavage of HIV gp120at rates sufficient to predict biologically relevant protection against the virus. The IgA hydrolytic reactionproceeded by noncovalent recognition of gp120 residues 421–433, a component of the superantigenic site ofgp120, coordinated with peptide bond cleavage via a serine protease-like mechanism. The Lys-432–Ala-433bond was one of the cleavage sites. Infection of peripheral blood mononuclear cells by a primary isolate ofHIV was neutralized by the IgA but not IgG fractions. The neutralizing activity was specifically inhibited byan electrophilic inhibitor of the catalytic activity. The existence of catalytic IgAs to gp120 in uninfected hu-mans suggests their role in resistance to HIV.

1541

INTRODUCTION

THE DEVELOPMENT OF ANTIBODY (Ab) responses to microbialantigens is usually characterized by adaptive diversifica-

tion of the variable (V) domains, accompanied by constant do-main class switching. These processes result in a transition fromthe initial synthesis of IgM Abs to adaptively matured IgG andIgA class Abs over the course of B cell differentiation. In ad-dition to numerous conventional antigenic epitopes known tostimulate adaptive Ab responses, the HIV coat protein gp120expresses a superantigenic site (gp120SAg), defined as a site rec-ognized by Abs present in the preimmune repertoire withoutthe requirement for adaptive sequence diversification of the Vdomains.1 Some humans remain free of infection despite re-peated exposure to HIV,2,3 and the clinical course of HIV-1 in-fection can be slow, with infected individuals progressing tothe symptoms of acquired immune deficiency syndrome at vary-ing rates.4 A role for IgG class Abs to gp120SAg in protectionagainst HIV infection has been proposed from observations that

uninfected individuals with increased gp120SAg binding IgGshave a reduced risk of contracting the infection.2 From syn-thetic peptide studies, it appears that the gp120SAg is a confor-mational determinant composed of peptide regions 231–260,331–360, and 421–440.5,6 The gp120 421–433 region is a par-ticularly noteworthy epitope because of its conserved sequencein diverse HIV strains and essential role in virion binding tohost CD4 receptors, the initial step in viral entry into host cells.Mutations in the 421–433 region7 and cleavage of the 432–433peptide bond8 result in loss of CD4 binding activity, and con-tacts between the 421–433 region and CD4 are evident by X-ray crystallography of gp120-soluble CD4 complexes.9

Certain Abs express the ability to catalyze peptide bond hy-drolysis. A single catalyst molecule can be reused in repeatedreaction cycles to cleave multiple antigen molecules (in com-parison, a noncatalytic Ab can at most inactivate the antigenstoichiometrically, e.g., 2 molecules gp120/molecule bivalentIgG). Early research on catalytic Abs was dominated by ex-pectations that immunization with analogs of intermediates

1Chemical Immunology Research Center, Departments of Pathology and Laboratory Medicine and Hemophilia and Thrombophilia Center,University of Texas-Houston Medical School, Houston, Texas 77030.

2Viral and Rickettsial Disease Laboratory, California Department of Public Health, Richmond, California 94804.3Division of Hematology, Mayo Clinic, Rochester, Minnesota 55905.

formed in the reaction pathway is necessary to induce adaptiveselection of catalytic sites over the course of V domain diver-sification. Later studies, however, showed that catalysis is a her-itable function expressed by naturally occurring Abs withoutthe requirement of experimental immune interventions.10 Wehave reported the ability of monoclonal and polyclonal IgMclass Abs to hydrolyze gp120.11 The hydrolytic reaction is fa-cilitated by noncovalent recognition of the 421–433 gp120 su-perantigenic region. Other studies have documented proteoly-sis by Abs obtained without exogenous immune interventionand following exposure to conventional antigens. Examples in-clude proteolytic Abs to vasoactive intestinal peptide (VIP),12

Arg-vasopression,13 thyroglobulin,14 Factor VIII,15 prothrom-bin,16 gp41,17 H. pylori urease,18 casein,19 and myelin basicprotein.20 Site-directed mutagenesis21 and X-ray crystallogra-phy22 have indicated the presence of activated nucleophilicamino acids within the combining sites of proteolytic Abs thatare similar to catalytic residues of conventional serine proteases.Furthermore, electrophilic phosphonates that were originallydeveloped as specific probes for enzymatic catalytic sites alsoreact covalently with the Ab nucleophilic residues.23,24 Inter-estingly, the nucleophilic and catalytic activities of adaptivelymatured IgG class Abs are lower than that of IgMs, and IgGswith proteolytic activity are encountered only rarely.25 Ac-cording to the clonal selection theory, antigen binding to the Bcell receptor (surface immunoglobulin complexed to Ig�, Ig�,and signal-transducing proteins) drives cellular proliferationand adaptive antibody maturation. It has been suggested thatefficient Ab proteolysis may be incompatible with B cell clonalselection, as rapid antigen hydrolysis by the receptor will resultin release of antigen fragments and loss of the B cell prolifer-ative signal.26

IgA class Abs are important mediators of immunologicaldefense at mucosal surfaces. Sexual transmission of HIV oc-curs across the mucosa. IgAs with proteolytic activity di-rected to an autoantigenic polypeptide26 and model proteasesubstrates27 have been described recently. Whether IgAs rec-ognize and cleave microbial antigens has not been explored.Here, we report observations that IgAs but not IgGs from hu-mans without HIV infection hydrolyze gp120 efficiently viarecognition of the protein as a superantigen. The IgAs neu-tralized a primary HIV strain in tissue culture, and an elec-trophilic 421–433 gp120 peptide analog blocked the neutral-izing activity.

MATERIALS AND METHODS

Abs

Polyclonal Abs were purified from saliva or serum derivedfrom peripheral venous blood of four human subjects with-out evidence of HIV infection or immunological disease (onefemale and three males, age 28–36 years, our laboratory sub-ject codes 2288–2291). Saliva was obtained following chew-ing of parafilm.28 The Abs were also analyzed as pools of theIgA and IgG fractions purified from 34 human subjects with-out HIV infection (17 females and 17 males, age 17–65 years;Caucasian 30, African–American 2, Asian 2; our laboratory

codes 679, 681–689, and 2058–2081). Monoclonal IgAs werepurified from sera of patients with multiple myeloma pro-vided by Dr. Robert Kyle (codes 2573–2587; 51–64 mgIgA/ml serum, determined as the size of the M-proteinband29). Blood and saliva collection was with informed con-sent as approved by the University of Texas Committee forProtection of Human Subjects. IgA was purified from sera orsaliva as in Mitsuda et al.27 by chromatography on a goatanti-human IgA-agarose gel using 0.1 M glycine, pH 2.7, forelution. IgG was purified using protein G-Sepharose using asstarting materials the unbound fraction from the anti-IgAcolumns.25 Protein concentrations were determined using amicroBCA kit (Pierce). Immunoblotting of sodium dodecylsulfate (SDS) electrophoresis gels was with peroxidase-con-jugated goat anti-human � chain, anti-human � chain, anti-human � chain, and anti-secretory component Abs(Sigma–Aldrich).11 Gel filtration of serum IgA (1.6 mg) andsalivary IgA (0.8 mg) previously purified by anti-IgA chro-matography (pooled from subject codes 2288–2291) was in6 M guanidine hydrochloride, pH 6.5, on a Superose-6 FPLCcolumn (0.2 ml/min) as described in Planque et al.25 Thenominal mass of eluted protein fractions was determined bycomparing retention time (Rt) values with IgM (900 kDa),thyroglobulin monomer (330 kDa), IgA (170 kDa), andbovine serum albumin (BSA) (67 kDa). IgA renaturation wasby dialysis against 50 mM Tris–HCl, 0.1 M glycine, pH 7.7,containing 0.1 mM 3-[(3-cholamidopropyl) dimethylammo-nio]-1-propanesulfonic acid at 4°C (Tris-Gly buffer; 2liters � 5, 4 days).

Proteolysis assays

Biotin was incorporated at Lys residues in gp120 (MN strain,Protein Science Inc.), soluble epidermal growth factor receptor(sEGFR), bovine serum albumin (BSA), HIV Tat (NIH AIDSResearch and Reference Reagent Program), and factor VIII C2fragment (kind gift from Dr. K. Pratt) at a stoichiometry of 1–2mol of biotin/mol protein.30 Protein hydrolysis was determinedby reducing SDS electrophoresis in duplicate.31 Following in-cubation with Abs in 20 �l Tris-Gly buffer containing 67 �g/mlgelatin, the reaction mixtures were boiled in SDS (2%) and 2-mercaptoethanol (3.3%), subjected to electrophoresis, blotted,and stained with streptavidin peroxidase. The detection of bi-otin allows measurement of cleavage rates from the extent ofdepletion of the intact gp120 band, but this method does notprovide information about the relative product concentrations,as the biotinylated-gp120 contains minimal amounts of biotin(�1 mol/mol gp120) and the products do not necessarily con-tain the biotin.11 The Ab concentrations employed in variousassays were adjusted to yield activity in the linear range of mea-surable hydrolysis. gp120 cleavage was determined by densi-tometry of the intact biotinylated gp120 band as [gp120]0 �([gp120]0 � (gp120Ab/gp120DIL)), where [gp120]0, [gp120Ab],and [gp120DIL] represent, respectively, the initial concentration,concentration in the Ab-containing reactions, and concentrationin reactions containing diluent instead of Ab [computed by mea-suring bands arbitrary volume units (AVU); pixel intensity �band area]. The minimum Ab concentration yielding detectablehydrolysis was the concentration at which a product band was

PLANQUE ET AL.1542

detected with intensity 3-fold greater than the background in-tensity within the same gel (determined in triplicate). For cleav-age site determination, gp120 was incubated with IgA (pooledfrom subject codes 2288–2291), the IgA was removed by bind-ing to an anti-human IgA column as described above, and theunbound fraction was lyophilized and redissolved in SDS elec-trophoresis buffer containing 2-mercaptoethanol. The gp120fragments identified by SDS electrophoresis were blotted onPVDF membranes, stained with Coomassie blue, and subjectedto N-terminal sequencing as described previously.11 The phos-phonate compounds employed were diphenyl N-(6-biotinami-dohexanoyl)amino(4-amidinophenyl)methanephosphonate(EP-hapten 1); N-(6-biotinamidohexanoyl)amino(4-amidinophenyl)methanephosphonic acid (nonelectrophilic hapten 2); diphenylN-(benzyloxycarbonyl)amino(4-amidinophenyl)methanephos-phonate (EP-hapten 3, corresponding to EP-hapten 1 withoutbiotin); gp120 residues 421–431 with the amidinophosphonatemimetic of residues 432–433 at the C-terminus (EP-421-433;sequence of residues 421–433, Lys–Gln–Ile–Ile–Asn–Met–Trp–Gln–Glu–Val–Gly–Lys–Ala; numbered according to strainMN sequence); and VIP containing the amidinophosphonate atthe Lys-20 side chain (EP-VIP). The synthesis and purity ofthese compounds have been described.32–34 In addition to theEP-hapten compounds, the following compounds were testedas protease inhibitors: EDTA, iodoacetamide, pepstatin A, and1,10-phenanthroline (all from Sigma-Aldrich). Hydrolysis ofB o c - G l u ( O B z l ) - A l a - A r g - a m i n o m e t h y l c o u m a r i n(Glu–Ala–Arg–AMC, Peptides International) was measured byfluorimetry (�ex 360 nm, �em 470 nm) with authentic 7-amino-4-methylcoumarin (AMC) employed to construct a standardcurve.25

Phosphonate binding

Purified IgA (pooled from subjects 2288–2291) was treatedwith EP-hapten 1, control hapten 2, EP-421-433, or EP-VIP.The formation of irreversible adducts was measured by reduc-ing SDS electrophoresis, electroblotting, staining with a strep-tavidin-peroxidase conjugate, and densitometry.30

HIV neutralization

Neutralization experiments were conducted as in Karle etal.35 using the primary HIV isolate (97ZA009; clade C, R5-de-pendent), phytohemagglutinin-stimulated human peripheralblood mononuclear cells (PBMCs) as hosts, and p24 determi-nation as the measure of infection. IgA or IgG preparationspooled from subjects 2288–2291 in PBS (10 mM sodium phos-phate, 137 mM NaCl, 2.7 mM KCl, pH 7.4) were incubated for1 h or as indicated with an equal volume of HIV [100 TCID50;final volume 0.2 ml RPMI 1640 containing 25% PBS, 0.25%fetal bovine serum (FBS), and 3% NHT Growth Factor (Zep-tometrix)], whereupon PBMCs were added and cultured furtherfor measurement of infection (FBS concentration in the cul-tures, 20%). Some assays were done following IgA treatmentwith EP-421-433 or EP-VIP (100 �M) followed by determina-tion of the HIV neutralizing activity. The broadly neutralizingmonoclonal IgG clone YZ23 was used as a positive control[mean effective concentration range resulting in 50% neutral-ization (EC50), 6.7 �g/ml, N � 20 assays].31

RESULTS

IgA catalytic activity

Each IgA preparation purified from the saliva and serum offour humans without HIV infection cleaved biotinylated-gp120(Bt-gp120), assessed by depletion of the parent gp120 band andappearance of lower mass fragments in electrophoresis gels(Fig. 1A). The Bt-gp120 product profiles observed using sali-vary and serum IgA as catalysts were essentially identical (prod-ucts with nominal mass 80, 55, 39, 32, 25, and 17 kDa). The80-kDa and 55-kDa bands generated in the initial stages of thereaction appeared to be digested further, as their intensities weredecreased at the later time points. The number of combiningsites per Ab molecule can vary (two sites/monomer IgA andIgG; four sites/salivary IgA dimer; approximate mass of the IgGmonomers, IgA monomers, and IgA dimers, respectively: 150,160, and 400 kDa). However, the number of combining sitesper unit mass of the monomers and aggregates is comparable(one combining site per �75–100 kDa), and normalization ofthe activity data to constant concentration (mg/ml) permits validcomparison of their activities. The mean proteolytic activity ofsalivary IgA from the four subjects was 15.4-fold greater thanserum IgA [means � SD, respectively, 128.6 � 23.3 and 7.2 �3.3 nM gp120/h/(mg IgA/ml); Fig. 1B]. From dose–responsecurves conducted with IgAs from subject #2289, the minimumserum and salivary IgA concentrations yielding detectable cat-alytic activity were 25 and 3 �g/ml, respectively. These con-centrations are about two orders of magnitude lower than thephysiological concentrations of IgA in serum (1.5–2.6 mg/ml)36

and saliva (110–300 �g/ml).37,38 Serum IgG fractions were devoid of detectable activity (Fig. 1C). Intravenous adminis-tration of IgG preparations pooled from uninfected humans (in-travenously administered immunoglobulin G, IVIG prepara-tions) has been considered for treatment therapy of HIVinfection.39 Like the human IgG prepared in our laboratory,commercial IVIG preparations did not cleave gp120 detectably(Gammagard S/D and Inveegam EN, Baxter; Intratect, BiotechPharma GmbH). Essentially identical results were obtained us-ing serum IgA and IgG purified from the pooled sera of 34HIV-seronegative humans (941 nM gp120 cleaved/h/mg IgA;undetectable gp120 cleavage at equivalent IgG concentration;reaction conditions as in Fig. 1A).

The electrophoretic homogeneity of the IgG purified as de-scribed here has been reported.40 Reducing SDS electrophore-sis of serum IgA obtained by affinity chromatography revealedthe heavy chain (60 kD) and light chain (25 kDa) subunits (Fig.2A). Salivary IgA contained these bands along with the addi-tional band stainable with anti-secretory component Ab (85kDa). All of the Coomassie blue-stainable protein bands in theIgA preparations were stainable by Abs to the � chain, �/�chain, or secretory component, indicating the absence of de-tectable non-IgA proteins. The observed IgA subunit bandswere not stained by anti-� or anti-� Abs, indicating the absenceof detectable IgGs or IgMs.

To test for contaminating conventional proteases, serum andsalivary IgA preparations purified by affinity chromatographyusing the anti-IgA column were subjected to further FPLC-gelfiltration in a denaturing solvent (6 M guanidine hydrochlo-ride). We chose to conduct the molecular mass-based fraction-

IgA GP 120ASES 1543

ation of our polyclonal IgAs because separations based on prop-erties such as molecular charge yield heterogeneous profileswith unpredictable fractionation of the individual Abs consti-tuting the mixtures. IgAs are known to form noncovalent andS–S bonded multimers.41 As reported previously,27 peaks cor-responding to polymeric, dimeric, and monomeric IgA were ev-ident (nominal mass, respectively, 915 kDa, 433 kDa, and 153

kDa). Each of these IgA fractions displayed reducing SDS gelelectrophoresis profiles identical to those of the starting affin-ity-purified IgA loaded on the column. Monomer IgA was thepredominant serum IgA species recovered from the column(82%; Rt 55.2 min). Ninety percent of the salivary IgA elutedas dimers (Rt 42.7 min) and higher order aggregates (Rt 33.7min). Following refolding by removal of guanidine hydrochlo-

PLANQUE ET AL.1544

FIG. 1. Cleavage of biotinylated gp120 by serum and salivary IgA from humans without HIV infection. (A) Streptavidin-per-oxidase-stained blots of reducing SDS gels showing time-dependent cleavage of Bt-gp120 (0.1 �M) incubated with pooled poly-clonal serum IgA (160 �g/ml) and salivary IgA (32 �g/ml) from four humans. Reaction volume, 0.02 ml. Diluent lane, gp120incubated with diluent instead of IgA. OE, overexposed lane showing Bt-gp120 incubated for 46 h with salivary IgA. (B) Com-parative gp120 cleaving activity of salivary IgA (32 �g/ml) and serum IgA (144 �g/ml) from four humans expressed per equiv-alent Ab mass. Reaction conditions: 17 h, 0.1 �M Bt-gp120. (C) Comparative gp120 cleaving activity of serum IgA and IgG(144 �g/ml) expressed per equivalent Ab mass. IVIG, commercial IgG preparations (the three data points correspond to the fol-lowing IVIG preparations: Intratect, Gammagard, Inveegam). Each IgA and IgG point represents Abs from a different human.

ride, the monomer serum IgA species recovered from the col-umn displayed gp120 cleaving activity (Fig. 2B) identical inmagnitude to the affinity-purified IgA preparation loaded on the column (respectively, 630 � 167 and 823 � 130 nMgp120/h/mg IgA), fulfilling the criterion of purification to con-stant specific activity. The refolded dimeric and higher ordersalivary IgA aggregates eluting from the column also displayedgp120 cleaving activity (Fig. 2B), confirming that the predom-inant form of secretory IgA is catalytically active. Salivary non-IgA proteases with large mass values corresponding to the ob-served catalytic species (433–915 kDa) are not described to ourknowledge. The denaturing gel filtration procedure dissociatesand removes lower mass contaminants that could bind nonco-valently to the affinity-purified IgA loaded on the column. Thepresence of non-IgA protease contaminants, therefore, does notexplain the observed proteolytic activity. The refolded salivaryIgA aggregates displayed gp120 cleaving activity 4.5-fold lowerthan the undenatured salivary IgA. A similar denaturant-in-duced loss of activity due to incomplete refolding into the na-tive protein conformation has been described for other prote-olytic antibody preparations.40

Interaction with electrophilic phosphonate hapten

EP-hapten 1 (Fig. 3A) was originally developed as an in-hibitor that binds irreversibly to nucleophiles found in the en-zymatic active site of serine proteases such as trypsin.23 The ir-reversible reactivity of this compound with proteolytic Abfragments and full-length Abs has also been reported.24,25 EP-

hapten 1 at a concentration of 1 mM inhibited the catalytic ac-tivity of salivary and serum IgA almost completely (Fig. 3B).Class-selective inhibitors of metalloproteases (EDTA, 2 mMand 1,10-phenanthroline, 1 mM), cysteine proteases (iodoac-etamide, 100 �M), and acid proteases (pepstatin A, 1 �M), didnot detectably inhibit the hydrolysis of gp120 by the pooledserum and salivary IgA. The concentrations of these compoundsused in these experiments are described to yield readily de-tectable inhibition of the corresponding proteases.42

Consistent with the predicted covalent mechanism of in-hibition, salivary and serum IgA preparations formed adductswith EP-hapten 1 stable to heating (100°C, 5 min) and de-naturation with SDS, corresponding to the dominant �60-kDa heavy chain adduct band and the weaker �25-kDa lightchain adduct band shown in Fig. 3C. Formation of the IgAadducts increased with increasing EP-hapten 1 concentration(at 0.1 and 1 mM EP-hapten 1, the intensity of serum IgAadducts was 2.95 � 0.01 and 86.31 � 1.14 � 103 AVU, re-spectively; the intensity of salivary IgA adducts was 2.4 �0.03 and 94.27 � 5.43 � 103 AVU, respectively). Similarly,EP-hapten 1 inhibition of gp120 cleavage by the IgAs wasdose dependent (at 0.1 and 1 mM EP-hapten 1, the inhibitionof serum IgA was 69 and 99%, respectively; inhibition ofsalivary IgA was 40 and 100%, respectively). The controlhapten 2 is structurally identical to EP-hapten 1 except forthe absent phenyl groups at the phosphorus atom, resultingin impaired electrophilic reactivity with enzymatic nucle-ophiles. Hapten 2 did not inhibit IgA-catalyzed gp120 cleav-age or form adducts with the IgAs.

IgA GP 120ASES 1545

FIG. 2. Electrophoretic homogeneity and gp120 cleavage by refolded polyclonal IgA following denaturing gel filtration.(A) Reducing SDS electrophoresis (4–20% gels) of human serum IgA and salivary IgA purified by affinity chromatographyon immobilized anti-IgA Ab. Serum IgA stained with Coomassie blue (lane 1), anti-� chain Ab (lane 2), and anti-�/� chainAb (lane 3); salivary IgA stained with Coomassie blue (lane 4), anti-� chain Ab (lane 5), anti-�/� chain Ab (lane 6), andanti-secretory component Ab (lane 7). (B) Streptavidin peroxidase-stained electrophoresis blots showing Bt-gp120 cleavageby salivary and serum IgA subjected to denaturing gel filtration in 6 M guanidine hydrochloride. See Materials and Meth-ods for gel filtration procedure. The 153-kDa and 433- to 915-kDa fractions obtained by gel filtration column of serum andsalivary IgA were renatured by dialysis against Tris–Gly buffer, pH 7.7 and incubated with Btgp120 (0.1 �M). Reactionconditions, IgA 32 �g/ml, 45 h.

Monoclonal IgA catalysis

Fifteen identically purified monoclonal IgAs from the serumof patients with multiple myeloma were analyzed for gp120cleaving activity (Fig. 4A). Twelve IgAs displayed detectablegp120-cleaving activity and two were without activity. Theelectrophoretic profile of the gp120 reaction products observedusing monoclonal IgA catalysts was essentially identical to thatobtained using polyclonal IgA. One of the monoclonal IgAswith gp120SAg hydrolyzing activity (code 2582) has previouslybeen reported to hydrolyze the amide bond between AMC andArg in Glu–Ala–Arg–AMC.27 The hydrolysis reaction proceedswithout the involvement of typical noncovalent interactions ac-companying recognition of antigenic epitopes, and similar pep-tide–AMC substrates have been employed as alternate sub-strates for other catalytic Abs.14,21 Figure 4B shows thetime-dependent hydrolysis of Glu–Ala–Arg–AMC by themonoclonal IgA and inhibition of the reaction by EP-hapten 3,consistent with a serine protease-like mechanism of catalysis.Our active site titration studies using this IgA indicated that in-hibition of the Glu–Ala–Arg–AMC hydrolysis occurs with astoichiometry of 2.4 molecules EP-hapten 3 per moleculemonoclonal IgA,27 a value close to the theoretical number ofbinding sites per IgA molecule, 2. In the present study, inclu-sion of excess Glu–Ala–Arg–AMC in the reaction mixture ofgp120 and IgA resulted in complete inhibition of gp120 hy-drolysis (Fig. 4C), indicating that both substrates are cleavedby the same catalytic site. Therefore, the active site titration re-sults cited above are also valid for the gp120SAg cleavage re-

action and provide additional evidence that trace contaminantis not the cause of gp120SAg hydrolysis.

Antigen selectivity and cleavage sites

Treatment of Bt-BSA, Bt-FVIII C2 domain, Bt-Tat, or Bt-sEGFR with human salivary IgA or serum IgA did not resultin noticeable depletion of electrophoresis bands correspondingto the full-length form of these proteins (Fig. 5). Under theseconditions, readily detectable Bt-gp120 cleavage was observed.

Synthetic peptides containing gp120 residues 421–433 arereported to inhibit noncovalent Ab binding to gp120SAg com-petitively.5,6 We have previously observed the irreversible bind-ing of catalytic IgMs by the electrophilic analog of gp120residues 421–433 containing the phosphonate diester and biotingroups (EP-421-433; top structure, Fig. 6A).11 In the presentstudy, inclusion of increasing concentrations of EP-421-433(10–100 �M) in the reaction mixtures resulted in progressiveinhibition of the cleavage of Bt-gp120 by salivary IgA (by21–85%) and serum IgA (by 41–91%). The control probe wasthe phosphonate–containing derivative of an irrelevant peptide(EP-VIP) that can inhibit the hydrolytic reaction by reacting co-valently with nucleophilic residues but is not anticipated to ex-press appreciable noncovalent binding strength for the Abs. EP-421-433 consistently inhibited IgA-catalyzed gp120 cleavagewith potency superior to EP-VIP (p � 0.01, Student’s t test;n � 4 repeat experiments; Fig. 6B). Similarly, EP-421-433 dis-played superior irreversible binding to the IgAs compared tocontrol EP-VIP, estimated by electrophoretic estimation of the

PLANQUE ET AL.1546

FIG. 3. EP-hapten 1 interactions with IgA. (A) EP-hapten 1 structure. The control non-electrophilic phosphonic acid hapten 2is structurally identical to hapten 1 except for the absent phenyl groups. (B) Inhibition of catalysis and irreversible binding byEP-hapten 1. gp120 (0.1 �M) was reacted with salivary IgA (2 �g/ml) or serum IgA (160 �g/ml) in the absence or presence ofEP-hapten 1 and control hapten 2 (1 mM) for 8 h before incubation with non-biotinylated gp120 for 16 h. Data are from SDSelectrophoresis gels stained with peroxidase-conjugated polyclonal anti-gp120. Percentage inhibition � [(100 � gp120 cleavedin the presence of inhibitor)/(gp120cleaved in the absence of inhibitor)] � 100]. Values are means of duplicates. (C) Strepta-vidin-peroxidase-stained blots of reducing SDS gels showing EP-hapten 1 and hapten 2-treated salivary IgA and serum IgA. Hand L denote heavy and light chain subunit bands, respectively.

adducts (serum IgA, by 4.6-fold; salivary IgA, by 16.7-fold)(Fig. 6C). Inclusion of the gp120 peptide 421–435 devoid ofthe phosphonate group in the reaction mixtures inhibited theformation of the IgA:EP-421-433 adducts (Fig. 6D). These ob-servations suggest a nucleophilic mechanism of IgA catalysis

in which noncovalent recognition of the 421–433 peptide re-gion contributes to the observed selectivity for gp120.

To identify the cleavage sites, non-biotinylated gp120 was di-gested by polyclonal salivary IgA and the gp120 fragments ob-tained by SDS electrophoresis were subjected to N-terminal aminoacid sequencing. Readily visible product bands at 55, 39, and 17kDa and a faint band at 32 kDa were evident after digestion for46 h (Fig. 7). The intensity of the 55-kDa band after digestion for9 h was greater than 46 h, indicating further cleavage of this prod-uct. Small amounts of the 55-kDa band were present in the stockgp120 preparation that had not been subjected to Ab treatment.No increase in the 55-kD band was evident following incubationwith diluent in the absence of IgA. The 55-kDa fragment yieldeda sequence corresponding to the N-terminus of gp120. The re-maining fragments yielded N terminal sequences correspondingto gp120 residues 84–88, 322–326, and 433–437, indicating cleav-age at the following peptide bonds: Val-83–Glu-84 (located in thegp120 C1 domain), Tyr-321–Thr-322 (V3 domain), and Lys-432–Ala-433 (C4 domain).

Neutralizing activity

We studied the effect of the Abs on infection of humanPBMCs by the HIV-1 strain 97ZA009 (clade C, chemokinecoreceptor R5 dependent). The sequence of gp120 residues421–433 in this virus strain and the recombinant gp120 em-ployed in catalysis studies are identical except for a conserva-

IgA GP 120ASES 1547

FIG. 4. Hydrolytic activity of monoclonal IgAs. (A) Scatter plot of Bt-gp120 cleavage by monoclonal IgAs. Bt-gp120, 0.1 �M;IgA, 75 �g/ml; reaction time, 21 h. Dashed line corresponds to background value (incubation with diluent instead of IgA) � 3standard deviations. (B) Inhibition of IgA-catalyzed Glu–Ala–Arg–AMC (EAR-AMC) hydrolysis by EP-hapten 3. The substrate(0.4 mM) was incubated with monoclonal IgA (16 �g/ml; code 2582) that had been treated with diluent or the indicated con-centrations of EP-hapten 3 for 20 h. Shown are mean � SD values (n � 3) of AMC release at various inhibitor concentrationsand least-square-fits to the equation [AMC] � V � t, where V and t represent, respectively, the initial velocity of AMC releaseand the time of incubation with the substrate. (C) Inhibition of IgA-catalyzed gp120 cleavage by Glu–Ala–Arg–AMC. Strepta-vidin-peroxidase-stained SDS gel blots (reducing conditions) showing Bt-gp120 (0.1 �M) incubated in diluent (lane 1) and mono-clonal IgA (80 �g/ml, from multiple myeloma subject 2582) in the absence (lane 2) or presence of Glu–Ala–Arg–AMC (lane 3;0.2 mM).

FIG. 5. Preferential cleavage of gp120 by IgA and sIgA. Bi-otinylated (Bt) proteins studied are gp120, soluble epidermalgrowth factor receptor (sEGFR), bovine serum albumin (BSA),C2 domain of human coagulation factor VIII (C2), and HIVTat. Shown are streptavidin-peroxidase stained blots of reduc-ing SDS gels of the proteins (0.1 �M) incubated (17 h) withserum IgA, salivary IgA (both 160 �g/ml), or diluent.

tive Arg/Lys substitution (KQIINMWQEVGR/KA). Pooledserum IgA and salivary IgA from uninfected donors displayedreproducible and dose-dependent neutralizing activity, respec-tively, in each of five and eight assays conducted with these Abpreparations (mean EC50 � SEM, serum IgA 62.0 � 38.3�g/ml; salivary IgA 48.5 � 14.5 �g/ml). Figure 8A reports theneutralizing activity of serum IgA and salivary IgA studied inparallel (identical virus preparation, identical host PBMC prepa-ration). Unlike the IgAs, serum IgG prepared in our laboratoryand three commercial IgG preparations (IVIGs; GammagardS/D, Inveegam EN, and Intratect) did not display detectableneutralizing activity. As expected, the neutralizing activity was maintained by treatment of the IgA with the irrel-evant probe EP-VIP, but inclusion of EP-421-433 in theIgA–virus mixture resulted in inhibition of the neutralizing ac-

tivity (Fig. 8B). Viral neutralization by salivary IgA was re-producibly observed following a comparatively short (1 h) in-cubation with HIV, whereas neutralization by serum IgA wasevident only upon prolonged IgA–virus incubations (24 h; Fig.8C). To confirm that the sensitivity of strain ZA009 to the IgAsis not due to some unusual Ab reactivity, we studied a secondprimary HIV isolate, 92BR020 (clade B, R5 dependent). Thisstrain was also neutralized by the salivary IgA preparation em-ployed in Fig. 8C (EC50, 5.0 � 2.1 �g/ml).

DISCUSSION

IgA molecules, like IgGs, are produced by terminally dif-ferentiated B cells that have been subjected to various processes

PLANQUE ET AL.1548

FIG. 6. Inhibition of gp120SAg hydrolysis and irreversible IgA binding by EP-421-433. (A) Structures of EP-421-433 and con-trol electrophilic peptide (EP-VIP). R1, amidinophosphonate mimetic of gp120 residues 432–433 linked to Gly-431 carboxylgroup; R2, amidinophosphonate group linked to Lys side chain amine. (B) Preferential inhibition of IgA-catalyzed gp120 cleav-age by EP-421-433. Salivary IgA (16 �g/ml) or serum IgA (160 �g/ml) was preincubated (6 h) with EP-421-433 or EP-VIP (100�M) and the reaction mixtures were incubated further for 16 h following addition of gp120 (0.1 �M). Inhibition of gp120 cleav-age determined as in Fig. 3. (C) Irreversible binding of EP-421-433 by serum IgA and salivary IgA. Shown are streptavidin-peroxidase-stained blots of reducing electrophoresis gels of IgA (80 �g/ml) incubated with EP-421-433, EP-VIP, or EP-hapten1 (10 �M; reaction time, 21 h). H and L denote heavy chain and light chain bands. (D) Inhibition of irreversible IgA: EP-421-433 binding by gp120 peptide 421–435. Salivary IgA (80 �g/ml) was treated with gp120 peptide 421–435 (100 �M) or diluentfollowed by addition of EP-421-433 (10 �M) and incubation for 21 h. EP-421-433 adducts were detected as in C. Plotted arethe aggregate intensities of the heavy and light chain subunits determined by densitometry.

responsible for adaptive regulation of the humoral immune re-sponse. Monoclonal IgAs from multiple myeloma patients areknown to contain adaptively diversified V domains.29 Unlikethe IgG preparations, polyclonal IgA from humans without HIVinfection and monoclonal IgAs from multiple myeloma patientscatalyzed the cleavage of gp120 potently and selectively. Theproteolytic activity was readily observed in the polyclonal IgAsheld at concentrations about two orders of magnitude lower thanthe physiological IgA concentrations in serum and saliva. Theproteolytic activity was evident in monomer and aggregate IgAspecies separated by denaturing gel filtration, a procedure that

removes noncovalently associated adventitious proteases.Serum IgA obtained by sequential chromatography using theanti-IgA affinity column and denaturing gel filtration columndisplayed essentially identical gp120SAg hydrolyzing activity,fulfilling the criterion of purification to constant specific activ-ity. Identically purified monoclonal IgAs displayed widelyvarying levels of activity and the activity was inhibited stoi-chiometrically by the electrophilic phosphonate compound.These observations do not support trace contamination withnon-IgA proteases as the explanation of gp120 cleavage. Theinhibition of catalysis by electrophilic phosphonates and their

IgA GP 120ASES 1549

FIG. 7. Identification of peptide bonds cleaved by salivary IgA. Shown are the Coomassie blue-stained SDS gel electrophore-sis lanes of gp120 (270 �g/ml) incubated for 9 h in diluent (lane 1) or with IgA (80 �g/ml) (lane 2). Lane 3 shows the elec-trophoretic profile of the gp120-IgA reaction mixture following more prolonged digestion (46 h). The yields of the indicatedamino acid were 0.3–1.5 pmol. A mixture of PTH-amino acids (2 pmol each; Applied Biosystems) was employed as standard(sensitivity of detection of individual amino acids, 0.04–0.10 pmol).

FIG. 8. HIV neutralization by Abs from HIV-seronegative humans. (A) Neutralizing potency of IgA and IgG Abs purifiedfrom pooled serum or saliva of four human subjects. IVIG, Gammagard S/D. HIV-1 strain, 97ZA009; host cells, phytohemag-glutinin-stimulated PBMCs. Abs were incubated with the virus for 24 h. Values are expressed as percent reduction of p24 con-centrations in test cultures compared to cultures that received diluent instead of the Abs (means � SD of four replicates). (B) In-hibition of IgA neutralizing activity (pooled from 34 donors) by EP-421-433. IgA purified from human serum (2 �g/ml) waspreincubated (0.5 h) with EP-421-433 (100 �M), control EP-VIP, or diluent, and the neutralizing activity was determined as inA. (C) Time-dependent HIV-neutralizing activity. HIV was preincubated with the salivary or serum IgA for 1 h and the neu-tralizing activity was measured as in A.

irreversible binding to the IgAs are consistent with a nucle-ophilic mechanism of catalysis similar to that of serine pro-teases. IgM Abs also catalyze peptide bond cleavage by a nu-cleophilic mechanism,25 and a recent report has identified anucleophilic Ser–Arg–Glu triad in a gp120SAg-hydrolyzing IgMas the putative catalytic site.22

Polypeptides unrelated to gp120 were not cleaved by theIgAs. EP-421-433, a probe for gp120SAg recognizing Abs, in-hibited the cleavage of gp120 and displayed superior irre-versible binding to the IgAs compared to the irrelevant EP-VIPprobe. The selective catalytic activity is attributable, therefore,at least in part to a nucleophilic attack on gp120 coordinatedwith noncovalent recognition of the 421–433 region. Threegp120 peptide bonds were cleaved by polyclonal IgA, includ-ing one located in the 421–433 region (residues 432–433). Theregions containing the other two cleavage sites have not beenlinked previously to the superantigenic properties of gp120. Thereaction product profiles using monoclonal and polyclonal IgAcatalysts were identical, suggesting that a single Ab reactivewith residues 421–433 can cleave bonds located outside this re-gion. Studies on monoclonal Abs that hydrolyze other polypep-tides have also indicated that a single Ab can cleave multiplepeptide bonds.17,43 The complexity of the reaction profile de-rives in part from the evident utilization of the initial cleavageproduct as a substrate for further digestion. As the initial cleav-age product will likely adopt a conformation distinct from thecorresponding region of full-length gp120, it may present Ab-sensitive peptide bonds that are inaccessible in the full-lengthgp120. According to the previously proposed split-site modelof catalysis,44 additional complexity can be introduced by hy-drolysis of peptide bonds located outside the epitope responsi-ble for initial noncovalent antigen-Ab binding. The model pro-poses that spatially neighboring but distinct Ab subsites areresponsible for noncovalent binding and catalysis, permittingthe formation of alternate ground state complexes containingdifferent peptide bonds positioned in register with the catalyticsite. When the Abs recognize a conformational epitope, the al-ternate cleavage sites must be spatial neighbors but they can bedistant in the linear sequence, resulting in complex product pro-files.

The infectivity of a primary HIV strain using PBMCs as hostswas reproducibly neutralized by the IgAs at concentrationscomparable to their physiological concentrations in blood andsaliva. EP-421-433 inhibited the neutralization. This suggeststhat recognition of gp120 residues 421–433 is important in themechanism of neutralization, consistent with the essential roleof the 421–433 region in HIV binding to host cells.7–9 Residues421–433 are largely conserved in diverse HIV strains and a re-combinant Ab to this region displays broad neutralization ofstrains belonging to different HIV clades.35 In contrast, mostAbs formed adaptively following HIV infection display limitedbreadth of neutralization as they are directed to the immun-odominant V3 region of gp120.45 Increased antigen neutraliza-tion attributable to the catalytic function of an Ab fragment hasbeen reported previously.46 HIV neutralization by serum IgAin the present study was evident only after comparatively pro-longed incubations with the virus, whereas salivary IgA repro-ducibly neutralized the virus more rapidly, consistent with itssuperior catalytic activity of the latter IgA preparation (by �15-

fold). From gp120 hydrolysis rate data fitted to the equationPt � S0[1 � e(�k[Ab0]t) (P, product concentration at time t; S0

and Ab0, respectively, initial gp120 and Ab concentrations; k,second-order rate constant kcat/Km corresponding to catalytic ef-ficiency; see Paul et al.11 for methods), the kcat/Km values forsalivary and serum IgA were 1.01 � 0.04 � 104 and 3.80 �0.02 � 102 min�1 M�1. If viral and monomer gp120 are hy-drolyzed equivalently and no inhibitors are present, 90% of vi-ral gp120 will be inactivated in 5 and 7 min, respectively, atthe physiological concentrations of IgA in saliva (0.30 mg/ml)and serum IgA (2.6 mg/ml). These rates will hold good untilthe concentration exceeds Kd (the viral gp120 concentrationused for the foregoing computation, 2 � 10�13 M, correspondsto 106 HIV copies/ml with 100 gp120 molecules/virion8). Com-mercially available pooled human IgG (IVIG) has been pro-posed previously for therapy of HIV infection.39 IgG prepara-tions purified in our laboratory and commercial IVIG did nothydrolyze gp120 or neutralize HIV at the concentrations testedhere. The anti-HIV action of locally produced IgAs in the fe-male reproductive tract and the potential utility of supplement-ing any such natural defense mechanism by exogenous appli-cation of salivary IgA (e.g., as a topical microbicide) remain tobe studied.

The observed IgA catalytic and neutralizing activities repre-sent the average properties of individual Abs contained in poly-clonal salivary and serum Ab fractions, and only limited infor-mation about the structural factors governing the activities canbe gleaned from these studies. Mucosal secretions containdimeric and polymeric IgA. Differences in the number of com-bining sites/Ab molecule could result in varying antigen-bind-ing strength due to avidity effects. However, this is an issueonly if multivalent binding to repeat epitopes within antigenmonomers or aggregates is feasible. gp120 does not containknown repeat epitopes and does not form aggregates in solu-tion. Avidity effects, therefore, are unlikely to contribute to theobserved activity differences between salivary IgA, serum IgA,and serum IgG. Ab V domains are responsible for the gp120SAg

recognition and hydrolysis functions.1,12 Nonetheless, structuraldifferences in the constant domain segment of these Ab classesexist, and the constant domains of IgG Abs have been observedto influence antigen binding interactions.47 The possibility mustbe left open, therefore, that the constant domain architecturecan influence catalytic site integrity via allosteric effects. Onthe other hand, if it is assumed that the constant domain archi-tecture is without consequence, differences in the gp120SAg-hy-drolyzing activity must be due to structural factors intrinsic tothe V domains. For example, differences in germline V geneusage or the patterns of V domain adaptive sequence diversifi-cation at mucosal and systemic sites may produce Ab popula-tions with distinct V domain sequences and varying ability tocomplete the steps necessary for gp120SAg hydrolysis, i.e., non-covalent gp120 recognition, nucleophilic attack, hydrolysis ofthe acyl-Ab covalent intermediate, and product release.

The existence of anti-gp120SAg-hydrolyzing IgAs in in-fection-free humans begs questions about the nature of im-munological processes driving their synthesis. Ab interac-tions with superantigens and conventional antigens arecharacterized, respectively, by more heavily weighted con-tributions from the framework regions (FRs) versus the com-

PLANQUE ET AL.1550

plementarity determining regions (CDRs).48,49 One possibil-ity is that the FR-encoded gp120SAg recognition function isretained fortuitously as the CDR sequences undergo diversi-fication to bind conventional, unrelated antigens. Alterna-tively, a non-HIV antigen that bears structural resemblanceto gp120SAg may drive adaptive improvement of thegp120SAg recognition over the course of differentiation ofIgA-producing but not IgG cells. This is conceivable becausethe FRs are susceptible to low-level sequence diversificationand limited numbers of CDR residues are also described ascontributing to gp120SAg recognition.48 The observation ofnucleotide sequence similarity between the gp120SAg region421–433 and a human endogenous retroviral sequence50 lendscredence to the notion that antigens may exist in uninfectedhumans that drive the production of anti-gp120SAg Abs. Thefeasibility of adaptively improved gp120SAg recognition issupported by studies on Abs found in the autoimmune dis-ease systemic lupus erythematosus. HIV-neutralizing Abs di-rected to the gp120 region 421–436 have been described inlupus patients.51 Clinical case studies have commented on thelow frequency of coexistent lupus and HIV infection.52 Incontrast, there is no evident amplification of gp120SAg bind-ing Abs in HIV infected subjects consistent with the beliefthat B cell superantigens induce cellular apoptosis.1,53 Pep-tide bond hydrolysis liberates considerably greater amountsof energy compared to noncovalent B cell receptor engage-ment, and the functional outcomes of gp120SAg hydrolysis by catalytic B cell receptors versus their noncovalent occupancymay not be identical. It may be hypothesized that the hydrolyticenergy can be utilized to induce a productive conformation inIgA (�) class receptors but not IgG (�) class receptors, trig-gering cellular proliferation and clonal selection of the catalyst-producing cells. These considerations suggest the theoreticalfeasibility of catalytic anti-gp120SAg IgAs, but the immuno-logical mechanisms underlying their synthesis remain to be elu-cidated.

In summary, our studies indicate that IgAs from uninfectedsubjects catalyze the cleavage of gp120 and neutralize the in-fectivity of a primary HIV strain in tissue culture. This is thefirst report suggesting the defensive function of catalytic IgAsdirected to the 421–433 region of gp120. The precise contri-bution of catalytic IgAs as a protective mechanism in initialinfection requires further study. Important factors that canimpact the suggested protective function are differences incatalytic IgA concentration at the site of infection, differen-tial recognition of diverse viral strains by the IgAs, and dif-ferences in the rate of entry and replication of the infectingviral strain. Our studies are also significant in regard to HIVvaccine development. The sequence of gp120 residues421–433 is comparatively conserved in various HIV strains,and induction of catalytic Abs may be proposed as the basisof novel vaccination strategies. Rare monoclonal catalyticAbs induced following immunization with an electrophilicanalog of gp120 have been reported.31 Provided the super-antigenic character of the EP-421-433 probe employed in thepresent study does not present hurdles in inducing an adap-tive immune response, this compound may be useful in di-recting the response to a functionally important epitope ofgp120.

ACKNOWLEDGMENTS

We thank Robert Dannenbring and Yogesh Bangale for tech-nical assistance. The authors have no conflicting financial in-terests. Funded by grants from National Institutes of Health[AI31268, AI058865, AI058684, AG025304, AI82515, andUL1 RR024148 (CTSA)] and the Texas Higher Education Co-ordinating Board.

REFERENCES

1. Berberian L, Goodglick L, Kipps TJ, and Braun J: Immunoglobu-lin VH3 gene products: Natural ligands for HIV gp120. Science1993;261:1588–1591.

2. Townsley-Fuchs J, Kam L, Fairhurst R, Gange SJ, Goodglick L,Giorgi JV, Sidell N, Detels R, and Braun J: Human immunodefi-ciency virus-1 (HIV-1) gp120 superantigenbinding serum antibod-ies. A host factor in homosexual HIV-1 transmission. J Clin Invest1996;98:1794–1801.

3. Devito C, Hinkula J, Kaul R, Kimani J, Kiama P, Lopalco L, BarassC, Piconi S, Trabattoni D, Bwayo JJ, Plummer F, Clerici M, andBroliden K: Cross-clade HIV-1-specific neutralizing IgA in mucosaland systemic compartments of HIV-1-exposed, persistently seroneg-ative subjects. J Acquir Immune Defic Syndr 2002;30:413–420.

4. Cao Y, Qin L, Zhang L, Safrit J, and Ho DD: Virologic and im-munologic characterization of long-term survivors of human im-munodeficiency virus type 1 infection. N Engl J Med 1995;332:201–208.

5. Goodglick L, Zevit N, Neshat MS, and Braun J: Mapping the Igsuperantigen-binding site of HIV-1 gp120. J Immunol 1995;155:5151–5159.

6. Karray S and Zouali M: Identification of the B cell superantigen-binding site of HIV-1 gp120. Proc Natl Acad Sci USA 1997;94:1356–1360.

7. Olshevsky U, Helseth E, Furman C, Li J, Haseltine W, and So-droski J. Identification of individual human immunodeficiencyvirus type 1 gp120 amino acids important for CD4 receptor bind-ing. J Virol 1990;64:5701–5707.

8. Pollard SR, Meier W, Chow P, Rosa JJ, and Wiley DC: CD4-bind-ing regions of human immunodeficiency virus envelope glycopro-tein gp120 defined by proteolytic digestion. Proc Natl Acad SciUSA 1991;88:11320–11324.

9. Huang CC, Tang M, Zhang MY, Majeed S, Montabana E, Stan-field RL, Dimitrov DS, Korber B, Sodroski J, Wilson IA, WyattR, and Kwong PD: Structure of a V3-containing HIV-1 gp120 core.Science 2005;310:1025–1028.

10. Gololobov G, Sun M, and Paul S: Innate antibody catalysis. MolImmunol 1999;36:1215–1222.

11. Paul S, Karle S, Planque S, Taguchi H, Salas M, Nishiyama Y,Handy B, Hunter R, Edmundson A, and Hanson C: Naturally occurring proteolytic antibodies: selective immunoglobulin M-catalyzed hydrolysis of HIV gp120. J Biol Chem 2004;279:39611–39619.

12. Paul S, Volle DJ, Beach CM, Johnson DR, Powell MJ, and MasseyRJ: Catalytic hydrolysis of vasoactive intestinal peptide by humanautoantibody. Science 1989;244:1158–1162.

13. Matsuura K and Sinohara H: Catalytic cleavage of vasopressin byhuman Bence Jones proteins at the arginylglycinamide bond. BiolChem 1996;377:587–589.

14. Li L, Paul S, Tyutyulkova S, Kazatchkine MD, and Kaveri S: Cat-alytic activity of antithyroglobulin antibodies. J Immunol 1995;154:3328–3332.

IgA GP 120ASES 1551

15. Lacroix-Desmazes S, Moreau A, Sooryanarayana, Bonnemain C,Stieltjes N, Pashov A, Sultan Y, Hoebeke J, Kazatchkine MD, andKaveri SV: Catalytic activity of antibodies against factor VIII inpatients with hemophilia A. Nat Med 1999;5:1044–1047.

16. Thiagarajan, P., Dannenbring, R., Matsuura, K., Tramontano, A.,Gololobov, G., and Paul, S: Monoclonal antibody light chain withprothrombinase activity. Biochemistry 2000;39:6459–6465.

17. Hifumi E, Mitsuda Y, Ohara K, and Uda T: Targeted destructionof the HIV-1 coat protein gp41 by a catalytic antibody light chain.J Immunol Methods 2002,269:283–298.

18. Hifumi E, Hatiuchi K, Okuda T, Nishizono A, Okamura Y, andUda T: Specific degradation of H. pylori urease by a catalytic an-tibody light chain. FEBS J 2005;272:4497–4505.

19. Odintsova ES, Buneva VN, and Nevinsky GA: Casein-hydrolyz-ing activity of sIgA antibodies from human milk. J Mol Recognit2005;18:413–421.

20. Ponomarenko NA, Durova OM, Vorobiev II, Belogurov AA Jr,Kurkova IN, Petrenko AG., Telegin GB, Suchkov SV, Kiselev SL,Lagarkova MA, Govorun VM, Serebryakova MV, Avalle B, Tor-natore P, Karavanov A, Morse HC 3rd, Thomas D, Friboulet A,and Gabibov AG: Autoantibodies to myelin basic protein catalyzesite-specific degradation of their antigen. Proc Natl Acad Sci USA2006;103:281–286.

21. Gao QS, Sun M, Rees AR, and Paul S: Site-directed mutagenesisof proteolytic antibody light chain. J Mol Biol 1995;253:658–664.

22. Ramsland PA, Terzyan SS, Cloud G., Bourne CR, Farrugia W,Tribbick G, Geysen HM, Moomaw CR, Slaughter CA, and Ed-mundson AB: Crystal structure of a glycosylated Fab from an IgMcryoglobulin with properties of a natural proteolytic antibody.Biochem J 2006;395:473–481.

23. Powers JC, Asgian JL, Ekici OD, and James KE: Irreversible in-hibitors of serine, cysteine, and threonine proteases. Chem Rev2002;102:4639–4750.

24. Paul S, Tramontano A, Gololobov G, Zhou YX, Taguchi H, KarleS, Nishiyama Y, Planque S, and George S: Phosphonate esterprobes for proteolytic antibodies. J Biol Chem 2001;276:28314–28320.

25. Planque S, Bangale Y, Song XT, Karle S, Taguchi H, PoindexterB, Bick R, Edmundson A, Nishiyama Y, and Paul S: Ontogeny ofproteolytic immunity: IgM serine proteases. J Biol Chem 2004;279:14024–14032.

26. Polosukhina DI, Buneva VN, Doronin BM, Tyshkevich OB, BoikoAN, Gusev EI, Favorova OO, and Nevinsky GA: Hydrolysis of myelinbasic protein by IgM and IgA antibodies from the sera of patients withmultiple sclerosis. Med Sci Monit 2005;11:BR266–BR272.

27. Mitsuda Y, Planque S, Hara M, Kyle R, Taguchi H, Nishiyama Y,and Paul S: Naturally occurring catalytic antibodies: Evidence forpreferred development of the catalytic function in IgA class anti-bodies. Mol Biotechnol 2007;36:113–122.

28. Krieger JW, Crowe M, and Blank SE: Chronic glutamine supple-mentation increases nasal but not salivary IgA during 9 days of in-terval training. J Appl Physiol 2004;97:585–591.

29. Kyle RA and Rajkumar SV: Monoclonal gammopathies of unde-termined significance: A review. Immunol Rev 2003;194:112–139.

30. Planque S, Taguchi H, Burr G, Bhatia G, Karle S, Zhou YX,Nishiyama Y, and Paul S: Broadly distributed chemical reactivityof natural antibodies expressed in coordination with specific anti-gen binding activity. J Biol Chem 2003;278:20436–20443.

31. Paul S, Planque S, Zhou YX, Taguchi H, Bhatia G., Karle S, Han-son C, and Nishiyama Y: Specific HIV gp120-cleaving antibodiesinduced by covalently reactive analog of gp120. J Biol Chem2003;278:20429–20435.

32. Nishiyama Y, Taguchi H, Luo JQ, Zhou YX, Burr G, Karle S, andPaul S: Covalent reactivity of phosphonate monophenyl esters withserine proteinases: An overlooked feature of presumed transitionstate analogs. Arch Biochem Biophys 2002;402:281–288.

33. Taguchi H, Burr G, Karle S, Planque S, Zhou YX, Paul S, andNishiyama Y: A mechanismbased probe for gp120-hydrolyzing an-tibodies. Bioorg Med Chem Lett 2002;12:3167–3170.

34. Nishiyama Y, Bhatia G, Bangale Y, Planque S, Mitsuda Y, TaguchiH, Karle S, and Paul S: Toward selective covalent inactivation ofpathogenic antibodies: A phosphate diester analog of vasoactiveintestinal peptide that inactivates catalytic autoantibodies. J BiolChem 2004;279:7877–7883.

35. Karle S, Planque S, Nishiyama Y, Taguchi H, Zhou YX, Salas M,Lake D, Thiagarajan P, Arnett F, Hanson CV, and Paul S: Cross-clade HIV-1 neutralization by an antibody fragment from a lupusphage display library. AIDS 2004;18:329–331.

36. Frazer JK and Capra JD: Immunoglobulins: Structure and function.In: Fundamental Immunology, 4th ed. (Paul WE, ed.). Lippincott-Raven, Philadelphia, 1999, pp. 37–74.

37. Kerr MA. The structure and function of human IgA. Biochem J1990;271: 285–296.

38. Raux M, Finkielsztejn L, Salmon-Ceron D, Bouchez H, Excler JL,Dulioust E, Grouin JM, Sicard D, and Blondeau C: Comparison ofthe distribution of IgG and IgA antibodies in serum and variousmucosal fluids of HIV type 1-infected subjects. AIDS Res HumRetroviruses 1999;15:1365–1376.

39. Olopoenia L, Young M, White D, Barnes S, Rahbar F, and Fo-mufod A: Intravenous immunoglobulin in symptomatic and asymp-tomatic children with perinatal HIV infection. J Natl Med Assoc1997;89:543–547.

40. Kalaga R, Li L, O’Dell JR, and Paul S: Unexpected presence ofpolyreactive catalytic antibodies in IgG from unimmunized donorsand decreased levels in rheumatoid arthritis. J Immunol 1995;155:2695–2702.

41. Brandtzaeg P, Farstad IN, Johansen FE, Morton HC, NorderhaugIN, and Yamanaka T: The Bcell system of human mucosae and ex-ocrine glands. Immunol Rev 1999;171:45–87.

42. Beynon RJ and Salvesen G: Appendix III. In: Proteolytic Enzymes:A Practical Approach (Beynon RJ, Salvesen G, eds.). IRL Press,Oxford, UK, 1989, pp. 241–249.

43. Sun M, Gao QS, Li L, and Paul S: Proteolytic activity of an anti-body light chain. J Immunol 1994;153:5121–5126.

44. Paul S: Natural catalytic antibodies. Mol Biotechnol 1996;5:197–207.

45. Moore J and Trkola A: HIV type 1 coreceptors, neutralizationserotypes, and vaccine development. AIDS Res Hum Retroviruses1997;13:733–736.

46. Berisha HI, Bratut M, Bangale Y, Colasurdo G, Paul S, and SaidSI: New evidence for transmitter role of VIP in the airways: Im-paired relaxation by a catalytic antibody. Pulmon Pharmacol Ther2002;15:121–127.

47. Morelock MM, Rothlein R, Bright SM, Robinson MK, GrahamET, Sabo JP, Owens R, King DJ, Norris SH, Scher DS, Wright JL,and Adair JR: Isotype choice for chimeric antibodies affects bind-ing properties. J Biol Chem 1994;269:13048–13055.

48. Graille M, Stura EA, Corper AL, Sutton BJ, Taussig MJ, Char-bonnier JB, and Silverman GJ: Crystal structure of a Staphylococ-cus aureus protein A domain complexed with the Fab fragment ofa human IgM antibody: Structural basis for recognition of B-cellreceptors and superantigen activity. Proc Natl Acad Sci USA2000;97:5399–5404.

49. Neshat MN, Goodglick L, Lim K, and Braun J: Mapping the B cellsuperantigen binding site for HIV-1 gp120 on a V(H)3 Ig. Int Im-munol 2000;12:305–312.

50. Nishiyama Y, Karle S, Planque S, Taguchi H, and Paul S: Anti-bodies to the superantigenic site of HIV-1 gp120: Hydrolytic andbinding activities of the light chain subunit. Mol Immunol2007;44;2707–2718.

51. Bermas BL, Petri M, Berzofsky JA, Waisman A, Shearer GM, andMozes E: Binding of glycoprotein 120 and peptides from the HIV-

PLANQUE ET AL.1552

1 envelope by autoantibodies in mice with experimentally inducedsystemic lupus erythematosus and in patients with the disease.AIDS Res Hum Retroviruses 1994;10:1071–1077.

52. Palacios R, Santos J, Valdivielso P, and Marquez M: Human im-munodeficiency virus infection and systemic lupus erythematosus.An unusual case and a review of the literature. Lupus 2002;11:60–63.

53. Goodyear CS and Silverman GJ: Staphylococcal toxin inducedpreferential and prolonged in vivo deletion of innate-like B lym-phocytes. Proc Natl Acad Sci USA 2004;101:11392–11397.

Address reprint requests to:Sudhir Paul

Chemical Immunology Research CenterDepartment of Pathology and Laboratory Medicine

University of Texas–Houston Medical School6431 Fannin

Houston, Texas 77030

E-mail: [email protected]

IgA GP 120ASES 1553