Harnessing the Glucosyltransferase Activities of Clostridium difficile for Functional Studies of Toxins A and B

Abstract

The incidence of Clostridium difficile infection (CDI) has been increasing within the last decade. Pathogenic strains of C. difficile produce toxin A and/or toxin B, which are important virulence factors in the pathogenesis of this bacterium. Current methods for diagnosing CDI are mostly qualitative tests that detect either the bacterium or the toxins. We have developed an assay (Cdifftox activity assay) to detect C. difficile toxin A and B activities that is quantitative and cost-efficient and utilizes a substrate that is stereochemically similar to the native substrate of the toxins (UDP-glucose). To characterize toxin activity, toxins A and B were purified from culture supernatants by ammonium sulfate precipitation and chromatography through DEAE-Sepharose and gel filtration columns. The activities of the final fractions were quantitated using the Cdifftox activity assay and compared to the results of a toxin A- and B-specific enzyme-linked immunosorbent assay (ELISA). The affinity for the substrate was >4-fold higher for toxin B than for toxin A. Moreover, the rate of cleavage of the substrate was 4.3-fold higher for toxin B than for toxin A. The optimum temperature for both toxins ranged from 35 to 40°C at pH 8. Culture supernatants from clinical isolates obtained from the stools of patients suspected to be suffering from CDI were tested using the Cdifftox activity assay, and the results were compared to those of ELISA and PCR amplification of the toxin genes. Our results demonstrate that this new assay is comparable to the current commercial ELISA for detecting the toxins in the samples tested and has the added advantage of quantitating toxin activity.

INTRODUCTION

Clostridium difficile is the leading identifiable cause of nosocomial diarrhea worldwide due to its virulence, multidrug resistance, spore-forming ability, and environmental persistence (3, 38, 46, 59). This bacterium has been implicated as the causative organism for 10 to 25% of the reported cases of antibiotic-associated diarrhea, 50 to 75% of antibiotic-associated colitis cases, and 90 to 100% of pseudomembranous colitis cases (4, 18). The toxigenic strains of C. difficile possess a 19.6-kb pathogenicity locus that encodes two notable proteins: toxins A (308 kDa) and B (269 kDa). These toxins are important virulence factors in the pathogenesis of C. difficile (22, 33, 37, 49, 58). Both toxins have the same enzymatic cleavage activity (13, 30, 31) and are cytotoxic to cultured cells; however, toxin B is 100- to 1,000-fold more potent than toxin A in most cell lines (29, 56, 58).

The C termini of these toxins have a β-solenoid structure that is involved in receptor binding (14, 26). The central regions of the toxins possess a cysteine protease activity, which cleaves the N-terminal region in the presence of inositol hexakisphosphate, releasing the N-terminally located glucosyltransferase domain into the cytosol of the mammalian host (16, 27, 44, 48, 50). The glucosyltransferase domain monoglucosylates low-molecular-weight GTPases of the Rho family (RhoA, -B, and -C, Rac, and Cdc42) in the host cytosol, using cellular UDP-glucose as the glucose donor (29, 30). This monoglucosylation interrupts the normal function of the Rho GTPases, leading to a variety of effects on intoxicated cells, such as apoptosis, cell rounding, actin cytoskeleton dysregulation, and altered cellular signaling (21, 27, 29, 30).

Currently, only one nonradioactive assay (the tissue culture cytotoxicity assay) is available for detection of the activities of the toxins. However, quantitative analysis of toxin activity by this method is tedious and requires the maintenance of a tissue culture system, which makes it costly in terms of time and effort. We have developed a quantitative assay (Cdifftox activity assay) that enables detection of C. difficile toxins A and B in a culture supernatant. The method is based on the inherent properties of these toxins to cleave p-nitrophenyl-β-d-glucopyranoside (PNPG), a chromogenic substrate with stereochemical characteristics similar to those of the natural substrate of the toxins, UDP-glucose. This assay is cost-efficient, sensitive, and quantitative and enables measurement of the cleavage activities of toxins A and B.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

C. difficile toxigenic strains ATCC 43255 (tcdA⁺ tcdB⁺), ATCC BAA-1805 (tcdA⁺ tcdB⁺; NAP1), ATCC 700057 (tcdA tcdB⁺), and ATCC BAA-1382 (tcdA⁺ tcdB⁺) were purchased from the American Type Culture Collection (Manassas, VA). Clinical isolates were obtained from stool samples from hospitalized patients with antibiotic-associated diarrhea suspected to be C. difficile positive (see below). The bacteria were grown in brain heart infusion (BHI)-based medium (Becton Dickinson and Company, Cockeysville, MD) or on BHI-agar containing cefoxitin (8 μg/ml) and d-cycloserine (250 μg/ml) and were incubated anaerobically in an atmosphere of 10% H₂, 5% CO₂, and 85% N₂ at 37°C in a controlled-atmosphere anaerobic chamber (Plas-Labs, Lansing, MI). The substrates were purchased from Biosynth International (Itasca, IL).

Sample storage conditions.

The clinical isolates were stored either short-term in chopped meat broth (BD Diagnostics, Franklin Lakes, NJ) at room temperature or in 15% glycerol stocks at −80°C. The purified toxins and eluents were stored at 4°C for a maximum of 1 month or until use, with no loss of activity.

Toxin assays. (i) Cdifftox activity assay.

The Cdifftox activity assay is performed with Cdifftox substrate reagent, composed of 15 mM PNPG, 50 mM Tris-HCl (pH 7.4), 50 mM NaCl, and 100 μM MnCl₂. The assay was performed in Costar sterile polystyrene 96-well plates (Corning Inc., NY) by adding to each well 200 μl of sample or culture supernatant fluid containing the toxins along with 100 μl of the substrate reagent. The plate was incubated at 37°C for 1 to 4 h, and each reaction was stopped by the addition to the well of 40 μl of 3 M Na₂CO₃. Cleavage of the substrate was monitored by measuring the absorbance at 410 nm, using a Spectra Max Plus 384 spectrophotometer (Molecular Devices, Sunnyvale, CA). To identify the best substrate for the assay, a number of substrates were evaluated, including PNPG, p-nitrophenyl-α-d-glucopyranoside, 4-aminophenyl-α-d-glucopyranoside, 4-aminophenyl-β-d-glucopyranoside, 5-benzyloxy-3-indoxyl-β-d-glucopyranoside, 5-bromo-6-chloro-3-indoxyl-β-d- glucopyranoside, 6-bromo-2-naphthyl-α-d-glucopyranoside, 6-chloro-3-indoxyl-α-d-glucopyranoside, 6-chloro-3-indoxyl-N-acetyl-β-d-glucosaminide, and 5-bromo-4-chloro-3-indoxyl-β-d-galactopyranoside. PNPG was selected as the substrate of choice because its cleavage by the toxins was the most efficient and sensitive and had the lowest background level. A molar extinction coefficient (ε) for p-nitrophenol of 17,700 M⁻¹ cm⁻¹ was used in the calculations (53). One unit of toxin activity was defined as the amount of toxin required to cleave one micromole of the PNPG substrate per hour under the experimental conditions.

(ii) Enzyme-linked immunosorbent assay (ELISA).

For comparison to the Cdifftox activity assay, the presence of toxins A and B in samples was also evaluated using the Wampole C. difficile Tox A/B II assay (TechLab, Blacksburg, VA). This assay was performed using the protocol provided by the manufacturer.

Purification of C. difficile toxins A and B.

To purify the toxins, C. difficile strain ATCC 43255 was cultured anaerobically for 5 days at 37°C in Spectra/Por dialysis bags (50 ml) with a molecular mass cutoff of 100 kDa (Spectrum Laboratories, Rancho Dominguez, CA). Purification of the toxins was performed according to established methods (39, 54, 57), with some modifications. Briefly, the culture was centrifuged at 10,000 × g for 10 min at 4°C, and the resulting supernatant was filtered through a 0.45-μm membrane filter (Millipore, Billerica, MA). To further eliminate low-molecular-mass proteins, the filtered supernatant was concentrated using a Pierce concentrator (Thermo Scientific, Rockford, IL) with a molecular mass cutoff of 150 kDa. The concentrated supernatant was precipitated by the addition of ammonium sulfate (450 g/liter), incubated overnight at 4°C with gentle stirring, and subsequently centrifuged at 6,000 × g at 4°C for 20 min. The precipitate was washed and dissolved in 50 mM Tris-HCl buffer (pH 7.4). The sample was loaded onto a fast-flow DEAE-Sepharose CL-6B (GE Healthcare Life Sciences, Piscataway, NJ) anion column preequilibrated with buffer D (50 mM Tris-HCl [pH 7.4] containing 50 mM NaCl) at a flow rate of 2 ml/min. The column was washed with buffer D (approximately 350 ml) until all unbound proteins were removed. Toxin A was eluted first with a linear gradient of NaCl (50 to 250 mM) in buffer D. The elution continued for toxin B with a NaCl gradient of 250 to 1,000 mM in buffer D, after a washing step with 250 ml of buffer D. The fractions (10 ml) were assayed for the presence of toxins by incubating 200 μl with 10 mM PNPG for 3 h at 37°C. The toxin-positive fractions were pooled and concentrated with 150-kDa concentrators for further purification.

The pooled fractions from the DEAE-Sepharose column were further purified by gel filtration chromatography. A 1-cm by 100-cm glass Econo column (Bio-Rad Laboratories, Gaithersburg, MD) was packed with Sephacryl S-300 high-resolution beads (GE Healthcare Life Sciences) and calibrated using the following standards purchased from Bio-Rad Laboratories: vitamin B₁₂ (1.35 kDa), myoglobin (17 kDa), ovalbumin (44 kDa), gamma globulin (158 kDa), and thyroglobulin (670 kDa). The concentrated toxins were applied to the column and eluted with buffer D at a flow rate of 0.5 ml/min. Fractions (5 ml) were assayed for the presence of the toxins, using a 200-μl volume, as described above. The purity of the purified toxins was evaluated by electrophoresis through a 5% acrylamide:bisacrylamide polyacrylamide gel (51). The protein concentrations in samples were determined using the Bradford assay (5), with bovine serum albumin as the standard.

Western blot analysis.

C. difficile toxins A and B (50 μg each) were separated in 5% polyacrylamide gels. The proteins were transferred from the gels onto Immun-Blot polyvinylidene difluoride (PVDF) membranes (Bio-Rad) by using a Trans-Blot cell (Bio-Rad). The membranes were incubated with individual mouse monoclonal antibodies specific for C. difficile toxins A and B as the primary antibodies (Abcam, Cambridge, MA). A WesternDot 625 Western blot kit (Invitrogen, Carlsbad, CA) was used to probe the membranes for the presence of each toxin. Briefly, the membrane was incubated with biotin-XX-conjugated goat anti-mouse IgG secondary antibody and, following washing, incubated with the Qdot 625 streptavidin conjugate according to the manufacturer's instructions. Imaging and analysis of the treated membrane were performed using a UVP BioDoc-It imaging system (Upland, CA).

Determination of K_m and V_max.

A series of experiments were performed to determine the Michaelis-Menten constant (K_m) and maximum velocity (V_max) for the PNPG cleavage activity of each toxin. To determine the amounts of the toxins necessary for the experiments, different amounts of each toxin ranging from 30 μg to 120 μg were evaluated with 10 mM PNPG as the substrate. A graph of toxin activity as a function of time was plotted, and the amount of each toxin that gave the best linear relationship but occurred slowly enough for the reaction to be monitored was chosen for the assay. Based on this analysis, 55 μg of toxin A and 100 μg of toxin B were used for the kinetics experiments. Each experiment was repeated four times and the average used for the analysis.

Effects of pH and temperature on toxin A and B activity.

The effects of pH and temperature on the activity of the toxins were investigated to determine the optimum temperature and pH for activity. For the pH experiment, the following buffers were used: glycine-HCl buffer (pH 2 to 3), citrate buffer (pH 4 to 6), Tris-HCl buffer (pH 7 to 10), disodium phosphate-sodium hydroxide buffer (pH 11 to 12), and KCl-NaOH (pH 13). Each pH experiment was initiated by incubating 100 μg of toxin A or toxin B with 10 mM PNPG, followed by incubation in the appropriate buffer at 37°C for 4 h. The reaction was monitored by measuring the absorbance at 410 nm.

The effect of the temperature of incubation on the PNPG cleavage activity was tested in 1.5-ml microcentrifuge tubes under the same conditions as those described above, except that the temperature of incubation was 4, 10, 15, 20, 25, 30, 35, 40, 45, or 50°C.

Inhibition assay.

To identify compounds that inhibit the activity of C. difficile toxins A and B, several compounds were tested, including sodium taurocholate, dimethyl sulfoxide, phenylmethylsulfonyl fluoride, and dimethyl formamide. Different concentrations of these agents (0, 50, 100, 200, and 300 mM) were added to 55 μg of either toxin A or toxin B in buffer D in a total reaction volume of 300 μl and incubated at 37°C for 10 min. After the toxin-inhibitor incubation period, 10 mM PNPG substrate was added and incubated at 37°C for 1 h. The absorbance at 410 nm was measured, and the percent inhibition was calculated as follows: % inhibition = (specific activity with inhibition/specific activity without inhibition) × 100.

Application of Cdifftox activity assay to culture supernatants from clinical C. difficile isolates.

Stool samples obtained from patients suspected to be infected by C. difficile were obtained from St. Luke's Hospital (Houston, TX) in an institutional review board-approved study. Single colonies, obtained independently from each patient's stool sample streaked onto BHI-agar medium containing cefoxitin (8 μg/ml) and d-cycloserine (250 μg/ml), were inoculated into 10 ml of BHI medium and incubated anaerobically at 37°C for 72 h, resulting in an optical density at 600 nm (OD₆₀₀) of about 1.3 to 1.4. After centrifugation at 10,000 × g for 10 min at 4°C, 250 μl of the supernatant was incubated with 50 μl of Cdifftox substrate reagent containing 30 mM PNPG at 37°C for 3 h. The assay was quantitated spectrophotometrically at an absorbance of 410 nm. The isolates were not typed to the strain level but were confirmed to be C. difficile based on PCR amplification of the genes that encode the toxins (tcdA and tcdB), as well as by toxin production. Culture supernatants from 18 clinical isolates and 4 ATCC strains (BAA-1805 [tcdA⁺ tcdB⁺; NAP1], 700057 [tcdA tcdB⁺], 43255 [tcdA⁺ tcdB⁺], and BAA-1382 [tcdA⁺ tcdB⁺]) were analyzed.

PCR amplification of C. difficile toxin genes.

The presence of the toxin genes (tcdA and tcdB) and the 16S rRNA gene in the genomes of the clinical isolates was confirmed by PCR amplification. Genomic DNA was isolated from 1 ml of culture at an OD₆₀₀ of 0.75, using a DNeasy kit (Qiagen, Valencia, CA). Amplification was performed using Phire Hot Start DNA polymerase (Finnzymes, Woburn, MA). The following primers were used: for toxin A, 5′-TGATGCTAATAATGAATCTAAAATGGTAAC-3′ (forward) and 5′-ACCACCAGCTGCAGCCATA-3′ (reverse); for toxin B, 5′-GTGTAGCAATGAAAGTCCAAGTTTACGC-3′ (forward) and 5′-CACTTAGCTCTTTGATTGCTGCACCT-3′ (reverse); and for 16S rRNA, 5′-ACACGGTCCAAACTCCTACG-3′ (forward) and 5′-AGGCGAGTTTCAGCCTACAA-3′ (reverse). The DNA was amplified with an initial denaturation of 98°C for 30 s and 36 cycles of 98°C for 10 s, 62°C for 10 s, and 72°C for 10 s, with a final extension of 72°C for 1 min. The PCR products were analyzed using 1.5% agarose gel electrophoresis.

Physicochemical analysis of C. difficile toxins A and B.

The ProtParam program on the ExPASy proteomics server (20) was used to evaluate the physical and chemical properties of toxins A and B. This analysis was performed computationally using the amino acid sequences with GenBank accession numbers YP_001087137.1 and YP_001087135.1 (52) for toxins A and B, respectively.

Data analysis.

All data were analyzed and plotted using GraphPad Prism, version 5.02, for Windows (GraphPad Software, San Diego, CA). The nonlinear regression method was used to calculate the K_m and V_max values. A paired t test was used to evaluate the performance of the new Cdifftox activity assay in detecting the presence of the toxins in comparison with ELISA. In all cases, statistical significance was defined as having a P value of <0.05.

RESULTS

Purification of C. difficile toxins A and B.

Clostridium difficile toxins A and B were purified 7-fold to characterize and evaluate their substrate cleavage specificities. The native toxins were purified from culture supernatant obtained from a toxin A- and B-positive strain (ATCC 43255) cultured in a dialysis bag with a 100-kDa molecular mass cutoff. The proteins in the culture supernatant were precipitated with ammonium sulfate, resuspended, and applied to a fast-flow DEAE-Sepharose anion-exchange chromatography column. After elution with a 50 mM to 1 M NaCl step gradient, two peaks were observed by UV detection and confirmed by Bradford protein assay (Fig. 1). The initial, narrow peak was determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) to contain a protein corresponding to the molecular mass of toxin A (308 kDa), and the second, broad peak was determined to contain a protein corresponding to toxin B (269 kDa) (data not shown). The Cdifftox activity assay was used to identify fractions that contained PNPG cleavage activity. The assay revealed two toxin-positive fractions that corresponded to the two protein peaks observed (Fig. 1A). The antibody-based ELISA, which is immunologically specific for both toxins A and B, identified the presence of toxins A and B in the fractions with PNPG cleavage activity (Fig. 1B). Specifically, all of the fractions that tested positive using the Cdifftox activity assay also tested positive using ELISA, and all of the fractions that were negative in the Cdifftox activity assay were also negative using the ELISA. These results indicate that the Cdifftox activity assay detects the activities of C. difficile toxins A and B.

Fig. 1.

Elution profiles of proteins in C. difficile culture supernatant separated by DEAE-Sepharose anion-exchange chromatography. Fractions (10 ml) were examined using the Cdifftox activity assay (A) and an antibody-based ELISA (B). The Cdifftox activity assay was performed by incubating 200 μl of each fraction in 50 mM Tris-HCl containing 50 mM NaCl (pH 7.4) with 10 mM PNPG substrate reagent at 37°C for 4 h. The assay was monitored by measuring the absorbance at 410 nm. The protein concentration was determined using the Bradford protein assay (Bio-Rad). The ELISA was performed using the Wampole C. difficile Tox A/B II assay (TechLab, Blacksburg, VA).

To complete the purification of C. difficile toxins A and B, the toxin-positive fractions eluted from the DEAE-Sepharose column were pooled, concentrated using a filter with a 150-kDa molecular mass cutoff, and applied to a Sephacryl S-300 gel filtration column. After elution with buffer D, three predominant peaks were observed by UV detection and confirmed by the Bradford protein assay (Fig. 2A). Examination of the fractions using the Cdifftox activity assay showed that the PNPG cleavage activity was present in two peaks of different molecular sizes. The fractions with PNPG cleavage activities were confirmed by ELISA to contain the toxins (Fig. 2B). Based on the elution profiles of the gel filtration standards used (data not shown), the fractions giving the first active peak corresponded to toxin A (308 kDa), and those giving the second active peak corresponded to toxin B (269 kDa). The fractions showing sufficient toxin activity were pooled and concentrated for further analysis by PAGE. The results of PAGE revealed a single visible band for each of the two pooled fractions, representing toxins A and B (Fig. 3A). This established that each toxin was purified to homogeneity. The total PNPG substrate cleavage activities of the toxins from each of the purification steps are shown in Table 1. The total enzyme units of cleavage activity for the toxins were enriched 158-fold. The final substrate cleavage activities of the purified toxins were 0.821 U/μg and 4.7 U/μg for toxins A and B, respectively.

Fig. 2.

Elution profiles of pooled C. difficile toxin-positive fractions purified by Sephacryl S-300 gel filtration chromatography. Fractions (5 ml) were examined using the Cdifftox activity assay (A) and an antibody-based ELISA (B). The Cdifftox activity assay was performed by incubating 200 μl of each fraction in 50 mM Tris-HCl containing 50 mM NaCl (pH 7.4) with 10 mM PNPG substrate reagent at 37°C for 4 h. The assay was monitored by measuring the absorbance at 410 nm. The protein concentration was determined using the Bradford protein assay (Bio-Rad). The ELISA was performed using the Wampole C. difficile Tox A/B II assay (TechLab, Blacksburg, VA).

Fig. 3.

Electrophoretic analysis of the purified and PNPG-active fractions of C. difficile toxins A and B. (A) PAGE analysis of C. difficile toxin A and B purification by anion-exchange and gel filtration chromatography. Proteins (50 μg each) were separated through a 5% polyacrylamide gel. Lanes: M, ProSieve QuadColor molecular weight marker (Lonza Rockland Inc., ME); 1, concentrated supernatant; 2, pooled and concentrated fractions from anion-exchange chromatography; 3, pooled and concentrated fractions from gel filtration chromatography (toxin A); 4, pooled and concentrated fractions from gel filtration chromatography (toxin B). The arrow indicates the location of toxins in the gel. (B) Western immunoblot analysis of C. difficile toxins A and B after gel filtration chromatography. Proteins (50 μg each) were separated by electrophoresis through a 5% polyacrylamide gel and transferred to PVDF membranes. Each membrane was probed using mouse monoclonal primary antibodies specific for toxin A or B. A WesternDot 625 Western blot kit (Invitrogen, Carlsbad, CA) was used for detection of the bound antibodies. Sup, crude culture supernatant; Tox A, toxin A; Tox B, toxin B.

Table 1.

Summary of C. difficile toxin A and B purification from crude culture supernatant

Purification step	Amt of protein (μg)			Total activity (units)a			Sp act (U/μg)			Purification (fold)b
	Both toxins	Toxin A	Toxin B	Both toxins	Toxin A	Toxin B	Both toxins	Toxin A	Toxin B
Crude culture supernatant	965,000			779,720			0.808			1
Concentration (150 kDa)	801,000			973,215			1.215			1.50
DEAE-Sepharose CL-6B chromatography		2,230	1,910		1,014	3,085		0.455	1.615	2.56
Sephacryl S-300 chromatography		1,410	805		1,158	3,775		0.821	4.689	6.82

^aOne unit of toxin activity was defined as the amount of toxin required to cleave one micromole of the PNPG substrate per hour under the experimental conditions.

^bFold purification was calculated using the combined specific activities of toxins A and B.

Characterization of toxin A and B activity.

To confirm that both toxins A and B cleave the PNPG substrate, Western immunoblot analysis was performed. Single bands were observed in each of the samples that had PNPG activity and contained either purified toxin A or B, according to their specific reactivity with monoclonal antibodies that recognize toxin A or B, respectively (Fig. 3B). Moreover, the toxin A-specific monoclonal antibody did not recognize toxin B, and the antibody specific for toxin B did not recognize toxin A. These results demonstrate that both toxins A and B cleave the PNPG substrate. This is consistent with the reported in vivo activity of these toxins, in that they have both been reported to cleave the same cellular substrate, UDP-glucose (29–31).

The effects of pH and temperature on the functional activities of the toxins were evaluated. The pH experiments were performed using 5 different buffers to establish a wide range of buffering capacities. Both toxins A and B demonstrated optimal PNPG cleavage activities within a pH range of 7 to 9 (Fig. 4). In contrast to toxin A, which showed significant activity within the pH range of 6 to 12, toxin B displayed a more narrow range of PNPG cleavage activity, with activity within the pH range of 7 to 10. This is consistent with the pathophysiological environment of the colon, where C. difficile causes disease. The pH of the colon varies from 6.4 ± 0.6 to 7.5 ± 0.4 (32). Both toxins showed activity optima at a temperature range of 35 to 40°C, with toxin A showing a broader range of activity than toxin B.

Fig. 4.

Effects of pH (A) and temperature (B) on PNPG cleavage activities of toxins A and B. For the pH experiment, the Cdifftox activity assay was performed by incubating 100 μg of toxin A or B with 10 mM PNPG at 37°C for 4 h in buffers at the various pH values shown. The following buffers were used for the pH values indicated: glycine-HCl buffer (pH 2 to 3), citrate buffer (pH 4 to 6), Tris-HCl buffer (pH 7 to 10), disodium phosphate-sodium hydroxide buffer (pH 11 to 12), and KCl-NaOH (pH 13). For the temperature experiment, the Cdifftox activity assay was performed by incubating 100 μg of toxin A or B in 50 mM Tris-HCl containing 50 mM NaCl (pH 7.4) with 10 mM PNPG at the indicated temperatures for 4 h. The assay was monitored by the absorbance at 410 nm. Error bars represent standard deviations between two replicate experiments.

The amino acid sequences of the toxins were analyzed using the ProtParam program (20) to assess their physicochemical characteristics. Based on the ProtParam analysis, toxin A has 588 total charged residues out of 2,710 residues, among which 54% and 46% are negatively and positively charged, respectively. Toxin B has more charged residues (597 of a total of 2,366 residues), and 66% and 34% are negatively and positively charged, respectively. These data support the lower isoelectric point (IEP) (4.42) estimated for toxin B than that for toxin A (5.51). The implication of this lower IEP for toxin B is a wide pH range for the maintenance of its overall negative charge at physiological pH. Toxin A is computed to be more stable, with an instability index (25) of 29.6, compared to that of 36.5 for toxin B. However, both toxins are estimated to have a relatively long in vitro half-life (30 h) based on the N-terminal end rule (1, 23, 55). These computational data suggest that toxin A should function in and tolerate a wider range of physiological and environmental conditions than toxin B.

To better define the activities of toxins A and B, a kinetic analysis was performed. The affinities and enzymatic cleavage abilities of each toxin were assessed with the PNPG substrate. Initially, to determine the amount of each toxin that cleaves the substrate at a measurable rate under the experimental conditions, different amounts of the toxins were evaluated at a constant substrate concentration. As expected, this resulted in a dose-dependent cleavage of the substrate with increasing toxin amounts. Increasing substrate concentrations also led to an increase in cleavage products as the incubation time increased. The activities of both toxins could be fit to a Michaelis-Menten curve, indicating a single active site reaction (Fig. 5). The Michaelis-Menten constant (K_m) values of the toxins for the PNPG substrate were determined by nonlinear regression to be 1.04 mM for toxin A and 0.24 mM for toxin B. The maximum velocity (V_max) for toxin A was 1.5 μmol/mg/min, whereas that for toxin B was 6.4 μmol/mg/min. These data indicate that the affinity of toxin B for the PNPG substrate is >4-fold higher than that of toxin A. Moreover, the rate of cleavage of the PNPG substrate is 4.3-fold higher for toxin B than for toxin A. These results agree with assays of the relative damage by toxins A and B to tissue culture cells, in which toxin B was found to be more potent than toxin A (30, 56, 58).

Fig. 5.

Michaelis-Menten plot for PNPG cleavage by C. difficile toxins A and B based on a nonlinear regression method. For toxin A, K_m = 1.04 ± 0.06 mM and V_max = 1.50 ± 0.03 μmol/mg/min. For toxin B, K_m = 0.24 ± 0.02 mM and V_max = 6.40 ± 0.12 μmol/mg/min. Error bars represent standard deviations for four replicate experiments.

C. difficile activity inhibition.

To further characterize the toxin-substrate interactions, we sought to identify molecules or compounds that could inhibit the activities of toxins A and B. After testing of several potential molecules (see Materials and Methods), sodium taurocholate was observed to inhibit the activities of both toxins. The addition of 300 mM sodium taurocholate reduced the activities of toxins A and B within 1 h of incubation, by 71% and 86%, respectively (Fig. 6). Interestingly, taurocholate and phosphatidylserine (both negatively charged lipids) have been reported to inhibit β-glucosidases in a noncompetitive manner (8, 24, 28, 43). These results support the idea that the cleavage of the PNPG substrate is due to the glucosyltransferase/hydrolase activities of the toxins.

Fig. 6.

Dose-response inhibition by sodium taurocholate of toxin A and B PNPG cleavage activities. These experiments were performed by incubating 55 μg of each toxin with the indicated amount of sodium taurocholate for 1 h at 37°C in 30 mM Tris-HCl buffer (pH 7.4) containing 50 mM NaCl and 10 mM PNPG. Error bars indicate standard deviations from three different experiments.

Analysis of toxin activity in C. difficile supernatant fluid.

To evaluate the capability of the new Cdifftox activity assay to detect C. difficile toxins A and B in culture supernatants, C. difficile was isolated from clinical stool samples. Cultures were prepared from 18 independent clinical isolates from different patients, and their supernatants were tested for the presence and activity of toxins A and B by using the Cdifftox activity assay and ELISA. All of the culture supernatants from the clinical isolates determined to be positive for the toxins by the Cdifftox activity assay were also positive by ELISA (Fig. 7). The toxin-negative culture supernatants were negative in both assays. Genomic DNA was isolated from each strain, and PCR amplification analysis was performed with specific primers to identify the genomes carrying the C. difficile tcdA (toxin A) and tcdB (toxin B) genes. The toxin gene-positive isolates matched those that were toxin positive by the Cdifftox activity assay and ELISA (data not shown). Paired t test analysis showed that both ELISA and the Cdifftox activity assay correlated significantly in detecting the presence of the toxins (P = 0.001). However, there was not always a correlation between the amount of ELISA signal and the Cdifftox activity. This was expected, as ELISA is not quantitative, whereas the Cdifftox activity assay is quantitative. Interestingly, some of the isolates that were confirmed by PCR to carry tcdA and tcdB and that tested positive with the Cdifftox activity assay initially tested negative with the ELISA. However, these isolates became ELISA positive after a longer incubation of the culture, suggesting that the Cdifftox activity assay is more sensitive than the ELISA. Taken together, these data illustrate that the Cdifftox activity assay is a sensitive and reliable method to detect and assess the functional activities of C. difficile toxins A and B in culture supernatants.

Fig. 7.

Comparison of testing of clinical isolates for the presence of C. difficile toxins A and B by using the Cdifftox activity assay and ELISA. Supernatants (250 μl) from isolated strains cultured in BHI medium were incubated with 5 mM PNPG for 3 h at 37°C. The assay was monitored by measuring the absorbance at 410 nm. The number of moles of glucose released was calculated using a molar extinction coefficient for p-nitrophenol of 17,700 M⁻¹ cm⁻¹ (53). ATCC strains: 1805, BAA-1805 (tcdA⁺ tcdB⁺; NAP1); 057, 700057 (tcdA tcdB⁺); 432, 43255 (tcdA⁺ tcdB⁺); 630, BAA-1382 (tcdA⁺ tcdB⁺). Clinical isolates: S1 to S14, culture supernatants from independent clinical isolates obtained from different patients, with the genotype tcdA⁺ tcdB⁺; C1 to C4, culture supernatants from independent clinical isolates with the genotype tcdA tcdB. Error bars represent the standard deviations between two replicate experiments. Paired t test analysis showed that both ELISA and the Cdifftox activity assay correlated significantly in detecting the presence of the toxins (P = 0.001). However, there was not always a correlation between the amount of ELISA signal and the Cdifftox activity. This was expected, as the ELISA is not quantitative, whereas the Cdifftox assay is quantitative.

DISCUSSION

Within the last decade, the incidence of C. difficile infection has been increasing, such that it is now the leading definable cause of nosocomial diarrhea (4). Potential factors that have contributed to this prevalence are the increasing use of intestinal flora-altering antibiotics, the emergence of hypervirulent strains of C. difficile, the propensity of C. difficile to produce recurrent illness refractory to treatment, suboptimal infection control practices, and the appearance of toxin-producing mutant strains with more potent activity (34, 35, 38, 41).

Current methods for diagnosing C. difficile infection are based on detection of the organism or the toxin genes and proteins or on the effect of the cytotoxin on tissue culture cells (2, 10–12). The only method that can provide information about the activities of the toxins is the cell cytotoxicity assay. Such limitations are problematic for diagnosis and studies of these toxins. We have developed a cost-efficient, sensitive, and reliable assay, designated the Cdifftox activity assay, that utilizes PNPG, which is similar to the native substrate of these toxins, as a chromogenic substrate. Perhaps as a result, the K_m values obtained for the toxins with the nonnative PNPG substrate (1.04 ± 0.06 mM and 0.24 ± 0.02 mM for toxins A and B, respectively) (Fig. 5) are relatively close to those reported for their native UDP-glucose substrate (0.14 mM and 0.18 mM for toxins A and B, respectively) (9).

We have successfully used this assay for the purification of C. difficile toxins A and B and simultaneously evaluated it by comparison to an antibody-based ELISA. Unlike commercial ELISAs that only detect the presence of a fragment or region of the toxins, the Cdifftox activity assay detects the presence of the toxins and quantitates their substrate cleavage activities. We recognize that the Cdifftox activity assay does not distinguish between toxins A and B, since both toxins cleave PNPG and act on the same cellular substrate in vivo (13, 30, 31). This lack of distinction is of little consequence, since both toxins are responsible for the pathogenesis of C. difficile infections (19, 33, 36).

Our data demonstrate that the PNPG cleavage activities of C. difficile toxins A and B are inhibited by sodium taurocholate, in a dose-dependent manner (Fig. 6). Taurocholate, which is one of the major bile acids found in humans (15, 40), is formed and secreted into the lumen of the small intestine by conjugation of cholic acid with taurine. The total bile acid concentration in the small intestine varies depending on diet and other metabolic conditions (15, 42, 45). However, only about 2 to 5% of the bile acids secreted in healthy humans enter the colon because the majority of the bile acids are reabsorbed in the ileum (15, 17). Our finding of inhibition of the C. difficile toxins by a major bile acid may explain why the pathology of C. difficile infection is restricted almost exclusively to the bile acid-poor colon, with relative sparing of the bile-rich small bowel.

Sodium taurocholate is known to noncompetitively inhibit mammalian β-glucosidases (8, 24, 28, 43). These enzymes, including glucosyltransferases, belong to a large family of enzymes that mediate a wide variety of functions, such as carbohydrate biosynthesis, metabolite storage, and cellular signaling (7). Glucosyltransferases transfer a monosaccharide from an activated nucleotide sugar donor to specific sugar residues, proteins, lipids, DNAs, or small-molecule acceptors. This transfer may occur by either inversion or retention of the configuration of the anomeric carbon (6, 47). Inhibition of toxin A and B activities by a molecule that also inhibits glucosidases suggests that the cleavage of the PNPG substrate utilized in the Cdifftox activity assay is due to the glucosyltransferase activities of the toxins. However, further confirmatory experiments are planned to test the activities of the toxin A and B glucosyltransferase domains. To our knowledge, the use of the glucosyltransferase activities of the A and B toxins to identify toxigenic C. difficile is unique and has not previously been reported.