of August 8, 2016.This information is current as

and Low Somatic MutationsMonoclonal IgA1 Presenting Short CDRH3 Specific Anti-HIV1 Broadly Neutralizing

−First Membrane Proximal External Region

Chanut, Christian Genin, Michel Cogné and Stephane PaulRochereau, Alexandre Girard, Fabienne Jospin, BlandineVerrier, Gael Champier, Nadine Vincent, Nicolas Fahd Benjelloun, Zeliha Oruc, Nicole Thielens, Bernard

ol.1600309http://www.jimmunol.org/content/early/2016/07/30/jimmun

published online 1 August 2016J Immunol 

MaterialSupplementary

9.DCSupplemental.htmlhttp://www.jimmunol.org/content/suppl/2016/07/30/jimmunol.160030

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2016 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

at Institut Pasteur-m

ediatheque scientifique on August 8, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

First Membrane Proximal External Region–SpecificAnti-HIV1 Broadly Neutralizing Monoclonal IgA1 PresentingShort CDRH3 and Low Somatic Mutations

Fahd Benjelloun,* Zeliha Oruc,† Nicole Thielens,‡,x,{ Bernard Verrier,x,‖ Gael Champier,#

Nadine Vincent,* Nicolas Rochereau,* Alexandre Girard,* Fabienne Jospin,*

Blandine Chanut,* Christian Genin,* Michel Cogne,† and Stephane Paul*

Mucosal HIV-1–specific IgA have been described as being able to neutralize HIV-1 and to block viral transcytosis. In serum and

saliva, the anti-HIV IgA response is predominantly raised against the envelope of HIV-1. In this work, we describe the in vivo

generation of gp41-specific IgA1 in humanized a1KI mice to produce chimeric IgA1. Mice were immunized with a conformational

immunogenic gp41-transfected cell line. Among 2300 clones screened by immunofluorescence microscopy, six different gp41-

specific IgAwith strong recognition of gp41 were identified. Two of them have strong neutralizing activity against primary HIV-1

tier 1, 2, and 3 strains and present a low rate of somatic mutations and autoreactivity, unlike what was described for classical

gp41-specific IgG. Epitopes were identified and located in the hepted repeat 2/membrane proximal external region. These Abs

could be of interest in prophylactic treatment to block HIV-1 penetration in mucosa or in chronically infected patients in

combination with antiretroviral therapy to reduce viral load and reservoir. The Journal of Immunology, 2016, 197: 000–000.

Immunoglobulin A are the major mediators of immunologicaldefense at mucosal surfaces. Studies focusing on the role ofmucosal env-specific secretory IgA (sIgA) present in genital

fluids from infected women highlight their ability to neutralizeHIV primary isolates in vitro and to inhibit HIV transcytosisacross epithelial cells (1, 2). This neutralizing property has alsobeen demonstrated for systemic and mucosal IgA in HIV-1+ andexposed seronegative (ESN) individuals (3, 4), but the magnitudeand frequency of HIV-specific sIgA seem to be very low duringchronic infection (5). The initial Ab response to HIV-1 can be

detected as circulating anti-gp41 IgG followed by production of

anti-gp120 Abs a few weeks later, but none of these responses

are able to efficiently neutralize the infecting strain (6). The

maturation of broadly neutralizing Abs (bNAbs) takes several

months (7) and goes through different steps of somatic mutations

and modifications (8, 9) before becoming more specific to known

sites of vulnerability of the HIV envelope such as the CD4

binding site, glycan V3 region, V1/V2 loop, and gp41, especially

the membrane proximal external region (MPER) domain (9, 10).

Neutralizing IgAs have been described at the mucosal level with

specificity against gp41-specific epitopes such as QARILAVERY,

Kennedy, or MPER (1, 11, 12). MPER is a highly conserved

region in gp41 involved in the early stages of HIV attachment

and infection (13). This region comprises the epitopes of well-

described human monoclonal IgG broadly neutralizing Abs such

as ELDKWA for 2F5, WF (N/D) IT for 4E10, and Z13 (14) for

10E8 (15). Recent studies have described the potency in target-

ing MPER for the generation of Abs with increased breadth,

potency, and high-affinity binding. Indeed, structural observa-

tions from the crystal structure of the interaction of modified

bNAbs and epitopes have provided clues to address gp41 and its

different fusion intermediate states. New broadly neutralizing

Abs like PGT122 and 35O22 have been also described, recog-

nizing an unknown epitope overlapping the gp41–gp120 inter-

face, and these Abs present potent neutralizing activity (9, 16).

Moreover, the role of cholesterol and the lipid context in the

mechanisms of broad neutralization by enhancing the presenta-

tion of MPER to B cells has been highlighted (17–19). Never-

theless, most of the studied bNAbs are IgGs (14). To date, very

few studies have described the role of broadly neutralizing anti-

MPER IgA and their involvement in the protection from and/or

neutralization of HIV (20, 21) during experimental anti-HIV

vaccine trials. The role of mucosal protective neutralizing IgA

against mucosal transmission of HIV is unclear and remains

controversial (5).

*Groupe d’Immunite des Muqueuses et Agents Pathogenes/EA3064, INSERM CIE31408 Vaccinology, Universites de Lyon, 42055 Saint-Etienne, France; †Universite deLimoges, CNRS UMR 7276, 87032 Limoges, France; ‡Universite Grenoble Alpes,Institut de Biologie Structurale, F-38044 Grenoble, France; xCommissariat a l’EnergieAtomique, Institut de Biologie Structurale, F-38044 Grenoble, France; {CNRS, Institutde Biologie Structurale, F-38044 Grenoble, France; ‖Institut de Biologie et Chimie desProteines, FRE3310/CNRS, Universite de Lyon, F69007 Lyon, France; and #B CellDesign, 87032 Limoges, France

ORCIDs: 0000-0001-6939-8721 (F.B.); 0000-0003-1314-6757 (Z.O.); 0000-0002-7354-0302 (N.T.); 0000-0002-8478-7095 (B.V.); 0000-0002-0089-9357 (G.C.); 0000-0001-6579-0275 (N.R.); 0000-0002-7251-8779 (A.G.); 0000-0001-7841-4973 (B.C.); 0000-0003-2065-8267 (C.G.); 0000-0002-8519-4427 (M.C.); 0000-0002-8830-4273 (S.P.).

Received for publication March 1, 2016. Accepted for publication July 3, 2016.

This work was supported by the Grenoble Instruct Centre (ISBG; UMS 3518 CNRS-CEA-UJF-EMBL) with support from FRISBI (ANR-10-INSB-05-02) and GRAL(ANR-10-LABX-49-01) within the Grenoble Partnership for Structural Biology.F.B. was supported by a fellowship from the Region Rhone-Alpes (France). Thiswork was also financed by research grants from Agence Nationale de Recherches surle SIDA, Sidaction, and Cluster 10 of Infectiology (ARC1).

Address correspondence and reprint requests to Prof. Stephane Paul, Groupe d’Im-munite des Muqueuses et Agents Pathogenes/EA3064, INSERM CIE3 1408 Vacci-nology, Universites de Lyon, 42055 Saint-Etienne, France. E-mail address: stephane.paul@chu-st-etienne.fr

The online version of this article contains supplemental material.

Abbreviations used in this article: bNAb, broadly neutralizing Ab; CDRH3, hydro-phobic H chain third CDR; CLP, cardiolipin; ESN, exposed seronegative; HC,H chain; HR, hepted repeat; KD, dissociation equilibrium constant; MPER, mem-brane proximal external region; RU, resonance unit; sIgA, secretory IgA; TCID50,tissue culture infective dose.

Copyright� 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1600309

Published August 1, 2016, doi:10.4049/jimmunol.1600309 at Institut Pasteur-m

ediatheque scientifique on August 8, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

The aim of our study was to elicit and to characterize recombinanthuman monoclonal IgA neutralizing Abs that target the C-terminalregion of gp41. To improve the immunogenicity of our Ag, theHEK 293-gp41MSD cell line that allows expression of a folded,transmembrane gp41 has been used for immunization (22). Trans-fected cells were injected into humanized chimeric mice a1KI toproduce gp41-specific humanized chimeric IgA1 (23). In this article,we describe the use of this mouse model to generate potent neu-tralizing IgA Abs. High-affinity Abs were produced as mAbs afterimmortalizing and selecting specific Ab-producing cells throughhybridoma derivation (24). Immunization in mice allowed thegeneration of six highly potent gp41-specific IgA-producing clones.From these clones, two Abs presented high recognition of gp41and potent neutralization activity in vitro against both HIV-1 labo-ratory–adapted virus and primary isolates from different clades. TheIgA1 clones F4.30 and C6.11 display high and low avidity to gp41,respectively, but high neutralizing potency against different HIV-1strains. These two Abs recognize a large conformational epitope thatoverlaps the highly conserved epitope of 4E10 and present low ratesof autoreactivity, unlike some previously described MPER-specificIgG and present a very low rate of somatic mutation (25).

Materials and MethodsReagents

The following materials were obtained from the NIH AIDS Research andReference Reagent Program, Division of AIDS, National Institute of Allergyand Infectious Diseases, National Institutes of Health: primary HIV-1 isolatesBAL, 92UG029, 92UG001, 92US660, HIV-1 G3, CAM1970, 92BR025,SF162, LAI, the HIV-1 subtype B (MN), and Env Peptides (15-mer) Com-plete Set. gp140 recombinant protein (strain 97/CN/54, clade C/B) and 3D6,4E10, and 2F5 mAbs were obtained from Polymun Scientific (Vienna, Aus-tria). The gp41 ectodomain from HXB2 strain was produced in Escherichiacoli in our laboratory. Human HEK 293 cell lines were obtained from theAmerican Type Culture Collection, and SP2/0 cells were obtained from UMRCentre National de la Recherche Scientifique 6101 (Limoges, France).HEK293-gp41MSD were developed and characterized in our research group(26). The a1KI mice were immunized and provided by the UMR CentreNational de la Recherche Scientifique 6101 (Limoges, France) (23, 26).

Immunization of mice and mAb production

a1KI mice (6–8 wk old) were immunized with 50 mg of naked DNA (plasmiddisplay hepted repeat (HR)1-PID-HR2 used to develop the HEK293-gp41MSD)in 2.5 ml of Ringer solution; this hydrodynamics-based immunization wererealized by i.v. injection via the tail vein. Next, two groups of mice wereimmunized according to two different protocols of immunization. At day 14,one group was immunized i.p. with 1 3 106 cells of HEK 293-gp41MSD inCFA (Sigma-Aldrich, St. Louis, MO) followed at day 45 by one i.p. injectionwith 1 3 106 cells in IFA (Sigma-Aldrich), one i.p. injection with 1 3 106

cells in PBS, and finally one i.v. boost with 1 3 106cells in PBS at day 60.Use of CFA/IFA adjuvants allowed the reduction of the number of humanHEK cells used for immunization to increase the number of accessible epi-topes in the emulsion and finally to facilitate the migration of APCs to thespleen. All mice were sacrificed 3 d after the last boost.

The second group of mice was immunized at day 15 using hydrodynamicmethod with 50 mg of naked DNA in 2.5 ml of Ringer solution followed by i.pinjection with 1 3 106 cells in CFA at day 45 and finally one i.v. boost with1 3 106 cells in PBS at day 60. Hydrodynamics enhances the DNA trans-fection efficiency, allowing the injected DNA to reach all the organs rapidly.The sera of each immunized mouse were tested by fluorescence microscopyon HEK293-gp41MSD cell lines. Splenocytes from selected animals wereimmediately fused with myeloma cells SP2/0 by using polyethyleneglycol-1000 as fusion reagent. The hybridoma cells were cultured in DMEMhigh glucose–L-glutamate medium (PAA, Velizy-Villacoublay, France) sup-plemented with 10% FCS (Life Technologies, Saint-Aubin, France), 0.1mmol/l hypoxanthine, 100 U/ml penicillin, and 100 mg/ml streptomycin in96-well plates. The hybridoma cells were diluted in round-bottom 96-wellplates by limiting dilutions to obtain 0.5 cells/well and cultured for 7 d.

Production and purification of anti-MPER IgA

Hybridomas were expanded in DMEM high glucose–L-glutamate medium(PAA) supplemented with 10% FCS (Life Technologies), 0.1 mmol/l hy-

poxanthine, 100 U/ml penicillin, and 100 mg/ml streptomycin and IL-6at 0.5 ng/ml (R&D Systems, Abingdon, U.K.). IgA were purified fromsupernatants by peptide M affinity chromatography (InvivoGen, Toulouse,France). Purified IgA were then concentrated on Amicon Ultra filters withmolecular mass cutoff of 50 kDa (Millipore, Lyon, France).

Epitope mapping of mAbs

ELISA was carried out using gp41 HXB2 clade B produced in E. coli,gp140 strain 97/CN/54 and Consensus HXB2 (clade B) env 15-mer peptidecomplete set that covers and overlaps the gp41 protein. An irrelevantpeptide representing the human hemagglutinin A was used as a negativecontrol. Each peptide was coated on a 96-well plate (Maxisorp; Nunc) at5 mg/ml in 0.1 mol/l sodium carbonate buffer (pH 9.6) overnight at 4˚C.The plates were washed with TBS (144 mmol/l NaCl and 25 mmol/lTris-HCl [pH 7.5]) containing 0.5% Tween 20 and blocked with the samebuffer and 2% BSA for 1 h at 37˚C. After washing, 100 ml of the generatedAbs (5 mg/ml) diluted in TBS were added to the wells and incubated for 2 hat 37˚C. After washing, detection was performed by using HRP-conjugatedgoat anti-human IgA and o-phenylenediamine substrate. The absorbance wasread at 492 nm. All assays were performed in triplicate (MultiskanMicroplatePhotometer, Thermo Scientific).

An ELISA for HIV specificity was also performed with coated freevirions and Abs as described previously (25).

Immunofluorescence on HEK293-gp41MSD

Binding of our anti-MPER IgA to the cell surface of HEK293-gp41MSD wasvisualized by immunofluorescence microscopy. HEK 293-gp41MSD cellswere cultured for 2 d in 96-well plates and fixed in cold acetone. IgAwereincubated for 30 min on fixed cells at a concentration of 5 mg/ml/well.After three washes with PBS containing 0.2% Tween 20, cells were incu-bated for 30 min with FITC-conjugated goat anti-human IgG or anti-humanIgA as appropriate (Abliance, Besancon, France). Different anti-gp41 humanmAbs 4E10, 2F5, and 3D6 at 10 mg/ml were used as positive controls and ananti–c-myc conjugated with FITC was used as positive control of gp41expression. The labeling was visualized by fluorescence microscopy (NikonTE 2000 Microscope, Burlingame, CA) using NIS-Elements software.

Binding to HEK293-gp41MSD by flow cytometry

Membrane binding to gp41 on HEK293-gp41MSD cells was also studied byflow cytometry. A total of 5 mg/ml IgA or positive control as 4E10 or 2F5were incubated 45 min at 37˚C. The cells were washed three times andthen incubated with R-PE goat anti-human IgA (Abliance, Besancon,France) for 30 min at 37˚C. After washing, cells were analyzed by flowcytometry. For multiple labeling, as in the binding assay, infected cellswere first labeled with anti–p24 KC57-FITC according to conditions rec-ommended by the supplier (Beckman Coulter, Villepinte, France). Next,the anti-MPER IgA were incubated 45 min at 37˚C and after three washeswere incubated with R-PE goat anti-human IgA (Abliance, Besancon,France). During acquisition of data, at least 10,000 events were analyzedon FACScan (BD Biosciences, Pont de Claix, France).

Measure of neutralizing activity

Measure of the neutralizing activity of purified gp41-specific Abs wasperformed using T cell line adapted strain LAI (clade B) or primary isolatesof clade A (92UG029), B (BAL, SF162, 92US660, and QHO), C (92BR025),G (HIV-1 G3), CRF02 AG (CAM1970), and D (92UG001). Neutralizingactivities were measured in duplicate and repeated three times. For LAI assay,CD4+/CXCR4+ SupT1 cells were incubated at 37˚C under 5% CO2 inDMEM supplemented with 10% FCS. Infection was performed in 96-well rounded bottom plates. Purified Abs (50 ml) were incubated with anequal volume of virus containing a 100 50% tissue culture infective doses(TCID50) for 2 h at 37˚C. Then, 3 3 105 SupT1 cells/well were added.At day 1, infected cells were washed twice with tissue culture medium.Supernatants were collected 7 d postinfection and p24 ELISAwas quantified(HIV-1 p24 Ag Capture Assay; Advanced Bioscience Laboratories). For theBaL strain, neutralizing activity of purified Abs was determined as describedpreviously (27).

Briefly, PBMCs were isolated from healthy donors (from EFS AuvergneLoire, France) and stimulated with PHA (5 mg/ml) and 200 U/mlrecombinant human IL-2 (Abcys, Paris, France). The cells were infectedwith a 100 TCID50 dose for 3 h at 37˚C in the presence of purified Abs, andafter 48 h, the cells were washed twice. Infected supernatants were col-lected 7 d postinfection, and p24 was quantified by ELISA (ABL). Neu-tralizing activity of purified Abs was also tested on stimulated PBMCinfected with CAM1970, 92UG029, 92US660, 92BR025, 93BR025, HIV-1G3, and 92UG001 primary HIV-1 isolates as described previously. The

2 gp41-SPECIFIC NEUTRALIZING IgA1

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

cells were infected with a 100 TCID50 dose in the presence of seriallydiluted Ab. The supernatants were collected 10 d postinfection, and p24measurement was performed. The percentage of neutralization was cal-culated as the reduction of p24 production. IC50 and IC80 were determinedas the lowest Ab concentration able to confer 50 and 80% of neutralization,respectively.

Production of pseudoviruses COT6.15 and mutants

Pseudoviruses were generated by transfecting the COT6.15env plasmid orderived mutant (28, 29) with pSG3Denv plasmid, using the Lyovec trans-fection reagent (InvivoGen) in HEK293 cells in a petri dish. The productionof pseudoviruses was confirmed by quantification of p24 produced, and theirinfectivity and measurement of TCID50s were quantified by infectingTZM-Bl cells (JC53-bl) with serial dilutions of the supernatant in triplicatein the presence of DEAE-dextran (30 mg/ml; Sigma-Aldrich, St. Louis). Theinfection was monitored 24 h later by evaluating luciferase activity, using theBright Glo reagent (Promega, Madison, WI) following the manufacturer’sinstructions. Luminescence was measured in a Berthold Tristar multilabelcounter during 5 s/well (Berthold THOIRY, France). The TCID50 was cal-culated as described previously (30). Wells with relative light unit readingsof 10 times that of the negative control were considered as positive.

Measurement of the avidity of anti-MPER IgA1

Biacore analyses were conducted on a BIAcore 3000 Instrument (GEHealthcare, Velizy-Villacoublay, France). Recombinant gp41 (HXB2), gp140(97/CN/54), and gp120c41 (HXB2; Abcys Eurobio, Courtaboeuf, France)were immobilized on a CM5 sensorchip (GE Healthcare) using aminecoupling chemistry. To determine avidity constants, increasing concentrationsof MPER IgA were injected over immobilized gp41 (1000 resonance units[RU]), gp140 (7500 RU), and gp120c41 (8200 RU) for 120 s at a flow rate of

20 ml/min in PBS containing 0.005% surfactant P20. Between injections, thesensorchip surface was regenerated by pulse injections (10 ml) of 10 mMHCl. The specific binding signal was obtained by subtracting the back-ground signal over a reference surface with no protein immobilized. Globalcurve fitting was carried out using BiaEvaluation software version 4.1.Before fitting, buffer blanks were subtracted from the data sets. The apparentdissociation equilibrium constants (KD) were calculated from the ratio of thedissociation and association rate constants.

Study of the autoreactivity of anti-MPER IgA

Autoreactivity to human epithelial (HEp-2) cells was determined by indirectimmunofluorescence on slides using FITC-conjugated goat anti-human IgApolyclonal Ab (Abliance, Besancon, France). IgA were all tested at25 mg/ml. Autoreactivity of Abs was also tested by specific ELISAwithpurified cardiolipin (CLP) (Thermofisher, Illkirch, France) as describedpreviously (31).

Competition assay

To study the ability of anti-MPER IgA to recognize the same epitope aswell-known anti-MPER IGg previously described, a competition assay wasperformed by ELISA.

The protein gp120c41 (HXB2; Abcys Eurobio, Courtaboeuf, France)were coated on a 96-well plate (Maxisorp; Nunc) at 5 mg/ml in 0.1 mol/lsodium carbonate buffer (pH 9.6) overnight at 4˚C. The plates werewashed with PBS containing 0.05% Tween 20 and blocked with the samebuffer and 4% BSA for 1 h at 37˚C. After washing, 100 ml of 2F5 and4E10 (0–20 mg/ml) diluted in PBS were added to the wells and incubated2 h at 37˚C. After washing, 100 ml of the competition Abs were added tothe wells at the concentration of 5 mg/ml, diluted in PBS+BSA 1%, andincubated 1 h at 37˚C. After washing, detection was performed using

FIGURE 1. Characterization of MPER-specific IgA1. (A) Coomassie blue staining and Western blot analysis of MPER-specific IgA. 10 mg of purified

IgAwere loaded on a 9% acryl/bis-acryl gel under nondenaturing conditions. Detection of IgAwas revealed with a polyclonal goat-anti-human IgA (Sigma-

Aldrich). Size of monomers, dimers, and polymers are indicated. (B) Indirect immunofluorescence of HEK 293gp41MSD-transfected cells with MPER-specific

IgA. Immunoreactivity of purified IgA (500 ng) was revealed by a rabbit polyclonal anti-human IgA labeled with FITC. No binding of MPER-specific IGAwas

observed on untransfected cells (ctrl-). 3D6 was used as positive control. (C) Neutralizing activity of MPER-specific IgA1 on primary and laboratory HIV-1

strains. 4E10 and 2F5 were used as positive controls. Dose-dependent neutralization was performed within a concentration range of 0.156–10 mg/ml. The

experiments were performed three times with similar results.

The Journal of Immunology 3

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

HRP-conjugated goat anti-human IgA (MP Biomedical, Strasbourg, France)and TMB substrate (Tebu bio, Le Perray en Yvelines, France). The absor-bance was read at 450 nm. All assays were performed in duplicate (Mul-tiskan Microplate Photometer; Thermo Scientific).

Sequencing and analysis

Sequencing was performed from purified DNA from pelted hybridomas toanalyze the H chain nucleotide sequence and the characteristics of CDRH3(B Cell Design, Limoges, France). (The characterization of Abs sequencewas performed using imgt.org/IMGT_vquest and ExPASy.org/ProtParamtools; http://www.ebi.ac.uk/Tools/msa/kalign/).

Statistical analysis

The statistical difference in vitro experiments between controls and Abs wasassessed using Student t test (Statistica 5.1 software; Statsoft, Maison-Alfort,France). A p value , 0.05 has been considered statistically significant. Thewhole results were obtained from the mean of at least a triplicate or twotriplicates. The significances were verified according to an ANOVA test withBonferroni posttests to compare the row means. The correlations were ob-tained after a Spearman correlation for nonparametric data.

ResultsTo induce gp41-specific neutralizing Abs (NAbs), immunization ofhumanized transgenic mice (23) were performed using HEK293-gp41MSD human cell line that allowed native shape expression ofa gp41 (22). After immunizations,.2300 hybridoma B cell cloneswere isolated and immortalized. Supernatants of all individual B cellclones were tested by immunofluorescence on HEK293-gp41MSD

and untransfected cells (Fig. 1B, Supplemental Fig. 1). Among thehybridomas, only six clones presented highly specific recognitionof the HEK293-gp41MSD cells. From those clones, IgA1 wereproduced and purified (Fig. 1A). Recombinant chimeric IgA areproduced both as monomers, dimers, and polymers (Fig. 1A). Theaverage concentration of IgA1 was with a range between 86 and295 ng/ml before purification.

Elicited IgA1 are specific to gp41 ectodomain

To confirm the specificity of purified IgA1, immunofluorescence onHEK 293-gp41MSD was performed. Only clones with high specificrecognition of HEK 293-gp41MSD were selected (Fig. 1B). Thisfirst screening allowed the selection of six IgA1 clones (F4.5,F4.6, F4.30, C6.11, C6.13, and SC16C13) (Fig. 1B).

F4.30 and C6.11 are highly potent and broadly neutralizingAbs of tier 1, 2, and 3 primary HIV-1 strains

The neutralizing activities of purified MPER-specific IgA1 weretested on different strains. Among them, F4.30 exhibited a strongneutralization activity with a high proportion of neutralized pri-mary isolates (.80% of tested strains) with very low IC80 and IC50

in the range of few nanograms per milliliter (Fig. 1C, Table I).F4.30 neutralized 92UG029, LAI, and COT6.15 strains with astrong and higher potency (IC80 , 1 mg/ml) than those obtainedfor 4E10 IgG or 2F5 IgA with LAI and those reported for 10E8.F4.30 neutralized BAL and 92UG001D strains more efficientlythan 4E10 with IC80 five times lower for BAL but higher than for2F5 IgA (3.3, 17.19, and 0.81 mg/ml, respectively). F4.30 showedstronger neutralization potency for 92UG001D strain with an IC80

of 8.62 mg/ml that is 47 and 4 times lower than those obtained for4E10 (410 mg/ml) and 2F5 (36 mg/ml), respectively. SF162 and92BR025 strains were weakly neutralized by F4.30, and the IC80

was lower than for 4E10. C6.11 neutralized efficiently the infec-tion of 70% of the tested strains with an IC50 , 1 mg/ml. C6.11neutralized 92US660 and CAM1970 with the same potency as4E10 and 2F5 IgA, whereas LAI, COT6.15, and HIVG3 wereneutralized with lower IC80. BAL, 92UG029, and 92UG001 weremoderately neutralized by C6.11, but its IC80 is lower than 4E10.C6.11 neutralized SF162 and 92BR025 with an IC80 of 13.27 andT

able

I.NeutralizingactivityofMPER-specificIgA

NAbsforIC

80/IC50(mg/m

l)

HIV-1

Strains

Tier

F4.30

C6.11

SC16C13

F4.6

C6.13

F4.5

2F5IgA

4E10IgG

IC80

IC50

IC80

IC50

IC80

IC50

IC80

IC50

IC80

IC50

IC80

IC50

IC80

IC50

IC80

IC50

BAL(CCR5;B)

1B

3.33a

2.08a

2.09a

0.46b

2.67a

1.67a

0.15b

0.09b

9.06a

4.05a

8.74a

2.23a

0.81b

2.39a

.10c

.10c

92UG029

(CXCR4;A)

2–3

0.15b

0.09b

2.13a

1.33a

0.49b

0.31b

2.33b

0.45b

2.04a

0.55b

8.85a

1.84a

2.15a

1.34a

.10c

9.82a

92UG001

(CCR5/CXCR4;D)

2–3

8.62a

5.39a

2.22a

1.38a

.10c

6.39a

.10c

.10c

9.33a

5.83a

8.69a

5.43a

.10c

.10c

.10c

.10c

92US660(CCR5;B)

10.13b

0.08b

0.16b

0.10b

0.17b

0.11b

0.13b

0.08b

0.13b

0.08b

0.14b

0.09b

0.13b

0.08b

0.16b

0.10b

HIV-1

G3(CCR5;G)

20.86b

0.54b

0.22b

0.14b

.10c

.10c

.10c

8.81b

0.19b

0.61b

0.26b

0.16b

0.87b

0.54b

1.02a

0.63b

CAM1970

(CCR5;CRF02AG)

2–3

0.16b

0.10b

0.15b

0.09b

0.19b

0.12b

0.22b

0.13b

2.40a

0.64b

.10c

0.49b

0.15b

0.09b

0.16b

0.10b

92BR025(CCR5;C)

1.10c

9.16b

.10c

0.58b

3.92a

2.45a

4.67a

2.92a

.10c

9.80c

.10c

2.42a

1.49c

0.93b

.10a

.10a

SF162(CCR5;B)

1A

.10c

6.34b

.10c

2.34a

.10c

.10c

8.93a

2.12a

.10c

.10c

9.89a

2.61a

.10c

2.33a

0.62b

0.39b

LAI(CXCR4;B)

10.23b

0.14b

0.72b

0.45b

.10c

2.17a

.10c

.10c

.10c

0.54b

.10c

8.96a

0.60b

0.38b

2.59a

1.62a

COT6.15

20.15b

0.09b

0.14b

0.09b

.10c

0.67b

2.72a

0.52b

2.61a

0.47b

2.28a

0.13b

.10c

0.52b

0.59b

0.10b

2F5A

IgA

and4E10IgG

areusedas

controls.

aIC

50orIC

80,

1mg/m

l.b1mg/m

l,

IC50orIC

80,

10mg/m

l.cIC

80orIC

50.

10mg/m

l.

4 gp41-SPECIFIC NEUTRALIZING IgA1

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

13.37 mg/ml, respectively, 20 times lower than 4E10 or 10E8 butclose to 2F5 IgA. SC16C13 neutralized .30% of the tested strainswith an IC80 , 137 mg/ml. 92UG029, 92US660, and CAM1970 wereneutralized with IC80 , 1 mg/ml (equivalent to 4E10 and 2F5). F4.6and C6.13 strongly neutralized only three of the tested strains andF4.5 presented weak neutralization except for 92US660 and HIV-1G3 that were neutralized with IC80 , 1 mg/ml (Fig. 1C, Table I).

gp41-specific IgA1 recognize different epitopes in the MPERregion

To characterize the neutralizing epitopes recognized by the morepotent IgA1 F4.30 and C6.11, the binding of purified IgA1 to peptideepitopes of the Env glycoprotein (clade B) and to recombinant gp41 orgp140 were measured (Supplemental Table I). The specificity ofIgA1 was directed to the MPER and HR2 regions. F4.30 and C6.11recognized the 2F5 neutralizing epitope on MPER. C6.11 recognizedthe two neutralizing epitopes ELDKWA and WFD/NIT on MPER,whereas F4.30 recognized the ELDKW motif and the HR2 region.All recombinant IgA1 were able to strongly recognize both the gp41ectodomain and the gp140 glycoprotein (Supplemental Table I).These epitopes were tested in a competition assay with 4E10 and2F5 based on recombinant gp120c41. C6.11 and F4.30 recognizedepitopes close to or that encompass those of 2F5 as indicated by therespective loss of 22.6 and 28.7% of their epitope binding capacitywith 2F5. 10 mg/ml. In contrast, the binding of F4.30 and C6.11 togp120C41 seems to be less altered by increasing concentrations of4E10 with a 3.4 and 10% decrease, respectively (Fig. 2A).

Anti-MPER IgA1s have a strong avidity for gp41

IgA1 abilities to recognize the gp41 glycoprotein were measuredusing surface plasmon resonance. Kinetic experiments were per-

formed using different targets as the gp41 ectodomain (clade B),the gp140 (clade C) and the gp120cgp41 (clade B). After immo-bilization of the target proteins, MPER-specific IgA1 (F4.30 andC6.11) as well as 4E10 or 3D6 IgGs at various concentrations(0.5–160 nM) were tested (Fig. 2B). F4.30 recognized gp120c41with a KD of 12 nM similar to 3D6 IgG (10.4 nM) whereas theKD of 4E10 was 0.34 nM (Fig. 2B, Supplemental Table II). Incomparison with 3D6 (KD of 14.9 nM) or 4E10 (KD of 10.9 nM)(Fig. 2B, Supplemental Table II), F4.30 recognized gp140 with alower avidity (KD of 18.8 nM). C6.11 recognized weakly thegp120c41 precluding determination of the KD value.

F4.30 and C6.11 are able to recognize both viral particles andinfected cells

To demonstrate the ability of IgA1 to efficiently recognize HIV-1particles, an ELISA was performed with coated viral particles ofboth laboratory and primary strains. As a negative control, an anti-OVA IgA1 was used. Free virions from BAL, CAM1970, LAI, andSF162 strains were strongly recognized by C6.11 and F4.30(Supplemental Table III). The anti-OVA IgA1 did not bind toCAM1970, BAL, and SF162 viral particles and showed a veryweak binding to the LAI strain (Supplemental Table III). Thebinding to free viral particles is probably the mechanism involvedin the neutralizing activity of IgA1. An infection assay with TCLALAI and SupT1 cell line was performed to demonstrate the abilityof IgA1 to specifically recognize HIV-1–infected cells (Fig. 2C).The percentage of p24+/IgA1+ cells was 12% with a mean fluo-rescence intensity ratio (infected/uninfected) of 5.1 for C6.11. F4.30 was not able to recognize .10% of infected cells with a meanfluorescence intensity of 4.3. No binding was observed with anti-OVA IgA1.

FIGURE 2. Interaction between MPER-specific IgA with gp41. (A) Competition assay for C6.11 and F4.30 binding to gp120c41 in the presence of 2F5

or 4E10. gp120c41-coated wells were preincubated with 2F5 or 4E10 at different concentrations (0–20 mg/ml). C6.11 and F4.30 were used at 5 mg/ml.

Assays were performed in triplicate. (B) Kinetic analysis of the interaction between MPER-specific IgA and recombinant gp120c41. C6.11 (5–80 nM),

F4.30 (0.5–16 nM), and 4E10 (0.5–8 nM) were injected over 8200 RU of immobilized gp120c41 in PBS containing 0.005% surfactant P20. The binding

signals shown were obtained by double referencing (subtraction of the reference surface and of buffer blank injection signals). Fits are shown as dotted lines

and were obtained by global fitting of the data using a 1:1 Langmuir binding model. (C) Binding of MPER-specific IgA on SupT1/LAI-infected cells.

Number of IgA+/p24+ SupT1 cells was analyzed by flow cytometry. 4E10 was used as positive control and labeling with anti-p24+ + anti-IgA-PE as

negative control. The ratio of the mean fluorescence of IgA+/p24+ SupT1 cells on the mean fluorescence of IgA2/p24+ SupT1 cells is indicated. Uninfected

cells were used as negative control.

The Journal of Immunology 5

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

Neutralizing epitopes of IgA1 are larger than 2F5 or 4E10 epitope.Mapping of neutralizing epitopes was performed by measuring theirneutralizing activities with HIV-1 COT6.15 pseudoviruses and com-pared with mutated pseudoviruses with alanine substitutions inMPER residues (662–682) as described previously (Fig. 3A) (32).As described for well-characterized MPER-specific IgG as 2F5 or4E10, IgA1 neutralizing activity was highly sensitive to substitutionof the W666 tryptophan and the K667 lysine residues that arecentral in the 2F5 recognition (14). For F4.30, a downstream-extended 2F5 epitope merged with the upstream 4E10 epitope667KNLWSWF D/N/S I675 seems to be crucial for neutralization.However, mutation of the T676 residue also reduced the neutral-izing potency of F4.30. C6.11 was also sensitive to the substitutionof W666, W670, and L669 residues also crucial in 2F5 neutrali-zation activity. C6.11 was also sensitive to the mutation of theC-Term residues of MPER (674D/N/SITKWLWYI682). The F4.30and C6.11 epitopes are probably larger than 4E10 and 2F5 epitopes.It was confirmed by the determination of an IC80 COT mutantversus IC80 COT wild-type ratio of for each Ala substitution.

F4.30 and C6.11 are not autoreactive/polyreactive

A property common to the previously characterizedMPER-specificmAbs 2F5 and 4E10 is their ability to cross-react with self-antigens(33). In addition, binding to both the cell membrane and the Envtrimer is thought to be important for optimal neutralization.Autoreactivity of IgA1 was first tested on HEp-2 cells by immu-nofluorescence (Fig. 3B). F4.30 and C6.11 did not recognize Hep-2cells contrary to what was observed and previously describedfor 2F5 or 4E10 (Fig. 3B). Immunoreactivity of IgA1 againstphospholipids was also tested with purified CLP. Unlike C6.11that did not recognize anionic phospholipids, such as CLP, F4.30

presented a very low level of binding to cardiolipin (SupplementalTable IV).

F4.30 and C6.11 show low levels of mutation and have a shortand low hydrophobic H chain third CDR

Wenext aimed to understand whether F4.30 and C6.11 share the samecharacteristics as 2F5, 4E10, or 10E8, which have a long and hy-drophobic H chain third CDR (CDRH3) and a high rate of somaticmutation. The H chains of F4.30 and C6.11 were sequenced and theCDRH3 of each clonewas analyzed and comparedwith 2F5 and 10E8(34). The sequences of the H chain of C6.11 and F4.30 are identicaland could come from the same original clone (Fig. 4A). However,the sequence of the L chains differs from two residues. F4.30 andC6.11 present a shorter CDRH3 with 9-aa residues, whereas 2F5 and10E8 (accession numbers JX645769/JX645770) contains 24 and 22residues respectively (Fig. 4B). The CDRH3 of F4.30 and C6.11 isless hydrophobic than 2F5 but more hydrophobic than 10E8 with ahydropathicity score of 20.833, 20.225, and 21.386, respectively.By comparing the sequence of the constant H chain of each Ab andthe VDJ allele precursor, we observed that both F4.30 and C6.11 Hchains come from the IGHV3-602 F with which they share 98.96%of identity, the IGHJ2*1 with 95.74% identity, and IGHD2-1*01F.The alignment of the variable domain of the H chains of F4.30 andC6.11 with 2F5, 10E8, and PG9 showed matches of 51.67, 40.83,and 49.17% respectively (data not shown).

DiscussionRecently, many studies have demonstrated the role of Nabs duringthe course of HIV-1 infection. New advances allowed to elicit Absdirected against different key epitopes, such as the quaternaryneutralizing epitopes of the viral envelope (35, 36). Passive ad-

FIGURE 3. (A) Screening of the neutralizing activity of MPER-specific IgA on COT6.15 pseudoviruses and COT6.15 mutants. MPER sequence is shown in

each panel. For each mAb and mutant, the value of SD is calculated for the neutralization from two independent experiments realized in triplicate. The critical

residues for neutralization are indicated in red. (B) Analysis of MPER-specific IgA autoreactivity on HEp-2 cells by immunofluorescence. F4.30, C6.11, anti-

OVA IgA negative control, and anti-DNA IgG positive control were used at 25 mg/ml. Assay were performed in triplicate. Original magnification 320.

6 gp41-SPECIFIC NEUTRALIZING IgA1

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

ministration of rare human envelope-specific monoclonal bNAbsto nonhuman primate models such as rhesus macaques can protectagainst SHIV or SIV challenge (37–39).Moreover, gp120-specific Abs seem to be one of the correlates of

protection in the recent RV144 vaccine trial (40, 41) even if themucosal immune response was not investigated. The presenceof systemic IgA (mainly monomeric) directed against certaingp120 specificities were associated with higher rates of infection.However, other specific IgA were not associated with infectionrisk (40, 42, 43). In contrast, the presence of IgA in saliva andserum, directed against different key regions of gp160 (CD4binding site, V2 loop or the MPER) and able to neutralize dif-ferent HIV-1 strains in vitro, has been associated to protection inESN (1, 12). Moreover, several studies have described the po-tential ability of mucosal sIgA to locally block infection (39, 44).Humanized mice are an alternative to humans for the production

of Abs (45) and may be useful for pharmacologic studies of anti-HIV drugs and passive immunization (46, 47). The interest inusing humanized mice is related to the wide repertoire of Ags theycan recognize. In this study, a1KI humanized mice were used toelicit neutralizing IgA1 specific to the MPER region of gp41.

The a1KI mice model allows the production of chimeric mo-nomeric, dimeric, and polymeric IgA1 with equivalent rate ofglycosylation similar to those encountered in human (48). Cross-linking of mIgA+ B cells residing in MALTs mediates local IgAsecretion responses in this model (49, 50). As for IgG, a B cellstage with mIgA expression is necessary to mount IgA Ab re-sponses (51). That both early and late B cell development can beensured by expression of membrane a H chain (HC) was shown inthe a1KI model where Sm was replaced with an Ig HC a1 gene(45). In such mice, a partial pre-B defect was reminiscent of datain mice with premature expression of membrane HC. BCR densitywas lower in a1KI than in wt mice, and was associated with in-creased abundance of both long-lived and short-lived plasma cells.In this model, a prime boost strategy using naked DNA as a prime

by hydrodynamic i.v. route was used to increase the immunogenicityof the gp41 (23). Using HEK293-gp41MSD cells as a boost gave riseto six clones able to specifically recognize gp41 in a conformationalshape. Among these clones, two Abs (F4.30 and C6.11) showed ahigh neutralizing potency for different strains of HIV. The neu-tralizing activity for F4.30 and C6.11 are very similar to the onereported for the 10E8 another bNAbs gp-41–specific IgG but higher

FIGURE 4. Sequencing of the C region of the H and L chain of F4.30 and C6.11. (A) Alignment of the sequences of the H chain of F4.30 and C6.11 and

the consensus sequence. Nucleotides that are different from the consensus sequence are highlighted in yellow, and the corresponding residues are in blue

and highlighted in yellow. (B) “Pearl necklace” representation of the amino acid sequence of the H chains of F4.30 and C6.11 and characterization of their

CDRH3. Amino acid sequence of the H chain of F4.30 and C6.11 as well as 2F5 and 10E8. CDRH1 are represented in red, CDR2 in orange, and CDRH3 in

purple. The length of the CDRH3 (number of amino acids) and the score of hydropathicity are mentioned under each sequence. (Obtained from IMGT/

VQUEST tools and ExPASy ProtPRWEMaram tool).

The Journal of Immunology 7

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

than for 4E10 or 2F5 (both in IgG and IgA backbone). It could bevery important to assess the breadth of the neutralizing activityagainst a wide range of pseudoviruses from National Institutes ofHealth panels. Moreover, the solubility of these two clones seemsto be higher than for 10E8 (52). F4.30 recognizes both MPERneutralizing epitopes ELDKWand WFD/NIT and the HR2 regionswith different efficiency. C6.11 bound the KWA motif in theELDKWA 2F5 neutralizing epitope. These data correlated withthe results obtained by surface plasmon resonance using immo-bilized gp41, gp140, and gp120c41 proteins. The high neutrali-zation activity of F4.30 is also correlated to a high level ofrecognition of free viral particles and the absence of autoreactivity.The high neutralization capacities in comparison with the well-known IgG 4E10 could be explained by a better recognition of thegp41 conformational shape by F4.30.Screening of the F4.30 binding site and competition assay suggest

that the epitope could be conformational or discontinuous and in-volve HR2 and other regions. In fact, the polymeric nature of IgAsuggests that they could bind large functional epitopes possiblylocated on adjacent proteins such as gp41 and gp120 during theinfection process or to HR1/HR2 domains in a fusion complex state.Other studies have described purified neutralizing IgA Abs from

patients directed against different key regions on the gp160 surfaceincluding the super antigenic site on gp120 and the CD4 binding site(53), which are also discontinuous epitopes. Furthermore, the re-cently described bNAb 35O22 recognizes and efficiently binds alarge epitope corresponding to the interface of gp120 and gp41(9, 16). Structural studies have described the ability to enhance theneutralization potency of well-known Abs like 10E8 by presentingMPER in a different context using a hydrocarbon stapled peptide thatreinforces unique architectures, such as the helix-kink-helix MPERmotif that presents a new fashioning of peptide structure (19). It willbe interesting to make such studies with F4.30 and C6.11.Numerous studies have also shown that the advent of heterologous

neutralizing Abs in a subset of HIV patient appears many yearspostinfection, what is explained by multiple rounds of modificationsand maturation that are exerted by a continually evolving pathogen(7, 54). These mutations lead to different features common to someneutralizing Abs including an exceptional high rate of somaticmutations and a long hydrophobic CDRH3 (8, 55). In fact, a longCDHR3 is observed for Abs that target and penetrate the envelopeglycan shield like PG9, PG16, and PGT145 or the 10E8 that targetsMPER (15, 56). However, 35O22 that recognizes a different site ofvulnerability on the glycoprotein envelope uses a novel mechanismof glycan-protein recognition, combining a protruding FW3 withCDRH1, H2, and H3 to form a “bowl” that holds glycan. FW3 andCDRH3 provide the top edges of the bowl and interact with theprotein surface of gp120, whereas CDRH1 and H2 are recessed andhold/recognize glycan. This structural mechanism of recognitioncontrasts with the extended CDRH3-draping glycan observed withother Abs that penetrate the glycan shield such as PG9 andPGT128. Whereas 35O22 Ab possesses a CDRH3 composed of 14aa that is shorter than 4E10 or the 2F5 Ab but presents an insertionof 8 aa in framework 3. This high rate of somatic mutation andframework 3 insertions is a feature of other HIV-specific bNAbs (9,16, 57, 58). F4.30 and C6.11 presented a low rate of mutation with ahigh score of similarity to the consensus VDJ alleles. This suggestsa weak maturation of the F4.30 and C6.11 clones and can beexplained by an exposition to a small number of immunogen var-iants because of the immunization protocol. Furthermore, F4.30 andC6.11 presented a short CDRH3 with a low hydrophobicity incomparison with 2F5, which does not exclude reactivity with acomplex glycosylated epitope but discards the idea of a deeplyhidden region buried within the glycan shield.

An important inconvenience of well-known MPER-specific Absas 2F5 and 4E10 is that they cross-react with self-antigens. F4.30and C6.11 do not present any autoreactivity or recognition of self-antigens as phospholipids. Studies of potent Abs such as 2G12 (59)described the capacity of monomeric or dimeric 2G12 to mediateAb-dependent cellular cytotoxicity or to activate the complementsystem in vitro (60, 61). The capacity of F4.30 to mediate Ab-dependent cellular cytotoxicity or Ab-dependent cell-mediatedvirus inhibition against infected cells remains to be investigated.This work strengthens our previous studies of the role of anti-MPER IgA present in parotid saliva of ESN and HIV+ individ-uals (1, 12, 28). As for M66.6 and M66 2F5-like Abs (62),MPER-specific IgA1 are similar but present different activities ofrecognition and binding for gp41. It may suggest that both Ablineages evolved under similar constraints for recognition andneutralization of HIV-1, whereas functional elements within CDR-like glycosylation may differ during this evolution and impact theirproperties.F4.30 and C6.11 IgA have the capacity to strongly neutralize.80%

of the HIV-1 strains tested. The avidity of IgA for its Ag was notcorrelated to neutralizing activity. C6.11 has a weak Ag aviditybut presents a high cross-clade neutralizing activity. The profile ofthe higher neutralizing IgA Abs is of intermediate avidity for therecognition of MPER, high recognition of free viral particles, andan absence of auto reactivity.These twomAbs (F4.30 and C6.11) are to our knowledge the first

recombinant neutralizing cross-clade IgA1. There are, however,remaining questions with the a1KI chimeric model such as therelevance with regard to a normal physiological development ofAbs, the proportion in those mice of LB1 that do not develop intomemory B cells, LB2, or B cells of the marginal zone, and theimpact of a prior IgA switch in the B cell compartment. a1KI is apromising chimeric model for the eliciting of new humanized Abs.Further understanding of the interaction of F4.30 with gp41 atthe atomic level would require crystallographic studies. MPER-specific IgA1 could be now tested in vivo to block the entry ofHIV-1 after recent viral exposure or to reduce, after systemicadministration, viral burden, and the massive CD4 cell depletionin HIV-1 reservoirs such as the intestine. Potent mAbs acquiretheir neutralization potency from their ability to block a func-tionally important site that is critical for viral entry as seen for theCD4 binding site Abs. Nonetheless, the breadth and potency ofF4.30 demonstrates a conserved site of gp41 is an important targetAg for HIV neutralization. The highly conserved MPER is a targetof potent, non-self-reactive neutralizing Abs, confirming the highinterest in MPER-based HIV vaccine design.

AcknowledgmentsWe thank Philip Lawrence for critical reading of the manuscript. We thank

Isabelle Bally for assistance and access to the surface plasmon resonance

platform of the Partnership for Structural Biology in Grenoble. We also

thank Dr. Elin Gray and Prof. Lynn Morris for providing the COT6.15

plasmids. This work used the platforms of the Grenoble Instruct centre

(ISBG, UMS 3518 CNRS-CEA-UJF-EMBL).

DisclosuresThe authors have no financial conflicts of interest.

References1. Vincent, N., E. Malvoisin, B. Pozzetto, F. Lucht, and C. Genin. 2004. Detection

of IgA inhibiting the interaction between gp120 and soluble CD4 receptor inserum and saliva of HIV-1-infected patients. AIDS 18: 37–43.

2. Matsuda, S., and M. Noda. 2000. Detection of IgA-binding sites on humanimmunodeficiency virus type-1 envelope glycoproteins, Gp120 and Gp41.Microbiol. Immunol. 44: 923–929.

8 gp41-SPECIFIC NEUTRALIZING IgA1

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

3. Kaul, R., F. Plummer, M. Clerici, M. Bomsel, L. Lopalco, and K. Broliden. 2001.Mucosal IgA in exposed, uninfected subjects: evidence for a role in protectionagainst HIV infection. AIDS 15: 431–432.

4. Hirbod, T., X. Kong, G. Kigozi, A. Ndyanabo, D. Serwadda, J. L. Prodger,A. A. Tobian, F. Nalugoda, M. J. Wawer, K. Shahabi, et al. 2014. HIVacquisitionis associated with increased antimicrobial peptides and reduced HIV neutralizingIgA in the foreskin prepuce of uncircumcised men. [Published erratum appearsin 2014 PLoS Pathog. 10: e1004515.] PLoS Pathog. 10: e1004416.

5. Mestecky, J., P. F. Wright, L. Lopalco, H. F. Staats, P. A. Kozlowski,Z. Moldoveanu, R. C. Alexander, R. Kulhavy, C. Pastori, L. Maboko, et al. 2011.Scarcity or absence of humoral immune responses in the plasma and cervico-vaginal lavage fluids of heavily HIV-1-exposed but persistently seronegativewomen. AIDS Res. Hum. Retroviruses 27: 469–486.

6. Tomaras, G. D., N. L. Yates, P. Liu, L. Qin, G. G. Fouda, L. L. Chavez,A. C. Decamp, R. J. Parks, V. C. Ashley, J. T. Lucas, et al. 2008. Initial B-cellresponses to transmitted human immunodeficiency virus type 1: virion-bindingimmunoglobulin M (IgM) and IgG antibodies followed by plasma anti-gp41antibodies with ineffective control of initial viremia. J. Virol. 82: 12449–12463.

7. Mouquet, H. 2014. Antibody B cell responses in HIV-1 infection. TrendsImmunol. 35: 549–561.

8. Finton, K. A., K. Larimore, H. B. Larman, D. Friend, C. Correnti, P. B. Rupert,S. J. Elledge, P. D. Greenberg, and R. K. Strong. 2013. Autoreactivity and ex-ceptional CDR plasticity (but not unusual polyspecificity) hinder elicitation ofthe anti-HIV antibody 4E10. PLoS Pathog. 9: e1003639.

9. Pancera, M., T. Zhou, A. Druz, I. S. Georgiev, C. Soto, J. Gorman, J. Huang,P. Acharya, G. Y. Chuang, G. Ofek, et al. 2014. Structure and immune recog-nition of trimeric pre-fusion HIV-1 Env. Nature 514: 455–461.

10. Klein, F., H. Mouquet, P. Dosenovic, J. F. Scheid, L. Scharf, andM. C. Nussenzweig. 2013. Antibodies in HIV-1 vaccine development and ther-apy. Science 341: 1199–1204.

11. Clerici, M., C. Barassi, C. Devito, C. Pastori, S. Piconi, D. Trabattoni, R. Longhi,J. Hinkula, K. Broliden, and L. Lopalco. 2002. Serum IgA of HIV-exposeduninfected individuals inhibit HIV through recognition of a region within thealpha-helix of gp41. AIDS 16: 1731–1741.

12. Benjelloun, F., R. Dawood, S. Urcuqui-Inchima, N. Rochereau, B. Chanut,B. Verrier, F. Lucht, C. Genin, and S. Paul. 2013. Secretory IgA specific forMPER can protect from HIV-1 infection in vitro. AIDS 27: 1992–1995.

13. Munoz-Barroso, I., K. Salzwedel, E. Hunter, and R. Blumenthal. 1999. Role ofthe membrane-proximal domain in the initial stages of human immunodeficiencyvirus type 1 envelope glycoprotein-mediated membrane fusion. J. Virol. 73:6089–6092.

14. Zwick, M. B., and D. R. Burton. 2007. HIV-1 neutralization: mechanisms andrelevance to vaccine design. Curr. HIV Res. 5: 608–624.

15. Huang, J., G. Ofek, L. Laub, M. K. Louder, N. A. Doria-Rose, N. S. Longo,H. Imamichi, R. T. Bailer, B. Chakrabarti, S. K. Sharma, et al. 2012. Broad andpotent neutralization of HIV-1 by a gp41-specific human antibody. Nature 491:406–412.

16. Huang, J., B. H. Kang, M. Pancera, J. H. Lee, T. Tong, Y. Feng, H. Imamichi,I. S. Georgiev, G. Y. Chuang, A. Druz, et al. 2014. Broad and potent HIV-1neutralization by a human antibody that binds the gp41-gp120 interface. Nature515: 138–142.

17. Apellaniz, B., E. Rujas, P. Carravilla, J. Requejo-Isidro, N. Huarte, C. Domene,and J. L. Nieva. 2014. Cholesterol-dependent membrane fusion induced by thegp41 membrane-proximal external region-transmembrane domain connectionsuggests a mechanism for broad HIV-1 neutralization. J. Virol. 88: 13367–13377.

18. Lai, R. P., M. Hock, J. Radzimanowski, P. Tonks, D. L. Hulsik, G. Effantin,D. J. Seilly, H. Dreja, A. Kliche, R. Wagner, et al. 2014. A fusion intermediategp41 immunogen elicits neutralizing antibodies to HIV-1. J. Biol. Chem. 289:29912–29926.

19. Bird, G. H., A. Irimia, G. Ofek, P. D. Kwong, I. A. Wilson, and L. D. Walensky.2014. Stapled HIV-1 peptides recapitulate antigenic structures and engagebroadly neutralizing antibodies. Nat. Struct. Mol. Biol. 21: 1058–1067.

20. Alfsen, A., P. Iniguez, E. Bouguyon, and M. Bomsel. 2001. Secretory IgAspecific for a conserved epitope on gp41 envelope glycoprotein inhibits epithelialtranscytosis of HIV-1. J. Immunol. Baltim. Md.: 1950 166: 6257–6265.

21. Tudor, D., and M. Bomsel. 2011. The broadly neutralizing HIV-1 IgG 2F5 elicitsgp41-specific antibody-dependent cell cytotoxicity in a FcgRI-dependent man-ner. AIDS 25: 751–759.

22. Dawood, R., F. Benjelloun, J. J. Pin, A. Kone, B. Chanut, F. Jospin, F. Lucht,B. Verrier, C. Moog, C. Genin, and S. Paul. 2013. Generation of HIV-1 potentand broad neutralizing antibodies by immunization with postfusion HR1/HR2complex. AIDS 27: 717–730.

23. Laffleur, B., V. Pascal, C. Sirac, and M. Cogne. 2012. Production of human orhumanized antibodies in mice. Methods Mol. Biol. 901: 149–159.

24. Akkina, R. 2013. New generation humanized mice for virus research: compar-ative aspects and future prospects. Virology 435: 14–28.

25. Haynes, B. F., M. A. Moody, L. Verkoczy, G. Kelsoe, and S. M. Alam. 2005.Antibody polyspecificity and neutralization of HIV-1: a hypothesis. Hum. An-tibodies 14: 59–67.

26. Duchez, S., R. Amin, N. Cogne, L. Delpy, C. Sirac, V. Pascal, B. Corthesy, andM. Cogne. 2010. Premature replacement of mu with alpha immunoglobulinchains impairs lymphopoiesis and mucosal homing but promotes plasma cellmaturation. Proc. Natl. Acad. Sci. USA 107: 3064–3069.

27. Vincent, N., A. Kone, B. Chanut, F. Lucht, C. Genin, and E. Malvoisin. 2008.Antibodies purified from sera of HIV-1-infected patients by affinity on the heptadrepeat region 1/heptad repeat region 2 complex of gp41 neutralize HIV-1 pri-mary isolates. AIDS 22: 2075–2085.

28. Morris, L., X. Chen, M. Alam, G. Tomaras, R. Zhang, D. J. Marshall, B. Chen,R. Parks, A. Foulger, F. Jaeger, et al. 2011. Isolation of a human anti-HIV gp41membrane proximal region neutralizing antibody by antigen-specific singleB cell sorting. PLoS One 6: e23532.

29. Gray, E. S., T. Meyers, G. Gray, D. C. Montefiori, and L. Morris. 2006. Insen-sitivity of paediatric HIV-1 subtype C viruses to broadly neutralising monoclonalantibodies raised against subtype B. PLoS Med. 3: e255.

30. Gray, E. S., P. L. Moore, R. A. Pantophlet, and L. Morris. 2007. N-linkedglycan modifications in gp120 of human immunodeficiency virus type 1 sub-type C render partial sensitivity to 2G12 antibody neutralization. J. Virol. 81:10769–10776.

31. Chamley, L. W., A. M. Duncalf, B. Konarkowska, M. D. Mitchell, andP. M. Johnson. 1999. Conformationally altered b 2-glycoprotein I is the antigenfor anti-cardiolipin autoantibodies. Clin. Exp. Immunol. 115: 571–576.

32. Gray, E. S., N. Taylor, D. Wycuff, P. L. Moore, G. D. Tomaras, C. K. Wibmer,A. Puren, A. DeCamp, P. B. Gilbert, B. Wood, et al. 2009. Antibody specificitiesassociated with neutralization breadth in plasma from human immunodeficiencyvirus type 1 subtype C-infected blood donors. J. Virol. 83: 8925–8937.

33. Haynes, B. F., M. A. Moody, L. Verkoczy, G. Kelsoe, and S. M. Alam. 2005.Antibody polyspecificity and neutralization of HIV-1: a hypothesis. Hum. An-tibodies 14: 59–67.

34. Ofek, G., F. J. Guenaga, W. R. Schief, J. Skinner, D. Baker, R. Wyatt, andP. D. Kwong. 2010. Elicitation of structure-specific antibodies by epitope scaf-folds. Proc. Natl. Acad. Sci. USA 107: 17880–17887.

35. Scheid, J. F., H. Mouquet, B. Ueberheide, R. Diskin, F. Klein, T. Y. Oliveira,J. Pietzsch, D. Fenyo, A. Abadir, K. Velinzon, et al. 2011. Sequence andstructural convergence of broad and potent HIV antibodies that mimic CD4binding. Science 333: 1633–1637.

36. Wu, X., A. Changela, S. O’Dell, S. D. Schmidt, M. Pancera, Y. Yang, B. Zhang,M. K. Gorny, S. Phogat, J. E. Robinson, et al. 2011. Immunotypes of a qua-ternary site of HIV-1 vulnerability and their recognition by antibodies. J. Virol.85: 4578–4585.

37. Hessell, A. J., E. G. Rakasz, P. Poignard, L. Hangartner, G. Landucci,D. N. Forthal, W. C. Koff, D. I. Watkins, and D. R. Burton. 2009. Broadlyneutralizing human anti-HIV antibody 2G12 is effective in protection againstmucosal SHIV challenge even at low serum neutralizing titers. PLoS Pathog.5: e1000433.

38. Mascola, J. R., and D. C. Montefiori. 2010. The role of antibodies in HIVvaccines. Annu. Rev. Immunol. 28: 413–444.

39. Mascola, J. R., S. S. Frankel, and K. Broliden. 2000. HIV-1 entry at the mucosalsurface: role of antibodies in protection. AIDS 14(Suppl. 3): S167–S174.

40. Haynes, B. F., P. B. Gilbert, M. J. McElrath, S. Zolla-Pazner, G. D. Tomaras,S. M. Alam, D. T. Evans, D. C. Montefiori, C. Karnasuta, R. Sutthent, et al. 2012.Immune-correlates analysis of an HIV-1 vaccine efficacy trial. N. Engl. J. Med.366: 1275–1286.

41. Rolland, M., and P. Gilbert. 2012. Evaluating immune correlates in HIV type 1vaccine efficacy trials: what RV144 may provide. AIDS Res. Hum. Retroviruses28: 400–404.

42. Shin, S. Y. 2016. Recent update in HIV vaccine development. Clin. Exp. VaccineRes. 5: 6–11.

43. Tomaras, G. D., G. Ferrari, X. Shen, S. M. Alam, H.-X. Liao, J. Pollara,M. Bonsignori, M. A. Moody, Y. Fong, X. Chen, et al. 2013. Vaccine-inducedplasma IgA specific for the C1 region of the HIV-1 envelope blocks binding andeffector function of IgG. Proc. Natl. Acad. Sci. USA 110: 9019–9024.

44. Belec, L., P. D. Ghys, H. Hocini, J. N. Nkengasong, J. Tranchot-Diallo,M. O. Diallo, V. Ettiegne-Traore, C. Maurice, P. Becquart, M. Matta, et al.2001. Cervicovaginal secretory antibodies to human immunodeficiency virustype 1 (HIV-1) that block viral transcytosis through tight epithelial barriersin highly exposed HIV-1-seronegative African women. J. Infect. Dis. 184:1412–1422.

45. Cogne, M., S. Duchez, and V. Pascal. 2009. [Transgenesis and humanization ofmurine antibodies]. Med. Sci. (Paris) 25: 1149–1154.

46. Denton, P. W., J. D. Estes, Z. Sun, F. A. Othieno, B. L. Wei, A. K. Wege,D. A. Powell, D. Payne, A. T. Haase, and J. V. Garcia. 2008. Antiretroviral pre-exposure prophylaxis prevents vaginal transmission of HIV-1 in humanized BLTmice. PLoS Med. 5: e16.

47. Balazs, A. B., J. Chen, C. M. Hong, D. S. Rao, L. Yang, and D. Baltimore. 2011.Antibody-based protection against HIV infection by vectored immunoprophy-laxis. Nature 481: 81–84.

48. Oruc, Z., C. Oblet, A. Boumediene, A. Druilhe, V. Pascal, E. Le Rumeur,A. Cuvillier, C. El Hamel, S. Lecardeur, T. Leanderson, et al. 2016. IgA structurevariations associate with immune stimulations and IgA mesangial deposition.J. Am. Soc. Nephrol. DOI: 10.1681/ASN.2015080911.

49. Leduc, I., M. Drouet, M. C. Bodinier, A. Helal, and M. Cogne. 1997. Membraneisoforms of human immunoglobulins of the A1 and A2 isotypes: structural andfunctional study. Immunology 90: 330–336.

50. Brandtzaeg, P., E. S. Baekkevold, I. N. Farstad, F. L. Jahnsen, F. E. Johansen,E. M. Nilsen, and T. Yamanaka. 1999. Regional specialization in the mucosalimmune system: what happens in the microcompartments? Immunol. Today 20:141–151.

51. Amin, R., C. Carrion, C. Decourt, E. Pinaud, and M. Cogne. 2012. Deletion ofthe a immunoglobulin chain membrane-anchoring region reduces but does notabolish IgA secretion. Immunology 136: 54–63.

52. Kwon, Y. D., I. S. Georgiev, G. Ofek, B. Zhang, M. Asokan, R. T. Bailer, A. Bao,W. Caruso, X. Chen, M. Choe, et al. Optimization of the solubility of HIV-1neutralizing antibody 10E8 through somatic variation and structure-based de-sign. J. Virol. 90: 5899–5914.

The Journal of Immunology 9

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from

53. Planque, S., M. Salas, Y. Mitsuda, M. Sienczyk, M. A. Escobar, J. P. Mooney,M. K. Morris, Y. Nishiyama, D. Ghosh, A. Kumar, et al. 2010. Neutralization ofgenetically diverse HIV-1 strains by IgA antibodies to the gp120-CD4-bindingsite from long-term survivors of HIV infection. AIDS 24: 875–884.

54. Derdeyn, C. A., P. L. Moore, and L. Morris. 2014. Development of broadlyneutralizing antibodies from autologous neutralizing antibody responses in HIVinfection. Curr. Opin. HIV AIDS 9: 210–216.

55. Prabakaran, P., J. Gan, Y. Q. Wu, M. Y. Zhang, D. S. Dimitrov, and X. Ji. 2006.Structural mimicry of CD4 by a cross-reactive HIV-1 neutralizing antibody withCDR-H2 and H3 containing unique motifs. J. Mol. Biol. 357: 82–99.

56. McLellan, J. S., M. Pancera, C. Carrico, J. Gorman, J. P. Julien, R. Khayat,R. Louder, R. Pejchal, M. Sastry, K. Dai, et al. 2011. Structure of HIV-1 gp120V1/V2 domain with broadly neutralizing antibody PG9. Nature 480: 336–343.

57. Walker, L. M., S. K. Phogat, P. Y. Chan-Hui, D. Wagner, P. Phung, J. L. Goss,T. Wrin, M. D. Simek, S. Fling, J. L. Mitcham, et al. 2009. Broad and potentneutralizing antibodies from an African donor reveal a new HIV-1 vaccine target.Science 326: 285–289.

58. Bonsignori, M., D. C. Montefiori, X. Wu, X. Chen, K. K. Hwang, C. Y. Tsao,D. M. Kozink, R. J. Parks, G. D. Tomaras, J. A. Crump, et al. 2012. Two distinctbroadly neutralizing antibody specificities of different clonal lineages in a singleHIV-1-infected donor: implications for vaccine design. J. Virol. 86: 4688–4692.

59. Platt, E. J., M. M. Gomes, and D. Kabat. 2012. Kinetic mechanism for HIV-1neutralization by antibody 2G12 entails reversible glycan binding that slows cellentry. Proc. Natl. Acad. Sci. USA 109: 7829–7834.

60. Trkola, A., M. Purtscher, T. Muster, C. Ballaun, A. Buchacher, N. Sullivan,K. Srinivasan, J. Sodroski, J. P. Moore, and H. Katinger. 1996. Human mono-clonal antibody 2G12 defines a distinctive neutralization epitope on the gp120glycoprotein of human immunodeficiency virus type 1. J. Virol. 70: 1100–1108.

61. Klein, J. S., A. Webster, P. N. Gnanapragasam, R. P. Galimidi, andP. J. Bjorkman. 2010. A dimeric form of the HIV-1 antibody 2G12 elicits potentantibody-dependent cellular cytotoxicity. AIDS 24: 1633–1640.

62. Ofek, G., B. Zirkle, Y. Yang, Z. Zhu, K. McKee, B. Zhang, G. Y. Chuang,I. S. Georgiev, S. O’Dell, N. Doria-Rose, et al. 2014. Structural basis for HIV-1neutralization by 2F5-like antibodies m66 and m66.6. J. Virol. 88: 2426–2441.

10 gp41-SPECIFIC NEUTRALIZING IgA1

at Institut Pasteur-mediatheque scientifique on A

ugust 8, 2016http://w

ww

.jimm

unol.org/D

ownloaded from