Purification of toxins and subunits

LT, LT-B, the non-toxic LT mutant LT(R192G/L211A), or dmLT, were produced from recombinant clones expressing LT, LT-B or dmLT based on LT from the human ETEC isolate E. coli H10407 as previously described [16, 17, 38]. Briefly, organisms were cultured overnight in 10-liters of Casamino Acids Yeast Extract medium, the cells were lysed, the lysate clarified by centrifugation, and LT-related proteins purified by galactose affinity chromatography. Proteins were stored lyophilized and freshly resuspended prior to use. His-tagged LT-A and His-tagged LT-A1 were prepared from solubilized inclusion bodies by HPLC with a nickel-affinity column as previously described [18]. The composition and purity of each protein was confirmed by SDS-PAGE and Limulus Amebocyte Lysate assay (Lonza, Inc.). The endotoxin content of the final products was <1 EU/mg.

Human Serum Samples

Human sera pooled 10 days after ETEC challenge was provided by David A. Sack (Johns Hopkins University). Irradiated human patient sera from Sierra Leone was provided by Robert F. Garry (Tulane University). The Sierra Leone samples were collected from suspected Lassa virus febrile patients or afebrile controls between December 2009 and April 2010 [39]. De-identified samples were irradiated and stored at -20°C. 37 samples were included in the final analysis from patients aged 7–74 years. Commercially available human serum was purchased as a control (Sigma).

Mouse Immunizations and sample collection

For these studies, we employed both in vitro (ELISA, CaCo-2 cell) and in vivo (patent mouse) assays following immunization by two different mucosal routes (intranasal, sublingual). Groups of 5 BALB/c mice each were intranasally immunized three times at weekly intervals with 10 μl saline containing 5 μg LT-A, LT-A1, LT-B, LT-A+LT-B (both at 5 μg) or dmLT. Alternatively, groups of mice were anesthetized with ketamine/xylazine and immunized with a single sublingual dose containing 5 μg LT-A, LT-A1, LT-B, LT-A+LT-B or LT in 10 μl of saline. All mice were sacrificed 21 days after the primary immunization. Serum was collected by cardiac puncture. Fecal samples were processed in protease buffer for supernatant antibody analysis as previously described [17]. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Tulane University Institutional Animal Care and Use Committee (Assurance Number A4499-01).

Detection of anti-LT or anti-subunit antibodies

Anti-LT, -A, or -B serum antibody ELISAs were performed on individual serum samples with wells coated with 0.1 μg of antigen and quantified with an external mouse IgG or IgA standard, as previously described [40, 41], or analyzed for serum endpoint titer using a cut-off value of two-fold the OD value of control serum. Human ELISAs were performed using goat anti-human IgG-AKP (Sigma).

Immunoblots were performed as previously described with gels loaded with 1 μg of unboiled or boiled LT in all lanes [18]. Following electrophoresis and transfer, blots were probed with human sera diluted 1:1000 in blocking buffer. Blots were developed using goat anti-human IgG-HRP (Invitrogen) and TrueBlue peroxidase substrate (KPL).

Neutralization of epithelial cell cAMP intoxication

Caco-2 cells (purchased from ATCC) were treated for 3 h with 0.1 μg trypsin-activated LT before the levels of intracellular cAMP were assessed using a cAMP assay kit (R&D Systems) [17, 18]. For neutralization of LT-induced cAMP, Caco-2 cells were treated with 10 μl containing 0.1 μg trypsinized LT pre-incubated overnight at 4°C with 10 μg GM1 monosialoganglioside (Sigma) or dilutions of rabbit anti-LTA, anti-LTB, anti-LT or negative control sera.

Neutralization of toxin-mediated intestinal fluid secretion

The patent (non-occluded gut) mouse assay was performed as previously described [18]. Mice were fasted overnight and then intragastrically fed 0.5 ml saline containing 25 μg LT pre-incubated for 30 min at room temperature with dilutions of sera (amounts corresponding to those used in cAMP assay appropriately scaled for the increased amount of LT). After 3 h, animals were sacrificed, and the entire intestine from duodenum to anus was carefully excised. Tissue and carcasses were separately weighed and individual gut/carcass mass ratios calculated.

Neutralization of toxin GM1 binding

Prevention of LT binding to GM1 was detected by ELISA as previously described [18]. Briefly, wells were coated with 10 μg of GM1 prior to addition of 0.1 μg LT with or without a 1 h pre-incubation with 1 μl serum at 25°C. LT binding was detected using anti-LT mouse serum prepared in our laboratory and anti-mouse IgG-AKP (Sigma). Absorbance was determined spectrophotometrically at 405 nm.

Neutralization of toxin epithelial cell entry and ADP-ribosylation

Confluent monolayers of Caco-2 cells in 24-well plates were treated for 3 h with 1 μg LT with or without pre-incubation with 10 μl antiserum at 4°C. Cells were immediately washed 3x, lysed, then analyzed by Western Blot analysis for ADP-ribosylated proteins using His-tagged macrodomain rAF1521 (a generous donation from Andreas Ladurner) followed by anti-His-horseradish peroxidase (HRP) antibodies (Qiagen) or LT B-subunit using rabbit anti-B serum prepared in our laboratory and anti-donkey-HRP (Santa Cruz).

Data analysis

Statistical analysis was performed using Prism (GraphPad Software, Inc.) for one-way ANOVA with Tukey’s Multiple Comparison post-test or Spearman’s Rank Order Correlation Coefficient (r) analysis. Unless otherwise indicated, P-values were coded as follows: *<0.05, **<0.01, ***<0.001.

ETEC-challenged human serum pool contains antibodies to both A- and B-subunits of LT

We first analyzed human immune serum for antibody response to LT holotoxin or subunits using a pool of sera from individuals challenged 10-days previously with ETEC isolate H10407 (O78:H11:K80 LT⁺ ST⁺), a prototypical strain of enterotoxigenic E. coli which reproducibly elicits diarrhea in human volunteer studies [42]. Commercially purchased human serum was included as a negative control. As seen in Fig 1, this ETEC-challenge serum tested positive for anti-LT, anti-A, and anti-B antibodies by ELISA (Fig 1A) in plates coated with native LT, LT-A or LT-B. The ELISA results indicated that antibodies to both the A-subunit and B-subunit were present in the sera of individuals infected with an LT-producing strain. The ELISA primarily detects antibodies against conformational epitopes and the use of A-subunit contaminated with traces of holotoxin or B-subunit could give a misleading impression of the presence of A-subunit antibodies. These studies used recombinantly produced LT-A and LT-B to reduce that possibility. In addition, we used immunoblots with boiled and unboiled preparations of LT to distinguish between the presence of antibodies against conformational and linear epitopes of LT in the H10407 challenge serum pool. In unboiled SDS-PAGE gels, LT runs as an 84 kD polymeric protein, pentameric B-subunit (56 kD), and LT-A (28 kD). When boiled and subjected to SDS-PAGE, LT separates into LT-A (28 kD) and monomeric LT-B (11.5 kD). When immunoblots with samples of unboiled LT were probed with H10407 challenge serum (Fig 1B), antibodies to LT, pentameric B (B5) and LT-A were observed, while predominantly anti-A antibodies were detected in the boiled LT preparation. This suggests that the anti-B response is against conformational epitopes, while the anti-A response is likely directed against both linear and conformational determinants. We confirmed this finding by probing immunoblots loaded with boiled or unboiled purified LT, LT-A, LT-A1, or LT-B with H10407 challenge serum (Fig 1C). Taken together, our results show that individuals infected with a wild-type ETEC strain developed antibodies to LT, LT-A, LT-A1 and LT-B.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. ETEC-challenged human serum pool contains antibodies to both A- and B-subunits of LT.

(A) ETEC challenge serum (pooled 10 days after oral H10407 challenge) anti-LT, anti-A, and anti-B antibody responses detected by ELISA (gray line, circles) compared to commercially purchased control sera (black lines, open circles) using dilutions of each sample. (B) ETEC-challenge serum (1) or control serum (2) immunoblot testing for anti-LT antibodies using unboiled LT-loaded lanes or boiled LT-loaded lanes. In unboiled SDS-PAGE gels, LT runs as an 84 kD polymeric protein, pentameric B-subunit (56 kD), and LT-A (28 kD). When boiled and subjected to SDS-PAGE, LT separates into LT-A (28 kD) and monomeric LT-B (11.5 kD). (C) ETEC-challenge serum or control serum anti-LT, anti-A, or anti-B responses detected with a modified Immunoblot using a slot blot apparatus to load 0.1 μg protein (LT, A, A1, or B) with raw images (top) and quantified band density of these images for unboiled, loaded proteins graphed (bottom).

https://doi.org/10.1371/journal.pone.0136302.g001

Serum samples from an ETEC-endemic area contain antibodies to both A- and B-subunits of LT

We next evaluated 37 human patient serum samples from Sierra Leone to evaluate antibodies in the setting of natural infection. While there is no epidemiological data specific to ETEC prevalence in this area, Sierra Leone is a developing country in sub-Saharan Africa, a region consistently associated within endemic ETEC [12, 24, 43]. In addition, current recommendations for travelers to Sierra Leone include precautions against contaminated food and water, identifying ETEC as the most common cause of travelers’ diarrhea [5]. These individual sera were collected from suspected Lassa virus febrile patients or afebrile controls, and while their ETEC exposure status is unknown, it is reasonable to assume that some or all of these individuals will have been infected with ETEC at some point. Individual serum samples were evaluated for the presence of antibodies to LT, LT-A and LT-B by ELISA. We first evaluated serial dilutions of individual patient serum in ELISA plates coated with LT, LT-A or LT-B to detect the presence of antibodies against the holotoxin and the individual subunits. The geometric mean titers of those sera were calculated and plotted by antigen (Fig 2A and S1 Fig) and approximately 70% of samples were positive for antibodies against one of the three antigens. Moreover, antibody responses to LT holotoxin significantly correlated with both anti-A (P<0.0001) and anti-B (P<0.0001) (Fig 2B), although there was heterogeneity within individual samples, including samples testing positive for anti-LT antibodies and both subunits or testing strongly for anti-LT antibodies and only anti-A- or anti-B antibodies. Similar to the H10407 challenge serum pool, antibodies to the B-subunit were predominantly conformational and were not apparent on immunoblots (using boiled LT), although bands corresponding to the A-subunit were readily detected (S1 Fig). Taken together, these results confirm our previous findings that individuals infected with ETEC developed antibodies to LT, LT-A, LT-A1 and LT-B.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Serum samples from an ETEC-endemic area contain antibodies to both A- and B-subunits of LT.

Human patient serum from Sierra Leone were evaluated for antibodies to LT, A-subunit, and B-subunit compared to commercially purchased control serum using endpoint titer analysis. (A) Anti-LT, anti-A, and anti-B ELISA antibody titers for all tested patient samples with geometric means indicated (lines). (B) Correlations between anti-LT and anti-A (open squares) or anti-B titers (gray diamonds) by Spearman’s Rank Order Correlation Coefficient (r) analysis. Antibody responses to LT holotoxin significantly correlated with both anti-A (P<0.0001) and anti-B (P<0.0001).

https://doi.org/10.1371/journal.pone.0136302.g002

Both A- and B-subunits are immunogenic and induce anti-LT antibodies following mucosal immunization, with maximal responses observed when both subunits are present

Analysis of sera from individuals experimentally or naturally infected with ETEC revealed that infection induces antibodies to both the A-subunit and B-subunit of LT. We next wanted to compare the antibody response after immunization with individual subunits alone or in combination. For these studies, we employed both in vitro (ELISA, CaCo-2 cell) and in vivo (patent mouse) assays following immunization by two different mucosal routes (intranasal, sublingual). For intranasal immunization, animals were immunized with LT-A, LT-A1, LT-B, LT-A and LT-B admixed together (LT-A+B), or the non-toxic A-subunit mutant dmLT. An SDS-PAGE gel of the purified proteins is shown in Fig 3A. Serum from these animals was collected and assayed in ELISA plates coated with native LT, LT-A or LT-B to examine the breadth of anti-LT responses induced by each immunogen. As seen in Fig 3B, LT-A1 was a poor immunogen and failed to induce antibodies by ELISA. By contrast, LT-A, which includes both A1 subunit and A2 domain, induced antibodies that recognized LT and LT-A. LT-B also induced antibodies to LT and to itself, but not LT-A. Surprisingly, the admix of LT-A and LT-B induced significantly higher responses against LT, LT-A and LT-B than either antigen alone, suggesting that the adjuvant function of LT-A was retained even when the molecules are dissociated. Specifically, LT-A is able to adjuvant the response to the B-subunit, even when the molecules are admixed and not present in the physiologically normal AB5 configuration. The anti-LT responses seen with the LT-A+B admix were equivalent to those seen with the non-toxic A-subunit mutant dmLT, which also retains immunogenic and adjuvant activity [17, 18]. Similar responses were observed when animals were immunized sublingually with LT-A, LT-B, the LT-A+B admix, or dmLT for both serum IgG and fecal IgA (Fig 3C). These results indicate that either the A- or B-subunit can be immunogenic when administered as a vaccine, but anti-toxin responses are maximized when both subunits are included in the immunization.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Both A- and B-subunits are immunogenic and induce anti-LT antibodies following mucosal immunization.

For these studies, we employed an in vitro ELISA assay following immunization by two different mucosal routes (intranasal, sublingual). For intranasal immunizations, BALB/c mice were immunized weekly for three weeks with 10 μl saline containing 5 μg of protein antigens A1, A, B, A+B, or dmLT. For sublingual immunizations, BALB/c mice were sublingually immunized once with 10 μl saline containing 5 μg of protein antigens A, B, A+B, dmLT, or LT. All mice were sacrificed 21 days after the primary immunization. The integrity of the antigens was demonstrated by SDS-PAGE (A). Anti-LT serum IgG (B,C) or fecal IgA (C) responses were detected by ELISA. Significance testing for ELISA panels was done using one-way ANOVA with Tukey’s Multiple Comparison post-test. P-values were coded as follows: *<0.05, **<0.01, ***<0.001.

https://doi.org/10.1371/journal.pone.0136302.g003

Anti-A or anti-B subunit antibodies can neutralize LT toxicity, with maximal responses observed when antibodies against both subunits are present

While LT-B facilitates cell binding and internalization of the toxin, LT-A mediates the intracellular enzymatic functions after proteolytic cleavage and disulfide bond reduction into the active A1 domain [15]. Within the cytoplasm of intestinal epithelial cells, the A1 domain ADP-ribosylates numerous cellular proteins, including Gsα, leading to irreversible activation of adenylate cyclase, cAMP accumulation, and deregulation of ion transport mechanisms that control luminal osmolarity and fluid secretion. Thus, there are a number of potential targets for protective antibodies within the steps leading to LT enterotoxicity. To determine how antibodies specific to each subunit might contribute to protection against LT toxicity, we employed two different assays—the patent mouse oral enterotoxicity assay and the CaCo-2 cell cAMP assay. For the patent mouse assay, animals were orally inoculated with an amount of native toxin sufficient to induce half-maximal fluid secretion after three hours (i.e., 25 μg LT). LT toxin was pre-incubated with rabbit anti-A serum, rabbit anti-B serum, negative control serum, or buffer. The rabbit polyclonal sera were produced in our laboratory against purified recombinant LT-A and LT-B. As seen in Fig 4A, both anti-A and anti-B were able to effectively neutralize the secretory activity of native LT in the patent mouse assay, indicating that both LT’s A-subunit and B-subunit are immunoprotective against LT-mediated diarrhea.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Anti-A or anti-B subunit antibodies neutralize LT toxicity, with maximal responses when both are present.

(A) Neutralization of toxin-mediated intestinal fluid secretion in the patent mouse assay after intragastric feeding with buffer (1) or 25 μg LT combined with anti-A (2), anti-B (3), negative control sera (4), or toxin alone (5). (B,E) Neutralization of epithelial cell cAMP intoxication in Caco-2 cells after 3 h treatment with 0.1 μg LT pre-incubated with positive control GM1 monosialoganglioside (GM1+LT), various dilutions and combinations of anti-serum, or toxin alone. (C) Blocking of toxin GM1 binding by ELISA detection using 0.1 μg LT toxin pre-incubated with anti-sera or alone. (D) Western blot of Caco-2 cells lysates after 3 h treatment with 0.1 μg LT toxin pre-incubated with anti-sera. Blots were probed for ADP-ribosylated proteins using rAF1521 macrodomains (bottom blot) then re-probed for presence of the B-subunit using anti-B rabbit serum (top blot). Boxes indicate proteins with altered ADP-ribosylation, potentially Gsα. Significance testing for secretion and cAMP panels was done using one-way ANOVA with Tukey’s Multiple Comparison post-test. P-values were coded as follows: *<0.05, **<0.01, ***<0.001.

https://doi.org/10.1371/journal.pone.0136302.g004

For the cAMP assay, CaCo-2 cells were treated with an amount of LT sufficient to cause maximal cAMP accumulation in 3 hours (i.e., 0.1 μg). For this assay, the LT was first pre-incubated with GM1 (positive control) or serial dilutions of rabbit anti-A or anti-B serum (Fig 4B). Interestingly, two different effects were observed. For anti-LT-A, neutralization of LT gradually diminished the further the serum was diluted, starting at 1:8 (100%) through 1:256 (0%). By contrast, anti-LT-B appeared to have a marked threshold where neutralization went from all to none in a single 2-fold dilution (64–128). This suggests that there is a threshold level of anti-B above which LT binding to the cell is blocked and any level below may not be effective.

We subsequently examined whether these antisera act to block binding or enzymatic activity of LT using ELISA plates coated with GM1 or an intracellular ADP-ribosylation assay. For both assays, LT was pre-incubated with rabbit anti-A, rabbit anti-B, negative control serum, or buffer prior to addition. Anti-B antibodies prevented GM1 binding (Fig 4C), entry of toxin into the cell (Fig 4D, top blot) and enzymatic activity (Fig 4D, lower blot). By contrast, anti-A antibodies only prevented the enzymatic activity of the toxin within epithelial cells. These results indicate antibodies to the B-subunit prevent toxin attachment and entry into intestinal cells, whereas antibodies to the A-subunit prevent enzymatic activity of the toxin within intracellular compartments.

We also examined potential synergy between anti-A and anti-B for toxin neutralization. For this assay, we combined anti-A and anti-B sera at dilutions that had not shown complete toxin neutralization in Fig 4B (e.g., 1:16 anti-A; 1:128–256 anti-B). As seen in Fig 4E, dilutions of anti-B at levels of 1:128–256 were unable to neutralize LT induction of cAMP, unless combined with anti-A. In addition, the neutralizing ability of anti-A was greatly enhanced by the presence of anti-B. Thus, anti-toxin responses are maximized when antibodies to both subunits are present.