CA1188986A - Isolation of escherichia coli heat-stable enterotoxin of porcine origin - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A heat-stable enterotoxin obtained from the Escherichia coli strain F11(P155) ATCC....... character-ized as the octadecapeptide H2N-Asn-Thr-Phe-Tyr-Cys-Cys-Glu-Leu-Cys-Cys-Asn-Pro-Pro-Cys-Ala-Ala-Gly-Tyr-COOH, and a process for preparing same in the substantially pure state. The substantially pure heat-stable enterotoxin is useful as an immunogenic agent in veterinary and in human medicine, and antiserum prepared therewith is useful as a reagent for determining the presence of said enterotoxin in biological fluids by immunochemical means.

Description

Field of Invention The present invention relates to a substan~ial-ly pure heat-stable enterotoxin obtained from Es~eric~ia ~o7:i strain Fll(P155) ATCC.~.. and to a process for its preparation.
Background of the Invention Various strains of Es~heric~hia ~3oZi with enterotoxigenic activity have been shown -to cause watery diarrhea in neonatal animals~ see e.g. Gyles in Ann. M.Y.
10 Acad. Sci. _76, 314 (1971), or Kohler in Am. J. Vet. Res.
29, 2203 ~1968) and Ann. N.Y. Acad. Sci. 176, 212 (1971), _ ._ or Smith et al. in J. Med. Microbiol. 3, 403 (1971) and ibid. 5, 243 (1972), or Whipp et al. in Infect. Immunity 12, 240 (1975), as w211 as in human infants, see e.g.
Ryder et al. in New England J. Med. 295~ 849 (1976) and in travellers, see e.g. Morris et al. in JO Clin.
Microbiol. _, 436 (1976), following colonization of the small bowel and release of enterotoxins. Two main groups oE such enterotoxins have been described so far: one of 20 those toxins is a heat-labile, immunogenic, high molecu-lar weight protein showing a mechanism of action and antigenic properties which are similar to the enterotoxin pxoduced by Vibrio choZerae, see e.g. Gyles in J. Infect.
Dis. 129, 277 (1974); another toxin would appear to be heat-stable, non-immunogenic, and of lower molecular weight, see e.g. Smith and Gyles in J. Med. Microbiol. 3, 387 (1970). It would appear from reports of a number of investigators that all enterotoxigenic strains of E. co1:,i are capable of producing the heat-stable enterotoxin 30 while only certain strains of porcine or of human origin can secrete the heat-labile enterotoxin, se~ e.g.
Robertson et al. in "Cholera and Related Diarrheas, 43rd Nobel Symposium", pp. 115 -126, Ouchterlony and Holmgren~
Eds~, S. Karger, Basel (1980).
The conditions required for the production and purification of the heat-labile enterotoxin have bee~

described, for example b~ Dorner in J. Biol. Chem. 250, 8712 (1975), or b~ Finkelstein et al. in J. Infect. Dis.
133, ~120 (1976), or by Shenkein et al. in In~ect.
Immunity 13, 1710 (1976~; Limjuco et al~ in ~.S. Patent 4,220,584, issued Sept~ 2, 1980 describe not only the preparation and purificatiorl of an immunogenic heat-labile enterotoxin from human and from porcine strains of E. coZi, but also describe the immunization of rabbits and of cows with said enterotoxin or with a toxoid pre-pared therefrom against subsequen-t challenges with the above enterotoxin.
However, the preparation and puri~ication of the heat-stable enterotoxin has been made difficult by the complexity of the various growth media used so far for its production and by the laborious and expensive pig gut loop test system used for its detection~ It is only since the -time that more rapid and less expensive testing methods for detecting the heat-stable enterotoxin have become av~ilable, for example the suckling mouse tes~
described by Dean et al. in J. Infect. Dis. 125, ~07 (1972) or the six-hour rabbit gut loop test described by Evans et al. in Infect. Immunity 7, 873 (1973), that some characteristics of the heat-stable enterotoxin have been described. Thus, Alderete et al., Infect. Immunity 19, 1021 (197~) described a 5100-dalton peptide (~7 residues) with heat-stable enterotoxic properties isolated from certain porcine strains of ~O co Zi which appeared to be homogeneous, contained six cysteine residues, and was found to have a glycine residue at the N~terminal.
3Q Kapitany et al., Infect. Immunity 26, 173 (1979) de-scribed the preparation and purification of two different enterotoxins from bovine and from porcine strains of . cozi~ respectively, both active in the nanogram range in the suckling mouse test and in the pig ligated loop test. Those two enterotoxins had different amino acid compositions, and both were heat-stable in the compara-\

tively crude state; however, the purified toxin of bovine origin lost most of its activity after 30 minutes at 100C while the purified preparation of porcine origin retained its activity under the same conditions. Staples et al., J. Biol.Chem. 255, 4716 (1980) isolated a heat stable enterotoxin from a strain of E. coZi of human origin and purified it to apparent homogeneity, with -the purest material being active at 2.7 ng doses in the suckling mouse assay; amino acid analysis demonstrated the presence of 18 amino acid residues in the molecule, six of which were identified as half-cystine. Lallier et al., Infect. Immuni~y 28~ 469 ~1980) have described the preparation and purification of a heat-stable enterotoxin from E. coZi strain Fll(P155) which was found to be active at nanogram levels in the suckling mouse test and in the six-hour rabbit gut loop test.
The heat-stable enterotoxin of this invention differs in amino acid composition from the enterotoxins described by Alderete et al., by Kapitany et al., and by Staples et al., all cited above. It is also different from the enterotoxin described by Lallier et al. cited above in being obtained by a different process as a sub-stantially pure compound with a well defined chemical structure.
S ~MARY OF THE NVENTION
The abbreviations used throughout this Appli-cation for the amino acids or the residues thereoE are generally based upon the recommendations of the IUPAC -IUB Commission on Biological Nomenclature, see Bio-30 chemistry 11, 1726 (1972). Asx represents asparagine - and/or aspartic acid~ and Glx stands for glutamine and/or glutamic acid. All amino acids have the natural or L-configuration.
Escherichia coZi strain Fll(P1553 is obtained from the non-enterotoxigenic avian strain Fll (serotype 018ab:K?:H14) by conjugation with the porcine strain P155 described by H.~. Smith et al. in J. Gen. Microbiol. 52, 319 (1968) and has the ability of producing both heat-labile and heat-stable enterotoxins. Samples o~ E. co Zi strain Fll(P155) have been deposited without restriction with the ~merican Type Culture Collection, 12301 Parklawn Drive, Rockville, Md., 20~52, and identified as ATCC.....
The above E. cozi strain Fll(P155) is grown in brain-heart infusion (Difco Laboratories, Detroit, Mich.) or in a semi-synthetic medium containing Casamino Acids (Trade Name Difco Laboratories, Detroit, Mich.~ together with sodium and potassium phosphate bufEers, ammonium chloride, sodium sulfate, and trace elements such as magnesium, manganese, and iron, similar to the media de-scribed by Finkelstein et al in J. Immunol. 96, 440 ; (1966~, at pH 6.8 - 8.0, preferably at pH 7.2 - 7.8, with addition of 0.0 - 0.6 per cent of yeast extract (Difco Laboratories, Detroit, Mich~) and with addition of 0.2 1.0 per cent of glucose, with agitation and forced aeration at 30 - 40~C, preferably at 37C, for periods ~0 of time of from 2 - 24 hours, preferably for 4 - 7 hours.
; Preferred conditions include the use of a semi-synthetic medium such as described above including 0.6 per cent of yeast extract and 0.2 per cent glucosel incubating at 37C with agitation and Eorced aeration Eor 7 hours, and controlling the pH so as to obtain pH 7.2 at the be-ginning and pH 7~8 towards the end oE the incubation period. Cell growth is monitored during incubation by measuring the optical density at 540 nm or by plate counts at periodic intervals. At the end of the incu-bation period the fermentation broth i5 centriEuged to remove cell growth and the supernatant is stored in the cold or frozen until assayed, either by a modification of the suckling mouse test described by Dean et al. in J.
Infect. Dis. 125, 407 (1972) or by a modifica-.ion of the six-hour rabbit gut loop test described by Evans et al.
in Infect. Immunity 7, 873 (1973). The yields of heat-stable enterotoxin obtained in the above manner vary from200-400 units per milliliter of supernatant.
Said last-named enterotoxin is precipitated by addition of acetone, or preferably of ammonium sulfate to the above cell free supernatant, allowed to settle, recovered by centrifuga~ion, and re-suspended in tris-(hydroxymethyl)aminomethane buffer, preferably in 0.05 M
buffer at pH 8.0, and filtered through a suitable ultra-filter membrane such as one have a cut-off of 10,000 to 30,000 daltons, preferably an Amicon PM-30 membrane manu-factured Amicon Co., U.S.A. The resulting filtrates are concentrated and the heat-stable enterotoxin contained therein is purified by gel filtration chromatography.
Individual fractions are assayed as described above, and fractions containing the heat-stable enterotoxin are pooled, freeze-dried, and further purified by high pressure liquid chromatography on alkylsilane or (substi-tuted al~yl)silane columns. It is preferred to carry out the above high pressure liquid chromatography in two stages, i.e. a first stage using à cyano(lower alkyl)-silane column equilibrated with a suitable buffer such as a phosphate-alkylamine buffer at pH of from 2.2 to 3.5, preferably at pH 3.0 and eluting with a linear gradient of an aprotic organic solvent such as a nitrile of a lower alkanoic acid; and a second stage, using the pooled active fractions obtained in the first stage, applying the same to an alkylsilane column equilibrated with a volatile buffer at pH of from 3~5 to 4.5, preferably at ; pH 4.15 such as an ammonium acetate buffer, and eluting 3~ with a linear gradient of a lower alkanol, e.g. methanol.
Heat-stable enterotoxic activity is detected in the indi-vidual fractions obtained in both stages by monitorin~
the UV absorbance, preferably at 210 nm and/or by bioassay as described above, active fractions are pooled, and solvents are removed by evaporation. The product obtained from the second stage of high pressure liquid chromatography as described above is freeze-dried and a sample thereof is found to be homogeneous in a further high pressure liquid chromatogram performed under the~
same conditions as described above for the second stage.
The substantially pure enterotoxin obtained as described above is a colourless fluffy substance ~hich is stable for 30 minutes at 100C and which is destroyed by heating to 121C for 15 minutes. It retains its entero-toxic activity when allowed to stand at room temperature for 24 hours at pH 2 - 10, but its activity is destroyed under the same conditions at pH 12. It is not affected by incubation at 37C for 4 hours with pronase or bac-terial alpha-amylase at pH 6.9, or with lipase at pH 8.1.
The constitution of the above heat-stable enterotoxin is established by amino acid analysis in conjunction with microsequencing, wi-th or without prior reduction and with incorporation of tritiated iodoacetic acid in the presence or absence of sodium dodecyl sulfate, or of 125I, and is found to be that of the octadecapeptide H2N-Asn-Thr-Phe-Tyr-Cys-Cys-Glu-Leu-Cys-Cys-Asn~Pro-Pro-Cys-Ala-Ala-Gly-Tyr-COOH. The above octadecapeptide is immunogenic, in that sera obtained from rabbits or goats immunized therewith neutralize the heat-stable entero-toxic activity of the cell-free supernatant prepared by culturing E. coZi strain ATCC........... as well as that of the substantially pure octadecapeptide prepared there-from.
Detailed Description o~ the Invention Æsehe~iehia co Zi ATCC.......... is prepared hy conjugating the non-enterotoxigenic avian strain of E. coZi Fll (serotype 018ab K?:H14) with the porcine strain P155 described by H.W. Smith et al. cited above.
The condition for incubation and growth medium are standard ones used for E~ coZi.
The media for growing E. eo Zi ATCCo~ may be either complex media such as brain~heart infusion obtained from Difco Laboratories~ Detroit, Mich., or semi-synthetic media similar to -those described by Finkelstein et al. cited above containing a casein hydrolysate, e.g. Casamino Acids (Trade Name, Difco Laboratories, Detroit, ~ich.~, together with sodium and potassium phosphate buffers, ammonium chloride, sodium sulfate, and trace elements as water-soluble salts such as magnesium chloride, manganese chloride, and ferric chloride. The inorganic and the organic constituents of the latter semi-syrlthe-tic media are sterilized separately by autoclaving, and mixed under sterile conditions before use. The pH of the above semi-synthe~ic media may vary from pH 6.8 to p~ 8.0, but it is most fre~uently found to be ahout pH 7~3O Furthermore, it has been found that the addition of small amounts of up to 0.6 per cent of yeast extract and of 0.2 - 1.0 per cent glucose increased the growth of E. co Zi ATCC....... and the production of the heat-stable enterotoxin obtained therefrom, with optimal results being obtained with the above semi-synthetic media -to which 0.6 per cent yeast extract and 0.2 per cent glucose had been added.
Incubation in the above media is carried out at - 40C, preferably at 37~C, for periods of time o~
from 2 24 hours, preferably for 4 - 7 hours, and preferably at pH 7.2 - 7.8. It is advantageous to carry out the incubation in three separate stages, viz., a first stage in which E. coIi ATCC~ o is incubated in brain-heart infusion (Difco~ with agitation at 37C for 4 hours; a second stage in which a sample of the culture obtained in the first stage is inoculated into the same semi-synthetic medium which is to be used for the ~hird stage and is incubated at 37C without agitation ior 12 18 hours, preferably overnight; and a third stage, in which a sample of the culture obtained in the second stage is inoculated into the same semi-synthe-tic medium used for the second stage and is lncubatecl a-t 37C with vigorous agitation, preferably at 400 - 500 rpm, and with forced aeration, preferably at about one vol~e of air per minute per volume of medium used, for 2 to 24 hours, -preferably for 4 to 7 hours; the semi-synthetic media used for the second and third stages advantageously con-taining 0.6 per cent yeast extract and 0.2 per cent glucose in addition -to -the constituents lis~ed above.
The pH of the fermentation broth is kept at 7.2 - 7.8, and it is advantageous to adjust the pH in such a manner during the third stage of incubation that the initial pH
is 7.2 and the final p~ is 7.8. Duriny the incubation period cell growth is monitored by measuring the optical density at 540 nm of the fermentation broth at periodic intervals~ and by occasional plate counts. At the end of the incubation period the fermentation broth i5 centri-fuged to remove cell growth, and the cell-free super-natant is stored in the cold until assayed. Supernatants obtained in this manner are found to contain 200 - 400 units of enterotoxic activity per milliliter, as de-termined by the modification of -the suckling mouse test -of Dean et al. referred to above.
The enterotoxic activity contained in the above cell-free supernatant is concentrated by precipi~ation with acetone, or preferably with ammonium sulfate. ~n a preferxed embodiment of the invention crystalline a~onium sulfate is 510wly added, with agitation at 1 - 10C, preferably at ~C, to a final concentration of about 90 per cent (NH4)2SO4, the resulting mixture is allowed to settle at 1 - 10C, preferably at ~C for 1~ ~ 24 hours, preferably overnight~ and the precipita-te is separated b~ centrifugation. The above precipitate, which contains more than 90 per cent of the total en-tero-toxic activity present in the cell free supernatant, is suspended in water or in a suitable buffer, preferably 0.05 M tris(hydroxymethyl)aminomethane buffer at pH 8.01 and filtered through an ultrafilter with a cut-off point of 10,000 - 30,000 daltons, preferably with a cut~off point of 30,000 daltons as such a filter gives satis~
factory fil~ration rates-without clogging and allows the-passage of substantially all of the enterotoxic activity while retaining about 90 per cent of inactive proteins.
The resulting ultrafiltrate is concentrated by flash evaporation or by freeze-drying, and the heat-stable enterotoxin contained in the concentrates obtained thereby shows a specific activity which is more than 200 times greater than that of the enterotoxin contained in the precipitate described above. Alternatively, by passing the cell-free supernatant through a non-ionic polymeric adsorbent, preferably of the type AMBERLIT~
XAD-2, and eluting it out of the column with an appropri-ate solvent mixture, preferably acidic a~ueous alkanol solution, similar results are obtained.
The concentrate obtained as above by flash evaporation is further purified by gel filtration on columns equilibrated with 5 per cent aqueous acetic acid, or the freeze-dried material obtained as above is dis-solved in 5 per cent aqueous acetic acid and further purified in the same manner. Suitahle adsorbents for gel ~iltration are chemically modified cross~linked dextrans, e.~. Sephadex G-25 (Pharmacia A.S~), or suitably cross-linked polyacrylamides, preferably Bio-Gel P 4 (Bio-Rad Laboratories, Richmond, Calif.) equilibrated with 5 per cent aqueous acetic acid. Elutioll with the same solvent and assaying of the fractions obtained thereby by the modification of the suckling mouse test referred to above shows that the enterotoxic activity is eluted in a single peak corresponding to one column volume. Pooling of active fractions followed by repeated lyophilization~
gives the purified heat-stable enterotoxin wlth a re-covery of about 90 per cent based on ~he cell-free super-natant, and with a specific activity about 3 5 times greater than that of the enterotoxin obtained by ultra-filtration as described above.
The purified heat~stable enterotoxin obtainedby gel filtration is further purified by high pressure liquid chromatography. It is further preferred to perform the high pressure liquid chromatography in two successive stages as follows.
In a first stage, a (substituted alkyl)silane e~uilibrated with a phosphate-alkylamine buffer at p~l 2.2 to 3.2 is used as the adsorbent in a reversed phase column, preferably cyanopropylsilane equilibrated with phosphate-triethylamine buffer at pH 3.0, and is eluted, wîth a linear gradient of an aprotic organic solvent such as a nitrile of a lower alkanoic acid, preferably aceto-nitrile, with monitoring the absorption of the eluates at 210 nm whereby the presence of the heat-stable entero-toxin is detected as a distinct and highly characteristic peak; fractions containing said peak are combined and assayed by the modification of the suckling mouse test referred to above and the results confirm the presence of the heat-stable enterotoXin in the above fractions, evaporation of the organic solvent followed by freeze-drying of the aqueous residue containing the phosphate-alkylamine buffer ~ives a mixture of the heat-stable enterotoxin with said buffer, substantially Eree from coloured contaminants.
In a second stage, said last-named mixture is applied to a reversed phase column containing as ad-sorbent an alkylsilane e~uilibrated with a vola-tile buEfer of pH 3.5 to 4.5, preferably octadecylsilane supported on microporous glass beads (Micro-Bondapak~
C18, Waters ~ssociates, Milford~ Mass.) equilibrated with~
0.01 M ammonium acetate buffer at pH ~.15, and :is eluted with a linear gradient of a lower alkanol, preferably methanol, with moni~oring the absorption of the eluates at 210 nm and assaying of the fractions within the area of peak absorption as descr~bed above. Pooling of the frac~ions with high absorption and with a high degree of enterotoxic activity followed by evaporation of the lower alkanol and freeze-drying of the residual aqueous solution gives the substantially pure heat-stable entero-toxin as a colourless fluffy substance. Said last-named material is found to be homogeneous when a sample thereof is subjected to the same procedure as described above for the second stage of high pressure liquid chromatography.
The over-all yield of the substantially pure heat-stable enterotoxin is about 6 - 8 per cent, based on the cell-free supernatant with a specific activity of about 7 ng per unit.
The substantially pure heat-stable enterotoxin retains its activity after heating in neutral solution to 100C for 30 minutes, but its activity is destroyed by heating to 121C for 15 minutes. It also retains its activity when kept at room temperature for 24 hours at pH 2 - 10, but its activity is destroyed under those con-ditions at pHIs above pH 10.
; 20 The above en-terotoxin is not affected by a number of enzymes such as pronase, bacterial alpha-amylase, and crystalline pancreatic lipase (all obtained from Calbiochem, San Diego, Calif.). Incubation of samples of the substantially pure heat-stable entero-toxin for 4 hours at 37C with 100 ~g/ml of pronase or of b~cterial alpha-amylase in 0.05 M phosphate buffer at pH 6.9, or with 100 mcg~ml of crystalline pancreatic lipase in 0.05 M tris(hydroxymethyl)aminomethane buffer at pH 8.1, followed by bioassay as described above showed that no detectable loss of activity had occurred.
The substantially pure heat-stahle enterotoxin is characterized by its amino acid composition as de-termined by amino acid analysis and its structure is established by determination of its amino acid sequence, as follows.
The amino acid analysis o-f the substantially pure heat-stable enterotoxin is carrled out by treating a sample (10 - 40 mcg) of the latter with 5.7 N hydro-chloric acid at 105C for 22 hours in the presence.of a reducing agent, preferably 0.1 per cent beta-mercapto-ethanol. The hydrolysate thus obtained is analyzed on a suitable commercially available amino acid analyzer~
Cysteine is preferably determined as cysteic acid after treatment of the substantially pure heat-stable en~ero-; toxin with performic acid in the manner described by Hirs in "Methods in Enzymology", vol. XI, pp. 53-62, Academic Press, New York, N.Y., 1967, followed by hydrolysis with 5.7 N hydrochloric acid at 105C for 22 hours.
The results o~tained in a number of determi-nations are shown below in Table I and ind;cate the presence of 18 amino acid residues; the latter n~ber is also confirmed by amino acid sequence studies as shown below.
The sequence of amino acid residues in the sub-stantially pure heat-stable enterotoxin is determined in a manner khown per se, by Edman degradation on a com-mercially available sequenator using an established program and a suitable carrier in the sequenator cup. An updated Beckman~ model 890B is the preferred sequanator, and the most suitable program is entitled "0.1 M Quadrol with Sl + S2 wash No. 121178", also supplied by Beckman Instxuments, Palo Alto~ Calif. The preEerred car.rier is a l/5-dimethyl-1,5-diaza-undecamethylene polymethobromide (Polybrene , Aldrich Chemical Co., Milwaukee, USA). It is furthermore preferred to perform the conversion step automatically immediately following the cleavage step, using a commercially available autoconver~er, preferably a Sequemat model P-6, for that purpose and carrying out the cleavage step with lo 5 N hydrochloric acid in methanol at 65C. The phenylthiohydantoin derivatives of the individual amino acids are then separated by liquid chromatography on a suitable adsorbent, preferably an ~ 13 -alkylsilane such as octadecylsilane supported on micro-porous glass ~ead~. In ~his manner the se~uence o~ all amino acids except cysteine is determined,--for example as shown in Fig. 5 When it is desired to determine the po~itions in the amino acid sequence occupied by cysteine residues, it is pre~erred to label the latter by treatment of the substan~ially pure heat-stable enterotoxin with a re-ducing agent, preferably dithiothreitol, and treating the reduced enteroto~in thus obtained with 3~-iodoacetic acid (New Englan~ Nuclear) in the manner described by Cxestfield et al. in J. Biol. ~hem. 238, 622 (1963).
The sequence of amino acid residues in the above entero-toxin which is now carboxymethylated and labelled with tritium on the cysteine residues is then determined in a manner known pe~ s~ and similar to that described above, using sperm whale apomyoglobin as an additional carrier as described by Seidail et al. in Proc. Natl. Acad. Sci.
U.S. 75, 3153 ~19783, and counting the individual fractions on a li~uid scintillation counter using a suitable scintillation 'cocktail' for determining the presence of tritium, preferably Aquasol (New England Nuclear). In this manner cysteine residues are localized in positions 5, 6, 9, 10 and 14 of the peptide chain, for example as shown in Fig. 4, which also shows that no carboxyme~hylation takes place when the above reaction with 3H-iodoacetic acid is carried out without prior treatment of the above enterotoxin with a reducing agent (see broken line in Fig. 4).
That -latter phenomenon may be due either to pairing of the cysteine residues, e.g. by formation of a dimer, or it may be due to stexic hindxance which makes at least one of the cysteine residues inaccessible to H-iodoacetic acid. In order to decide this point the sub-stantially pure heat-stable entero-toxin is treated for a very brief pèriod of time with iodine (125I), i.e. for 5 seconds to prevent oxidation of cysteine residues, using the chloramine-T method described by Greenwood et al. in Biochem. J. 89, 114 (1963), to obtain the corresponding iodinated enterotoxin which is purified by chromatography on a suitable cross-linked polyacrylamide, preferably Bio-Gel P-~. Chromatography of a mixture of said iodinated enterotoxin and of the reduced carboxymethyl-ated enterotoxin obtained as described above on a suitable cross-linked polyacrylamide, preferably Bio-Gel~ P-4, shows that the former is eluted later than the latter, thus eliminating the possibility that the substantially pure heat-stable enterotoxin could be a dimer.
The above data establish the structure of the substantially pure heat-stable enterotoxin o~ this invention as that of the octadecapeptide of formula (1) H2N-Asn-Thr-Phe-Tyr-Cys-Cys-Glu-Leu-fys Iys 10 (1) HOOC-Tyr-Gly-~la-Ala-Cys-Pro-Pro-Asn However, some of the data shown in Fig. 5 seem to indicate that a minor component may be present, giving rise to an (N-2) sequence which would appear to be present in certain sequence runs. Furthermore, there are also less prominent features in some sequence runs which could indicate the additional presence of (N-l) and (N-3) sequences.
The yields of individual amino acids in each cycle of the sequence determination are illustrated in Fig. 6r and the average repetitive yield is calculated as about 85 per cent. This low average repetitive yield may be due to the abrupt drop in yields follo~7ing the cysteine residues at positions 5 and 6, as also shown in Fig. ~.
The substantially pure heat-stable enteroto~in of formula (1~ of this invention is immunogenic and is useful for the preparation of antisera wh;ch neutralize its enterotoxigenic activity. Such antisera contain specific antibodies to the above entero-toxin and are useful for determining its presence in biological fluids by immunochemical assay methods.
The above enterotoxin of formula (1~ is furthermore useful for the preparation of vaccines which are useful in veterinary and in human medicine for immunization against E. ~o~i infections, in particular for the protection of the subjects against diarrhea caused by said E. coZi infections. Such vaccines are prepared in a manner known per se by mixing the above enterotoxin with a suitable pharmaceutically acceptable vehicle or carrier and are used for purposes of immuni-zation by injecting the vaccine so prepared subcutane-ously, intradermally, or intramuscularly. The above enterotoxin of formula (1) is also useful for the prepa-ration of toxoids which may be prepared therefrom in a manner known per se, for example by carefully controlled heating or by treatment with formaldehyde, and such toxoids are equally useful for the same purposes of immunization as discussed above.
When the above enterotoxin of formula ~1) or a toxoid prepared therefrom is employed as an immunizing a~ent in ma~nals, for example in laboratory animals such as rabbits, or in domestic animals such as plgs, goats, sheepl or bovines, or when it is used in human medicine~
it is preferably employed in combination with pharma-ceutically acceptable carriers or vehicles, the pro-portion of which is determined by the solubility and chemical nature of the compound of formula ~1), by the chosen route of administration, and by standard biologi-cal practice. For parent~ral administration by the sub-cutaneous, intradermal, or intramuscular routes the enterotoxin of formula (1) or the toxoids prepared therefrom may be used in the ~orm o~ sterile solutions or suspensions in pharmaceu~ically acceptable li~uid carriers such as water, ethanol, propylene glycol, or polyethylene glycol, containing other solutes or sus-pending agents, for example enough saline or glucose to make the solution isotonic, bile salts, acacia, gelatin, sorbitan monooleate, polysorbate 80 or "Tween ~0" (trade name for oleate esters of sorbitol an~ its anhydrides copolymerized with ethylene oxide), and the like.
Aqueous sterile solutions or suspensions are preferred, and suitable preservatives, for example methyl or propyl parahydroxybenzoate may be added as well as other solutes, for example sufficient sodium chloride or glucose to make the solution or suspension isotonic. ~he enterotoxin of formula (1) or a toxoid prepared there-from may also be administered in solutions or suspensions in sterile liquid carriers other than water, for example in suitable vegetable or animal oils, with or without the use of other solutes or of suspending agents as listed 20. above. The above compounds may also be used in finely divided solid form together with one or more finely divided pharmaceutically acceptable solid carriers, for example polyeth~lene glycol ~Carbowax~ 1540~, lactose, or very finely divided silica (Cab~0-Sil ), and may be applied directly to the skin or to a superficial lesion therein.
The dosage of the enterotoxin of formula (1) or of a toxoid prepared therefrom will vary with the form of administration and with the particular hos~ under treatment. For purposes oE immunization it is generally preferred to administer doses which will afford effective results without causing any harmful or deleterious side effects, and to administer such doses several times with intervals of 2 - 20 days until the desired titer of anti-bodies has been reached. As immunity against E. ~ozi is known to be transferred from the mother to -the fetus i-t is sometimes advantageous to immunize the animal or human mother at some time ante partum, preferably 2 - 8 weeks ante partum, by administering the enterotoxin of formu]a [1) or a toxoid prepared therefrom in the manner de-scribed above, whereby the neonatal animal or the human infant will be protected against E. coZi infections, and especially against the diarrhea caused by such in-fections. This procedure is especially valuable in pigs where an unusually high proportion of newly-born piglets are often seen to die from diarrhea. The dosage of the enterotoxin of formula (1) or of a ~oxoid prepared therefrom which will effect immunization upon repeated administration as described above is in the range of from about 1.0 ~g to abowt lQ00 ~g per kilogram body weight, although variations will occur as dlscussed above. How-ever, a dosage level which is in the range of from about 10 mcg to about 500 mcg per kilogram body weight is most desirably employed in order to achieve effective results.
The following non-limitative Examples will serve to lllustrate this invention.

A sample of Escherichia coZi Fll (P155~ ATCC
....... is inoculated into brain heart infusion (Dico ) and is incubated at 37C with agitation for ~ hours. An aliquot of said culture (0.5 ml) is used to inoculate 300 ml of the production medium containing the followiny constituents (in g/l): Na2HPO4r 5.0; K2HPO4, 5~0; N~4Cl, 1.18; Na2SO4, 0.089; MgC12.6H2O, 0.042; MnC12.4H~O, 0~004; FeC13.4H2O, 0 005; Casamino Acids (Difco~, 30.0;
yeast extract (Difco ), 6.0; glucose, 2.0; pH 7.3; the mixture so obtained is incubated at 37C without agi-tation overnight, and 250 ml of the culture thus obtained is used to inoculate 5 liters of the same production medium as above in a model 19 fermentor (New Brunswick Scientific Co.), equipped with an agitator, an aerator, an automatic pH controller, and an automatic foam con-troller. The resulting mixture is incubated at 37C for 7 hours with agitation at 500 rpm and aeration at 5 liters per minute, keeping pH 7.3 + 0.1. Samples are withdrawn at hourly intervals, and growth is determined by measuring optical density a~ 540 nm and by occasional plate counts. ~t the end of the incubation period the fermentation mixture is centrifuged at 10,000 x g for 30 minutes, and samples of the resulting cell-free super-natant are stored at -20C until assayed by the suckling mouse test which shows an enterotoxic activity of 400 units per milliliter for the cell-free supernatan-t, or 90 units per 10 cells; somewhat lower results are obtained when yeast extract ana/or glucose are omitted from the production medium.
Ammonium sulfate is added in small portions with vigorous agitation at 4C to the above cell-free supernatant until a final concentration of 650 g/1 is reached, the mixture is allowed to settle overnight, centrifuged at 12,Q00 x g for 45 minutes, the precipitate is suspended in 0.05 M tris(hydroxymethyl)aminomethane buffer at pH 8.0, the suspension is filtered through an ; Amicon~ PM-30 membrane, and the ultrafiltrate is concen-trated 10 times by flash evaporation. The above precipi-tate contains about 1.2 x 106 units of enterotoxlc activity, or about 90 per cent of the total activity present in the cell-free supernatant, ancl the above concentrate obtained from the ultrafiltrate contains abou-t 90 per cent of the activit~ present in the precipi-tate, or approximately 1.0 x 106 units with a specific activity of about 50 ng per unit, and is lyophilized.
XAMP~E 2 The lyophilized concentrate obtained as de-scribed in Example 1 is dissolved in 5~ aqueous acetic acid (3 ml) and the resulting solution is applied -~o a column (2.5 ~ 70 cm) of Bio-Gel~P-4 previously equi-librated with 5~ aqueous acetic acid. The column is eluted with 5~ acetic acidS the eluates are collected in 10 ml fractions with monîtoring of optical density at 270 nm, and each fraction is assayed for enterotoxic activity by the suckling mouse test. Active fractions are found to be eluted in a peak corresponding to one column volume, preceded by another peak of optical density at 270 nm whicll does not contain any enterotoxic activity. The active fractions are pooled, acetic acid is removed by lyophilization followed by disso~ving the solids thus obtained in water and lyophilizing again, and this procedure is repeated three times to obtain a solid which contains approximately 9 x 105 units of enterotoxic activity, for a recovery of about 90 per cent of the total activity of the concentrate obtained as described in Example 1, with a specific activity of about 14 ng per unit.

The solid obtained as described in Example 2 is dissolved in 0.2 M phosphate-triethylamine buffer at pH 3.0 and is further purified by high pressure liquid chromatography using a Waters Associates (Milford, Mass ) system consisting of two model 6000A pumps, a model UK6 injector, and a model 660 programmer for gradient elution.
Components are detected by monitoring the optical density of the eluates at 210 nm, using a Schoeffel model 770 variable wavelength flow spectrophotometer. The above solution of the solid obtained as described in Example 2 is applied to a reversed phase column of cyanopropyl-silane (0.4 x 30 cm3 equilibrated with a 0.2 M phosphate-triethylamine buffer at pH 3.0 and the flow rate is setat 1.5 ml/min. The column is eluted with a 25-rninute linear gradient of from 0% to ~0~ acetonitrile, 1 ml fractions are collected, and a distinct peak of ab-sorption a~ 210 nm is observed at 60~ acetonitrile concentration after 26 minu~es of elution time. The fractions contalning said absorption peak are assayed by the suckling mouse test and active fractions are pooled.
Acetonitrile is removed by flash evaporation, and the aqueous residual solution containing a mixture of the enterotoxin with the above phosphate-trie~hylamin~ buffer is lyophilized. The above procedure removes coloured \ impurities and is illustrated in Fig. 1, with the solid lines showing the acetonitrile gradient and the optical density of the eluates at 210 nm, and the broken line showing the results of the suckling mouse assay.

The lyophilized mixture of the enterotoxin with phosphate-triethylamine buffer obtained as described in Example 3 is dissolved in 0.01 M ammonium acetate buffer at pH 4.15 and is applied to a reversed pha~e column of octadecylsilane supported on microporous glass beads ~Micro-Bondapak C18, Waters Associates, Mil-ford, Mass.) previously e~uilibrated with 0.01 M ammonium acetate buffer at pH 4.15 using the same high pressure liquid chromatography equipment as described in Example 3. The above column is eluted with a linear gradient (1 hour) of rom 0% to 100% methanol at a flow rate of 2.0 ml/min.
and the optical density of the luates is monitored at 210 nm. ~ractions of 1 ml each are collected, and a disti~ct peak of absorption at 210 nm is observed at 80%
methanol concentration after 50 minutes of elution time, accompanied on either side by two minor peaks oE imp~lri-ties~ The fractions containing said distinct peak of absorption at 210 nm are assayed by the suckling mouse test, active fractions are found to be located a-t the center of said absorption peak and are pooled, and the pooled fractions are lyophi~ized to give 720 ~g of the substantially pure heat-stable enteroto~in as a fluffy colourless substance with a total activity of 80,000 to 100,000 units corresponding to an over-all recovery of 6 - 8 per cent calculated on the activity of the cell-free supernatant, and with a specific activity of about 7 ng per unit. The procedure is illustrated in Fig. 2 showing the absorption at 210 nm and the methanol gradient as solid lines, and the entero~oxic activity as determined by the suckling mouse assay as a broken line.
A sample of the above substantially pure heat-stable enterotoxin (20 ~g) is applied to a Micro-Bondapak~ C18 column (0.7 x 30 cm) and is sub-jected to high pressure liquid chromatography under exactly the same conditions as described above. The results are shown in Fig. 3 which demonstrates that the compound is essentially pure and homogeneous.
A sample of the above substantially pure heat-stable enterotoxin (7 ~g~ is dissolved in distilled water and the solution is found to contain about 1000 units by the suckling mouse assay. The solution is then heated to 100C for 30 minutes, assayed again as above, and is found to have the same activity as above. No enterotoxic activity is found after heating an aliquot of the solution prepared in the first instance to 121C for 15 minutes. All solutions are made up to their original volumes ater hea-ting.
A sample of t~e above substantially pllre heat-s~able enterotoxin is dissolved in water and the solution is assayed. Aliquots thereof are adjusted to pH 2 - 12 by addition of 2 M hydrochloric acid or 2 M
sodium hydroxide and are allowed -to stand at room temper-ature for 24 hours. Assays performed on the above aliquots after adjustment to pH 7.0 show that the above enterotoxin is stable at p~ 2.0 - 10.0 but that it losPs activity at higher pH's and that all activity is de stroyed at pH 12 under the above conditions.
Samples of the above substantially pure heat-stable enterotoxin of about 1000 units each are incubated for 4 hours at 37C in solution in 0.05 M phospha-te buffer at pH 6.9 with pronase and with bacterial alpha-amylase, respectively, and in 0.05 M tris(hydroxyme~hyl)-aminomethane buffer at pH 8.1 with crystalline pancreatic lipase. All enzymes are obtained from Calbiochem, San Diego, Calif., and their final concentrations are 100 ~g/ml. Assaying the above solutions before and after incubation shows that no activity had been lost.

The following test svstems for the detection of enterotoxic activity are used.
Suckling Mouse Test Swiss albino mice, 1 - 3 days old, are sepa-rated from their mothers jus-t before use and are randomly divided into yroups of three animals each. Samples con-taining heat~stable enterotoxic activity are assayed by making two-fold serial dilutions in physiological saline containing one drop of 2~ Evans Blue dye and 0.1 ml of each sample is administered via the oroesophagal route.
The mice are maintained at room temperature for 4 hours, then killed and their responses determined by the method of Dean et al., J. Infect. Dis. 125, 407 (1972). The last dilution which gives a response ~reater than 0.09 is considered as the end point and is expressed as the number of units per 0.1 milliliter.
Six-Hour Rabbit ~ejunal Loop Test ___ Young adult rabbits weighing abou-t 1.5 kg each are starved for 24 hours and jejunal loops are prepared therefrom in the manner described by I,ariviere et al~ in Can. J. Comp. Med. 36, 319 (1972). Samples to be assayed (2 ml each) are injected into 4 ~ 6 cm loops, and the rabbits are sacrificed 6 hours thereaftex. Results are evaluated according to the method of Evans et al. Infect.
Immun. 7, p.373, 1973 and only such results are accepted as positive which show a response with more than 0.5 ml of fluid per centimeter of loop, and negative results are also accepted from control loops.
EXA~LE 6 Amino acid analysis of the substantially pure heat-stable enterotoxin obtained as described in Example 4 is performed on samples of 10 - 40 ~g each, by treating the respective sample with 5.7 N hydrochloric acid at 105C for 22 hours in the presence of 0.1~ beta-mercapto-ethanol. The resulting hydrolysate is analyzed on a modified Beckman model 120C amino acid analyzer in which the modification consisted in using a microcolumn (37 x 0.6 cm) containing a Beckman W3 resin in the manner described by Fauconnet et al. in AnalO Biochem. 91, 403 (1978). For the determination of cysteine as cysteic acid the sample oE the above enterotoxin is first treated with performic acid and then hydrolyzed as described above, in the manner described by Hirs cited above, and the hydrolysate thus obtained is analyzed in the same manner as above. The results obtained in a number o runs, with and without performic acid oxidation of the sample, are shown in Table I, with the analysis values expressed according to the underlined amino acid.
TABLE I
20 Amino Acid Analysis Nearest Found by Integer Sequence Asx 1.6 2 2 Thr 1.0 Glx 0.3 Pro 2.4 2 2 Gly 1.0 Ala 2.0 2 2 Leu 1.0 Tyr 1.7 2 2 Phe 0.8 Cys 5.1 ~ 5 Presumed total: 18 18 The sequence of amino acid residues in the sub-stantially pure heat-stable enterotoxin is determined on an updated Beckman model 890B sequenator using 0.3 M
Quadrol (for N,N,N',N'-tetrakis(2-hydroxypropyl~-ethylenediamine) buffer at pH 9.0 and 3 mg POLYBRENE
(for 1,5-dimethyl-1,5-diaza-undecamethylene polymetho-bromide, Aldrich) in the cup. A sample of the dipeptide Leu-Val (100 nanomoles) is added and four simulated cycles are performed in order to block any aldehydes which could interfere with the subsequent sequence determination. A sample of the substantially pure heat-stable enterotoxin obtained as described in Example 4 (approximately 150 ~g) is then added to the sequenator cup and double coupling is performed in the first cycle only. The program used is Beckman's "0~1 M Quadrol with Sl + S2 wash No. 121178". The sequenator is equipped with a Beckman Sequemat~ model P-6 autoconverter and thus all conversions are done automatically immediately following the cleavage step, using 1O5 N hydrochloric acid in methanol at 65C. The phenylthiohydantoin derivatives of the individual amino acids are then sepa-rated by high pressure liquid chromatography on a column t0.46 x 25 cm) containing Altex 5 ~ ultrasphere ODS
(trade name for octadecylsilane supported on microporous glass beads of 5 ~ diameter~ as adsorhent, using a Waters Associate~s model 204 liquid chromatograph equipped with a Wisp 710 autoinjector, a data Module 730 integrator-plotter, and a 720 system controller. The phenylthio-hydantoin derivative of norleucine is added as aninternal standard (approxima-tely 10 8 moles are used for this purpose), and in this manner the positions of all amino acid residues in the sequence are established except those of the cysteine residues, and except tha-t the phenylthiohydantoin derivatives of glutamic and aspar~ic acids were deteeted as their respective methyl esters. The results of a nurnber of sequence determi-nations carried out in the manner described above are shown in Fig. 5.

The positions of the cys-teine residues in the amino acid sequence of the substantially pure h~at-stable enterotoxin obtained as described in Example 4 are determined as followsO
A sample of approximately 45 ~g of the above enterotoxin is reduced with dithiothreitol and carboxy-me-thylated with 3H-iodoacetic acid (New England Nuclear) in the manner described by Crestfield et al. cited aboveO
The carboxymethylated and tritium-labelled enterotoxin is then sequenced on an updated Beckman 890B sequenator using the Beckman "0.3 M Quadrol" program with 3.0 mg "POLYBRENE " in the cup and 2.5 mg sperm whale apomyo-globin as additional carrier as described by Seidah et al. cited above. The thiazolinones obtained at each sequenator cycle are counted directly on a Beckman liquid scintillation counter using "Aquasol 1 (New England Nuclear) as the scintillation 'cocktail' for tritium~
Results are shown in Fig. 4, which indicates that cysteine residues occupy the positions 5, 6, 9, 10 and 14 in the peptide chain o~ the above enterotoxin.
A n~ er of experiments are also carried out in the same manner as described above but omitting the initial reduction of the above enterotoxin with dithio-threitol. Although such experiments are carried out under more stringent conditions than described above, i.e. by prolonging the reaction with 3H-iodoacetic acid to four hours by increasing the reacti,on temperature to the boiling point, and in the presence or absence of sodium dodecyl sulfate, no carboxymethylation with 3H-iodoacetic acid takes place as evidenced by the broken line shown in Fig~ 4.
The localisation of the cysteine residues at &~

positions 5, 6, 9, 10 and 14 as shown in Fig. 4 in combi-nation with the results of sequencing described in Example 7 and illustrated in Fig. 5 establishes the structure of the substantially pure heat-stable entero-toxin obtained as described in Example 4 as -that of the octadecapeptide of formula (1).

A sample of the substantially pure heat-stable enterotoxin obtained as described in ~xample 4 (approxi-mately 18 ~Ig) is iodinated with 125I using the chlor-amine~T method described by Greenwood et al. cited above and a reaction time of 5 seconds to avoid possible oxidation of cysteine residues. The resulting product is separated from salts and other impurities by gel fil-tration on a column (1 x 48 cm) of Bio-Gel P-4 using 0.1 M ammonium bicarbonate buffer at pH 8.0 after addition of 2 mg sperm whale apomyoglobin as a carrier, to obtain a solution of the iodinated enterotoxin pre-sumably iodinated on the tyrosine residues. Said last-named solution is applied to a column (1 x 48 cm3 of Bio~Gel P-4 and eluted with 0.1 M NH4HPO3 pH 8Ø
When the same chromatography is carried out under exactly the same conditions as above with a sample of the tritium-labelled carboxymethylated enterotoxin the elution position of the iodinated peptide is found to be after that of the tritium-labelled carboxymethylated enterotoxin, thus excluding the possibility of the Pxistence of th~ substantially pure heat-stable entero-toxin of Example 4 in the form of a dimer, and confirming its structure as that of the octadecapeptide of formula .

Claims (7)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. A process for preparing the octadecapeptide of formula (1) (1) which comprises the following steps (a) incubating Escherichia coli F11(P155) ATCC....... in brain-heart infusion with agitation at 37°C for 4 hours to obtain a first stage culture;
inoculating a semi-synthetic medium containing a casein hydrolysate, sodium and potassium phosphate buffers, ammonium chloride, sodium sulfate, and trace elements together with yeast extract and glucose with a sample of said first stage culture and incubating at 37°C for 12 -18 hours to obtain a second stage culture; inoculating the same semi-synthetic medium as used in the second stage with a sample of the second stage culture and incubating with agitation at 37°C and with forced aeration at pH 7.2 - 7.8 for 4 - 7 hours; and separating the growth from the fermentation mixture, to obtain a cell-free supernatant;
(b) adding ammonium sulfate to said cell-free supernatant to a final concentration of 90 per cent (NH4)2SO4, allowing the mixture to settle at 1°- 10°C for 12 - 24 hours, and separating the precipitate;
(c) suspending said precipitate in a liquid selected from water and buffers at pH 8.0 and filtering through an ultrafilter with a cut-off point of 10,000 -30,000 daltons, to obtain an ultrafiltrate; and concen-trating said ultrafiltrate to obtain a concentrate;

(d) purifying said concentrate by gel fil-tration on a gel selected from the group consisting of chemically modified cross-linked dextrans and cross-linked polyacrylamides, to obtain a purified heat-stable enterotoxin;
(e) further purifying said purified heat-stable enterotoxin by high pressure liquid chromatography on a (substituted alkyl)silane equilibrated with a phosphate-alkylamine buffer at pH of 2.2 to 3.2, eluting with a linear gradient of an aprotic solvent selected from nitriles of lower alkanoic acids, monitoring the absorption of the eluates at 210 nm, assaying the eluates for enterotoxic activity with high absorption, pooling the fractions with high enterotoxic activities, evapo-rating the solvents and freeze-drying the aqueous residue, to obtain a mixture of the heat-stable entero-toxin with said phosphate-alkylamine buffer; and (f) further purifying said last-named mixture by high pressure liquid chromatography on an alkylsilane equilibrated with a volatile buffer at pH of 3.5 to 4.5, eluting with a linear gradient of a lower alkanol, monitoring the absorption of the eluates at 210 nm, assaying the eluates with high absorption for enterotoxic activity, pooling the fractions with high enterotoxic activities, evaporating the solvents and freeze-drying the aqueous residue, to obtain said octadecapeptide of formula (1).

2. A process for preparing the octadecapeptide of formula (1) (1) which comprises the following steps (a) incubating Escherichia coli F11(P155) ATCC....... in brain-heart infusion with agitation at 37°C for 4 hours to obtain a first stage culture;
inoculating a semi-synthetic medium containing a casein hydrolysate, sodium and potassium phosphate buffers, ammonium chloride, sodium sulfate, and trace elements together with yeast extract and glucose with a sample of said first stage culture and incubating at 37°C for 12 -18 hours to obtain a second stage culture; inoculating the same semi-synthetic medium as used in the second stage with a sample of the second stage culture and incubating with agitation at 37°C and with forced aeration at pH 7.2 - 7.8 for 4 - 7 hours; and separating the growth from the fermentation mixture, to obtain a cell-free supernatant;
(b) passing said cell-free supernatant through a non-ionic polymeric adsorbent and eluting same with an appropriate solvent to obtain a concentrate;
(c) purifying said concentrate by gel fil-tration on a gel selected from the group consisting of chemically modified cross-linked dextrans and cross-linked polyacrylamides, to obtain a purified heat-stable enterotoxin;
(d) further purifying said purified heat-stable enterotoxin by high pressure liquid chromatography on a (substituted alkyl)silane equilibrated with a phosphate-alkylamine buffer at pH of 2.2 to 3.2, eluting with a linear gradient of an aprotic solvent selected from nitriles of lower alkanoic acids, monitoring the absorption of the eluates at 210 nm, assaying the eluates for enterotoxic activity with high absorption, pooling the fractions with high enterotoxic activities, evapo-rating the solvents and freeze-drying the aqueous residue, to obtain a mixture of the heat-stable enterotoxin with said phosphate-alkylamine buffer; and (e) further purifying said last-named mixture by high pressure liquid chromatography on an alkylsilane equilibrated with a volatile buffer at pH of 3.5. to 4.5, eluting with a linear gradient of a lower alkanol, monitoring the absorption of the eluates at 210 nm, assaying the eluates with high absorption for entero-toxic activity, pooling the fractions with high entero-toxic activities, evaporating the solvents and freeze-drying the aqueous residue, to obtain said octadeca-peptide of formula (1).

3. A process as claimed in Claim 1 or 2, in which the ultrafilter has a cut-off point of about 30,000 daltons.

4. A process as claimed in Claim 1 or 2, in which the cross-linked polyacrylamide is Bio-Eel? P-4.

5. A process as claimed in Claim 1 or 2, in which the (substituted alkyl)silane is cyanopropylsilane, the phosphate-alkylamine buffer is phosphate-triethyl-amine buffer at pH 3.0, and the aprotic solvent is acetonitrile.

6. A process as claimed in Claim 1 or 2, in which the alkylsilane is octadecylsilane and the volatile buffer is ammonium acetate buffer at pH 4.15.

7. The substantially pure heat-stable entero-toxin of Escherichia coli F11(P155) ATCC......., said enterotoxin being the octadecapeptide of the formula (1) (1) characterized by a high pressure liquid chromatography retention time of 50 minutes as shown in Figure 3, when prepared by the process defined in Claim 1 or by an obvious chemical equivalent.