Sub-Domains of Ricin’s B Subunit as Targets of Toxin Neutralizing and Non-Neutralizing Monoclonal Antibodies

The B subunit (RTB) of ricin toxin is a galactose (Gal)−/N-acetylgalactosamine (GalNac)-specific lectin that mediates attachment, entry, and intracellular trafficking of ricin in host cells. Structurally, RTB consists of two globular domains with identical folding topologies. Domains 1 and 2 are each comprised of three homologous sub-domains (α, β, γ) that likely arose by gene duplication from a primordial carbohydrate recognition domain (CRD), although only sub-domains 1α and 2γ retain functional lectin activity. As part of our ongoing effort to generate a comprehensive B cell epitope map of ricin, we report the characterization of three new RTB-specific monoclonal antibodies (mAbs). All three mAbs, JB4, B/J F9 and C/M A2, were initially identified based on their abilities to neutralize ricin in a Vero cell cytotoxicty assay and to partially (or completely) block ricin attachment to cell surfaces. However, only JB4 proved capable of neutralizing ricin in a macrophage apoptosis assay and in imparting passive immunity to mice in a model of systemic intoxication. Using a combination of techniques, including competitive ELISAs, pepscan analysis, differential reactivity by Western blot, as well as affinity enrichment of phage displayed peptides, we tentatively localized the epitopes recognized by the non-neutralizing mAbs B/J F9 and C/M A2 to sub-domains 2α and 2β, respectively. Furthermore, we propose that the epitope recognized by JB4 is within sub-domain 2γ, adjacent to RTB’s high affinity Gal/GalNAc CRD. These data suggest that recognition of RTB’s sub-domains 1α and 2γ are critical determinants of antibody neutralizing activity and protective immunity to ricin.


Introduction
Ricin, a natural product of the castor bean plant (Ricinus communis), is one of the most lethal protein toxins known [1,2]. In its mature form, ricin consists of two distinct subunits, RTA and RTB, joined by a single disulfide bond. RTA (32 kDa) is an RNA N-glycosidase that irreversibly inactivates eukaryotic ribosomes through hydrolytic cleavage of a conserved adenosine residue within in the sarcin-ricin loop (SRL) of the 28S rRNA [3,4]. RTB (34 kDa) is a galactose-and N-acetylgalactosamine (Gal/GalNac)specific lectin that mediates attachment, endocytosis, and trafficking of RTA from the plasma membrane to the trans-Golgi network (TGN) and then the endoplasmic reticulum (ER) [5]. Once in the ER, RTA is transported via a process known as retrotranslocation, across the ER membrane and into the cytoplasm where it refolds into its enzymatically active conformation and initiates ribosome depurination [6]. Ricin's potency is due in large part to RTB's ability to adhere to and be internalized by virtually all mammalian cell types [7].
In an effort to identify the regions (or sub-domains) of RTB that are important in eliciting protective immunity to ricin, we recently produced and characterized a collection of two neutralizing and four non-neutralizing RTB-specific murine monoclonal antibodies (mAbs) (Fig. 1) [25,32]. The epitopes recognized by the four nonneutralizing mAbs (JB11, CB12, SA3, and TFTB-1) were identified by pepscan analysis [32]. Three bound within RTB's sub-domain 1b, while the fourth bound within sub-domain 2a (Fig. 1). The epitope recognized by 24B11, one of the two neutralizing mAbs we characterized, was tentatively identified through the use of a phage-displayed peptide library, as being adjacent to the Gal-specific CRD in sub-domain 1a (Fig. 1) [25]. The epitope recognized by SylH3, the other neutralizing mAb we characterized, has not been definitively mapped, although a preponderance of evidence would suggest that it is adjacent to the Gal/GalNac-specific CRD in sub-domain 2c [32]. Other RTBspecific mAbs with potent toxin neutralizing activity have been reported recently, although the location of their respective epitopes was not disclosed [30].
One of the objectives of our research program is to generate a complete B cell epitope map of RTB, and to then use this information in the design of an RTB-based subunit vaccine that could be used by public health and military sectors as a countermeasure against the threat of ricin as a biological weapon [1,33,34]. The partial B cell epitope map we previously constructed using a limited collection of mAbs ( Fig. 1) suggests that toxin-neutralizing mAbs bind exclusively to epitopes within sub-domains 1a and 2c, whereas non-neutralizing mAbs cluster within the interior sub-domains (1b, 1c, 2a, 2b). However, such a conclusion is premature without additional epitope assignments, particularly within sub-domains 1c and 2b. Towards this end, we have screened an additional collection of ricin-specific B cell hybridomas and now describe the characterization of two new non-neutralizing and one new neutralizing RTB-specific mAbs. One of the two non-neutralizing mAbs (C/M A2) recognizes an epitope within sub-domain 2b, while the other (B/J F9) binds an epitope tentatively situated within 2a. The highly potent ricin-neutralizing mAb, JB4, recognizes an epitope within sub-domain 2c that overlaps (or is identical to) SylH3's binding site. These data significantly refine the current B cell epitope map of RTB and further support our hypothesis that neutralizing B cell epitopes are restricted to sub-domains 1a and 2c, an observation that is of considerable importance in vaccine design.

Chemicals, Biological Reagents and Cell Lines
Ricin toxin (Ricinus communis agglutinin II), Ricinus communis agglutinin I (RCA-I), ricin toxin A subunit (RTA), and ricin toxin B subunit (RTB) were purchased from Vector Laboratories (Burlingame, CA). Ricin was dialyzed against phosphate buffered saline (PBS) at 4uC in 10,000 MW cutoff Slide-A-Lyzer dialysis cassettes (Pierce, Rockford, IL), prior to use in cytotoxicity studies. GlutaMax TM , fetal calf serum and goat serum were purchased from Gibco-Invitrogen (Carlsbad, CA). Ph. D. TM -12 phage display peptide library kit was purchased from New England BioLabs (Beverly, MA). A ClonaCell HY TM kit for hybridoma production was purchased from STEMCELL Technologies (Vancouver, BC, Canada). Unless noted otherwise, all other chemicals were obtained from Sigma-Aldrich (St. Louis, MO). Vero, THP-1, and the murine myeloma cell line P3X63.Ag8.653 were purchased from the American Type Culture Collection (Manassas, VA). Cell culture media were prepared by the Wadsworth Center Media Services facility. Unless otherwise noted, all cell lines and hybridomas were maintained in a humidified incubator at 37uC with 5% CO 2 .  [46]. Depicted are RTA (grey), RTB (black), ricin's N-linked mannose side chains (yellow sticks), and lactose moieties (white sticks) situated with CRD 1a and 2c. Specific known (and tentative) B cell epitopes recognized by RTB-specific mAbs are color-coded. (Lower panel). Linear depiction of RTB's subdomain organization, including the leader/linker sequences (1l, 2l). The location of epitopes recognized by neutralizing and non-neutralizing mAbs are indicated above the figure as triangles and squares, respectively. The color of the shapes corresponds to the mAbs listed below. The asterisks indicate the putative location of mAb epitopes. See Table 1 for a description of specific characteristics of each mAb. doi:10.1371/journal.pone.0044317.g001

Mouse Strains, Animal Care and Immunizations
Female BALB/c mice approximately 8-10 weeks of age were purchased from Taconic Labs (Hudson, NY). Animals were housed under conventional, specific pathogen-free conditions and were treated in compliance with the Wadsworth Center's Institutional Animal Care and Use Committee (IACUC) guidelines. For hybridoma production, female BALB/c mice were primed i.p. with ricin toxoid (RT, 50 mg or 2 uM per mouse in 0.4 ml PBS) on day 0, and then boosted by the same route with RT (50 mg) on days 10 and 20. RT was produced as described previously [35].

B-cell Hybridoma Production
Four days after the second boost with RT (50 mg), mice were euthanized, and total splenocytes were fused with the myeloma cell line P3X63.Ag8.653, using polyethylene glycol (PEG) as described previously [27]. The resulting hybridomas were seeded in methylcellulose and cloned as per the instructions in the ClonaCell -HY TM hybridoma cloning manual (STEMCELL Technologies, Vancouver, BC, Canada). Hybridomas secreting antibodies of interest were expanded and cultured in either RPMI medium containing 10% fetal calf serum, oxaloacetate, pyruvate, and insulin (OPI), 8 mM GlutaMax TM , and penicillin-streptomycin, or in medium A (STEMCELL Technologies) before being transitioned to CD Hybridoma, a serum-free, protein-free, antibioticfree medium (Gibco-Invitrogen, Carlsbad, CA).

Vero Cell Cytotoxicity Assays
Vero cell cytotoxicity assays were performed as previously described [25,26]. Briefly, Vero cells were trypsinized, adjusted to approximately 5610 4 cells per ml, and seeded (100 ml/well) into white 96-well plates (Corning Life Sciences, Corning, NY), and allowed to adhere overnight. Vero cells were then treated with ricin (0.01 mg/ml; 154 pM), ricin:mAb mixtures, or medium alone (negative control) for 2 hr at 37uC. The cells were washed to remove non-internalized toxin or toxin:mAb mixtures, and were then incubated for 48 hr. Cell viability was assessed using CellTiter-GLO reagent (Promega, Madison, WI). All treatments were performed in triplicate, and 100% viability was defined as the average value obtained from wells in which cells were treated with medium only.

Passive Protection Studies
Individual mAbs (60 mg) were diluted into endotoxin-free PBS and then administered in a final volume of 0.2 ml to female BALB/c mice (ages 8-10 weeks) by i.p. injection. Twenty-four hours later, the mice were injected with ricin (2 mg; 100 mg/kg ) by the i.p. route, which is roughly equivalent to 10 LD 50 s [22,36,37].
Survival was monitored over a 2-6 day period. In addition, hypoglycemia was used as a surrogate marker of intoxication [26,38]. Blood (,5 ml) was collected from the tail vein of the animals at 18-24 hr intervals. Blood glucose levels were measured with an Avia ACCU-CHEK handheld blood glucose meter (Roche, Indianapolis, IN). Mice were euthanized when they became overtly moribund and/or blood glucose levels fell below 25 mg/dl. For statistical purposes, readings at or below the meter's limit of detection of ,12 mg/dl were set to that value.

Antibody Affinity Measurements and Competition Analysis
Affinity of antibodies for ricin toxin was determined by surface plasmon resonance (SPR) using a Biacore 3000 (GE Healthcare) instrument. Ricin was attached to a CM5 chip at a density of 550 to 650 RU. HEPES-buffered saline with EDTA and surfactant P20 (HBS-EP; 10 mM HEPES, pH 7.4, 150 mM NaCl, 3.4 mM EDTA, 0.005% of the surfactant P20 from GE Healthcare) was employed as the running buffer at a flow-rate of 30 ml/min. Serial dilutions of each antibody were made in HBS-EP, pH 7.4 from 600 nM to 18.75 nM, with each concentration series having at least one cycle of a buffer alone injection. Injection times were 3-4 minutes with dissociation times of 10 minutes. Regeneration of the chip surface was performed at a flow-rate of 50 ml/min by two 30 s pulses of 10 mM glycine, pH 1.5. The regeneration was followed by a 2 min stabilization period. All kinetic experiments were run a 25uC. Kinetic constants were obtained by analysis using the BIA evaluation software.
Antibody competitive binding assays by Biacore were performed with HBS-EP, pH 7.4 as the running buffer at a flow-rate of 10 ml/min. The first mAb was injected until saturation was achieved (i.e., when no significant additional rise in resonance units (RU) was observed after antibody injection.) The second competing mAb was then injected using a 2-min injection time. The amount of second mAb bound to the chip, in RU, was calculated as the RU value at 15 s after the injection minus the RU value at 15 s preceding the start of the injection. The chip surface was regenerated by short pulses with 10 mM glycine, pH 1.5, until the RU values had returned to baseline.

Ricin Apoptosis Assays
THP-1 human monocytes (5610 5 cells/200 ml) were subjected to ricin (2.5 mg/ml : 38 nM) in the presence or absence of anti-RTB mAbs (20 mg/ml : 133 nM) for 5 h in an incubator in 96well Microtest TM U-bottom tissue culture treated plates (BD) at 37uC and 5% CO 2 . After incubation, the cells were collected by centrifugation and then re-suspended in 1x binding buffer (200 ml), 100 ml of which was stained with 5 ml of Annexin V-FITC, as recommended by the manufacturer (Annexin V-FITC Apoptosis Detection Kit II from BD Pharmingen, cat# 556570). Samples were assayed for early apoptosis using a FACS Calibur (BD Biosciences). Results were reported as % cells positive for Annexin V-FITC. A minimum of 10,000 cells were analyzed per sample.

Ricin Binding Assays
To determine if mAbs prevent ricin binding to asialofetuin (ASF), Nunc Maxisorb F96 microtiter plates (ThermoFisher Scientific) were coated with ASF (0.4 mg/well) (EY Laboratories, San Mateo, CA) in PBS (pH 7.4) for 18 hr at 4uC. Plates were washed with PBS containing Tween-20 (PBS-T; 0.05% v/v), blocked with 2% goat serum in PBS-T (0.05% v/v) and then overlayed with biotinylated ricin (50 ng/ml : 770 pM) and IgG mAbs (20 mg/ml : 133 nM) for 1 hr. The plates were washed to remove unbound toxin, labeled with avidin-HRP (0.4 mg/ml) and developed using TMB, as described above for ELISAs. To determine if mAbs bind the galactose binding pockets of ricin, plates were coated with ASF for 18 hr at 4uC. Plates were washed with PBS-T (0.05% v/v), blocked with 2% goat serum in PBS-T (0.05% v/v), overlaid with ricin (10 mg/ml) for 1 hr, then with mAbs (10 mg/ml) for 1 hr. Plates were then labeled with IgG-HRP and developed TMB, as described above for ELISAs.
Epitope Mapping by Phage Displayed Peptide Library C/M A2 and irrelevant IgG isotype control antibody MOPC21 were immobilized onto Falcon Polystyrene Tissue Culture Dishes at 4uC overnight (10 mg of antibody in 1.5 ml PBS : 44 nM), and then blocked with 2% PBS-BSA at 4uC for 2 hr. Phage library was diluted in 1 ml PBS to a concentration of 1.5610 11 pfu/ml and added to the MOPC 21 plate for 1 hour. The supernatant from the plate was transferred directly to the C/M A2 plate and allowed to bind for 1 hr. The plates were then washed with PBS-T (0.1% v/v), and a 1 ml solution of RTB (100 mg/ml : 3 mM) was added for 1 hour to elute phages from the variable site of the antibodies. The supernatant was collected and amplified in ER2738 Escherichia coli for 4.5 hours. The bacteria was spun down, and phage in the supernatant was precipitated by adding a 1:5 volume of 20% w/v PEG8000, 2.5 M NaCl solution and incubating at 4uC overnight. The next morning, the precipitated phages were spun down, the supernatant removed, and 1 ml of PBS was added to re-dissolve the phages. Phage titer was then determined by serial dilution and plating onto LB/Agar/IPTG/X-gal plates. This amplified phage stock from Round 1 was then diluted down to the same concentration as the starting library concentration (1.5610 11 pfu/ml) and served as the input to Round 2. Rounds 2 and 3 were largely the same procedure as Round 1, except that in Round 2 incubation times were 30 minutes with C/M A2, and an overnight elution with RTB. In Round 3, C/M A2 binding proceeded for 15 minutes, again with an overnight RTB elution. Unamplified Round 3 supernatant was plated on LB/Agar/IPTG/X-gal plates, and individual plaques were picked and amplified. These clonal phage stocks were then used in phage ELISAs and for DNA isolation with the Qiaprep Spin M13 Kit. B/J F9 phage display was performed as for C/M A2.
Phage Binding ELISA C/M A2 or B/J F9, MOPC21, and RTB were coated (RTB 10 mg/ml : 294 nM and Abs 10 mg/ml : 67 nM) in 96 well plates and then blocked with 2% PBS-BSA. Phages (at 1610 10 pfu/ml) were added for 1 hr, and then washed with PBS-T (0.1% v/v). Anti-M13-HRP was then added for 1 hr. The wells were again washed with PBS-T (0.1% v/v), and TMB Substrate was added. After ,10 min, the reaction was quenched with 1 M phosphoric acid. The plate was then read on a plate reader at 450 nm.

Peptide Inhibition ELISA
A 96 well plate was coated overnight with ricin (1 mg/ml : 15 nM) and then blocked with 2% BSA for 2 hours. A8 and C4 peptides at serially diluted concentrations were incubated with C/ M A2 (5 mg/ml : 33 nM) for 1 hr. The peptide-antibody solutions were then allowed to bind to ricin for 1 hr, at which point the wells were washed with PBS-T (0.1% v/v). Goat anti-mouse IgG-HRP was then added for 1 hr, and the wells were washed again with PBS-T (0.1% v/v). TMB substrate was added for ,10 min, and then quenched with 1 M phosphoric acid. The plate was then read on a plate reader at 450 nm.

C4 Monomer Cytotoxicity Inhibitio
C4 peptide at 10 mM was incubated with serially diluted concentrations of C/M A2 for 2 hr at room temperature. C/M A2 alone at the same serially diluted concentrations was also incubated. Ricin toxin was then added to a final concentration of 10 ng/ml or 154 pM. After a 5 min incubation, the mixtures were applied to tissue culture 96 well plate that was seeded the night before with 1610 4 Vero cells per well. The toxin-antibodypeptide mixtures were allowed to incubate with the cells for 2 hours at 37uC, after which the mixture was replaced with fresh media. 48 hr later, cell viability was assessed with Cell-Titer Glo.

Western Blots
Ricin was prepared at a concentration of 300 mg/ml or 4.6 mM in a solution of 0.05 mM EDTA, and then diluted 1:2 with 2X Laemmli sample buffer. Reduction of the disulfide bond linking RTA and RTB was achieved by the addition of b-mercaptoethanol (BME) to a final concentration of 2.3 mM. Reduced and nonreduced ricin samples were denatured by boiling for 6 min before being loaded on a 12% SDS-PAGE gel. The gel was soaked in transfer buffer (Bjerrum -Schafer -Nielsen) for 60 min, and then transferred to a nitrocellulose membrane at 10 V for 30 min using the semi dry unit (BioRad). Membrane was washed in PBS-Tween 20 (0.1%v/v) 365 min and blocked in 2% goat serum overnight at 4 C. Membrane was washed again in PBS-T (0.1%v/ v) and primary antibodies were applied in 10 ml/membrane of blocking buffer 1-5 mg/ml (6 nM -33 nM) for 1 hr. Secondary antibody (goat-anti-mouse-HRP) was also applied for 1 hr. Membrane was washed 3620 min in PBS-T (0.1% v/v) and developed using ECL Western blotting substrate (Pierce) as per the supplied instructions.

Statistical Analysis and Software
Statistical analysis was carried out with GraphPad Prism 5 (GraphPad Software, San Diego, CA). The open-source molecular visualization software PyMOL (DeLano Scientific LLC, Palo Alto, CA) was used for epitope modeling.

Identification and Characterization of Additional RTBspecific mAbs with Ricin Neutralizing Activity
We screened a collection of ,1000 hybridomas produced from RT immunized BALB/c mice in an effort to identify additional RTB-specific mAbs capable of neutralizing ricin toxin. The screen yielded three mAbs of interest, JB4, C/M A2 and B/J F9. The remaining hybridomas were not further characterized as they did not secrete antibodies that reacted with RTB or they failed to neutralize ricin. In a Vero cell cytotoxicity assay, JB4 and B/J F9 each demonstrated a dose-dependent capacity to neutralize ricin. With IC 50 s of approximately 0.80 (5.3 nM) and 0.04 mg/ml (0.27 nM) respectively, JB4 and B/J F9 are as effective (if not slightly more effective) than the previously described mAb SylH3 at inactivating ricin in a Vero cell-based assay ( Table 1; Fig. 2 A,  B). C/M A2 also neutralized ricin in a dose-dependent manner, but less effectively than either JB4 or B/J F9 ( Table 1; Fig. 2B). SPR analysis using a Biacore instrument revealed that JB4, C/M A2 and B/J F9 each bound ricin holotoxin with affinities roughly equal to (or slightly greater than) other previously described RTBspecific neutralizing mAbs, including SylH3 and 24B11 ( Table 1). We also used Biacore to perform mAb competition studies, as done previously by our laboratory [27,39]. We found that when tested in series JB4, C/M A2 and B/J F9 were largely unaffected in their abilities to bind ricin holotoxin when the Biacore chips were first saturated with one or both of the other mAbs (data not shown). On the other hand, the binding of JB4 to ricin was virtually eliminated when the toxin was first exposed to SylH3, indicating that the two mAbs recognize the same or a similar epitope on RTB (Table S2).
We recently demonstrated that THP-1 cells are highly susceptible to ricin intoxication and that ricin uptake into these cells is mediated in part by the mannose receptor (MR; CD206) [40]. Therefore, we also examined the ability of JB4, C/M A2 and B/J F9 to protect THP-1 cells from ricin-induced apoptosis THP-1 cells were incubated with ricin or ricin-mAb complexes for 5 hours at 37uC before being subjected to Annexin V-FITC staining and flow cytometry. Annexin V staining was ,7 times greater on ricin-only treated cells as compared to control, untreated cells, thereby confirming ricin's capacity to initiate apoptosis in THP-1 cells (Fig. 3). Incubation of ricin with the previously described RTB-specific neutralizing mAbs, SylH3 and 24B11, blocked ricin-induced apoptosis to a large degree, whereas the addition of the non-neutralizing mAbs TFTB-1 and CB12 had no effect on toxin-induced cell killing (Fig. 3). The new mAbs were then tested in turn: JB4 performed as effectively (if not slightly better than) SylH3 and 24B11 in neutralizing ricin in this assay, further demonstrating that is a potent toxin neutralizing mAb. In contrast, neither C/M A2 nor B/J F9 protected THP-1 , or TFTB-1 were assessed for their capacity to protect Vero cells from the cytotoxic effects of ricin. TFTB-1 is a non-neutralizing RTB-specific IgG 1 mAb that binds a linear epitope within sub-domain 2a (residues 169-184). TFTB-1 served as a negative control for these studies. Ricin (10 ng/ml; 154 pM) was incubated for 1 hr with each mAb at the indicated concentrations (starting at 5 mg/ml; 33 nM), and then applied in triplicate to Vero cells grown in 96-well microtiter plates. Cell viability was assessed 48 hr later. Each symbol (with SEM) represents the average of at least three replicate wells. This experiment was repeated at least 3 times. (Panels C, D) Passive protection studies in which groups of BALB/c mice (n = 5 per group) were injected (i.p.) with the indicated mAbs (60 mg mAb per animal) and then challenged 24 hr later by the i.p. route with 10xLD 50 s of ricin (2 mg per animal). Shown (y-axis) is percent survival as a function of time (x-axis). Mice that were challenged with ricin, but not mAb-treated were considered positive controls for this experiment and indicated by the (+ R) symbol. Mice that were sham challenged were considered negative controls and are indicated by the (2 R cells from ricin-induced apoptosis (Fig. 3). While this outcome was not surprising in the case of C/M A2, which was relatively ineffective at neutralizing ricin in the Vero cell assay (described above), it was unexpected in the case of B/J F9, which was as effective as JB4 at neutralizing ricin in the Vero cell assay ( Table 1). Based on the results of THP-1 cell-killing assays, we predicted that JB4, but not B/J F9 and C/M A2, would be capable of neutralizing ricin in vivo.
Therefore, we compared JB4, B/J F9 and C/M A2 for their ability to passively protect mice against a systemic ricin challenge. Groups of BALB/c mice were administered individual mAbs (60 mg) by i.p. injection, and then challenged by the same route with 10xLD 50 s of ricin toxin 24 hr later. Immunity to ricin was assessed by two means: onset of hypoglycemia, a well-established surrogate marker of ricin toxicosis [38] and mean time to death. JB4 fully protected mice against ricin-induced death (Fig. 2C), although the animals did experience an acute reduction in blood glucose at 24-72 hr post challenge (Fig. 2E). In this respect, JB4 is comparable to other previously described RTB-specific mAbs, including SylH3 [32]. B/J F9 and C/M A2, on the other hand, were not sufficient to protect mice against ricin intoxication. Only a single mouse (1/5) administered B/J F9 survived ricin challenge and no mice (0/5) treated with C/M A2 survived toxin exposure (Fig. 2C, D). Moreover, following ricin challenge, mice treated with B/J F9 or C/M A2 experienced declines in blood glucose levels that were indistinguishable from the control, toxin-only treated animals (Fig. 2E, F). The mAbs did counteract the effects of ricin to some degree in vivo, as evidenced by the fact that B/J F9 (p,0.05) and C/M A2 (p,0.05) each delayed time to death, as compared to control, toxin-only treated animals. These data demonstrate that among the three new mAbs identified in this study only JB4 was capable of protecting mice against ricin intoxication.

Differential Capacities of JB4, B/J F9 and C/MA2 to Block Ricin Attachment to Cell Surface Receptors
It is postulated that RTB-specific mAbs neutralize ricin primarily by interfering with toxin attachment to cell surface glycoprotein and glycolipid receptors [41]. We used a quantitative solid-phase binding assay to assess the capacities of JB4, B/J F9 and C/M A2 to block ricin attachment to terminal Gal/GalNAc residues (see Materials and Methods). Biotin-labeled ricin (50 ng/ ml) was incubated with JB4, C/M A2, B/J F9, SylH3, or TFTB-1 at a range of concentrations (0.25-20 mg/ml) and then applied to 96-well microtiter plates coated with ASF, a surrogate glycoprotein receptor [25]. As shown in Figure 4A, JB4 reduced ricin attachment to ASF in a dose-dependent manner, with an estimated IC 50 of ,1 mg/ml (7 nM). B/J F9 and C/M A2, on the other hand, were only moderately effective at blocking ricin attachment in this assay. Indeed, even at 20 mg/ml (133 nM), B/J F9 and C/M A2 only reduced ricin binding to ASF by ,50% and ,40%, respectively (Fig. 4A, B). Nonetheless, B/J F9 and C/M A2 were each more effective than TFTB-1, which had no detectable capacity to block ricin attachment.
To test whether JB4 interferes with ricin attachment to cell surfaces by physically binding an epitope within one (or both) of RTB's two galactose binding pockets situated in sub-domains 1a or 2c, we performed a modified ELISA in which 96-well microtiter plates were first coated with ASF and then secondarily coated with ricin. It has previously been argued that RTB's two galactose binding pockets are occupied under these conditions, because ricin is bound to the solid substrate solely by virtue of its ability to bind ASF [16,25]. Using this assay, we found JB4's  capacity to bind RTB was unaffected when the toxin was docked to ASF (Fig. 4C), thereby suggesting the mAb recognizes an epitope adjacent to but not within RTB's galactose binding pockets. The binding of C/M A2 and B/J F9 to RTB were similarly unchanged when ricin was immobilized via ASF (Fig. 4C, D).

Localization of the Epitopes on RTB Recognized by JB4, B/J F9 and C/M A2
We reasoned that epitope specificity must account for the disparate capacities of JB4, B/J F9 and C/M A2 to neutralize ricin in vitro and in vivo, because all three mAbs are IgG 1 s and all three bind ricin holotoxin with similar affinities. As a first step towards identifying the regions on RTB recognized by JB4, B/J F9 and C/ M A2, we compared the reactivity profiles of each mAb with ricin holotoxin and RTB. JB4 bound ricin holotoxin considerably better than RTB, suggesting that JB4 recognizes an epitope whose conformation (or accessibility) is influenced by RTB's association with RTA (Fig. 5 A, C). Not surprisingly, JB4's binding profile is strikingly similar to that of SylH3 [32]. B/J F9 and C/M A2, on the other hand, each bound RTB and ricin holotoxin equally well, demonstrating that their respective epitopes are not influenced by the proximity or association of RTB with RTA ( Fig. 4A-D). As expected, none of the mAbs bound purified RTA to any appreciable degree (Fig. 4 E,F).
As an additional strategy to localize the epitopes on RTB recognized by JB4, B/J F9, and C/M A2, we examined by ELISA the binding of each mAb to the lectin Ricinus communis agglutinin I (RCA-I). RCA-I is a tetrameric glycoprotein consisting of two ricin-like heterodimers whose B subunit (RCB) shares 84% sequence identity with RTB [42,43]. The utility of RCA-I as a tool in epitope discrimination is exemplified by the differential reactivities of two previously described mAbs, TFTB-1 and CB12 [32]. As shown in Figure 6 A,B, TFTB-1 binds equally well to RTB and the RCA-I B subunit (RCB), whereas CB12 recognizes RTB but not RCB. This result is consistent with the fact that TFTB's epitope is completely conserved between RTB and RCB, whereas only 9/14 residues within CB12's epitope are present on RCB (Fig. 6G). Examination of JB4, B/J F9, and C/M A2 by ELISA revealed that all three mAbs bound RCB to some degree, but considerably less well than they bound RTB (Fig. 5C-F). In fact, JB4 (and SylH3) were particularly inept at binding RCB. While these data alone are not sufficient to enable us to pinpoint the exact epitopes recognized by JB4, B/J F9, and C/M A2, they do suggest that the mAbs each bind to a region of RTB that is not fully conserved with RCB (Fig. 6).
To further differentiate the mAbs, we examined reactivity by Western blot analysis under conditions in which RTB was (i) solubilized in Laemmli sample buffer, (ii) solubilized in Laemmli sample buffer and boiled (''denatured''), and (iii) solubilized in Laemmli sample buffer with BME and boiled (''denatured and reduced''). We rationalized that the failure of a mAb to bind RTB in its reduced form would indicate that the mAb recognizes an epitope that is constrained by one (or more) of RTB's four intramolecular disulfide bonds spanning residues C 20 -C 39 , C 63 -C 80 , C 151 -C 164 , C 190 -C 207 . As shown in Fig. 7A, C/M A2 bound equally well to RTB in its native, denatured, and denatured/ reduced forms, suggesting that C/M A2 binds a linear (continuous) epitope on RTB. Recognition of RTB by B/J F9, on the other hand, was greatly diminished when RTB's disulfide bonds were reduced, suggesting B/J F9, like 24B11, binds a conformation-dependent (discontinuous) epitope (Fig. 7B,C). JB4 recognized all three forms of RTB, although binding was considerably diminished when RTB was reduced (Fig. 7D).

Refined Epitope Mapping using Pepscan Analysis and Phage-displayed Peptide Library
The fact that C/M A2 bound RTB in its reduced form (see Fig. 6A) suggested that this mAb binds a linear epitope. To localize this epitope, C/M A2 was subjected to pepscan analysis using a collection of 15-mer peptides that overlap by 8 amino acids and span the full length of RTB [32]. While neither JB4 nor B/J F9 reacted with any peptide in the array (data not shown), C/M A2 reacted with two peptides; A8 and C4 (Fig.  8A). A8 corresponds to residues T50-L64 (TLKRDNTIRSNGKCL) spanning sub-domains 1b and 1c, while peptide C4 corresponds to residues C190-I204 (CLTSDSNIRETVVKI) in sub-domain 2b. To determine whether C/M A2 binds to one peptide (''epitope'') preferentially over the other, we performed a competition ELISA in which C/ M A2 was pre-incubated with either C4 or A8 peptides at various concentrations before being applied to immobilized ricin in a 96 well plate. As shown in Fig. 8B, peptide C4 inhibited the binding of C/M A2 to immobilized ricin in a dose-dependent manner, whereas peptide A8 peptide did not. Peptide C4 was also able to block C/M A2's ricin neutralization activity in a Vero cell cytotoxicity assay (Fig. 8C). Based on these data, we conclude that C/M A2's epitope likely constitutes residues C190-I204 in sub-domain 2b.
As an alternative strategy to epitope identification, we subjected a phage displayed peptide library to affinity-enrichment using C/M A2 as ''bait'', using a protocol recently established in our laboratory (Vance and Mantis, manuscript submitted). After three rounds of panning, 20 phages were isolated by serial dilution on agar plates. Eighteen of the 20 phages bound specifically to C/M A2 by ELISA (Fig. S1). Analysis of the peptides encoded by these phages revealed 13 unique sequences that shared a common DxNxR motif (Fig. 8D). The DxNxR motif is present in the C4 peptide (CLTSDSNIRETVVKI) and constitutes residues D 194 , N 196 and R 198 of RTB. Moreover, there was hydrophobic residue between N and R in every phage we sequenced, which likely reveals the importance of the hydrophobic I 197 between N 196 and R 198 in RTB. There was no universal consensus sequence at the position corresponding to residue 195, although 10/13 phage analyzed contained contain a polar or charged residue in that position similar to residue S 195 .
We also subjected the phage displayed peptide library to affinity-enrichment using B/J F9 as bait, on the off chance that its epitope, although likely discontinuous in nature, could be reconstituted by surface display. We isolated 24 phages after 3 rounds of panning, 20 of which demonstrated specificity for B/J F9. Sequence analysis revealed 5 different peptide sequences that shared a considerable degree of similarity (Fig. S2). Of note, all 5 sequences contained at least two tryptophans, and three of them had 3 tryptophan residues. This is unusual considering that the observed frequency of tryptophan in the phage library is ,2% of the random codons. Interestingly, there are only 3 tryptophans exposed on the surface of RTB (W37, W93,W160) and all three are situated in a region that is different from RCB. Discussion RTB has two essential roles in ricin cytotoxicity: it mediates toxin attachment to glycoprotein and glycolipids on cell surfaces, and it facilitates the entry and retrograde transport of RTA from the plasma membrane to the ER. Sub-domains 1a and 2c mediate attachment events; the specific regions of RTB involved in trafficking have yet to be identified. Because RTB is a relatively small protein (34 kDa), we have previously postulated that its association with virtually any antibody (150 kDa for IgG) would have a profound effect on ricin's ability to bind to host cell receptors and/or engage host cell proteins associated with retrograde transport [44]. Surprisingly, this appears not to be the case. The majority of RTB-specific mAbs that have been produced and characterized to date are in fact non-neutralizing or weakly neutralizing in vitro and in vivo, despite their high affinities for ricin holotoxin [15,16,23,30,32]. Moreover, we have estimated in mice that RTB-specific, ricin-neutralizing antibodies constitute only a small fraction of the antibody pool elicited in response to ricin holotoxin or RTB immunization [32]. The results of the Figure 7. Known and proposed epitopes on RTB recognized by neutralizing and non-neutralizing RTB-specific mAbs. Alignment of the B chain sequences from ricin (PDB 2AAI) and RCA-I (PDB 1RZO) using BLAST (www.ncbi.nlm.nih.gov). Boxed areas indicate regions of differences between RTB and RCB. Bolded text (highlighted by underlying brackets) delineates identified epitopes recognized by RTB-specific mAbs. The name of the corresponding mAb that binds the indicated epitope is written below the brackets. Regions 1-4, as indicated by vertical arrows, reflect regions of difference between RTB and RCB where mAb binding sites have yet to be ascribed. doi:10.1371/journal.pone.0044317.g007 current study represent a continuation of our ongoing effort to generate a comprehensive B cell epitope map of RTB and to identify the regions of the protein responsible for eliciting protective antibody responses.
JB4 is only one of a handful of RTB-specific mAbs that have been shown to be capable of conferring passive immunity to ricin in a mouse model [16,30,32,41]. While the exact epitope recognized by JB4 remains unknown, we speculate that it is localized within RTB's sub-domain 2c, possibly even encompassing residues 247-254 (Figs. 6G; 9). JB4 is an extremely potent inhibitor of ricin binding to galactosides, which strongly argues that it recognizes a region of RTB that is in close proximity to one of RTB's two CRDs. The binding of JB4 to ricin was not affected by 24B11, a mAb that binds an epitope immediately adjacent to the CRD in sub-domain 1a (Figs. 1, 6G; 9 A,B; [25]). Finally, JB4 reacted poorly with RCB by ELISA, suggesting that JB4's epitope is not conserved between the two very closely related Ricinus communis proteins. The only notable region of difference between RTB and RCB within sub-domain 2c corresponds to residues 247-254 [25,32]. On the other hand, we cannot exclude the possibility that even a single amino acid difference at sites other than sub-domain 2c may account for the differential capacity of JB4 to bind RTB and RCB. Further studies will be required to definitively localize JB4's epitope on RTB.
It is interesting to note that JB4 recognizes an epitope that is similar if not identical to that recognized by the recently characterized mAb SylH3 [32]. Competition assays by ELISA and SPR revealed that the two mAbs almost completely inhibit one another from binding to ricin (Table S2). The two mAbs are not, however, identical, as they were isolated from two independent hybridoma fusions and their V H and V L sequences are different (M. Pauley, personal communication). Moreover, JB4 possesses a slightly higher affinity for ricin than does SylH3 (Table 1). Nonetheless, we think it is significant that JB4 and SylH3, the only protective mAbs we identified from a combined screen of ,4000 RTB-specific hybridomas, bind the same or a similar epitope. While this result may be fortuitous, we think it more likely that there are a very limited number of neutralizing ''hotspots'' on RTB and that the epitope(s) recognized by JB4 and SylH3 constitute one of the more immunodominant.
Antibody C/M A2 was originally identified based on its ability to partially neutralize ricin in a Vero cell cytotoxicity assay. Further analysis revealed however, that C/M A2 was unable to inactive ricin in a THP-1 based apoptosis assay, nor was it able to confer any measurable protection against ricin challenge in a mouse model. Although the reason for the difference in C/M A2 neutralizing activities remains unclear, we speculate that it could be due to the MR. In THP-1 cells ricin uptake occurs through two pathways: galactose-dependent, RTB-mediated endocytosis and mannose-dependent, MR-mediated endocytosis [40,45]. In Vero cells, ricin uptake occurs solely through galactosedependent, RTB-mediated endocytosis. The fact that C/M A2 neutralizes ricin in a Vero cell assay, but not a THP-1 cell assay, suggests it blocks galactose-dependent, RTB-mediated, but not mannose-dependent MR-mediated uptake of ricin into cells. How this relates to ricin toxicity in vivo remains unclear, because we recently reported that mice lacking the MR are more sensitive to ricin that their wild type counterparts [40]. This finding is consistent with the MR playing a role in scavenging ricin from circulation, and not serving as an alternative route of entry into host cells. Further studies are required to fully resolve the importance of the MR, and macrophages in general, in the uptake of ricin into cells in vivo.
The observation that C/M A2 recognizes RTB in its reduced and denatured forms enabled us to use two parallel approaches, pepscan analysis and phage displayed peptide library, to define C/ M A2's epitope in great detail. These two approaches, as well as differential reactivity with RCB, strongly implicate residues C190-I204 within sub-domain 2b as being C/M A2's target on RTB. If correct, then C/M A2 would be the first mAb known to bind RTB's sub-domain 2b. Interestingly, when modeled by PyMol on the surface of ricin, C/M A2's epitope is relatively close to the CRD in sub-domain 2b (Fig. 9). However, this distance is apparently too far to occlude the CRD, as evidenced by the failure of C/M A2 to block RTB-galactoside binding.
B/J F9 was also initially selected based on its relatively potent ricin neutralizing activity in the Vero cell cytotoxicity assay. It was therefore quite surprising that B/J F9 failed to confer any passive immunity to ricin in a mouse model. This disconnect likely reflects the shortcomings associated with relying on Vero cells as a primary measure of ricin cytotoxicity, as discussed above. Unfortunately, we were unsuccessful in pinpointing B/J F9's epitope, although several lines of evidence suggest it localizes within one of two regions; sub-domain 1b (indicted by arrow 2, Fig. 6G) or subdomain 2a (indicated by arrow 3, Fig. 6G). This conclusion is based on the differential binding of B/J F9 with RCB and RTB by ELISA, limited reactivity of B/J F9 with RTB in its reduced and denatured forms by Western blot, and finally, peptide sequences obtained from affinity enrichment of phage displayed peptide library. The peptides displayed by phage that were capable of binding B/J F9 were enriched in tryptophan residues (Fig S2). Sub-domain 1b and sub-domain 2a each contain surface displayed tryptophan, as well as loops formed by an intra-molecular disulfide bonds. Considering the differential binding of B/J F9 to RCB and ricin was not drastically different, we propose that B/J F9 binds sub-domain 2a rather than 1b, as the sequence similarity between RTB and RCB is greater in sub-domain 2a.
We used PyMOL to model the mAb epitopes in 3D. Fig. 9 shows the epitopes of the mAbs as surface and secondary structures; (9 A, B) -previously characterized mAb 24B11, (9 C,D) C/M A2 and putative epitope of JB4, and, (9 E,F) the putative loop of B/J F9. If B/J F9 binds the proposed loop, then its epitope is located in sub-domain 2a, spatially even further away from the galactose binding site than C/M A2 (Movies S1 and Figure 9. Delineation of the C/M A2 epitope using pepscan analysis and phage display. (A) C/M A2 (10 mg/ml: 67 nM) was examined by ELISA for the ability to bind to an RTB peptide array consisting of 37 15-mers (A1-D1, x-axis), each overlapping its neighbors by 8 amino acids 32]. C/M A2 reacted with peptides C4 and A8, corresponding to residues T50-L64 (TLKRDNTIRSNGKCL) spanning sub-domains 1b and 1c and residues C190-I204 (CLTSDSNIRETVVKI) in sub-domain 2b, respectively. Data is representative of 2 independent experiments. (B) A competition ELISA in which 96well microtiter plate coated with RTB was probed with C/M A2 in the presence of increasing concentrations of peptides C4 and A8. (C) A competition cytotoxicity assay in which ricin (10 ng/ml) was incubated for 1 hr with C/M A2 mAb at the indicated concentrations in the presence or absence of soluble C4 peptide (10 mM) before being applied in triplicate to Vero cells grown in 96-well microtiter plates. Cell viability was assessed 48 hr later. Data is representative of 2 independent experiments. (D) A phage displayed 12-mer peptide library was subjected to affinity enrichment against immobilized C/M A2 mAb, as described in Materials and Methods. The 13 unique sequences identified following 3 rounds of C/M A2 selection are aligned vertically and compared to peptides C4 (top) and A8 (bottom). All 13 phage-derived sequences contained a DxNxR motif (as well as a hydrophobic residue in the 4 th position) that aligned exactly with the C4 peptide. doi:10.1371/journal.pone.0044317.g009 S2) and in the same sub-domain as a previously characterized non-neutralizing mAb TFTB-1 [32].
In summary, characterization of three novel anti-RTB mAbs has enabled us to refine our previously constructed B cell epitope map of RTB (Movies S1 and S2). On the whole, our data are consistent with a model in which neutralizing Abs target subdomains 1a and 2c, and non-neutralizing Abs target sub-domains 1b, 2a, and 2b. We have not yet identified any mAbs that bind epitopes within sub-domain 1c. However, based on the fact that it lies between the two other non-neutralizing sub-domains (1b and 2a), and is not involved in galactose binding per se, we think it unlikely that 1c is a target for neutralizing Abs. The results of this study not only advance our understanding of immunity to ricin, but have important implications for the rational design of RTBbased subunit vaccines. We are currently attempting to express individual sub-domains in RTB with the assumption that subdomains 1a and 2c are likely highly efficient at eliciting neutralizing Abs, whereas 1b, 1c, 2a, and 2b are not. These same constructs will be used to generate additional sub-domain specific mAbs as a strategy to further establish a B cell epitope map of RTB.

Supporting Information
Figure S1 C/M A2-specific binding of phage clones. ELISA showing the ability of representative phage clones to bind to C/M A2 (gray bars), irrelevant IgG Ab MOPC (black bars), RTB (striped bars) or BSA (white bars). Three of these phages bound strongly and specifically to C/M A2, whereas the fourth bound only weakly to C/M A2, and also partially recognized RTB. Of the twenty overall clones isolated, eighteen bound specifically to C/M A2, and all eighteen displayed the DxNxR motif. Correspondingly, the two phages that did not recognize C/ M A2 specifically, did not contain the DxNxR motif. (TIF) Figure S2 B/J F9 specific peptide sequences determined using phage display. Phage display was carried out against mAb B/J F9, and DNA from 24 clones was isolated and sequenced. Of those, 8 unique sequences were shown by ELISA to bind strongly and specifically to B/J F9 (data not shown). Five of these sequences show significant homology, each containing a WxWxP motif (bolded), as well as other conserved residue types (italicized). Two additional peptides also showed partial homology with the motif, while an eighth peptide had no significant homology. Interestingly, all eight peptides have at least one tryptophan residue, and several have multiple tryptophans. This proves the importance of tryptophan in B/J F9 recognition of RTB, as the presence of tryptophan in the random peptide library is expected to be much lower. (TIF)  Movie S1 3D rotation of new B cell epitopes on secondary structure of RTB. PyMOL modeling of ricin, with mAb (putative) epitopes and regions of difference between RTB and RCB highlighted. 24B11 (green), C/M A2 (orange), JB4/ SylH3 (marine blue), B/J F9 (magenta), RTA (wheat), RTB (grey), regions of amino acid sequence difference between RCB and RTB (cyane), disulfide bonds (red), mannose side chains (olive green), lactose (yellow).

(MOV)
Movie S2 3D rotation of new B cell epitopes on surface structure of RTB. PyMOL modeling of ricin, with mAb (putative) epitopes and regions of difference between RTB and RCB highlighted. 24B11 (green), C/M A2 (orange), JB4/SylH3 (marine blue), B/J F9 (magenta), RTA (wheat), RTB (grey), regions of amino acid sequence difference between RCB and RTB (cyan), disulfide bonds (red), mannose side chains (olive green), lactose (yellow). (MOV) Figure 10. Modeling of proposed B cell epitopes on RTB. We used PyMOL and PBD accession AAI2 to model using surface (left, panels A,C,E) and secondary structure (right, panels B,D,F) depictions of the epitopes recognized by indicated mAbs. Panels A and B: depiction of the epitope recognized by 24B11 (green); panels C and D; depiction of the epitopes recognized by C/M A2 (orange) and JB4/SylH3 (marine blue); panels E and F the epitopes recognized by B/J F9 (magenta). Highlighting common to all panels include RTA (grey), RTB (black), regions of amino acid differences between RCB and RTB (cyan), disulfide bonds (red), mannose side chains (olive green), and lactose (yellow) situated within CRDs. doi:10.1371/journal.pone.0044317.g010