Search tips
Search criteria 


Logo of iaiPermissionsJournals.ASM.orgJournalIAI ArticleJournal InfoAuthorsReviewers
Infect Immun. 2009 November; 77(11): 4859–4867.
Published online 2009 August 24. doi:  10.1128/IAI.00117-09
PMCID: PMC2772530

Neutralizing Monoclonal Antibodies Directed against Defined Linear Epitopes on Domain 4 of Anthrax Protective Antigen[down-pointing small open triangle]


The anthrax protective antigen (PA) is the receptor-binding subunit common to lethal toxin (LT) and edema toxin (ET), which are responsible for the high mortality rates associated with inhalational Bacillus anthracis infection. Although recombinant PA (rPA) is likely to be an important constituent of any future anthrax vaccine, evaluation of the efficacies of the various candidate rPA vaccines is currently difficult, because the specific B-cell epitopes involved in toxin neutralization have not been completely defined. In this study, we describe the identification and characterization of two murine monoclonal immunoglobulin G1 antibodies (MAbs), 1-F1 and 2-B12, which recognize distinct linear neutralizing epitopes on domain 4 of PA. 1-F1 recognized a 12-mer peptide corresponding to residues 692 to 703; this epitope maps to a region of domain 4 known to interact with the anthrax toxin receptor CMG-2 and within a conformation-dependent epitope recognized by the well-characterized neutralizing MAb 14B7. As expected, 1-F1 blocked PA's ability to associate with CMG-2 in an in vitro solid-phase binding assay, and it protected murine macrophage cells from intoxication with LT. 2-B12 recognized a 12-mer peptide corresponding to residues 716 to 727, an epitope located immediately adjacent to the core 14B7 binding site and a stretch of amino acids not previously identified as a target of neutralizing antibodies. 2-B12 was as effective as 1-F1 in neutralizing LT in vitro, although it only partially inhibited PA binding to its receptor. Mice passively administered 1-F1 or 2-B12 were partially protected against a lethal challenge with LT. These results advance our fundamental understanding of the mechanisms by which antibodies neutralize anthrax toxin and may have future application in the evaluation of candidate rPA vaccines.

Bacillus anthracis has long been recognized as a serious public health threat, given the ease with which B. anthracis spores can be disseminated via aerosol and due to the high mortality rate that accompanies spore inhalation. These fears were realized in the fall of 2001, when B. anthracis spores were circulated through the U.S. postal system, resulting in five deaths, 22 known cases of infection, and the possible exposure of more than 30,000 people (7). Although routine vaccination of civilians against anthrax is neither necessary nor desirable, certain segments of the population, notably emergency first responders and research laboratory personnel, remain at risk of exposure and are in need of an effective preexposure vaccine. In the United States, the only licensed anthrax vaccine, Anthrax Vaccine Adsorbed (AVA) or Biothrax, has been subject to controversy for years, due to safety concerns (3), and is not considered an ideal vaccine due to its protracted vaccination schedule (six injections over 18 months). AVA, which consists of formalin-treated culture filtrate from an attenuated strain of B. anthracis adsorbed to aluminum hydroxide, is also inherently difficult to manufacture and to standardize. A recombinant anthrax vaccine, manufactured by VaxGen, was considered to be a leading candidate to replace AVA and to supply the Strategic National Stockpile with 75 million doses, but it was recently discontinued because of concerns regarding vaccine production and stability. For these reasons, there is now a concerted effort to develop a new anthrax vaccine based on a defined antigenic composition.

Current anthrax vaccines are focused on protective antigen (PA), the primary determinant in AVA that is involved in immunity to anthrax infection (21). PA is the receptor-binding subunit common to both lethal factor (LF), and edema factor (EF), which, upon PA binding form lethal toxin (LT) and edema toxin (ET), respectively. The anthrax toxins are responsible for the high mortality rates associated with inhalational B. anthracis infection. PA is secreted by B. anthracis as an 83-kDa protein (PA83) that binds two known anthrax toxin receptors (ATRs), namely, tumor endothelium marker 8 (TEM-8) and capillary morphogenesis protein 2 (CMG-2) (5, 6, 19). Following receptor engagement, the amino-terminal 20-kDa region of the PA (PA20) is proteolytically cleaved by a furinlike protease, exposing a region of PA that allows for the formation of the holotoxin. The membrane-bound 63-kDa PA (PA63) spontaneously oligomerizes into heptamers and then associates either with LF, a zinc-dependent protease specific for certain members of the mitogen-activated protein kinase kinase family (12, 13), or EF, a calmodulin-, Ca2+-dependent adenylate cyclase (20), to form LT and ET.

It is well established that animals immunized with recombinant PA, or passively administered anti-PA antiserum, are protected against LT/ET exposure or spore challenge (reviewed in references 16 and 7). For example, we recently reported that LT-challenged mice treated with goat anti-PA antiserum (either alone or in combination with antibiotics) demonstrated increased survival rates compared to untreated mice (18). However, the antibody response to PA is complicated: total anti-PA titers, as measured by enzyme-linked immunosorbent assays (ELISAs), do not correlate well with protection (35). This is likely due to the fact that a large number of anti-PA antibodies are directed against non-neutralizing (or possibly even toxin-enhancing) epitopes (29, 41). Pioneering work by Little et al. (23) suggests that at least 20 antigenic determinants exist on PA. At present, only a few of these epitopes has been described in detail; several have been shown to be targets of monoclonal antibodies (MAbs) capable of effective neutralization in vivo or in vitro (1, 8, 10, 17, 44, 45). The fact that key neutralizing epitopes on PA remain widely unidentified poses a significant barrier to the evaluation of the efficacy of vaccines based on full-length and truncated PA subunits.

PAs can be divided into four functional domains. Domain 1 (residues 1 to 258) contains the furin recognition site (164RKKR167), as well as the region of PA that is recognized by LF and EF (30). Domain 2 (residues 259 to 487) and part of domain 3 (residues 488 to 595) are implicated in heptamerization, pore formation, and translocation of EF/LF across endosomal membranes (4, 27, 28, 34). Domain 4 (residues 596 to 735) constitutes the region of the PA involved in receptor recognition and attachment (39). Antibodies against domain 4 are postulated to be the most effective in neutralizing LT and ET (1), since they are proposed to interfere with PA binding to ATR. However, the only neutralizing epitope that has been characterized in detail is the one recognized by MAb 14B7 (24, 37, 40). This MAb contacts the face of domain 4 that is involved in receptor recognition, and has been proposed to encompass residues 671 to 721. Certainly, additional neutralizing epitopes exist on domain 4 (1, 46). In particular, Abboud and Casadevall (1) suggested a linear epitope, immediately adjacent to or possibly overlapping the 14B7 binding site, as the target of neutralizing antibodies. The characterization of antibodies that are both linear epitope dependent and neutralizing will be useful in the evaluation specific correlates of immunity; such MAbs can be incorporated into functional screening assays of potential vaccine candidates.

The goal of our study was to identify additional B-cell epitopes, within domain 4 of PA, that constitute the targets of neutralizing MAbs. Toward this end, we screened a collection of B-cell hybridomas, produced from PA83-immunized mice, for MAbs capable of binding to a domain 4-specific peptide array. We identified two domain 4-reactive MAbs, 1-F1 and 2-B12. 1-F1 recognized a peptide spanning residues 692 to 703; this epitope maps to a region of domain 4 that overlaps with the binding site recognized by the well-characterized neutralizing antibody 14B7 (residues 671 to 721) (23, 40). 2-B12, on the other hand, recognized a peptide spanning residues 716 to 727; this epitope maps adjacent to the core 14B7 binding site and at a location not previously identified as being a target of neutralizing antibodies. Both MAbs neutralized LT in vitro and partially protected mice against LT challenge. These results advance our fundamental understanding of the mechanisms by which antibodies neutralize anthrax toxin and may have future application in the evaluation of candidate rPA vaccines.


Cell lines.

The J774A.1 murine macrophagelike cell line was obtained from the American Type Culture Collection. Cells were cultured in Dulbecco modified Eagle medium 1× high glucose with GlutaMAX (Invitrogen Life Technologies, Carlsbad, CA) supplemented with 10% fetal bovine serum and penicillin-streptomycin at 37°C with 5% CO2 (18).

Recombinant anthrax toxin proteins.

Histidine-tagged, recombinant PA83 and PA32 were obtained from the Wadsworth Center Protein Expression Core (18). PA32 is a truncated form of PA83 containing only domains 3 and 4 and was previously shown to be properly folded and capable of binding to ATR (11). Recombinant LF (rLF) was obtained from BEI Resources (Manassas, VA; catalog no. NR-4367). Functional LT was produced by mixing rLF and rPA83 at a 1:1 (wt/wt) ratio in sterile phosphate-buffered saline (PBS). Toxin stocks were qualified by dose-response titrations, as previously described (18), and used at 2 50% tissue culture effective concentration(s) (TCEC50) for in vitro toxin neutralization assays. For in vivo studies, mice were challenged with 200 μg of LT by intraperitoneal injection, as described below. Recombinant proteins were stored at −20°C as single-use aliquots at a concentration of 1 mg/ml in PBS, except for the rLF obtained from BEI Resources, which was frozen at a concentration of 500 μg/ml in 1% bovine serum albumin. PA83 was biotinylated by using EZ-Link Sulfo-NHS-LC Biotin (Pierce Biotechnology, Inc., Rockford, IL).

Hybridoma production and screening.

Female BALB/c mice (Taconic Laboratories, Hudson, NY) ~8 weeks of age were immunized with PA83-MIS (50 μg) administered by intraperitoneal injection on days 0, 14, and 24 and boosted with a single injection of PA83 (100 μg) without adjuvant 3 days prior to B-cell fusion. PA83-MIS consists of rPA (PA83) covalently coupled to a nontoxic muramyl dipeptide derivative, as described previously (18). The mice were sacrificed 3 days after the final boost, and total splenocytes were harvested at the Wadsworth Center Immunology Core Laboratory for MAb production using a ClonaCell-HY hybridoma cloning kit (Stem Cell Technologies, Vancouver, British Columbia, Canada).

Hybridoma supernatants were screened by ELISA for reactivity against PA83 and PA32. Nunc Maxisorb 96-well microtiter plates were coated with PA83 or PA32 (0.1 μg/well) overnight at 4°C. Prior to the addition of hybridoma supernatant (0.1 ml/well), the plates were blocked with Blocker Casein (Pierce Biotechnology, Inc.) and washed three times with PBS containing 0.05% Tween 20. The plates were probed with horseradish peroxidase (HRP)-labeled goat anti-mouse immunoglobulin G (IgG) (Invitrogen Life Technologies) and developed with SureBlue TMB 1-Component microwell peroxidase substrate (KPL, Inc., Gaithersburg, MD). Microtiter plates were analyzed by using a SpectraMax 250 microtiter plate reader (Molecular Devices, Union City, CA). Antibodies that reacted with PA32 were considered to be specific for domains 3 and 4 and were further evaluated for LT-neutralizing activity (see below) and for reactivity with a PA peptide array library (see below).

PA peptide array ELISAs were performed essentially as described above. A collection of 31 overlapping peptides, 12 amino acids (aa) in length and each overlapping by 4 aa, spanning PA83 (GenBank accession no. AAA22637) were custom synthesized by New England Peptide (Gardner, MA). Lyophilized peptides were solubilized in dimethyl sulfoxide (DMSO) at 10 mg/ml and then diluted to 1:5 (vol/vol) into PBS and frozen at −20°C as single-use aliquots. The sequences of the individual domains 3 and 4 peptides are shown in Table Table1.1. For ELISAs, Nunc Maxisorb 96-well microtiter plates were coated with peptides overnight at 4°C, washed with PBS, and then blocked with Blocker Casein before being treated with hybridoma supernatants or human AVA serum diluted 1:100 into PBS. As negative controls, wells were coated with peptides spanning the signal sequence region of PA or amine-only derivatives only.

Domain 3 and 4 peptide sequences

Selected hybridomas were produced in ascites by the Wadsworth Center Immunology Core Laboratory, and MAbs were purified by standard endotoxin-free protein A purification by the Wadsworth Center Protein Expression Core. The subclass of each MAb was determined by ELISA. Nunc Maxisorb plates were coated with PA83 and then reacted with each MAb for 1 h. The plates were then washed and incubated with HRP-conjugated goat anti-mouse immunoglobulin specific (IgG1, IgG2a, IgG2b, and IgG3) antibodies (Southern Biotechnology Associates, Birmingham, AL).

MAb affinity determination and competitive binding assays by SPR.

The affinities of individual MAbs for PA83 were determined by surface plasmon resonance (SPR) by using a Biacore 3000 instrument (GE Healthcare, Troy, NY). PA83 was immobilized on a CM5 chip using NHS/EDC amine coupling and with a ligand density of ~4,000 response units (RU). Chips were equilibrated for 3 min in running buffer (HBS-EP; 0.01 M HEPES, 0.15 M NaCl, 3 mM EDTA, 0.005% surfactant P20) prior to injection of individual MAbs at a constant flow rate of 30 μl/min. Dissociation was monitored for 10 min. Association (Ka) and dissociation (Kd) rate constants were calculated as previously described (33). For competitive binding studies, individual MAbs (50 μg/ml) were injected four to five times (10 μl/injection) into the microfluidics cartridge to saturate ligand bound to the sensor surface. Subsequently, individual MAbs (50 μg/ml) were injected (20 μl/injection) in series. All MAbs were diluted in HBS-EP (pH 7.4).

In vitro toxin neutralization assay.

PA83 reactive antibodies were evaluated for their ability to neutralize LT in vitro, with an MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide]-based cytotoxicity assay, as described previously (18). Neutralization activity was defined as amount of MAb (μg/ml) required to protect 50% of the cells in culture from the effects of LT, a value commonly referred to as the TCEC50. J774A.1 cells were seeded in 96-well plates (~6 × 104 cells/well) and incubated until they achieved 70 to 80% confluence. LT (60 ng; ~2 TCEC50) was mixed with various dilutions of MAb and then incubated for 4 h at 37°C before being applied to the J774.1 cells. Cell viability was determined with a cell growth determination kit (Sigma-Aldrich). As controls, cells were treated with PBS (cell viability control) or LT (cell death control). The percent protection was determined as follows: {1 − [(PBS − ΜΑb)/(PBS − LT)]} × 100.

Receptor CMG-2 solid-phase binding assay.

Nunc Maxisorb plates (96 well) were coated overnight with soluble recombinant CMG-2 receptor (1 μg/well in PBS containing 2 mM MgCl2+), kindly provided by Robert Liddington (Burnham College). Biotinylated PA83 (1 μg) was mixed with MAb supernatants for 1 h at room temperature and then applied to CMG-2-coated wells. The plates were incubated for 1 h at room temperature, washed to remove unbound biotinylated PA83, and then incubated with avidin-HRP, as described above for ELISAs.

Mouse passive protection studies.

For in vivo passive protection studies, female BALB/c mice (Taconic Laboratories), 6 to 8 weeks of age, were challenged by intraperitoneal injection with LT (100 μg in 200 μl of PBS) administered to the lower left quadrant of the abdomen. Approximately 5 min later, the animals were injected in the lower right quadrant with saline only (controls) or with MAbs diluted into saline. After toxin challenge, mice were scored for signs of distress and illness (e.g., muscle weakness, difficulty breathing, and reduced responsiveness to stimuli) every 2 h for the first 24 h and thereafter every 8 h. Mice were euthanized when they reached a score of 9 on a behavioral scale, with 9 representing moribund (adapted from Swearengen [42]). Death was used as an endpoint in the present study, and all mice were monitored for a total of 15 days. All animals used in the present study were housed under conventional, specific-pathogen-free conditions and were treated in strict compliance with the guidelines established by the National Institutes of Health and the Institutional Animal Care and Use Committee at the Wadsworth Center.


More than two decades ago, Little et al. (23) proposed that at least 20 distinct antigenic regions exist on PA. Since that original report, only a few of these antigenic regions have been characterized in any detail. Identification of the exact epitopes on PA that are recognized by neutralizing MAbs has proven challenging since several of the best-characterized MAbs (e.g., 14B7) are directed against conformation-dependent epitopes. However, there is evidence that linear B-cell epitopes on PA, specifically within domain 4, are the target of neutralizing antibodies (1). Identification of these epitopes is critical to our understanding of the key residues on PA that are involved in protective immunity to anthrax toxin and may have important implications for the evaluation of the efficacies of future anthrax vaccines.

Identification of MAbs that recognize linear epitopes on domain 4.

We produced a collection of B-cell hybridomas from the spleens of BALB/c mice immunized with an adjuvanted derivative of rPA (see Materials and Methods). Hybridoma supernatants were screened simultaneously by ELISA for reactivity with PA83 and PA32. PA32 is a recombinant protein consisting only of PA domains 3 and 4 (11). PA32 has been previously shown to be properly folded and capable of binding to ATRs on cell surfaces. From a total of 384 cloned hybridomas that were screened by this method, 39 (~10%) produced MAbs that reacted with PA83. Six of these recognized PA32, suggesting that they were specific for domain 3 or 4. These six MAbs were further subjected to a peptide array ELISA consisting of 12-aa peptides with 4-aa overlaps, spanning domains 3 and 4 of PA83 (Table (Table1).1). Two MAbs were specific for peptides from domain 3, whereas three MAbs were reactive with peptides from domain 4 (Fig. (Fig.1).1). One MAb did not react with the peptide array library and was not studied further. One domain 3-reactive MAb, 2-A7, and two domain 4-reactive MAbs, 1-F1 and 2-B12, were further evaluated.

FIG. 1.
Reactivity of 1-F1, 2-A7, and 2-B12 with peptides spanning domains 3 and 4 of PA. MAbs 2-A7 (A), 1-F1 (B), and 2-B12 (C) were applied to 96-well microtiter plates coated with overlapping 12-mer peptide library spanning domains 3 (A) and 4 (B and C) of ...

MAbs 1-F1 and 2-B12 recognize distinct epitopes within domain 4.

A number of key characteristics of the three MAbs chosen for further study are outlined in Table Table2.2. MAb 2-A7 was determined to be an IgG2a that recognized a peptide spanning residues 532 to 543 within domain 3 of PA83, an area not expected to contribute to the neutralization of the toxins (28). Biacore analysis revealed that 2-A7 recognizes rPA with a dissociation constant (Kd) of approximately 1.6 × 10−9 M. 1-F1 was determined to be an IgG1 that recognized a peptide spanning residues 692 to 703, a region of domain 4 known to be an important target of neutralizing antibodies (1, 23, 40). 1-F1 bound PA83 with an affinity similar to that of 2-A7. 2-B12, also an IgG1, recognized a peptide spanning residues 716 to 727. BIAcore analysis revealed that 2-B12 recognized rPA with a Kd of approximately 1.2 × 10−10 M.

Physical characteristics of MAbs specific for PA domains 3 and 4

We used PyMOL to model the locations of the putative linear epitopes recognized by MAbs 2-A7, 1-F1, and 2-B12 on the three-dimensional structure of PA83 in association with the cell surface receptor CMG-2 (Fig. (Fig.2A).2A). This modeling confirmed that all three putative epitopes mapped to the surface PA83 and were not buried within the tertiary structure of the protein. The epitopes recognized by 1-F1 (red) and 2-B12 (blue) were situated within close proximity to the interface between PA83 (yellow) and CMG-2 (purple). The putative epitope recognized by 1-F1 (residues 692 to 703) is located between the small and large loops of domain 4 and is nested within the region of PA83 that is recognized by the neutralizing MAb 14B7 (Fig. (Fig.2B).2B). The putative epitope recognized by 2-B12 (residues 716 to 727) is only partially located within the large loop (residues 704 to 723) and yet overlaps by 6 aa with the region of domain 4 recognized by 14B7 (Fig. (Fig.2B)2B) (21, 36). Based on this information, we predicted that 14B7 would competitively inhibit (completely or partially) 1-F1 and 2-B12 from binding to PA83.

FIG. 2.
Modeling the location of the epitopes on PA recognized by 2-A7, 1-F1, and 2-B12. (A) The location of the linear epitopes recognized by MAbs 1F1, 2A7, and 2B12, modeled on the PA83 structure by using PyMOL ( and PDB accession 1T6B ...

We initially used competitive ELISAs to test these predictions. However, these assays were confounded by the need to use secondary-enzyme-linked reagents whose activities potentially complicated interpretation of relative amounts of MAb binding (data not shown). Biacore analysis, on the other hand, provides a means to measure MAb-target interactions in real time without the need for secondary labels or detection reagents (2). Indeed, Marvaud et al. utilized this technology to determine whether murine MAbs bound to similar regions on Clostridium perfringens Iota toxin (25). BIAcore sensor chips were coated with PA83 and then exposed to MAbs individually or in series. When tested individually, MAbs 14B7, 1-F1, 2-B12, and 2-A7 were each capable of associating with PA83 immobilized on the sensor chips, as evidenced by measurable changes in SPR. The changes, expressed as RU, were as follows; 14B7, 434 RU; 1-F1, 254 RU; 2-B12, 220 RU; and 2-A7, 300 RU. When tested in series, 14B7 blocked the ability of 1-F1 and 2-B12 to bind to PA83, since there was no measurable increase in RU upon subsequent addition of these MAbs (Fig. (Fig.3).3). On the other hand, the binding of the domain 3-specific MAb, 2-A7, was not affected by 14B7, as indicated by an increase in RU of >300 (Fig. (Fig.3).3). Identical results were obtained when 1-F1 and 2-B12 were applied to the chip in reverse order (i.e., 14B7 followed by 2-B12 and then 1-F1). These data demonstrate that 1-F1 and 2-B12 (but not 2-A7) recognize epitopes within or adjacent to the region of PA83 bound by 14B7.

FIG. 3.
14B7 blocks 1-F1 and 2-B12 from associating with PA83. A representative sensorgram generated by BIAcore analysis demonstrates that 14B7 prevents 1-F1 and 2-B12 from binding to PA83. PA83 was immobilized on a CM5 chip by using NHS/EDC amine-coupling, as ...

MAbs 1-F1 and 2-B12 neutralize LT in vitro.

We used a well-established murine macrophage cell-based toxicity assay to determine the capacity of 1-F1, 2-B12, and 2-A7 to neutralize LT in vitro (18). In this assay, the TCEC50 for 1-F1 was ~0.002 μg/ml (Fig. (Fig.4).4). At concentrations of ≥0.007 μg/ml, 1-F1 completely protected J774 cells from the effects of LT. This is virtually identical to the protection profile observed for 14B7 (data not shown). The TCEC50 for 2-B12 was determined to be ~0.004 μg/ml. However, maximal protection conferred by 2-B12 never exceeded 90%, suggesting that it is slightly less effective than 14B7 and 1-F1 at neutralizing LT in vitro. 2-A7 was completely ineffective at neutralizing LT, confirming that domain 3 antibodies are generally not protective.

FIG. 4.
1-F1 and 2-B12, but not 2A7, neutralize LT in vitro. LT (60 ng; ~2 TCEC50) was mixed with serial twofold dilutions of 1-F1 (□), 2-B12 (○), and 2-A7 (•) at the indicated concentrations and then applied to J774.1 cells grown ...

Domain 4-specific MAbs disrupt PA binding to CMG-2.

Antibodies directed against domain 4 are proposed to neutralize LT and ET by preventing the association of PA83 with ATRs, such as CMG-2, on the surfaces of host cells (40, 43). 14B7, for example, inhibits PA83's binding to host cells by >95% (23). We used an in vitro binding assay to determine the effects of 1-F1 and 2-B12 on the association of soluble rPA with plate-bound CMG-2. Biotin-labeled rPA83 was incubated with individual MAbs (14B7, 1-F1, 2-B12, or 2-A7) and then applied to microtiter plates coated with CMG-2. 14B7 was used as a positive control for the studies, and all data were normalized to these values. 1-F1 was as effective as 14B7 in preventing the interaction between rPA83 and CMG-2. 2-B12, on the other hand, was only 73% as effective as 14B7 in disrupting rPA-receptor engagement (Fig. (Fig.5).5). The inability of 2-B12 to completely block PA binding to CMG-2 was not due to limiting amounts of MAb, as evidenced by the fact that a 20-fold dilution of 2-B12 was sufficient neutralize LT in vitro (Fig. (Fig.5).5). As expected, 2-A7, which is directed against domain 3, demonstrated little capacity to disrupt PA binding to CMG-2.

FIG. 5.
MAb-mediated disruption of PA binding to ATR. Biotinylated PA83 was mixed with indicated MAbs (100 μg each) for 1 h before being applied to wells of a 96-well microtiter plate coated with recombinant CMG-2. The plates were then probed with avidin-HRP ...

MAbs 1-F1 and 2-B12 protect mice against LT challenge.

We used a previously established LT challenge model (18, 41) to examine whether 1-F1 and 2-B12 are protective in vivo. In this model, BALB/c mice were administered LT (100 μg) by intraperitoneal injection and 5 min later were treated with individual MAbs (550 μg/animal) by the same route. Using this model, we have previously shown that control animals injected with LT plus saline die within 48 h, whereas animals treated with anti-PA83 polyclonal IgG survive beyond 15 days (18). Consistent with our previous findings, all five LT-challenged, saline-treated control animals died within 24 h (Fig. (Fig.6).6). Mice treated with 2-A7 died at a rate indistinguishable from that for the control animals, confirming, at least for the present, that MAbs against domain 3 do not afford any protection against LT in vivo. In contrast, 1-F1 and 2-B12 each conferred partial protection against LT. 1-F1 conferred 60% protection by 20 h and 40% protection by 80 h. 2-B12 conferred 80% protection by 20 h and 60% protection by 80 h. Although we were unable to establish the level of protection for 14B7 in this model due to limited reagent availability, we expect that individual administration of 14B7 should result in levels of protection similar to those reported here (8, 24). Our results confirm that 1-F1 and 2-B12 are effective in protecting mice against lethal LT challenge.

FIG. 6.
1-F1 and 2-B12 protect mice against LT. BALB/c mice (n = 3 to 5/group) were administered LT (100 μg) by intraperitoneal injection and 5 min later treated with individual MAbs (550 μg/animal) as described in Materials and Methods. ...

Reactivity of human anti-AVA serum with linear epitopes on domain 4.

We wanted to determine whether the linear B-cell epitopes on domain 4 of PA identified in the present study are in fact recognized by humans. To test this, serum from a single individual immunized repeatedly with AVA was subjected to domain 4 peptide array ELISA. This serum had been previously established as having high LT neutralizing activity in vitro (data not shown). The ELISA revealed that the serum antibodies reacted with five distinct peptides, although reactivity against only three (peptides 83, 91, and 92) of the five peptides was >5% above background (Fig. (Fig.7).7). The most reactive peptides, 91 and 92, span residues 692 to 711 (PNYKVNVYAVTENTIlNPS) of PA83, which defines the linear epitope recognized 1-F1 (Fig. (Fig.2).2). These data suggest the linear epitope defined by MAb 1-F1 is in fact a target of neutralizing antibodies in humans and that serum antibody reactivity with this peptide may possibly serve as an indicator of immune status of individuals immunized with AVA and or candidate rPA vaccines.

FIG. 7.
Human anti-AVA serum reacts with the linear epitope on PA recognized by 1-F1. Serum from a single AVA-vaccinated individual was tested for reactivity against the PA83 domain 4 peptide array, as described in Fig. Fig.1.1. The solid black bars represent ...


Domain 4 (residues 596 to 735) constitutes the region of PA involved in receptor recognition and attachment (39). The region from aa 679 to 693 is especially important in binding to the host cell surfaces because these residues form a small loop that directly interacts with TEM-8 and CMG-2. Although it has been appreciated for decades that domain 4 in general, and the small loop in particular, is the target of neutralizing antibodies (23), the specific B-cell epitopes within this region of PA that are responsible for eliciting protective immunity have not been fully defined. One exception is the conformation-dependent epitope recognized by 14B7 (24, 37, 40). We have now identified and characterized two additional neutralizing MAbs directed against domain 4. MAb 1-F1 binds to a linear epitope immediately adjacent to the region of domain 4 that is recognized by 14B7. 2-B12, on the other hand, binds a linear epitope within the large loop of domain 4, a region not previously known to be the target of neutralizing antibodies. Identification of these epitopes is significant because it furthers our understanding of the key residues on PA that are involved in protective immunity to anthrax toxin and may have important implications for the evaluation of the efficacies of future anthrax vaccines.

14B7 and 1-F1 are similar in a number of key respects, including IgG subclass, affinity for PA, and capacity to block PA-ATR interactions, as well as efficacy in LT neutralization. 1-F1 differs from 14B7 in that it recognizes a linear epitope (residues 692 to 703) adjacent to the small loop on domain 4 (residues 679 to 692), whereas 14B7 recognizes a conformation-dependent epitope, which has been difficult to map precisely. Early analysis indicated that 14B7 binds to a region encompassing residues 671 to 721 (24). Rosovitz et al. (37) significantly refined that model by revealing that the critical residues of domain 4 recognized by 14B7 overlap, but are distinct from, those that contact the ATRs. For example, mutagenesis of Asp683 reduced PA binding to cell surfaces by more than 1,000-fold but had little effect on the ability of 14B7 to bind to PA. Detailed computer modeling of 14B7 docked to PA (40), along with other recent studies, has further narrowed the putative epitope to ~14 aa (residues 680 to 692) (22, 24, 37, 40). In light of this information, it is interesting that the epitope recognized by 1-F1 lies just outside of the region of domain 4 thought to make direct contact with cell surfaces. Nonetheless, we propose that 1-F1, like 14B7, neutralizes LT by physically interfering with the ability of PA to associate with CMG-2. This conclusion is based our observation that 1-F1 was as effective as 14B7 in blocking the association of PA with CMG-2, in an in vitro binding assay. The identification of 1-F1 underscores the importance of this face of domain 4 as a target of neutralizing antibodies. 14C7 and 3B6, two other IgG1 MAbs that compete with 14B7 for binding to PA, were described by Little et al. (23). However, the exact epitopes recognized by the MAbs were never identified. Given domain 4's small size, it is likely that 14C7 and 3B6 bind to residues closely adjacent to, if not identical, to those recognized by 1-F1, 2-B12, or 14B7 (1, 8, 40).

The epitope recognized by 2-B12, as determined by peptide array reactivity, spans residues 716 to 727 and is situated within the large loop of domain 4 (residues 704 to 722) immediately adjacent to (and overlapping with) the region recognized by 14B7. Indeed, the proximity of the two epitopes was confirmed by competitive binding assays, in which 14B7 was able to completely block 2-B12's ability to associated with PA83. Although the large loop of domain 4 is not postulated to be involved in direct physical association with ATRs (43), the observation that 2-B12 partially inhibited (~70%) PA83 from binding to CMG-2 suggests that this MAb neutralizes LT by preventing toxin-receptor interactions. This mechanism is not implausible, if we consider that IgG's mass is almost 10 times greater than that of domain 4. It is unclear, however, whether steric hindrance alone accounts for 2-B12's capacity to neutralize LT in vitro and in vivo. In a mouse model, for example, 2-B12 was as effective as 1-F1 in preventing LT-induced death. 2-B12 is somewhat unusual, in that its affinity for PA is ~10-fold greater than the affinities measured for 14B7 and 1-F1. It is possible, therefore, that the greater affinity of 2-B12 for its target, compared to the affinities of 14B7 and 1-F1, compensates for the fact that 2-B12 binds to a region of PA outside the receptor recognition domain. Several groups have recently documented the importance of antibody affinity in the neutralization of anthrax toxins (1, 46).

In the present study, we also identified and characterized a domain 3-specific MAb, 2-A7. 2-A7 recognizes a linear epitope (residues 532 to 543) that lies outside the region of domain 3 that is involved in PA oligomerization and heptamer formation (28). 2-A7 had no discernible capacity to neutralize LT in vitro or in vivo. However, the affinity of 2-A7 for PA was comparable to that the affinities observed for 14B7 and 1-F1. These data underscore the importance of epitope specificity as a primary determinant of antibody-mediated immunity to anthrax toxins. At least in the challenge model used in the present study, Fc-mediated clearance of toxin-antibody complexes does not contribute substantially to protection. This fact may account in part for the repeated observations that anti-PA antibody titers, as determined by ELISA, often fail to correlate with immunity to anthrax. On the other hand, serum reactivity with specific antigenic determinants or linear epitopes, such as those defined by 1-F1 and 2B12, could be useful as predictors of protection. In this regard, it was very interesting that serum antibodies from an AVA-immunized individual reacted with linear peptides corresponding to the epitope on domain 4 recognized by the neutralizing MAb 1-F1. Peptide arrays are increasingly being applied to infectious diseases (31, 32) and will likely have important applications for biodefense vaccines as an additional means to assess neutralizing antibody titers in the sera of vaccinated individuals.

Considering that rPA will most certainly be a constituent of the next-generation anthrax vaccine (9, 14, 15, 15, 18), much remains to be done in terms of identifying the specific regions of PA that elicit neutralizing (and non-neutralizing) antibodies, and to better understand how these antibodies interfere (or not) with toxin attachment, oligomerization, and translocation (17, 36, 45). For example, in our screen to identify domain 4 MAbs, we identified 33 MAbs that reacted with domains 1 or 2. Preliminary characterization of those MAbs has indicated that roughly 60% of them either fail to neutralize LT or actually enhance LT-induced cell death; the remainder demonstrates various levels of protection in vitro (C. Kelly-Cirino, unpublished data). In our opinion, a complete and systematic identification of both conformation-dependent and conformation-independent epitopes is necessary if protective immunity to anthrax toxins is to be fully understood and if the efficacy of candidate human rPA vaccines is to be properly evaluated. Although beyond the scope of the present study, MAbs 1-F1 and 2-B12 could be further investigated as possible therapeutics for the treatment of anthrax intoxication.


We extend our special thanks to Stuart Balaban (Immunology Core, Wadsworth Center) for hybridoma and ascites production and to Jane Kasten-Jolly (Immunology Core, Wadsworth Center) for assistance with the Biacore analysis. We thank Stephen Leppla (National Institutes of Health) for kindly providing MAb 14B7, Robert Liddington (Burnham Institute for Medical Research) for the soluble recombinant CMG-2 receptor, and Freyja Lynn (National Institute of Allergy and Infectious Diseases) for assistance in obtaining rLF. We also acknowledge Karen Chave of the Northeast Biodefense Center's (NBC) Protein Core Laboratory for recombinant PA83 and PA32 and Frank Gelder (Virionyx Corp.) for providing the rPA83-MIS.

This study was supported in part by a New Opportunities Award from the NBC (U54-AI057158-Lipken) to N.J.M.

This manuscript is dedicated to the memory of Nick M. Cirino. He will be remembered for his passion for science, collaborative research and, most of all, his friendship.


Editor: J. B. Bliska


[down-pointing small open triangle]Published ahead of print on 24 August 2009.


1. Abboud, N., and A. Casadevall. 2008. Immunogenicity of Bacillus anthracis protective antigen domains and efficacy of elicited antibody responses depend on host genetic background. Clin. Vaccine Immunol. 15:1115-1123. [PMC free article] [PubMed]
2. Alfthan, K. 1998. Surface plasmon resonance biosensors as a tool in antibody engineering. Biosensors Bioelectronics 13:653-663. [PubMed]
3. Anonymous. 2002. The anthrax vaccine: is it safe? Does it work? Institute of Medicine, National Academy Press, Washington, DC.
4. Benson, E. L., P. D. Huynh, A. Finkelstein, and R. J. Collier. 1998. Identification of residues lining the anthrax protective antigen channel. Biochemistry 37:3941-3948. [PubMed]
5. Bradley, K. A., J. Mogridge, M. Mourez, R. J. Collier, and J. A. Young. 2001. Identification of the cellular receptor for anthrax toxin. Nature 414:225-229. [PubMed]
6. Bradley, K. A., and J. A. Young. 2003. Anthrax toxin receptor proteins. Biochem. Pharmacol. 65:309-314. [PubMed]
7. Brey, R. N. 2005. Molecular basis for improved anthrax vaccines. Adv. Drug Deliv. Rev. 57:1266-1292. [PubMed]
8. Brossier, F., M. Levy, A. Landier, P. Lafaye, and M. Mock. 2004. Functional analysis of Bacillus anthracis protective antigen by using neutralizing monoclonal antibodies. Infect. Immun. 72:6313-6317. [PMC free article] [PubMed]
9. Campbell, J. D., K. H. Clement, S. S. Wasserman, S. Donegan, L. Chrisley, and K. L. Kotloff. 2007. Safety, reactogenicity, and immunogenicity of a recombinant protective antigen anthrax vaccine given to healthy adults. Hum. Vaccines 3:205-211. [PubMed]
10. Chen, Z., M. Moayeri, Y. H. Zhou, S. Leppla, S. Emerson, A. Sebrell, F. Yu, J. Svitel, P. Schuck, M. St. Clair, and R. Purcell. 2006. Efficient neutralization of anthrax toxin by chimpanzee monoclonal antibodies against protective antigen. J. Infect. Dis. 193:625-633. [PubMed]
11. Cirino, N. M., D. Sblattero, D. Allen, S. R. Peterson, J. D. Marks, P. J. Jackson, A. Bradbury, and B. E. Lehnert. 1999. Disruption of anthrax toxin binding with the use of human antibodies and competitive inhibitors. Infect. Immun. 67:2957-2963. [PMC free article] [PubMed]
12. Collier, R. J., and J. A. Young. 2003. Anthrax toxin. Annu. Rev. Cell Dev. Biol. 19:45-70. [PubMed]
13. Duesbery, N. S., C. P. Webb, S. H. Leppla, V. M. Gordon, K. R. Klimpel, T. D. Copeland, N. G. Ahn, M. K. Oskarsson, K. Fukasawa, K. D. Paull, and G. F. Vande Woude. 1998. Proteolytic inactivation of MAP-kinase-kinase by anthrax lethal factor. Science 280:734-737. [PubMed]
14. Fasanella, A., F. Tonello, G. Garofolo, L. Muraro, A. Carattoli, R. Adone, and C. Montecucco. 2008. Protective activity and immunogenicity of two recombinant anthrax vaccines for veterinary use. Vaccine 26:5684-5688. [PubMed]
15. Gauthier, Y. P., J.-N. Tournier, J.-C. Paucod, J.-P. Corre, M. Mock, P. L. Goossens, and D. R. Vidal. 2008. Efficacy of a vaccine based on protective antigen and killed spores against experimental inhalational anthrax. Infect. Immun. 77:1197-1207. [PMC free article] [PubMed]
16. Grabenstein, J. D. 2008. Vaccines: countering anthrax: vaccines and immunoglobulins. Clin. Infect. Dis. 46:129-136. [PubMed]
17. Gubbins, M. J., J. D. Berry, C. R. Corbett, J. Mogridge, X. Y. Yuan, L. Schmidt, B. Nicolas, A. Kabani, and R. S. Tsang. 2006. Production and characterization of neutralizing monoclonal antibodies that recognize an epitope in domain 2 of Bacillus anthracis protective antigen. FEMS Immunol. Med. Microbiol. 47:436-443. [PubMed]
18. Kelly, C. D., C. O'Loughlin, F. B. Gelder, J. W. Peterson, L. E. Sower, and N. M. Cirino. 2007. Rapid generation of an anthrax immunotherapeutic from goats using a novel non-toxic muramyl dipeptide adjuvant. J. Immune Based Ther. Vaccines 5:11. [PMC free article] [PubMed]
19. Lacy, D. B., D. J. Wigelsworth, H. M. Scobie, J. A. Young, and R. J. Collier. 2004. Crystal structure of the von Willebrand factor A domain of human capillary morphogenesis protein 2: an anthrax toxin receptor. Proc. Natl. Acad. Sci. USA 101:6367-6372. [PubMed]
20. Leppla, S. H. 1982. Anthrax toxin edema factor: a bacterial adenylate cyclase that increases cyclic AMP concentrations of eukaryotic cells. Proc. Natl. Acad. Sci. USA 79:3162-3166. [PubMed]
21. Leppla, S. H., J. B. Robbins, R. Schneerson, and J. Shiloach. 2002. Development of an improved vaccine for anthrax. J. Clin. Investig. 110:141-144. [PMC free article] [PubMed]
22. Leysath, C. E., A. F. Monzingo, J. A. Maynard, J. Barnett, G. Georgiou, B. L. Iverson, and J. D. Robertus. 2009. Crystal structure of the engineered neutralizing antibody M18 complexed to domain 4 of the anthrax protective antigen. J. Mol. Biol. 387:680-693. [PMC free article] [PubMed]
23. Little, S. F., S. H. Leppla, and E. Cora. 1988. Production and characterization of monoclonal antibodies to the protective antigen component of Bacillus anthracis toxin. Infect. Immun. 56:1807-1813. [PMC free article] [PubMed]
24. Little, S. F., J. M. Novak, J. R. Lowe, S. H. Leppla, Y. Singh, K. R. Klimpel, B. C. Lidgerding, and A. M. Friedlander. 1996. Characterization of lethal factor binding and cell receptor binding domains of protective antigen of Bacillus anthracis using monoclonal antibodies. Microbiology 142(Pt. 3):707-715. [PubMed]
25. Marvaud, J. C., T. Smith, M. L. Hale, M. R. Popoff, L. A. Smith, and B. G. Stiles. 2001. Clostridium perfringens Iota toxin: mapping of receptor binding and Ia docking domains on Ib. Infect. Immun. 69:2435-2441. [PMC free article] [PubMed]
26. Maynard, J. A., C. B. Maassen, S. H. Leppla, K. Brasky, J. L. Patterson, B. L. Iverson, and G. Georgiou. 2002. Protection against anthrax toxin by recombinant antibody fragments correlates with antigen affinity. Nat. Biotechnol. 20:597-601. [PubMed]
27. Miller, C. J., J. L. Elliott, and R. J. Collier. 1999. Anthrax protective antigen: prepore-to-pore conversion. Biochemistry 38:10432-10441. [PubMed]
28. Mogridge, J., M. Mourez, and R. J. Collier. 2001. Involvement of domain 3 in oligomerization by the protective antigen moiety of anthrax toxin. J. Bacteriol. 183:2111-2116. [PMC free article] [PubMed]
29. Mohamed, N., J. Li, C. S. Ferreira, S. F. Little, A. M. Friedlander, G. L. Spitalny, and L. S. Casey. 2004. Enhancement of anthrax lethal toxin cytotoxicity: a subset of monoclonal antibodies against protective antigen increases lethal toxin-mediated killing of murine macrophages. Infect. Immun. 72:3276-3283. [PMC free article] [PubMed]
30. Molloy, S. S., P. A. Bresnahan, S. H. Leppla, K. R. Klimpel, and G. Thomas. 1992. Human furin is a calcium-dependent serine endoprotease that recognizes the sequence Arg-X-X-Arg and efficiently cleaves anthrax toxin protective antigen. J. Biol. Chem. 267:16396-16402. [PubMed]
31. Neuman de Vegvar, H. E., and W. H. Robinson. 2004. Microarray profiling of antiviral antibodies for the development of diagnostics, vaccines, and therapeutics. Clin. Immunol. 111:196-201. [PubMed]
32. Nguyen, M. L., S. R. Crowe, S. Kurella, S. Teryzan, B. Cao, J. D. Ballard, J. A. James, and A. D. Farris. 2009. Sequential B-cell epitopes of Bacillus anthracis lethal factor bind lethal toxin-neutralizing antibodies. Infect. Immun. 77:162-169. [PMC free article] [PubMed]
33. Nowakowski, A., C. Wang, D. B. Powers, P. Amersdorfer, T. J. Smith, V. A. Montgomery, R. Sheridan, R. Blake, L. A. Smith, and J. D. Marks. 2002. Potent neutralization of botulinum neurotoxin by recombinant oligoclonal antibody. Proc. Natl. Acad. Sci. USA 99:11346-11350. [PubMed]
34. Petosa, C., R. J. Collier, K. R. Klimpel, S. H. Leppla, and R. C. Liddington. 1997. Crystal structure of the anthrax toxin protective antigen. Nature 385:833-838. [PubMed]
35. Reuveny, S., M. D. White, Y. Y. Adar, Y. Kafri, Z. Altboum, Y. Gozes, D. Kobiler, A. Shafferman, and B. Velan. 2001. Search for correlates of protective immunity conferred by anthrax vaccine. Infect. Immun. 69:2888-2893. [PMC free article] [PubMed]
36. Rivera, J., A. Nakouzi, N. Abboud, E. Revskaya, D. Goldman, R. J. Collier, E. Dadachova, and A. Casadevall. 2006. A monoclonal antibody to Bacillus anthracis protective antigen defines a neutralizing epitope in domain 1. Infect. Immun. 74:4149-4156. [PMC free article] [PubMed]
37. Rosovitz, M. J., P. Schuck, M. Varughese, A. P. Chopra, V. Mehra, Y. Singh, L. M. McGinnis, and S. H. Leppla. 2003. Alanine-scanning mutations in domain 4 of anthrax toxin protective antigen reveal residues important for binding to the cellular receptor and to a neutralizing monoclonal antibody. J. Biol. Chem. 278:30936-30944. [PubMed]
38. Santelli, E., L. A. Bankston, S. H. Leppla, and R. C. Liddington. 2004. Crystal structure of a complex between anthrax toxin and its host cell receptor. Nature 430:905-908. [PubMed]
39. Singh, Y., K. R. Klimpel, C. P. Quinn, V. K. Chaudhary, and S. H. Leppla. 1991. The carboxyl-terminal end of protective antigen is required for receptor binding and anthrax toxin activity. J. Biol. Chem. 266:15493-15497. [PubMed]
40. Sivasubramanian, A., J. A. Maynard, and J. J. Gray. 2008. Modeling the structure of MAb 14B7 bound to the anthrax protective antigen. Proteins 70:218-230. [PubMed]
41. Staats, H. F., S. M. Alam, R. M. Scearce, S. M. Kirwan, J. X. Zhang, W. M. Gwinn, and B. F. Haynes. 2007. In vitro and in vivo characterization of anthrax anti-protective antigen and anti-lethal factor monoclonal antibodies after passive transfer in a mouse lethal toxin challenge model to define correlates of immunity. Infect. Immun. 75:5443-5452. [PMC free article] [PubMed]
42. Swearengen, J. R. 2005. Biodefense: research methodologies and animal models. CRC Press, Inc., Boca Raton, FL.
43. Varughese, M., A. V. Teixeira, S. Liu, and S. H. Leppla. 1999. Identification of a receptor-binding region within domain 4 of the protective antigen component of anthrax toxin. Infect. Immun. 67:1860-1865. [PMC free article] [PubMed]
44. Zarebski, L. M., K. Vaughan, J. Sidney, B. Peters, H. Grey, K. D. Janda, A. Casadevall, and A. Sette. 2008. Analysis of epitope information related to Bacillus anthracis and Clostridium botulinum. Expert Rev. Vaccines 7:55-74. [PMC free article] [PubMed]
45. Zhang, J., J. Xu, G. Li, D. Dong, X. Song, Q. Guo, J. Zhao, L. Fu, and W. Chen. 2006. The 2β2-2β3 loop of anthrax protective antigen contains a dominant neutralizing epitope. Biochem. Biophys. Res. Commun. 341:1164-1171. [PubMed]
46. Zhou, B., C. Carney, and K. D. Janda. 2008. Selection and characterization of human antibodies neutralizing Bacillus anthracis toxin. Bioorg. Med. Chem. 16:1903-1913. [PMC free article] [PubMed]

Articles from Infection and Immunity are provided here courtesy of American Society for Microbiology (ASM)