Cloning, sequencing, and expression in ficoll-generated minicells of an Escherichia coli heat-stable enterotoxin gene

Cloning, sequencing, and expression in ficoll-generated minicells of an Escherichia coli heat-stable enterotoxin gene

PLASMID 20,42-53 (1988) Cloning, Sequencing, and Expression in Ficoll-Generated Minicells of an Escherichia co/i Heat-Stable Enterotoxin Gene H. ...

1MB Sizes 4 Downloads 39 Views




Cloning, Sequencing,

and Expression in Ficoll-Generated Minicells of an Escherichia co/i Heat-Stable Enterotoxin Gene



*Department ofMicrobiology, University of Texas Southwestern Medical Center, Dallas, Texas 75235, -fDepartamento de Genetica, CINVESTA V, Mexico, D. F. Mexico, and SCentro de Investigation sobre Ingenieria Genetica y Biotecnologia, UNAM, Cuernavaca, Mar. Mexico

Received March 3, 1988; revised May 20, 1988 The gene encoding a heat-stableenterotoxin of Escherichia coli was cloned as a 960-bp fragment from a plasmid isolated from a Mexican strain of human origin. Deoxyribonucleotide sequencing unveiled a 2 16-bp open reading frame similar to that of a previously sequenced ST-toxin gene. The gene is preceded by a proposed binding site for the CAMP-mediated positive regulator (CAP) that is part of a 23-bp inverted repeat. The proposed CAP site is followed by a 6A, IT, and 6A deoxyribonucleotides. Minicells containing the toxin gene, which were isolated from Ficoll gradients, shown to preserve the localization of intracellular and pcriplasmic enzymes, allowed the detection of a biosynthetically radiolabeled polypeptide with an apparent M, 8400. The data suggestthat the enterotoxin genesestA2, estA3, and estA4 are very similar, even in clinical strains isolated from distinct geographical locations; that the transcription of heat-stable enterotoxin genes is controlled by the CAMP-mediated positive regulatory system, and that the heat-stable enterotoxins are initially synthesized as 72 amino acid precursors to yield the extracellular active 18-l 9 amino acid polypcptides. 0 1988 Academic Press. Inc.

Two types of enterotoxins encoded by plasmids of enterotoxigenic Escherichia coli (ETEC)2 have been shown to cause secretory diarrhea in animals and humans: the heat-la-

bile (LT) and the heat-stable (ST) enterotoxin families. ST toxins purified from bacterial supernatants of ETEC have been shown to be of two types: a methanol soluble toxin (ST,+ or ST,) that causes positive responses in the suckling mouse model and in porcine jejunal Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under Accession loops (Whipp et al., 198 1) and a methanol insoluble toxin (STn or STrr) that is positive only No. JO3311. ’ Presentaddress:Harold C. Simmons Arthritis Research in porcine jejunal loops. Gene estA1, cloned Center, Department of Internal Medicine, University of and sequenced from a plasmid of bovine Texas Southwestern Medical Center, Dallas, TX 75235. ETEC strain, B41, was shown to be in trans2Abbreviations used: ETEC, enterotoxigenic Escheposon Tnl68 1 (So and McCarthy, 1980); richia coli; Ap’, ampicillin resistance;Tc’, tetracycline resistance; Cm’, chloramphenicol resistance; SDS, sodium estA2 was cloned from a human ETEC strain dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; (CRL 25090) isolated in Bangladesh (DeWilde TCA, trichloroacetic acid PBS, phosphate-bufferedsaline; et al., 1981); estA3 was cloned from strain DATD, N,N’-diallyltartardiamide; ST,, methanol soluble 153837-2, a human ETEC also isolated in heat-stable enterotoxin; ST,, methanol insoluble heatBangladesh (Moseley et al., 1983). Subsestable enterotoxin; estA and estB are the respective genes. quently, Tn 168 1 has been shown to be in Alleles are denoted by a number, i.e., STI\.t for the polyETEC strains of bovine, avian, and porcine peptide and esfA1 for the gene. We prefer these abbreviations and nomenclature as opposed to the one that refers origins (Sekizaki et al., 1985). The inferred seto the human or animal origin of the ST-ETEC strain quences at the carboxy termini of two of these (e.g., ST-H for human and ST-P for porcine) because it gene products have been demonstrated to be relates to the physical properties of the polypeptides and to the amino acid and nucleotide sequencehomologies of accurate by the determination of the amino acid sequence of the extracellular active enthe toxins and genes,respectively. 0147-619X/88


Copyright Q 1988 by Academic Press, Inc. All rights of reproduction in any form reserved.




terotoxins: the 18 amino acids of ST.+r, from purified culture supematants of human (Ronnberg et al., 1983; Takao et al., 1983; Thompson and Gianella, 1985),bovine (Saeed et al., 1984), and porcine &azure et al., 1983; Sekizaki et al., 1985) strains and the 19 amino acids of STAe3,in supernatants of human strains (Aimoto et al., 1982). The metabolic processing of ST* has been difficult to study because the toxins do not react with conventional protein stains and by themselvesare not immunogenic. In this paper we describe the cloning and sequencing of an estA4 isolated from an 80-MDa plasmid derived from a Mexican ETEC strain of human origin (Stieglitz et al., 1980) and the comparison of this sequenceto other sequencedestAs. Minicells isolated by centrifugation in Ficoll gradients, a method that is shown to respect the natural topology of intracellular and periplasmic enzymes, were used to expressestA4. Their use enabled the detection of an intracellular precursor of estA4. MATERIALS


Bacterial strains, plasmids, and media. The E. coli strains used were 55-4 (pro22, met63, Nal”) (Stieglitz et al., 1980), JM83 (Am, Alacpro, rpsL, Thi, 480d, AlacZi’kfl5) (Vieira and Messing, 1982), RR1 (F-, hsdS20, ara-Z4, proA-2, lacY1, galK2, rpsL20, ~~1-5, mtl-1, supE44, X-) (Covarrubias et al., 198l), JK8306 (HfrH, thi, Acya, rpsl) (Martinez-Cadena et al., 1981), Xl 122 (minA, minB, Thr, Leu, Thi, Su, X’) (Frazer and Curtiss, 1975) and Xl 144, a cysteine auxotroph induced by nitrosoguanidine treatment of strain Xl 122. The source of the ST gene was the 80-MDa plasmid pYKOO7 (Stieglitz et al., 1980) isolated from a clinical human ST+ E. coli strain (serotype 0166:H27). The cloning vehicles were the ColEl derivatives pBR328 (Ap’, Tc’, Cm”), pBR329-2 (Ap’, Tc’, Cm”) (Covarrubias et al., 1981) and pUC8 (Vieira and Messing, 1982). Plasmid pBR328A (Tc~ was obtained by treatment of pBR328 with EcoRI and PstI, followed by S1 nucleasetreatment and ligation of the blunted ends; this DNA was used to selectTc’ transformants of strain 55-4. T broth





and L medium (Martinez-Cadena et al., 1981) were used to determine the production of ST enterotoxin and for plasmid isolations and transformations, respectively. ML broth (Curt&, 1965) [NH&l (0.5%), NH4NO3 (O.l%), Na2S04 (0.2%), &HP04 (0.9%), w2P04 (0.3%), MgS04 * 7H20 (O.Ol%), glucase (0.5%), Difco (Detroit, MI) vitamin free casamino acids (0.5%), supplemented with vitamins when needed] was used to grow the minicell-producing bacteria; the minicells were radioactively labeled in Difco cysteine assay media. When needed, penicillin (200 &ml), tetracycline (25 pg/ml), and/or chloramphenicol (25 pg/ml) were added to the broths or agar plates. DNA manipulations and cloning of estA4. Plasmid pYKOO7 was isolated as described previously (Stieglitz et al., 1980) by lysing the cultures with lysozyme-EDTA-Triton X- 100, alkali denaturing, and absorbing the singlestranded DNA to nitrocellulose powder; the double-stranded DNA was precipitated from the solution with ethanol-sodium acetate. Cloning vehicles and recombinant plasmids were isolated from large (500 ml) or small (5 ml) cultures by lysis with NaOH-SDS after lysozyme spheroplasting of the cultures as described previously (Guzman-Verduzco and Kupersztoch, 1987). Supercoiled plasmid DNA was separatedfrom linear and open circular DNAs by ethidium bromide-CsCl isopycnic density centrifugation (Guzman-Verduzco et al., 1983), and the DNA solutions were digested with restriction endonucleases and ligated following the recommendations of the manufacturers. Restriction analysis of DNA fragments was done by agaroseor polyacrylamide gel electrophoresis. When necessary, the DNA restriction fragments were recovered from the gels by excising the fluorescent bands and electroeluting the DNA using ISCO cups (Lincoln, NE). The electrophoresis chambers were filled with TEA buffer, pH 8.1 (50 mM Tris, 0.1 IIIM EDTA, and 100 mM sodium acetate), and the cup was filled with TE buffer ( 10 mM Tris, 1 IIIM EDTA); the gel slice was submergedin the large compartment of the cup and electroelution at 4°C was at



100 V for 1.5 h. Two hundred microliters was withdrawn from the small chamber and electroelution continued 1.5 h more. Another 200 ~1was withdrawn from the small chamber and mixed with the first 200 ~1 and the solution was extracted with an equal volume of butanol. The DNA was precipitated from the solution with 2.5 vol of ethanol and the DNA resuspended in TE buffer to a concentration of 1 rg/pl. Bacteria were transformed by the CaClz-RbCl method as described previously (Guzman-Verduzco and Kupersztoch, 1987). E&ST was cloned as diagrammatically representedin Fig. 1. PstI-restricted pYKOO7 and pBR328 DNAs were ligated (Fig. 1, arrow 1). The plasmid (pHS-1) of a Cm’ transformant of strain RR1 was characterized as ApS Tc’ Cm’ ST+ and purified. PstI digestion of pHS1 revealed two fragments of 1500 and 11,000 bp, respectively, derived from pYKOO7 which were recloned into pBR328; the ST activity was localized in the 1l,OOO-bp fragment (pYK102, Fig. 1, arrow 2). EcoRI treatment of pYKl02 followed by religation eliminated a 4600-bp fragment (1180 bp of the vehicle, 3420 of the insert) to yield pYKl03 (Tc’ST+) (Fig. 1, arrow 3). A 4000-bp PstI-BamHI fragment resulting from a PstI and partial BamHI digestion of pYKl03 was cloned into pBR329-2 treated with the sameenzymes(Fig. 1, arrow 4) resulting in pYK 104 (Ap’ Cm’ Tc” ST+) from which a 2850-bp S&I fragment was isolated and ligated into the San site of pBR328 (Fig. 1, arrow 5) to yield pYKlO5 (Ap’ Cm’ Tc” ST+). A 1390-bp HincII fragment (HincII cuts the San site) from pYKlO5 was ligated into the HincII site of pUC8. Ap’ transformants of JM83, selectedas white colonies on L-Xgal-Ap plates, were assayedfor ST and when cut with BamHI both orientations of the HincII insert were detected (Fig. 1, arrows 6 and 7). Plasmid pYK109 was isolated from Ap’ lac- transformants of JM83 by ligating the 960-bp BamHI fragment of pYK107 into the BamHI site of pUC8 (Fig. 1, arrow 8). As previously reported for pYKOO7 (Martinez-Cadena et al., 1981), the enterotoxin activity of all the ST+ plasmids illustrated in Fig. 1 was shown to be under the


control of the cyclic-AMP-linked positive regulatory system and all exhibited thermoactivation of the periplasmic ST-like activity previously reported (Guzman-Verduzco et al., 1983). Even though 554 (pYK109) produces about 50 times more toxin than 554 (pYKOO7), we were unable to reproduce the published staining of the toxin with Coomassie blue G-250 (Pesti and Lukacs, 1981) after SDS-PAGE of culture supematants. DNA sequencing.The sequence of the STDNA was defined by the chemical cleavage method of Maxam and Gilbert (1980) and confirmed by the dideoxy-chain termination method as described by Sanger et al. (1977). Purified pYK109 (Fig. 1) DNA was cleaved separately with EcuRI or HindIII. The linearized DNA was dephosphorylated with alkaline phosphatase and end-labeled by incubation with [cz-~*P]ATP and Tq polynucleotide kinase. The EcoRI-cleaved radiolabeled DNA was then digested with Hind111 and the HindIII-cleaved radiolabeled sample was recleaved with EcoRI. The two fragments were then purified by PAGE, electroeluted, and sequenced.The sequencingdata obtained by this strategy unveiled an AvaII site 322 basesfrom the Hind111 site. This AvuII site was used to sequence from within the 960-bp BamHI fragment. Plasmid pYKl09 DNA wascut with AvaII, dephosphorylated, 32P-labeled as described above, and cleaved separately with Hind111or EcoRI. The 1245-bp A&I-EcoRI and the 1390-bpAvaII-Hind111 fragments derived from plasmid pUC8 were separated by PAGE from the 640-bp AvuII-EcoRI and the 322-bp AvuII-Hind111 fragments of the estA BumHI insert. The two smaller estA radioactive bands localized by radioautography of the gel were excised and sequenced.The sequencing experiments were repeatedthree times and confirmed by cloning the 960-bp BumHI-ST+ fragment into M 13Mpl8 and resequencedby the dideoxy-chain termination method using as primers a synthetic 20-nucleotide oligomer (Guzman-Verduzco and Kupersztoch, 1987) complementary to the 3’ terminus of the toxin in one direction and in the other, the universal Ml3 primer.

pYK 102




pYK 105


FIG. I. Cloning of the ST enterotoxin: purified plasmid DNAs were treated with the restriction endonucleases indicated; the DNA fragments eliminated during the cloning steps are shown as dotted lines and the solid lines indicate the regions that were subcloned. These fragments were ligated to vector DNA pretreated with the restriction endonudeases represented in the figure and the ligation mixture was used to transform recipient bacteria. Transformants using the pBR piasmids were selected on L agar plates supplemented with antibiotics whose resistance markers were carried by the cloning vectors (steps 1 to 6). When the vector was pUC8, transformants were selected on L-Xgalampicillin agar plates; the candidates with insertions were further analyzed. The wavy line shown in plasmid pYK104 and pYK.105 is vector DNA carried in the subcloning steps.



14 DNA



Minicell isolation and analysis of plasmidencoded polypeptides. Cultures of minicell

producing strains were grown overnight at 37’C in ML broth. The suspensionswere diluted loo-fold in the same media and incubated at 37°C for 18 h. The bacteria were harvested by centrifugation (GSA rotor, Sorvall, DuPont Co., Wilmington, DE) 6598, 5 min, 4°C and the minicells in the supernatant were pelleted in the same rotor, 15 min at 11,9OOg, 4°C. The pellet was resuspendedand homogenized in 20 ml of BSG (Cur&s, 1965) (NaCl, 0.85%; KH2P04, 0.03%; Na2HP04, 0.06%; and gelatin, 0.00 1%)per liter of initial culture. Two and one half milliliters of the suspension was applied to 27 ml of a 5-l 5% Ficoll-400 gradient (Pharmacia Fine Chemicals, AB, Uppsala, Sweden) in water or a lo-35% sucrose gradient in 0.1 M NaCl, 0.0 1 M TrisHCl, 0.005 M EDTA, pH 7.5 (TES). The gradients were centrifuged 20 min at 4°C in an HB4 rotor at 4 124g. The upper band was collected by suction from the top of the tube, diluted twofold in BSG, and harvested by centrifugation in a SS34rotor at 26,89Og,4°C for 15 min. The pellet was resuspended in BSG and the gradient purification was repeated twice. The final minicell pellet was resuspended in cysteine assaymedia, and the concentration adjusted to an ODZeOof 2 (2 X 10” minicells/ml); typically 500 ml of culture yielded 3 ml of minicells (ODzeO= 2). One hundred microliters was used to determine bacterial contamination by viable count; the preparation was used for experiments only when the cellular contamination was lessthan 1 colony-forming unit per lo6 minicells. The minicells were preincubated at 37°C for 30 min, followed by the addition of [35S]cysteine (25 &i/ml final) (New England Nuclear, Boston, MA). The incubation was continued for the times indicated, after which cold cysteine (0.5% final) was added and the minicell suspension was frozen in a dry ice-acetone bath in 500~~1aliquots and stored at -70°C until further use. The aliquots were thawed and washed twice by centrifugation in the same media (with cold cysteine) at 4°C in an Eppendorf centrifuge (Model 54 14, Eppendorf


Gemtebau, Hamburg, Germany). The pelleted minicells were resuspendedin 100 ~1of cracking buffer (10 mM NaH2P04, 6 M urea, 1% SDS, 1%2-mercaptoethanol, and 0.1% bromphenol blue) and heated for 5 min at 93°C. Aliquots ( 10 ~1)were used to determine TCA precipitable counts on GF/C Whatman glass fiber filters that had been presoaked in a 0.1 M L-cysteine solution in 10%TCA. The radioactivity in the filters was counted by scintillation. The [35S]cysteine-labeledpeptides were separatedby applying 30 &well of the samples to a 1.5-mm-thick 15% slab polyacrylamide gel (Shapiro et al., 1967) [substituting N,N’diallyltartardiamide (DATD) for bisacrylamide, polyacrylamide:DATD (40:4)], dissolved, and polymerized in 0.1 M NaH2P04, pH 7.2, 0.1% SDS, 6 M urea, 0.05% ammonium persulfate, and 0.024% TEMED. The gel was electrophoresed at 5-6 V/cm for 16- 18 hrs in 0.1 M NaH2P04, pH 7.2 and 0.1% SDS. The gels were stained and fixed with Coomassie blue G-250 (Pesti and Lukacs, 1981) in a 12% TCA solution for 3 h, followed by destaining in water. The gels were prepared for fluorography using dimethyl sulfoxide and 2,5-diphenyloxazole and exposedto Kodak X-Omat film at -70°C. Determination of ST activity. The ST assay was done by the suckling mouse model as previously described (Martinez-Cadena et al., 1981). Thermoactivation of periplasmic ST was at 65°C for 30 min (Guzman-Verduzco et al., 1983).The effectof CAMP on ST activity was assayedsupplementing the growth media of strain JK8306, host of the plasmids under study, with 3 n’IM CAMP as described previously (Martinez-Cadena et al., 1981). Fractionation of spheroplasts and periplasm.

Osmotic shock treatment of cells was previously described (Guzman-Verduzco et al., 1983): cells were resuspendedin 20% sucrose, 0.03 M Tris, pH 8, followed by the addition of EDTA to 10e3M; the suspension was centrifuged; and the pellet resuspendedrapidly in distilled water. The sampleswere centrifuged; the supernatant and pellet are referred to as osmotic shock fluid (periplasm) and spheroplast, respectively. Alternatively, spheroplasts





Detection of ST, enterotoxin precursors in minicells. The most widely used support to purify minicells from minicell-producing bacteria is sucroseas a gradient dissolved in a variety of buffers (reviewed in Frazer and Curt&, 1975).We examined the effectof someof these buffers and of sucrose concentrations on the localization of intracellular (&galactosidase) and periplasmic (alkaline phosphatase) enRESULTS zymes. Bacteria were grown under the conSequence of EntST. We sequenced the ditions describedfor minicell purification. The cloned 960-bp fragment that includes estA4 suspensions were pelleted and the bacteria (Fig. 2). The open reading frame (positions were converted to spheroplasts.The periplasm 496 to 711) is shown as triplets, and under- and the spheroplasts were separated by cenneath are the inferred amino acids (ST&. The trifugation and the enzyme activities tested in hypothetical binding site for the positive tran- both fractions. As shown in Table 1, the major scriptional regulator CAMP receptor protein- effect of the buffers (TES and PBS)was on the CAMP (CAP), the proposed -35, - 10, + 1, and localization of P-galactosidase;after 80 min of the ribosomal binding site (SD) are also in- incubation (estimated time to perform the pudicated. A potential transcriptional termina- rification of minicells), a significant percentage tion loop (Ter) is shown at position 7 16-749. of the total P-galactosidasewas displaced to The proposed CAP site is in the base of an the periplasm. The same buffers containing inverted repeat (Fig. 3) that is followed by 6A, sucrosehad a similar effect. Sucrosedissolved 1T, and 6A deoxyribonucleotides. The role of in M-9 media also displaced more than 50% these elements in the transcription of estAI of the /3-galactosidaseto the periplasm, while mediated by the positive effecters CAMP and M-9 alone did not. CRP is unknown. An alternative support was sought to purify The hydropathic profile (Kyte and Doolitle, minicells. 20% Ficoll in either M-9 media or 1982) of the inferred amino acid sequenceof water did not affect the localization of either STAm4(Fig. 4) shows that only the first 20 enzyme (Table 1). Lower Ficoll concentrations amino acids share the properties of a mem- also did not alter their localization (data not brane spanning domain; the same region has shown). We chose to use Ficoll dissolved in the characteristics of a signal peptide (Von water since the gradients can be formed by Heijne, 1985). This segment of STA4 is fol- freeze-thawing a solution of Ficoll without the lowed by a sequence(positions 23 to 43) that generation of a salt gradient that results when is rich in basic amino acids and serine and the same procedure is performed on an M-9lacks aromatic amino acids; it is predicted Ficoll solution. The correlation between index (Garnier et al., 1978) that this region has a of refraction (n) and specific gravity (D:$ was helical structure. The plasmid-encoded extra- determined experimentally for Ficoll-water cellular STAm3 starts at its amino terminus with solutions and they are related by the equation asn (Aimoto et al., 1982) and, due to the se- n = 0.2089 D:8 + 1.264. The localization of quence identity in this region, STAm4 should P-galactosidaseand alkaline phosphatase was yield the same amino terminus; thus, the pre- also determined in minicells isolated in cursor should be cleaved between mets3 and aqueous sucrose and Ficoll gradients; the asns4.The resulting peptide is identical to the minicells were converted to spheroplasts by amino acid sequence determined by Aimoto lysozyme-EDTA treatment since the osmotic et al. (1982). Six cysteines are part of the 19 shock procedure (Guzman-Verduzco et al., amino acids that constitute the biologically 1983) was not as effective. Similarly to strain active extracellular enterotoxin. J5-4, 58.3% of the total P-galactosidasewas were formed by the lysozyme-EDTA treatment of bacteria (Guzman-Verduzco et al., 1983). Computerized data analysis. DNA sequence analysis and the conformational projections of peptides were done using DNAstar (Madison, WI) and Microgenie (Beckman Instruments, Palo Alto, CA).



FIG. 2. Deoxyribonucleotide sequence (estA4) and inferred amino acid sequence (ST,..+)of the 960-bp BumHI fragment that codes for STA4 and comparison with other methanol soluble heat-stable enterotoxin genes(estA2, DeWilde et al., 1981, complete sequence),estAl (So and McCarthy, 1980,their positions 136 to 534), and their inferred polypeptides (ST,.* and ST,.,). Deoxyribonucleotides are numbered above estA4 and amino acids under ST,., . The asterisks in estAl and estA2 frame the compared sequences.Shown in upper caseare the basesin estA2 and estA1 that differ from estA4, and in lower caseare insertions in estA1 (between estAl positions 449 and 450 and 458 and 459) and estA2 (between positions 427 and 428). Overand underlined are the hypothetical crp-CAMP recognition site (CAP) and the proposed -35, -10, +I Shine-Delgamo sequences(S.D.), the start codon, and the end of the open reading frame. The open reading frame is shown as triplets from position 496 (ATG) to position 711. The proposed gene ends with two termination codons (positions 7 12-7 17) followed by a putative transcriptional termination loop (underlined TER). Empty spacesindicate identities of deoxyribonucleotides and amino acids. Capatalized in ST,, are the conserved amino acids and in lower case are represented the amino acids that differ in ST,., and STA.2.The solid underlined peptide is identical to the sequence of the extracellular toxin (Aimoto et al., 1982). Extracellular ST,., starts with amino terminal asn$,. A gene, estA3, almost identical estil, has been sequenced(Moseley et al., 1983);it differs from esfA4in nucleotides484 (deletion of T) and 550 (C substitutes G); the latter change replaces ala,!, for pro.

recovered in the periplasm of minicells purified in sucrose gradients. Both preparations showed more than 86% of the alkaline phosphatase activity in the periplasm. The inferred amino acid sequence of ST, shown in Fig. 2, and the amino acid sequence

determined in the extracellular ST (Aimoto et al., 1982;Saeedet al., 1984;Takao et al., 1983; Thompson and Gianella, 1985) show a remarkably high content of cysteine (6 of 18 or 19 amino acids). This property was used to enrich the labeling of ST* with radioactive




(Stieglitz et al., 1980). The open reading frame is precededby a proposed Pribnow box located from - 19 to - 12 bases from the putative T G TA transcription start point; its (Pribnow, 1975) CG TA sequence,TCAAAT, is partially homologous CO to the consensussequenceTs,&T45&&,T96 G T-400 TA (McClure, 1985; Rosenbergand Court, 1979). 380-A T TA The proposed -35 sequence,TTGCGC, is 18 CO bp upstream from the Pribnow box and shares AT TA homology with the consensus sequence T 1 A 6 (T82Ts4G7sA&4&5) recognized by the comTA monly used u factor of the E. coli RNA polyA T-410 A G merase (McClure, 1985). The proposed ribo370-T T AT somal binding site, CGGAGG, is four bases TTTTTTCGGTCGCCGAAAAGCATCACAAAAAAAATAAAAAA I from the ATG met codon and differs only in 3Ao 3Ao 420 the first base from the consensus AGGAGG FIG. 3. A 23-bp inverted repeat that includes part of the (Shine and Dalgarno, 1975). Sixty bases upproposed CAP site (dark letters). In the coding strand shown,the CAP site is followed by 6A, lT, and 6A residues. stream, a proposed CAP site (TAGTATGATGTTCATCACAAA) shows similarity to The Gibbs free energy was calculated to be - 11.8 kcal. the consensus sequences for CAP (AT2As9A




cysteine. Figure 5 shows the time dependent incorporation of [‘%]cysteine into minicells purified in sucroseor Ficoll gradients. It is seen that Ficoll-purified minicells incorporated [35S]cysteinemore efficiently than those purified in sucrose gradients. Fic&gradient-purified minicells horn stmins X1 144 (pBR328A) and X1 144 (pYK103) (the samepBR328A with a 6400 bp ST+ fragment) were incubated separately with [35S]cysteine and fractionated in spheroplast and periplasm and the polypeptides separatedby PAGE. The lysates of the ST+-producing spheroplast (Fig. 6, lane 2) showed a band with an apparent M, 8.4; this band was absent from the ST- lysates. No ST-related peptide was detected in the periplasm. After a 1-min pulse with [‘%Icysteine, followed by a cold chase,the 8.4 kDa chased very inefficiently (data not shown); attempts to localize the STAd by SDS-PAGE of minicell supematants were unsuccessful. DISCUSSION

The methanol soluble heat-stable enterotoxin is one of the few fully exported polypeptides.encoded naturally by E. coli; we cloned and sequenced an estA# gene (Figs. 1 and 2) from a human ETEC strain isolated in Mexico


NA61T55TSO); this site is in the base of the inverted repeat illustrated in Fig. 3 and is followed by 6A, lT, and 6A residues. This sequence is not common to other CAP sites (Barbaud and Schwartz, 1984). The extracellular ST,., (Alderete and Robertson, 1977) and STAM4 (Martinez-Cadena et al., 1981) activities have been shown to be affected by CAMP and the CAMP receptor protein. estA4-mRNA was not detected in Acrp

40 2 2


u F 2


e I





I 0





FIG. 4. Hydropathic profile of ST,, (kindly performed by J. Kyte and R. Doolittle). The plot was made with a window of seven residues; the hydropathic index cutoff value for a membrane spanning sequence is 11.2. Only the first 20 amino acids are thought to interact with a membrane.




Alkaline phosphatase % activity Treatment (time) None G-4 (b) TES (a) 0 40 min Sucrose 10%in TES (a) 0 40 80 PBS (b) 0 80 min Sucrose 10%in PBS (b) 0 80 min Sucrose20% in PBS (b) 0 80 min M-9 (b) 80 min Sucrose 10%in M-9 (b) 0 40 Ficoil20% in M-9 (b) 0 40 80 Water 80 min (a) FicolI20% in water (a) 80 min

&Galactosidase % activity





82.6 90.1

15.7 9.9

16.3 16.5

83.7 83.5

81.1 90.8

18.9 9.2

32.0 81.3

68.0 18.8

87.9 90.7 90.1

12.1 9.3 9.9

26.7 35.7 41.3

73.3 64.3 58.7

93.5 89.4

6.5 10.6

29.8 37.4

70.2 62.7

91.4 90.8

8.6 9.8

21.5 20.0

78.5 80.0

92.3 90.2

7.7 9.8

36.0 54.9

64.0 45.1





89.2 87.2

10.8 12.8

46.4 59.8

53.6 40.2

88.0 82.5 89.4

12.0 17.5 10.6

18.4 19.7 18.6

81.7 80.3 81.4









Note. Bacteria from exponentially growing cultures (5 X 10scells/ml) of strain 554 were harvested by centrifugation and resuspendedin the appropriate solutions for the times indicated. Spheroplast and periplasm were obtained and enzymatic activities assayedasdescribedunder Materials and Methods. (a) and (b) refer to data obtained in two different experiments.

mutants, but a low level was seenin Acyu mutants (Guzman-Verduzco et al., 1985). The inverted repeat, the A/T rich regions that follow the CAP site, and/or the specific sequence of this CAP site could account for the leakiness observed in Acya mutants in the in vivo transcription of estil (Guzman-Verduzco et al.,

1985). A similar region has been observed downstream from the CAP site of tea (colicin El gene) (G. Weinstock, personal communication). The transcriptional role of the inverted repeat that contains part of the proposed CAP site, and the A/T rich polynucleotides that follows it, is currently under investigation.




the same conformation (Chou and Fassman, 1974; Garnier et al., 1978); however, near the carboxy terminus of the signal peptide, the predictions are different for the odd and even numbered ST* toxins. Nevertheless, the hydropathic profile (Kyte and Doolitle, 1982) for all STAsshow only one membrane spanning region that is very similar (Fig. 4). Downstream from the first cleavage site, from positions 23 to 53, the molecules show a highly conservedhelical structure (Chou and Fassman, 1974; Garnier et al., 1978). The cleavagethat results in the extracellular toxins should take place between met53and asns4for STAm2, STAm3, and STAAand between asns4and ser55 for ST*., . The amino acids in this region Ob 120 90 60 30 0 (positions 23 to 53) and the first three resIncubation Time bin) idues of the mature molecules are the least FIG. 5. Comparison of [%]cysteine incorporated into conserved regions in the known STAs(Fig. 2). minicells isolated in Ficoll (solid circles) or sucrose(open circles) gradients. From the same culture of strain X 1144, This observation suggeststhat proteolysis of minicells were isolated as described under Materials and the precursors is not defined by specific amino Methods, adjusted to the same optical density, and incubated in parallel with 50 &i/ml of [35S]cysteine;at the indicated times, duplicate lOO-~1aliquots were used to determine TCA precipitable counts.

The sequencedestAgenesare very A/T rich (>65%) and within the inferred coding region, STAm2, STAe3,and STAm4 are very similar while ST*-, is the least homologous. The proposed transcription and translation elements are also conserved (Fig. 2), and again STAqIshows the least similarity. It is striking that estA3 and estA4differ only by two bases(deletion of T4s4 and substitution of C for G& even though the strains were isolated in Bangladesh and Mexico, respectively. The region adjacent to position 550 is of particular interest since it has been shown with a fused STAm2-LTBthat the first processing site is between alar9 and glnzo (Fig. 2) (L. M. Guzman-Verduzco and Y. M. Kupersztoch, manuscript in preparation) and that the cleavedproduct accumulates intracellularly (Guzman-Verduzco and Kupersztoch, 1987). The change of pro,9 (ST& for ala (ST*$ causesa change in the adjacent region of the molecule as inferred from the Chou and Fassman (1974) and Gamier et al. (1978) protein structural predictions. The first 14 amino acids of all ST,.,polypeptides predict




3000-c FIG. 6. Fluorography of PAGE [35S]cysteine-labeled peptides obtained from spheroplastsof minicells harboring plasmids pYK103 (ST’) (lines 2 and 4) and pBR328A (ST-) (lines 1 and 3). The same number of counts were applied to each lane. Samples 1 and 2 were treated with cracking buffer containing 2-mercaptoethanol. Prestained molecular weight markers were ovalbumin (43 kDa), (Ychymotrypsinogen (25.7 kDa), @-la&globulin (18.4 kDa), lysozyme (14.3 kDa), cytochrome C (12.3 kDa), bovine trypsin inhibitor (6.2 kDa), and insulin (3.0 kDa).



acid residues, but by the structure of the polypeptide that they constitute. Minicells have been extensively used to characterize the expression of cloned genes. Someof the polypeptides analyzed in minicells are processedand interact with cellular components as they reach their destination. However, we found that ,&galactosidase,an intracellular enzyme, appears in the periplasm of minicells isolated on sucrose gradients. The use of Ficoll gradients decreasesthe effect of hypertonic solutions on the topology of cytoplasmic enzymes. Therefore, Ficoll gradient minicells were used to radioactively label the proteins encoded by the esrA4 gene. The use of [35S]cysteineas a tracer enhanced the fluorographic detection of a peptide coded by estA4;this peptide showedan electrophoretical mobility equivalent to 8400 Da (Fig. 6). The inferred amino acid sequenceof STAA(Fig. 2) should result in a 7409 Da peptide. Previously, Lathe et al. (1980), using estA1 in an in vitro transcription-translation system,detected two species with inferred molecular weights of 10,000 and 7000 Da; these polypeptides were thought to be the precursor-product of STAML. The open reading frame of esrAI results in a 7938 dalton polypeptide speciesand the known molecular weight of the extracellular ST*-, is 1979Da; the electrophoretic migration of the radiolabeled ST*., peptides was faster than the smallest molecular weight marker used by these authors. It has been reported that the conventional SDS-PAGE system (Hashimoto et al., 1983)doesnot show a linear relationship between low molecular weight peptides and electrophoretic mobility. Even when systemsrecently described for the fractionation of small molecular weight polypep tides (Hashimoto et al., 1983; Kyte and Rodriguez, 1983) are used to analyze ST*-, and STA4, the migration is not linear to molecular weight (K. Rasheed,L. M. Guzman-Verduzco, and Y. M. Kupersztoch, in press). The precursor-product relationship suggestedby the in vitro transcription-translation studies (Lathe et al., 1980) thus remains to be correlated more stringently to the size and amino acid compositions of the two proteins. _r The 8400-Da species detected in [“‘S]-

cysteine-labeledminicells is poorly chasedout. Analysis of 35S-radiolabeledperiplasmic and supernatant proteins by SDS-PAGE did not allow the detection of an estA4-encodedproduct of the expected mobility. The inability of minicells to completely process ST indicates that an alternative system should be sought to understand the metabolic steps that pre-proSTAundergoesto reach the extracellular media as the mature ST*. Although the available information suggeststhat STAsundergo two independent processing steps to reach the exterior of the cell, the second intermediate has not been identified; until then, the model remains hypothetical. ACKNOWLEDGMENTS We thank Nigel Harford for supplying the poster of his abstract, where STA.Zwas first described, Robert Munford for the critical review of this manuscript, and Cindy Bas elski for typing it. This work was supported in part by USPHS Grant No. AI-21698 and by a grant from the Consejo National de Ciencia y Tecnologia, Mexico.

REFERENCES AIMOTO,S., TAKAO, T., SHIMONISHI,Y., HARA, S., TAKEDA, T., TAKEDA, Y., AND MIWATANI, T. (1982). Amino acid sequenceof a heat stable enterotoxin produced by human enterotoxigenic E. coli. Eur. J. Biochem. 129,25?-263. ALDERETE,J. F., AND ROBERTSON, D. C. (1977). Repression of heat-stable enterotoxin synthesis in enterotoxigenie Escherichia coli. Infect. Immunol. 17,629-633. BARBAUD,O., AND SCHWARTZ,M. (1984). Positive control of transcription initiation in bacteria. Annu. Rev. Genet. 18, 173-206. CHOW,P. Y., AND FASSMAN,G. D. (1974). Prediction of protein conformation. Biochemistry 13, 222-245. COVARRUBIAS,L., CERVANTES,L., COVARRUBIAS,A., SOBERON, X., VICHIDO,I., BUNCO, A., KUPERSZTOCHPORTNOY,Y. M., AND BOLIVAR,F. (1981). Construction and characterization of new cloning vehicles. V. Mobilization and coding properties ofpBR322 and several deletion derivatives including pBR327 and pBR328. Gene 13,25-35. CURTIS& R. III (1965). Chromosomal aberrations associated with mutations to bacteriophage resistancein E. coli. J. Bacterial. 89, 28-45. DEWILDE,M., YSEBEAR,M., AND HARTFORD,N. (1981). DNA sequenceof the STA.r enterotoxin gene from an E. coli strain of human origin. In “Molecular Biology, Pathogenicity and Ecology of Bacterial Plasmids” (S. B. Levy, R. C. Clowes, and E. L. Koenig, Eds.), p. 596. Plenum, New York. FRAZER,A. C., AND CURTISS,R. III (1975). Production,




polymerase binding site in an early T7 promoter. Proc. Natl. Acad. Sci. USA 72,784-788. RONNBERG,B. T., WADSTROM,T., AND JERNVALL,J. T. (1983). Structure of a heat stable enterotoxin produced by a human strain of Escherichia coli. FEBS Lett. 155, 183-186. GUZMAN-VERDUZCO, L. M., FONSECA, R., AND KuROSENBERG, M., AND COURT,D. (1979). Regulatory sePERS~TOCH-PORTNOY, Y. M. (1983).Thermoactivation quencesinvolved in the promotion and termination of transcription. Annu. Rev. Genet. 13, 319-353. of a periplasmic heat stable enterotoxin of Escherichia coli. J. Bacterial. 154, 146-151. SAEED,A. M. K., MAGNUSON,N. S., SRIRANGANATHAN, GUZMAN-VERDUZCO,L. M., AND KUPERSZTOCH, Y. M. N., BURGER,D., AND COSAND,W. (1984). Molecular (1987).Fusion of two Escherichia coli enterotoxins: SThomogeneity of heat-stable enterotoxins produced by bovine enterotoxigenic Escherichia coli. Infect. ImLTs. J. Bacterial. 169, 5201-5208. GUZMAN-VERDUZCO,L. M., STIEGLITZ,H., AND Kumunol. 45,242-247. PERSZTOCH, Y. M. (1985). Transcriptional regulation SANGER,F., NICKLEN, S., AND CAULSON,A. R. (1977). DNA sequencing with chain terminators inhibitors. of the ST enterotoxin of Escherichia coli. “Proceedings of the 2 1stConference,U.S.-Japan Cooperative Medical Proc. Natl. Acad. Sci. USA 74,5463-5467. Science Program, Cholera Panel,” in press. SEKIZAKI,T., AKASHI, H., AND TERAKADO,N. (1985). HASHIMOTO, F., HORIGOME, T., KANBAYASHI, M., Nucleotide sequence of the genes for Escherichia coli YOSHIDA,K., AND SUGANO,H. (1983). An improved heat stable enterotoxin I of bovine, avian and porcine method for separation of low molecular weight polyorigins. Amer. J. Vet. Res. 46, 909-912. peptides by electrophoresis in sodium dodecyl sulfate SHAPIRO,A. L., VINUELA, E., AND MAIZEL, J. V. (1967). polyacrylamide gel. Anal. Biochem. 129, 192-199. Molecular weight estimation of polypeptide chains by KYTE, J., AND DOOLITLE,R. F. (1982). A simple method electrophoresis in SDS-polyacrylamide gels. B&hem. for displaying the hydropathic character of a protein. J. Biophys. Res. Commun. 28,8 15-820. Mol. Biol. 157, 105-137. SHINE, J., AND DALGARNO,L. (1975). Determinant of KYTE, J., AND RODRIGUEZ,H. (1983). A discontinuous cystron specificity in bacterial ribosomes. Nature (Lonelectrophoretic system for separating peptides on polydon) 254,34-38. acrylamide gels.Anal. Biochem. 133, 5 15-522. So, M., AND MCCARTHY, B. J. (1980). Nucleotide seLATHE,R., HIRTH, P., DEWILDE,M., HARFORD,H., AND quence of the bacterial transposon Tn 1681 encoding a LECOCQ,J.-P. ( 1980).Cell free synthesisof enterotoxin heat-stable (ST) toxin and its identification in enteroof E. coli from a cloned gene. Nature 284,473-474. toxigenic E. coli strains. Proc. Natl. Acad. Sci. USA 77, LAZURE,C., SEDAH,N. G., CHRETIEN,M., LALLIER, R., 401 l-4015. AND ST. PIERRE,S. (1983). Primary structure deter- STIEGLITZ,H., FONSECA,R., OLARTE,J., AND KUPERSZmination of Escherichia coli heat stable enterotoxin of TOCH-PORTNOY, Y. M. (1980). Linkage of heat-stable porcine origin. Canad. J. Biochem. Cell. Biol. 61,787enterotoxin activity and ampicillin resistancein a plas792. mid isolated from an Escherichia coli strain of human origin. Infect. Immunol. 30,6 17-620. MARTINEZ-CADENA, M. G., GUZMAN-VERDUZCO, L. M., STIEGLITZ,H., AND KUPERSZTOCH-PORTNOY, Y. M. TAKAO, T., HITOUJI, T., AIMOTO, S., SHIMONISHI,Y., HARA, S., TAKEDA, T., TAKEDA, Y., AND MIWATANI, ( 198I). Catabolite repression of Escherichia coli heatT. (1983). Amino acid sequenceof a heat stable enterstable enterotoxin activity. J. Bacterial. 145,722-728. otoxin isolated from enterotoxigenic Escherichia co/i MAXAM, A. M., AND GILBERT, W. (1980). Sequencing 18D. FEBS Lett. 152, l-5. and labeled DNA with base specific chemical cleavage. In “Methods in Enzymology” (L. Grossman and K. THOMPSON,M. R., AND GIANELLA,R. A. (1985). Revised amino acid sequence for heat-stable enterotoxin proMoldave, Eds.), Vol. 65, pp. 499-560. Academic Press, duced by an Escherichia coli strain ( 18D) that is pathoNew York. genic for humans, Infect. Immunol. 47, 834-836. MCCLURE,W. R. (1985). Mechanism and control oftranscriptional initiation in prokaryotes. Annu. Rev. VIEIRA, J., AND MESSING,J. (1982). The pUC plasmids, an M 13mp7-derived system for insertion mutagenesis Biochem. 54, 17l-204. and sequencing with synthetic universal primers. Gene MOSELEY,S. L., HARDY, J. W., HUQ, M. I., ECHEVERRIA, 19,259-268. P., AND FALKOW,S. (1983). Isolation and nucleotide sequencedetermination of a geneencoding a heat stable VON HEIJNE,G. (1985). Signal sequences.The limits of variation. J. Mol. Biol. 184, 99-105. enterotoxin of Escherichia coli. Infect. Immunol. 39, WHIPP,S. C., MOON,H. W., ANDARGENZIO,R. A. (1981). 1167-l 174. Comparison of enterotoxin activities of heat-stable enPESTI,L., AND LUKACS,K. (198 1). Staining technique for terotoxins from class I and class 2 Escherichia coli of peptides of Escherichia coli heat stable enterotoxin. Inswine origin, Infect. Immunol. 31, 245-25 1. fect. Immunol. 33, 944-947. PRIBNOW,D. (1975). Nucleotide sequence of an RNA Communicated by Stuart B. Levy properties and utility of bacterial minicells. Curr. Top. Microbial. Immunol. 69, l-84. GARNIER, J., OSCXJTHORPE,D., AND ROBSON, B. (1978). Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. 120, 97-120.