THE JOURNAL OF INFECTIOUS DISEASES • VOL. 133, SUPPLEMENT • © 1976 by the University of Chicago. All rights reserved.
MARCH 1976
Purification of Heat-Labile Enterotoxin from Escherichia coli 078:Hll by Affinity Chromatography with Antiserum to Vibrio eholerae Toxin Zipora Dafni and John B. Robbins
From the Division of Bacterial Products, Bureau of Biologics, Food and Drug Administration, Bethesda, Maryland
Unsuccessful attempts have been made to isolate highly purified heat-labile enterotoxin from Escherichia coli by physicochemical methods, including isoelectric focusing, gel filtration with cross-linked dextrans and agarose, anion exchange chromatography, and zone electrophoresis [1-3]. Affinity chromatography, including immunoadsorbents, is a useful technique for isolation of highly purified proteins [4-6]. Because the enterotoxins of E. coli and Vibrio cholerae crossreact immunologically [7-9], we prepared an affinity immunoadsorbent column with antiserum to the toxin of V. cholerae for the specific isolation of the heat-labile enterotoxin of E. coli. Materials and Methods
Bacteria. E. coli 078:Hll, strain HI0407, that was from a human source and produced both heat-stable and heat-labile enterotoxins, was donated by Dr. S. Formal (Walter Reed Army Research Hospital, Washington, D.C.). The cultivation medium was free of macromolecules and consisted of the minimal medium of Davis [10], We thank Ms. Shirley Crymes for excellent technical assistance, Mr. David Rogerson for the fermentations, and Dr. R. Bradley Sack for performing the adrenal cell assay. Drs. Samuel Formal and Rachel Schneerson gave valuable advice and support. Please address requests for reprints to Dr. Zipora Dafni, Department of Medicine, Baltimore City Hospital, 4940 Eastern Avenue, Baltimore, Maryland 21224.
S138
2% casamino acids (Difco, Detroit, Mich.), 0.5% filter-sterilized yeast extract dialyzate (Difco), 5 mg of FeCl a · 6H 20/liter, and 4 mg of MnCI 2·4H20/liter. A 50-liter fermenter was used. Two liters of an overnight growth inoculum were added, and the following conditions were maintained for 18 hr: aeration (25 liters/min), slow agitation (200 rpm), and pH ~7.0. The bacteria were removed first by centrifugation in a Sharples lab model centrifuge (Equipment Division, Tennessee Salt Chemical Corp., Wynnewood, Pa.) and then by filtration through a membrane (pore size, 0.45 nm). The culture filtrate was concentrated by vacuum dialysis through Visking tubing (Union Carbide, Chicago, III.) at the rate of approximately 1 liter/hr in a cold room (4 C). The resulting concentrated culture fluid was dialyzed against 0.15 M NaCl, 0.01 M Tris, and 0.001 M EDTA, pH 7.4 (TBS) overnight in a cold room with two changes of buffer (14 liters). The filtrate was centrifuged at 9,000 g for 2 hr at 4 C. Both the original filtrate and the concentrate were active in the rabbit ileal loop assay [11]. We found that phosphate ions cause aggregation and eventual insolubilization of the toxin from strain H10407. Lyophilization of partially purified toxin after dialysis against ammonium acetate, NH 4HCOa, and dilute Tris-buffered NaCl or water resulted in an insoluble product. Burro antiserum to V. cholerae toxin. Purified enterotoxin of V. cholerae (prepared under contract for the National Institute of Allergy and
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
Concentrated culture filtrate of Escherichia coli 078:Hll, strain HI0407, was applied to an affinity column prepared with IgG antibodies to the toxin of Vibrio cholerae. Elution of the retained material with 3 M KCNS yielded a nonenterotoxic protein that precipitated with antiserum to V. cholerae toxin and had three major protein components on sodium dodecyl sulfate gels. After treatment with 2-mercaptoethanol, two protein components were observed. Elution of the affinity column with 5 M guanidine yielded an enterotoxic protein that precipitated with antiserum to V. cholerae toxin. After treatment with 2-mercaptoethanol, only one protein component, with mobility identical to that of the slower component of the eluate (treated with 3 M KCNS and 2-mercaptoethanol), was observed.
Purification of E. coli Enterotoxin
S139
Table 1.
Antitoxin titers of burro serum as determined over a 125-fold range of dilutions of serum. Estimated testing level (AU/ml)*
Factor Calculated AU/ml Potency ratiot
5
I
1:5
1:25
6,580 0.89
7,360 1.0
6,560 0.89
4,950 0.67
NOTE. The titrations of antitoxin were done for us by Dr. John P. Craig [13]. * Values were based on the predetermined antitoxin titer of 6,950 antitoxin units (AU) /ml, Antitoxin titer was determined by the LB method with use of each of the four dilutions of antitoxin titrated against appropriate falling 0.I5-log decrements of toxin [13]. t The titer obtained at a serum dilution of I AU/ml was arbitrarily assigned a potency ratio of 1.0. Titers obtained at other levels were calculated accordingly. If antitoxin neutralizes toxin in constant proportion, the potency ratio is constant at all levels tested. With lower avidity, the potency ratio is reduced at lower concentrations of antitoxin.
tation with 37% saturated (NH 4hS04 and by anion exchange chromatography on DEAE-cellulose equilibrated with 0.02 M potassium phosphate (pH 7.4). The IgG was concentrated by vacuum dialysis and equilibrated against 0.1 M sodium citrate (pH 6.5). Sepharose 4B was activated with 75 mg of cyanogen bromide/rnl of packed gel (pH 10.5-11.0) at 4 C for 12-15 min. The gel was washed sequentially with 10 volumes of distilled water, 0.1 M HCI, 0.1 M NaHC0 3 and 0.1 M sodium citrate (pH 6.5). The concentration of IgG was adjusted to 40 mg/rnl, and the IgG was added to an equal volume of the activated Sepharose 4B; the mixture was tumbled overnight in the cold room and poured into a 5- X 50-cm column. The column was washed with 0.2% gelatin in PBS and then with TBS. Approximately 95% of the IgG was bound to the Sepharose. The immunoadsorbent was stored in TBS with 0.01 % NaN 3 at 4 C. The column was washed thoroughly with sterile TBS before use. The concentrated filtrate was passed through the column, which was then washed with TBS. When the optical density (280 nm) of the effluent reached base line, the eluting solvent (either 3 M KCNS or 5 M guanidine dissolved in TBS) was applied. The eluted protein was concentrated, dialyzed against TBS, centrifuged at 15,000 g for 1 hr at 4 C, and stored frozen. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). A modification of the method of Weber and Osborn was used [14]. The gel for electrophoresis contained 0.1 M sodium phosphate (pH 7.0), 7.5 % polyacrylamide, and 0.1 % sodium dodecyl sulfate (SDS). The samples were dissolved in 2 % SDS, brought to 100 C for 1 min, and applied to the gel in 50% sucrose containing 0.05% bromphenol blue. Some samples were treated at 100 C for 1 min with 2-mercaptoethanol (final concentration, 1% ). Enterotoxin assay. Activity of enterotoxin was assayed in the rabbit ileal loop and in Y-l adrenal cells in tissue culture systems [11, 15]. Results
Characterization of enterotoxin. The crude concentrate not bound to the immunoadsorbent
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
Infectious Diseases, Bethesda, Md., lot no. 0572) was used as the immunogen. For initial immunization 25 ug of cholera toxin, emulsified with 1 ml of phosphate-buffered saline (PBS), 0.7 ml of Arlacel A (Atlas Chemical Industries, Wilmington, Del.) and 1.3 ml of Marcol 70 (Exxon Co., U. S. A.), and 10 mg of heat-killed, dried bacille Calmette-Guerin. Additional injections (given at weekly intervals) contained the same proportions of PBS, Arlacel A, and Marcol 70, with doses of cholera toxin increasing to 50 ug, The burro was bled on days 5 and 8 after the last immunization. The characteristics of these samples of serum were as follows. (l) The specificity was such that two precipitin lines were formed with cholera toxin (lot no. 0572). (2) The total antibody content was approximately 14 mg/rnl of serum. (3) The serum contained 6,950 antitoxin units/rnl as determined by the rabbit intracutaneous assay method [12]. (4) In the rabbit ileal loop assay, dilutions of the antiserum of > 1 :1,000 neutralized cholera toxin. (5) The antiserum precipitated culture filtrates and affinity chromatography eluates of E. coli (strain H10407); the serum also yielded halos on antiserum-containing agar with toxigenic strains of E. coli. The antitoxin titers of the serum, which demonstrate its avidity, are listed in table 1. Immunoadsorbent. IgG was isolated from the burro antiserum to cholera toxin by precipi-
Dajni and Robbins
8140
ant. The guanidine-eluted protein peak contained only two bands on SDS-PAGE; these bands yielded one rapidly migrating component after treatment with 2-mercaptoethanol (figure 1, tubes 6 and 7). This single band corresponded to the central migrating component of the KCNS eluant. Enterotoxin activity, as determined by the ileal loop assay, is removed by treatment of the entire culture supernate with 3.0 M KCNS, but not with 5.0 M guanidine. Accordingly, it is not surprising that only the material eluted with guanidine yielded activity in both the ileal loop and adrenal cell assays. Discussion
The elution of antigens bound to immunoadsorbents requires conditions that may alter the
Figure 1. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of preparations of Escherichia coli 078:H11 enterotoxin and purified enterotoxin of Vibrio cholerae (lot no. 0572) in 0.1 M sodium phosphate (pH 7.0), 0.1% sodium dodecyl sulfate, and 7.5% polyacrylamide. Electrophoresis was carried out at 4 mAl gel for 11 hr at 25 C. Contents of tubes are as follows: I, concentrate of culture filtrate; 2, first peak from affinity column with enterotoxic activity; 3, second peak from affinity column without enterotoxic activity; 4, fraction eluated with 3 M KCNS; 5, fraction treated with 3 M KCNS and 2-mercaptoethanol; 6, fraction eluted with 5 M guanidine; 7, fraction treated with 5 M guanidine and 1% 2-mercaptoethanol; 8, cholera toxin; 9, cholera toxin treated with 2-mercaptoethanoi.
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
yielded two protein peaks. The first, which eluted in the void volume, contained enterotoxic activity that was not removed after repeated passes through the column. No enterotoxic activity was detected in the second peak. SDS-PAGE of these two fractions yielded multiple bands, and immunoelectrophoresis showed multiple components reactive with polyvalent antisera to E. coli (figure 1, tubes 2 and 3). A protein peak, released by elution with 3 M KCNS, revealed three major bands on SDSPAGE. Treatment of this fraction with 2-mercaptoethanol resulted in the disappearance of the slowest (highest-molecular-weight) component, without detectable change in the migration of the two faster components (figure 1). A slightly different pattern of elution from the immunoadsorbent was obtained with 5 M guanidine as elu-
Purification of E. coli Enterotoxin
References
1. Lariviere, S., Gyles, C. L., Barnum, D. A. Preliminary characterization of the heat-labile enterotoxin of Escherichia coli F11 (PI55). J. Infect. Dis. 128:312-320, 1973. 2. Jacks, T. M., Wu, B. J., Braemer, A. C., Bidlack, D. E. Properties of the enterotoxic component . in Escherichia coli enteropathogenic for swine. Infec. Immun. 7: 178-189, 1973.
3. Soderlind, 0., Mollby, R., Wadstrom, T. Purification and some properties of a heat-labile enterotoxin from Escherichia coli. Zentralbl. Bakteriol. [Orig. A] 229: 190-204, 1974. 4. Cuatrecasas, P., IIliano, G. Purification of neuraminadases from Vibrio cholerae, Clostridium perfringens, and influenza virus by affinity chromatography. Biochem. Biophys. Res. Commun. 44: 178-184, 1971. 5. Hughes, M., Thomson, R. 0., Knight, P., Stephen, J. The immunopurification of tetanus toxoid. J. Appl. Bacteriol. 37:603-621, 1974. 6. Cukor, G., Readio, J. D., Kuchler, R. J. Affinity chromatography purification of diphtheria toxin. Biotechnol. Bioeng, 16:925-931, 1974. 7. Smith, N. W., Sack, R. B. Immunologic crossreactions of enterotoxins from Escherichia coli and Vibrio cholerae. J. Infect. Dis. 127: 164-170, 1973. 8. Gyles, C. L. Relationships among heat-labile enterotoxins of Escherichia coli and Vibrio cholerae. J. Infect. Dis. 129:277-283, 1974. 9. Gyles, C. L. Immunological study of the heatlabile enterotoxins of Escherichia coli and Vibrio cholerae. Infec. Immun. 9:564-570, 1974. 10. Davis, B. D., Mingioli, E. S. Mutants of Escherichia coli requiring methionine or vitamin B12 . J. Bacteriol. 60: 17-28, 1950. 11. Kasai, G. J., Burrows, W. The titration of cholera toxin and antitoxin in the rabbit ileal loop. J. Infect. Dis. 116:606-614, 1966. 12. Craig, J. P., Eichner, E. R., Hornick, R. B. Cutaneous responses to cholera skin test in man. I. Responses in unimmunized American males. 1. Infect. Dis. 125:203-215, 1972. 13. Craig, J. P. Cholera toxins. In S. Kadis, T. C. Montie, and S. J. Ail [ed.]. Microbial toxins. Vol. 2A. Academic Press, New York, 1971, p. 189254. 14. Weber, K., Osborn, M. The reliability of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. BioI. Chern. 244:4406-4412, 1969. 15. Donta, S. T., Moon, H. W., Whipp, S. C. Detection of heat-labile Escherichia coli enterotoxin with the use of adrenal cells in tissue culture. Science 183:334-335, 1974. 16. Holmgren J., Soderlind, 0., Wadstrom, T. Crossreactivity between heat-labile enterotoxins of Vibrio cholerae and Escherichia coli in neutralization tests in rabbit ileum and skin. Acta Pathol. Microbiol. Scand. [B] 81:757-762, 1973.
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
activity of the bound component. KCNS was found to inactivate the crude toxins of E. coli and V. cholerae. Accordingly, it was expected that the material eluted with KCNS would not have enterotoxic activity. The material eluted with guanidine was enterotoxic; however, guanidine did not elute the third (fast-migrating) component in the material eluted with KCNS. The method that we have described permits isolation of enterotoxin from culture filtrates in a single step. Weare studying the efficacy of this isolation technique with other enterotoxin-producing strains of E. coli. Heterogeneity among the E. coli enterotoxins is suggested by their interaction with antisera to V. cholerae toxin and with antisera to crude preparations from several strains of E. coli [9, 16}. It is possible that there are both commonly shared and unique subunits among enterotoxins from various strains of E. coli. These subunits may differ in their crossreactivity with antisera to V. cholerae toxin. An advantage of this method is the single-step procedure for isolation of enterotoxin from the crude culture filtrate with reagents that are readily available. A second advantage is the stability of the immunoadsorbent, which may be used many times. A disadvantage is that only the material eluted with guanidine has enterotoxic activity. A possible explanation for the exclusion of some of the crude enterotoxic material from the column would be that, although the material is biologically active and antigenically identical to the bound material, the excluded enterotoxin has formed a high-molecular-weight aggregate that is poorly bound by the insolubilized antibody.
S141
MARCH 1976
Purification of Heat-Labile Enterotoxin from Escherichia coli 078:Hll by Affinity Chromatography with Antiserum to Vibrio eholerae Toxin Zipora Dafni and John B. Robbins
From the Division of Bacterial Products, Bureau of Biologics, Food and Drug Administration, Bethesda, Maryland
Unsuccessful attempts have been made to isolate highly purified heat-labile enterotoxin from Escherichia coli by physicochemical methods, including isoelectric focusing, gel filtration with cross-linked dextrans and agarose, anion exchange chromatography, and zone electrophoresis [1-3]. Affinity chromatography, including immunoadsorbents, is a useful technique for isolation of highly purified proteins [4-6]. Because the enterotoxins of E. coli and Vibrio cholerae crossreact immunologically [7-9], we prepared an affinity immunoadsorbent column with antiserum to the toxin of V. cholerae for the specific isolation of the heat-labile enterotoxin of E. coli. Materials and Methods
Bacteria. E. coli 078:Hll, strain HI0407, that was from a human source and produced both heat-stable and heat-labile enterotoxins, was donated by Dr. S. Formal (Walter Reed Army Research Hospital, Washington, D.C.). The cultivation medium was free of macromolecules and consisted of the minimal medium of Davis [10], We thank Ms. Shirley Crymes for excellent technical assistance, Mr. David Rogerson for the fermentations, and Dr. R. Bradley Sack for performing the adrenal cell assay. Drs. Samuel Formal and Rachel Schneerson gave valuable advice and support. Please address requests for reprints to Dr. Zipora Dafni, Department of Medicine, Baltimore City Hospital, 4940 Eastern Avenue, Baltimore, Maryland 21224.
S138
2% casamino acids (Difco, Detroit, Mich.), 0.5% filter-sterilized yeast extract dialyzate (Difco), 5 mg of FeCl a · 6H 20/liter, and 4 mg of MnCI 2·4H20/liter. A 50-liter fermenter was used. Two liters of an overnight growth inoculum were added, and the following conditions were maintained for 18 hr: aeration (25 liters/min), slow agitation (200 rpm), and pH ~7.0. The bacteria were removed first by centrifugation in a Sharples lab model centrifuge (Equipment Division, Tennessee Salt Chemical Corp., Wynnewood, Pa.) and then by filtration through a membrane (pore size, 0.45 nm). The culture filtrate was concentrated by vacuum dialysis through Visking tubing (Union Carbide, Chicago, III.) at the rate of approximately 1 liter/hr in a cold room (4 C). The resulting concentrated culture fluid was dialyzed against 0.15 M NaCl, 0.01 M Tris, and 0.001 M EDTA, pH 7.4 (TBS) overnight in a cold room with two changes of buffer (14 liters). The filtrate was centrifuged at 9,000 g for 2 hr at 4 C. Both the original filtrate and the concentrate were active in the rabbit ileal loop assay [11]. We found that phosphate ions cause aggregation and eventual insolubilization of the toxin from strain H10407. Lyophilization of partially purified toxin after dialysis against ammonium acetate, NH 4HCOa, and dilute Tris-buffered NaCl or water resulted in an insoluble product. Burro antiserum to V. cholerae toxin. Purified enterotoxin of V. cholerae (prepared under contract for the National Institute of Allergy and
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
Concentrated culture filtrate of Escherichia coli 078:Hll, strain HI0407, was applied to an affinity column prepared with IgG antibodies to the toxin of Vibrio cholerae. Elution of the retained material with 3 M KCNS yielded a nonenterotoxic protein that precipitated with antiserum to V. cholerae toxin and had three major protein components on sodium dodecyl sulfate gels. After treatment with 2-mercaptoethanol, two protein components were observed. Elution of the affinity column with 5 M guanidine yielded an enterotoxic protein that precipitated with antiserum to V. cholerae toxin. After treatment with 2-mercaptoethanol, only one protein component, with mobility identical to that of the slower component of the eluate (treated with 3 M KCNS and 2-mercaptoethanol), was observed.
Purification of E. coli Enterotoxin
S139
Table 1.
Antitoxin titers of burro serum as determined over a 125-fold range of dilutions of serum. Estimated testing level (AU/ml)*
Factor Calculated AU/ml Potency ratiot
5
I
1:5
1:25
6,580 0.89
7,360 1.0
6,560 0.89
4,950 0.67
NOTE. The titrations of antitoxin were done for us by Dr. John P. Craig [13]. * Values were based on the predetermined antitoxin titer of 6,950 antitoxin units (AU) /ml, Antitoxin titer was determined by the LB method with use of each of the four dilutions of antitoxin titrated against appropriate falling 0.I5-log decrements of toxin [13]. t The titer obtained at a serum dilution of I AU/ml was arbitrarily assigned a potency ratio of 1.0. Titers obtained at other levels were calculated accordingly. If antitoxin neutralizes toxin in constant proportion, the potency ratio is constant at all levels tested. With lower avidity, the potency ratio is reduced at lower concentrations of antitoxin.
tation with 37% saturated (NH 4hS04 and by anion exchange chromatography on DEAE-cellulose equilibrated with 0.02 M potassium phosphate (pH 7.4). The IgG was concentrated by vacuum dialysis and equilibrated against 0.1 M sodium citrate (pH 6.5). Sepharose 4B was activated with 75 mg of cyanogen bromide/rnl of packed gel (pH 10.5-11.0) at 4 C for 12-15 min. The gel was washed sequentially with 10 volumes of distilled water, 0.1 M HCI, 0.1 M NaHC0 3 and 0.1 M sodium citrate (pH 6.5). The concentration of IgG was adjusted to 40 mg/rnl, and the IgG was added to an equal volume of the activated Sepharose 4B; the mixture was tumbled overnight in the cold room and poured into a 5- X 50-cm column. The column was washed with 0.2% gelatin in PBS and then with TBS. Approximately 95% of the IgG was bound to the Sepharose. The immunoadsorbent was stored in TBS with 0.01 % NaN 3 at 4 C. The column was washed thoroughly with sterile TBS before use. The concentrated filtrate was passed through the column, which was then washed with TBS. When the optical density (280 nm) of the effluent reached base line, the eluting solvent (either 3 M KCNS or 5 M guanidine dissolved in TBS) was applied. The eluted protein was concentrated, dialyzed against TBS, centrifuged at 15,000 g for 1 hr at 4 C, and stored frozen. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). A modification of the method of Weber and Osborn was used [14]. The gel for electrophoresis contained 0.1 M sodium phosphate (pH 7.0), 7.5 % polyacrylamide, and 0.1 % sodium dodecyl sulfate (SDS). The samples were dissolved in 2 % SDS, brought to 100 C for 1 min, and applied to the gel in 50% sucrose containing 0.05% bromphenol blue. Some samples were treated at 100 C for 1 min with 2-mercaptoethanol (final concentration, 1% ). Enterotoxin assay. Activity of enterotoxin was assayed in the rabbit ileal loop and in Y-l adrenal cells in tissue culture systems [11, 15]. Results
Characterization of enterotoxin. The crude concentrate not bound to the immunoadsorbent
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
Infectious Diseases, Bethesda, Md., lot no. 0572) was used as the immunogen. For initial immunization 25 ug of cholera toxin, emulsified with 1 ml of phosphate-buffered saline (PBS), 0.7 ml of Arlacel A (Atlas Chemical Industries, Wilmington, Del.) and 1.3 ml of Marcol 70 (Exxon Co., U. S. A.), and 10 mg of heat-killed, dried bacille Calmette-Guerin. Additional injections (given at weekly intervals) contained the same proportions of PBS, Arlacel A, and Marcol 70, with doses of cholera toxin increasing to 50 ug, The burro was bled on days 5 and 8 after the last immunization. The characteristics of these samples of serum were as follows. (l) The specificity was such that two precipitin lines were formed with cholera toxin (lot no. 0572). (2) The total antibody content was approximately 14 mg/rnl of serum. (3) The serum contained 6,950 antitoxin units/rnl as determined by the rabbit intracutaneous assay method [12]. (4) In the rabbit ileal loop assay, dilutions of the antiserum of > 1 :1,000 neutralized cholera toxin. (5) The antiserum precipitated culture filtrates and affinity chromatography eluates of E. coli (strain H10407); the serum also yielded halos on antiserum-containing agar with toxigenic strains of E. coli. The antitoxin titers of the serum, which demonstrate its avidity, are listed in table 1. Immunoadsorbent. IgG was isolated from the burro antiserum to cholera toxin by precipi-
Dajni and Robbins
8140
ant. The guanidine-eluted protein peak contained only two bands on SDS-PAGE; these bands yielded one rapidly migrating component after treatment with 2-mercaptoethanol (figure 1, tubes 6 and 7). This single band corresponded to the central migrating component of the KCNS eluant. Enterotoxin activity, as determined by the ileal loop assay, is removed by treatment of the entire culture supernate with 3.0 M KCNS, but not with 5.0 M guanidine. Accordingly, it is not surprising that only the material eluted with guanidine yielded activity in both the ileal loop and adrenal cell assays. Discussion
The elution of antigens bound to immunoadsorbents requires conditions that may alter the
Figure 1. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of preparations of Escherichia coli 078:H11 enterotoxin and purified enterotoxin of Vibrio cholerae (lot no. 0572) in 0.1 M sodium phosphate (pH 7.0), 0.1% sodium dodecyl sulfate, and 7.5% polyacrylamide. Electrophoresis was carried out at 4 mAl gel for 11 hr at 25 C. Contents of tubes are as follows: I, concentrate of culture filtrate; 2, first peak from affinity column with enterotoxic activity; 3, second peak from affinity column without enterotoxic activity; 4, fraction eluated with 3 M KCNS; 5, fraction treated with 3 M KCNS and 2-mercaptoethanol; 6, fraction eluted with 5 M guanidine; 7, fraction treated with 5 M guanidine and 1% 2-mercaptoethanol; 8, cholera toxin; 9, cholera toxin treated with 2-mercaptoethanoi.
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
yielded two protein peaks. The first, which eluted in the void volume, contained enterotoxic activity that was not removed after repeated passes through the column. No enterotoxic activity was detected in the second peak. SDS-PAGE of these two fractions yielded multiple bands, and immunoelectrophoresis showed multiple components reactive with polyvalent antisera to E. coli (figure 1, tubes 2 and 3). A protein peak, released by elution with 3 M KCNS, revealed three major bands on SDSPAGE. Treatment of this fraction with 2-mercaptoethanol resulted in the disappearance of the slowest (highest-molecular-weight) component, without detectable change in the migration of the two faster components (figure 1). A slightly different pattern of elution from the immunoadsorbent was obtained with 5 M guanidine as elu-
Purification of E. coli Enterotoxin
References
1. Lariviere, S., Gyles, C. L., Barnum, D. A. Preliminary characterization of the heat-labile enterotoxin of Escherichia coli F11 (PI55). J. Infect. Dis. 128:312-320, 1973. 2. Jacks, T. M., Wu, B. J., Braemer, A. C., Bidlack, D. E. Properties of the enterotoxic component . in Escherichia coli enteropathogenic for swine. Infec. Immun. 7: 178-189, 1973.
3. Soderlind, 0., Mollby, R., Wadstrom, T. Purification and some properties of a heat-labile enterotoxin from Escherichia coli. Zentralbl. Bakteriol. [Orig. A] 229: 190-204, 1974. 4. Cuatrecasas, P., IIliano, G. Purification of neuraminadases from Vibrio cholerae, Clostridium perfringens, and influenza virus by affinity chromatography. Biochem. Biophys. Res. Commun. 44: 178-184, 1971. 5. Hughes, M., Thomson, R. 0., Knight, P., Stephen, J. The immunopurification of tetanus toxoid. J. Appl. Bacteriol. 37:603-621, 1974. 6. Cukor, G., Readio, J. D., Kuchler, R. J. Affinity chromatography purification of diphtheria toxin. Biotechnol. Bioeng, 16:925-931, 1974. 7. Smith, N. W., Sack, R. B. Immunologic crossreactions of enterotoxins from Escherichia coli and Vibrio cholerae. J. Infect. Dis. 127: 164-170, 1973. 8. Gyles, C. L. Relationships among heat-labile enterotoxins of Escherichia coli and Vibrio cholerae. J. Infect. Dis. 129:277-283, 1974. 9. Gyles, C. L. Immunological study of the heatlabile enterotoxins of Escherichia coli and Vibrio cholerae. Infec. Immun. 9:564-570, 1974. 10. Davis, B. D., Mingioli, E. S. Mutants of Escherichia coli requiring methionine or vitamin B12 . J. Bacteriol. 60: 17-28, 1950. 11. Kasai, G. J., Burrows, W. The titration of cholera toxin and antitoxin in the rabbit ileal loop. J. Infect. Dis. 116:606-614, 1966. 12. Craig, J. P., Eichner, E. R., Hornick, R. B. Cutaneous responses to cholera skin test in man. I. Responses in unimmunized American males. 1. Infect. Dis. 125:203-215, 1972. 13. Craig, J. P. Cholera toxins. In S. Kadis, T. C. Montie, and S. J. Ail [ed.]. Microbial toxins. Vol. 2A. Academic Press, New York, 1971, p. 189254. 14. Weber, K., Osborn, M. The reliability of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. BioI. Chern. 244:4406-4412, 1969. 15. Donta, S. T., Moon, H. W., Whipp, S. C. Detection of heat-labile Escherichia coli enterotoxin with the use of adrenal cells in tissue culture. Science 183:334-335, 1974. 16. Holmgren J., Soderlind, 0., Wadstrom, T. Crossreactivity between heat-labile enterotoxins of Vibrio cholerae and Escherichia coli in neutralization tests in rabbit ileum and skin. Acta Pathol. Microbiol. Scand. [B] 81:757-762, 1973.
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on September 5, 2015
activity of the bound component. KCNS was found to inactivate the crude toxins of E. coli and V. cholerae. Accordingly, it was expected that the material eluted with KCNS would not have enterotoxic activity. The material eluted with guanidine was enterotoxic; however, guanidine did not elute the third (fast-migrating) component in the material eluted with KCNS. The method that we have described permits isolation of enterotoxin from culture filtrates in a single step. Weare studying the efficacy of this isolation technique with other enterotoxin-producing strains of E. coli. Heterogeneity among the E. coli enterotoxins is suggested by their interaction with antisera to V. cholerae toxin and with antisera to crude preparations from several strains of E. coli [9, 16}. It is possible that there are both commonly shared and unique subunits among enterotoxins from various strains of E. coli. These subunits may differ in their crossreactivity with antisera to V. cholerae toxin. An advantage of this method is the single-step procedure for isolation of enterotoxin from the crude culture filtrate with reagents that are readily available. A second advantage is the stability of the immunoadsorbent, which may be used many times. A disadvantage is that only the material eluted with guanidine has enterotoxic activity. A possible explanation for the exclusion of some of the crude enterotoxic material from the column would be that, although the material is biologically active and antigenically identical to the bound material, the excluded enterotoxin has formed a high-molecular-weight aggregate that is poorly bound by the insolubilized antibody.
S141
