Search

The binding of ETEC strains expressing different colonization factors to the human enterocyte-like cell line Caco-2 and to isolated human enterocytes were determined. Strains expressing CFA/I, CS2, CS4+CS6, CS5+CS6, CS7, CFA/III+CS6 and PCFO166 adhered well to both types of cells whereas bacteria producing CS1, CS6 only, or CS17 did not adhere to either Caco-2 cells or to jejunal human enterocytes, suggesting that similar binding structures are present in both cell types. However, in some instances, binding of bacteria to the two types of cells differed, e.g. bacteria expressing CS3 or PCFO9 bound well to human enterocytes but not to Caco-2 cells. Conversely, bacteria producing PCFO20 or PCFO159 only adhered to Caco-2 cells and not to jejunal enterocytes. With few exceptions, this inability to bind to human enterocytes was the same for cells from all parts of the small intestine. This study contradicts previous reports suggesting that the binding structures of Caco-2 cells are identical to those of human enterocytes.

SummaryOne hundred and eighty–one stool specimens from patients with various types of diarrhoea (135 patients) or from non-diarrhoeal controls (23 acute medical patients, 23 inflammatory bowel disease in remission) were investigated using a colony–blot DNA hybridization assay for the presence of Verocytotoxin–producing (VTEC), enteroaggregative (EAggEC) and diffusely adherent (DAEC)Escherichia coli. Twelve patients had probe–positive EAggEC in the stool and 8 of these had diarrhoea, 6 following recent travel. Eight patients had DAEC, 7 of whom had travellers diarrhoea. Six of 10 (60%) travellers with gastroenteritis, but without a recognized enteric pathogen, were positive for EAggEC (4) or DAEC (2). Five of 10 (50%) travellers with gastroenteritis related to a recognized enteric pathogen also had DAEC identified in their stool. Of the 23 acute medical control patients 11 had been abroad, 4 of these were immigrants and had EAggEC. VTEC were not found and, with one exception, immunoassays for antibodies toE. coliO 157 and O 2 lipopolysaccharides were negative.

Sequences regulating production of fimbriae were cloned from two enterotoxigenic Escherichia coli strains. One cloned region, from E. coli 025.H42, controlled expression of coli surface-associated (CS) antigen 4, whereas the function of the other, from E. coli 0167.H5, was unclear. Both regulators were related to the cfaD gene that controls expression of colonization factor antigen I (CFA/I) although low stringency conditions were required to show significant hybridization between cfaD and the regulatory fragment from E. coli 0167. The cloned regulatory genes promoted expression of CFA/I, CS1, CS2 and CS4 antigens but the levels of production in the presence of the 0167 regulator were lower than those promoted by the CS4 regulator of cfaD.

SUMMARYEnterotoxigenicEscherichia coli(ETEC) from Burma, central Africa (Rwanda and Zaire) and Peru, were screened by enzyme-linked immunoassays for the colonization factor antigens (CFAs) and putative colonization factors (PCFs): CFA/I, CFA/II, which consists of three coli surface-associated (CS) antigens, CS1, CS2 and CS3, CFA/III, CFA/IV (CS4, CS5, CS6), CS7, PCFO9, PCFO159.H4, PCFO166, and CS17. The highest proportion of ETEC with identifiable colonization factors (71%) were found in the strains from Burma, which were mainly positive for CFA/I (38%), but strains producing CFA/II (4%), CFA/IV (11%), CS7 (10%), CS17 (4%), PCFO159.H4 (2%) and PCFO166 (2%) were also found. Sixty-nine percent of the ETEC from central Africa were positive for known colonization factors. While CFA/I positive strains were important (12%), a higher number of ETEC producing CFA/IV (33%) and CS17 (24%) were found. Fifty-two percent of the Peruvian strains produced identifiable colonization factors. The largest group of strains produced antigens of the CFA/IV complex (17%), while ETEC producing CFA/II (6%), CFA/III and CS6 (2%), CS7 (6%), PCFO9 (6%), PCFO166 (8%) and CS17 (7%) were also found. These surveys show that there is a considerable variation in the proportions and types of colonization factor found in different geographical areas. From 29 to 48% of the ETEC did not possess an identifiable colonization factor. These were particularly of the LT only producing type. These results have important implications for vaccine formulation.

Summary: Escherichia coli strain 334 is a human enterotoxigenic strain of serotype 015:H11 which had previously been shown to produce “attachment pili”. These fimbriae were compared with other colonization factors. From strain 334 a mannose-resistant haemagglutination positive colony 334A and a mannose-resistant haemagglutination negative variant 334C were isolated. By electron microscopy the fimbriae of strain 334A were shown to have a helical structure resembling coli-surface-associated antigen (CS5) fimbriae. An antiserum was raised to strain 334A and absorbed with a fimbriae-negative variant of that strain, 334C. By immuno-electron microscopy this antiserum was shown to coat fimbriae of strain 334A but not CS5 fimbriae produced by strain E17018A. Conversely, CS5 antiserum did not coat the fimbriae produced by strain 334A. No antigenic cross-reaction was detected between these intact fimbriae when anti-strain 334A serum and CS5 antiserum were used in immunodiffusion tests. By enzyme-linked immunosorbent assays (ELISAs) the fimbriae of strain 334A were shown to be antigenically unrelated to most other human ETEC adhesins, namely colonization factor antigens (CFA/I, CFA/III and CFA/IV), coli-surface-associated antigens (CS1, CS2, CS3, CS4, CS6 and CS17) and putative colonization factors (PCF0159:H4 and PCF0166). However, a heated suspension of strain 334A reacted weakly with CS5 antiserum in an ELISA. By SDS-PAGE the fimbriae of strain 334A were shown to consist of subunits of similar size to CS5 subunits, that is about 21·5 kDa. Western immunoblotting revealed that the subunits of 334A and CS5 fimbriae shared common epitopes. Expression of 334A fimbriae was controlled by a plasmid which also coded for enterotoxin production, a system similar to that reported for a number of other human ETEC attachment factors. Strain 334A fimbriae have therefore been designated as CS7 fimbriae. Surveillance with CS7 anti-serum of ETEC with no known attachment factors showed two other strains of serotype O15:H11, one strain of serotype O103:H49 and 20 strains of serotypes O114:H49 and H- to possess CS7 fimbriae.

Production of coli surface-associated antigen 1 (CS1) by Escherichia coli strain E24377 of serotype 0139.H28 was controlled by a plasmid that also encoded heat stable and heat labile enterotoxins and CS3. The presence of a regulatory sequence was detected on this plasmid by hybridization with the cfaD gene that regulates expression of colonization factor antigen I fimbriae and is at least 96% homologous with the rns sequence controlling production of CS1 or CS2 fimbriae by strains of serotype 06.H16 of appropriate biotype. A separate plasmid, pDEP20, carrying the structural genes for CS1 synthesis was identified and transformed into E. coli strain HB101 or a derivative of strain E24377 without large plasmids. Transformants carrying pDEP20 did not produce CS1 fimbrial antigen, but antigen expression was obtained when a cloned cfaD gene or a wild-type plasmid carrying the rns sequence was introduced. Transposon mutagenesis with Tn1000 identified a 3.7 kbp region of pDEP20 essential for production of CS1 fimbriae. Genes encoding production of CS1 fimbriae were cloned on a 9.9 kbp BamHI fragment and were expressed in the presence of the cfaD sequence. A strain producing both CS1 and CS2 antigens was constructed by introduction of the cloned cfaD gene into a strain of serotype 06.H16 biotype C carrying plasmid pDEP20.

Two enterotoxigenic Escherichia coli strains of serotype 0.25.H42 that produced coli surface associated antigens CS4 and CS6 hybridized with a probe containing the cfaD sequence that regulates expression of colonization factor antigen CFA/I. Transformation of a cloned cfaD gene into some derivatives of the strains that were negative for CS4 and CS6 resulted in expression of CS4 but not CS6. By hybridization the sequence that regulated CS4 production in the wild type 025 strains was located on a plasmid that also encoded the CS6 antigen. The structural genes for the CS4 antigen were on a separate plasmid. The 025 strains carried a third plasmid encoding enterotoxin production which was therefore unlinked to regulation sequences or genes encoding CS antigens.

The roles of the subcomponents of colonization factor antigen II, the coli surface antigens CS1, CS2, and CS3, as colonization factors and protective antigens was studied in a nonligated rabbit intestine model (RITARD). Infection with enterotoxigenic Escherichia coli (ETEC) carrying CS3 alone or CS1 plus CS3 induced diarrhea in most (80%) of the rabbits, whereas nonenterotoxigenic strains expressing CS1 or CS2 rarely induced diarrhea. Strains carrying CS1, CS2, or CS3 alone were all shed in stools for a significantly longer period than normal fecal flora-type E. coli. Initial infection with ETEC positive for CS1 plus CS3 induced significant protection against disease caused by reinfection with a highly diarrheagenic dose of the homologous strain; rabbits previously infected with serotype-heterologous, nontoxigenic bacteria carrying CS1 only were also protected against this challenge, whereas no such protection was induced by serogroup-homologous E. coli carrying CS2 only. Animals previously infected with CS1-, CS3-, or CS1-plus-CS3-positive bacteria excreted the CS1-plus-CS3 challenge strain for a significantly shorter period than did "nonimmunized" rabbits, whereas initial infection with bacteria carrying CS2 only did not result in such reduced shedding. Monoclonal antibodies against CS1, CS2, or CS3 all protected against experimental infection with ETEC carrying the corresponding CS factor. These results suggest that all the subcomponents of colonization factor antigen II are colonization factors and may induce anticolonization immunity.

Enterotoxigenic Escherichia coli (ETEC) of serotype O114:H21, which produced only heat-labile enterotoxin (LT), gave mannose-resistant hemagglutination (MRHA) with bovine erythrocytes. One strain, E20738A, was shown to possess fimbriae of approximately 7.5 nm diameter. On SDS-PAGE two possible fimbrial polypeptides of molecular masses 17.5 and 15.5 kDa were seen; the 17.5-kDa band was the most prominent. Loss of LT and MRHA together from strain E20738A was associated with loss of a 100-MDa plasmid. An absorbed anti-strain E20738A serum reacted specifically with the 17.5- and 15.5-kDa polypeptides and bound to the intact fimbriae. This antiserum reacted positively in an ELISA with LT-positive E. coli strains of serogroups O8, O15, O48, O114, and O146. The antiserum did not react with ETEC carrying known colonization factors. The term coli-surface-associated antigen (CS) 17 has been used to describe the fimbriae.

Enterotoxigenic Escherichia coli (ETEC) are an important cause of diarrhoeal disease in infants in developing countries and in travellers to these countries. These E.coli belong to a large number of different serotypes but have common pathogenic mechanisms namely the ability to adhere to the epithelial surface of the small intestine by means of colonization factors and to produce a heat stable (ST) or a heat labile (LT) enterotoxin, or both. Animal studies and volunteer experiments have shown that colonization factors and some toxins induce protective immunity and could be used to make vaccines1. However the effectiveness of a vaccine incorporating colonization factors will depend on the identification of these antigens on the majority of ETEC strains which cause diarrhoea in man.

Adhesion of bacteria to epithelial cells of the host is the first and probably the most important step in the pathogenesis of bacterial infections. In humans, adhesion of enterotoxigenic E.coli (ETEC) to the mucosa of the small intestines is mediated by serologically distinct and host-specific fimbrial antigens on the bacterial surface. The first fimbrial antigen to be identified in human ETEC strains was colonization factor antigen I (CFA/I)1. CFA/I is a single fimbrial antigen found on ETEC producing only heat-stable (ST) enterotoxin and on ETEC producing both heat-labile and heat-stable enterotoxin (LT, ST) of at least 15 serogroups. Later CFA/11 was detected2 and was subsequently shown to consist of three distinct coli surface associated (CS) antigens, designated CS1, CS2 and CS33. Nearly all CFA/II-positive E.coli produce CS3, but those of serotype 06.H16 or 06. H- also express either CS1 or CS2 depending on their biotype: biotype A strains produce CS1 whereas those of biotypes B, C and F express CS23. Production of CS1 and CS3 has been reported in only one other ETEC strain. This strain was of serotype 0139.H28 and differs from other 0139.H28 strains that are CS3 only producers4.

Enterotoxigenic Escherichia coli, ETEC, colonize the human intestine by means of fimbrial structures called colonization factor antigens, CFAs. On human ETEC strains three different types of CFAs have been described, i.e. the colonization factor antigens I, (CFA/I), CFA/II and CFA/IV. In addition a number of “new” putative colonization factor antigens have been identified, e.g. PCF0166, CS7, CS17 and CFA/III [M.M McConnell, this volume] although their prevalence and role in colonization have not been fully assessed.

A DNA sequence, homologous to the cfaD gene of CFA/I region 2, was identified on CFA/I region 1. This sequence is designated cfaD′. It differs from the cfaD gene in containing two deletions and a stop codon. The cfaD′ sequence therefore can only encode a truncated CfaD-like protein. The CfaD protein may be a DNA binding protein and functions as a positive regulator of CFA/I fimbriae expression. A regulatory function for the cfaD′ is not likely since deletion of the cfaD′ sequence does not affect production of CFA/I fimbriae in E. coli K-12 strains. That the cfaD′ sequence is present on CFA/I wild-type plasmids isolated from CFA/I strains of different serotypes, obtained at various geographical locations, suggests, however, that this DNA region is not completely without a function.

SUMMARYStrains ofSalmonella typhimuriumof predominantly Middle Eastern origin, but distributed from England to India, were found to carry at least three types of resistance plasmid. The most important was initially identified as an FIplasmid by compatibility tests, but differs from the F factor on the one hand and the FIfactors R162 and CoIV on the other. The three groups of FIplasmids can be distinguished by their compatibility reactions with the MP10 plasmid ofS. typhimurium(Smith, Humphreys, Grindley, Grindley & Anderson, 1973) and group H1factors: the F factor is unilaterally incompatible with group H1(Smith, Grindley, Humphreys & Anderson, 1973; Anderson, 1975b); the FIfactors are compatible with MP10 and group H1and FImefactors are incompatible with MP10 but compatible with H1. The majority ofS. typhimuriumcultures belonged to phage type 208; most of those that did not, belonged to types related to 208. Only a minority of their FImeplasmids were autotransferring. The remainder were mobilizable by F-like plasmids, and by group H1and H2factors, but not by thefi–I1facter δ, or by plasmids of the I2, B, P, W, N andcom7 groups. The compatibility reactions of the autotransferring F1meplasmids were identical with those of the non-transferring members of the group, and both were large, single-copy plasmids.TheS. typhimuriumstrains of this series carried A or AK, and SSu resistance determinants: small, probably multicopy, non-transferring plasmids similar to those originally described in phage type 29 ofS. typhimurium(Anderson & Lewis, 1965b).TheseS. typhimuriumcultures probably represent a clone of wide geographical distribution. The accurate epidemiological study of such clonal outbreaks requires, in addition to phage typing, precise identification of the plasmids harboured by the epidemic strains, and may have to be carried to the molecular level.FImeplasmids were identified in other drug-resistant salmonellas, notably in a strain ofS. wienwhich caused large outbreaks of mainly paediatric infection in Algeria, and also spread to Britain. An FImeplasmid was found inS. typhiphage type 44 from Algeria, in which the phage-restricting properties of the plasmid are responsible for the specificity of the type.

Enterotoxigenic Escherichia coli (ETEC) with colonisation factor antigens I and II (CFA/I and CFA/II) and PCF8775 were examined in the ‘salting-out’ test to assess the hydrophobicity of the individual adhesion antigens. The fimbrial antigens CFA/I and the coli surface antigens CS1, 2, 4 and 5 were all strongly hydrophobic. The fibrillar fimbrial antigen CS3 was weakly hydrophobic; this antigen is apparently non-fimbrial.

We examined seven enterotoxigenic Escherichia coli strains which produced colonization factor antigen I (CFA/I). Four of these strains were from South Africa (three serotype O78:H12 and one serotype O63:H-), one was from Ethiopia (O78:H12), and two were from Bangladesh (O78:H11 and O78:H12). Plasmids coding for CFA/I were mobilized from six of these strains by using resistance or enterotoxin factors. No plasmid was mobilized from the serotype O78:H12 Bangladesh strain. The transconjugants obtained from crosses with the O78 strains also produced heat-stable enterotoxin (ST), and additional investigations showed that CFA/I and ST were coded for by a single non-autotransferring plasmid. These plasmids were fertility inhibition negative, did not restrict any of the coliphages with which they were tested, and were incompatible with each other. Four had molecular weights of approximately 60 X 10(6), and one had a molecular weight of 52 X 10(6). Like the other CFA/I plasmids, the CFA/I plasmid transferred from the O63:H- strain coded for ST, but this plasmid also coded for heat-labile enterotoxin. In most other respects the properties of this plasmid were similar to those of the CFA/I-ST plasmids previously described. The molecular weight of this plasmid was 65 X 10(6). The IncT R-factor Rtsl was marked with a transposon for tetracycline resistance and then transferred into the two Bangladesh wild-type strains. Plasmids which coded for tetracycline resistance, CFA/I, and ST were transferred from these strains. These plasmids were incompatible with Rtsl and with the CFA/I-ST plasmids described above and were recombinants between Rtsl and a CFA/I-ST plasmid. Their properties are also described.

Summary: Two plasmids transferred from enterotoxigenic Escherichia coli (ETEC) of serotype 06. H16 and biotypes A and C coded for mannose-resistant haemagglutination (MRHA) and production of heat-stable enterotoxin (ST) and heat-labile enterotoxin (LT). Both plasmids were non-autotransferring being mobilized most efficiently by the R plasmid R100-1. They were similar in their genetic properties being incompatible with each other and plasmids of the Inc group FI. The wild-type strains produced the colonization factor antigen II (CFA/II) which was made up of different coli surface antigens (CS). The biotype A strains produced CS1 and CS3 while the biotype C strains produced CS2 and CS3. These three antigens have the ability to cause MRHA. When plasmids coding for MRHA were transferred to K12 strains, the degree of haemagglutination was markedly reduced and only CS3 was produced. When both plasmids were transferred back into biotype A strains, good MRHA was restored and the strains produced CS1 and CS3. In a biotype C strain CS2 and CS3 were formed. The production of the antigens was compared by enzyme-linked immunosorbent assay (ELISA). The strains were also examined by electron microscopy where it was found that CS1 and CS2 were fimbrial antigens while CS3 was not.

Immunoglobulins, prepared from polyclonal rabbit antisera raised against Escherichia coli fimbrial antigens, colonization factor antigen (CFA)/I, and coli-surface-associated antigens (CS)1, CS2 and CS4, were used to assess antigenic cross-reactions between these four fimbrial types by Western immunoblotting. Antibodies in a serum, prepared against CS4, cross-reacted strongly with the fimbrial subunits of CFA/I, CS1 and CS2. Antibodies in sera prepared against CFA/I and CS1 gave weak reactions with CS1 or CFA/I respectively and also with CS2 and CS4, while the antiserum prepared against CS2 did not react. CS4 antiserum also reacted with the CS17 fimbrial subunit, but not with the subunits of fimbrial antigens: CFA/III, CS5, putative colonization factor (PCF) 0159: H4 or PCF0166.

Sequences encoding the CS6 antigen of colonisation factor antigen (CFA)IV were cloned on a 3kb ClaI fragment. The recombinant plasmid pDEP5 coded for surface expression of CS6 measured by ELISA and production of CS6 polypeptides was detected in E. coli minicells. The genes for the CS1, CS2 and CS3 components of colonisation factor antigen CFA/II were cloned together on a length of DNA corresponding to about 17kb. CS3 was always expressed but production of either CS1 or CS2 depended on the serotype and biotype of the host strain. Separate subclones were obtained that expressed CS3 or CS1 and CS2.

We examined a strain of Escherichia coli, serotype O159.H34, of human origin which produced heat-stable and heat-labile enterotoxins, was resistant to ampicillin, and produced colicin. By conjugation and transformation experiments plasmids coding for enterotoxin production (Ent), enterotoxin production and ampicillin resistance (Ap-Ent), ampicillin resistance (Ap), and colicin production were isolated. Both the Ent and Ap-Ent plasmids were autotransferring and belonged to the F-incompatibility complex. However, the Apr Ent+ transconjugants showed differences in their levels of resistance and in their abilities to propagate F-specific phages and to transfer resistance. The results suggested there was transposition from the small Ap plasmid to the Ent plasmid. The Ap-Ent plasmids were larger than the enterotoxin factor and when treated with restriction endonuclease BamHI showed an additional fragment not present in the enterotoxin plasmid. The insertion of ampicillin resistance probably occurred at different sites on the enterotoxin plasmid, resulting in the observed variation in phenotype.

A total of 134 enterotoxigenicEscherichia coli (ETEC) of serogroups O25, O27, O148, and O159 were tested in the enzyme-linked immunosorbent assays for the colonization factor antigens I (CFA/I), CFA/II (coli surface antigens CS1, 2 and 3) and putative colonization factor (PCF) 8775 (CS4, 5 and 6). CS6 was detected without CS4 or CS5 in 94% of the strains of serogroup O25, 86% of strains of serogroup O27, 87% of strains of serogroup O148, and 29% of strains of serogroup, O159. The frequency with which CS6 occurs in ETEC of common serotypes without the antigens CS4 or CS5 suggests that it might be a colonization factor.

The CS2 fimbrial antigen was isolated by salt and isoelectric precipitation and by column chromatography. The purified antigen was free of other fimbrial proteins present on the same bacterial strain. Analysis of the N-terminal amino acid sequence indicated extensive homology with the CFA1 fimbrial antigen, which was surprising since the two proteins do not show any immunological cross reactivity. It could therefore be concluded that the N-terminal parts of these fimbrial proteins are not located on the surfaces of the proteins and might be involved in conservation of the structural integrity of these proteins.

The molecular properties of enterotoxin (Ent) plasmids from 12 Escherichia coli strains of human origin were examined. Ten strains belonged to the O78 serogroup, and the remainder were of serogroup O7 or O159. Eleven plasmids coded for heat-labile enterotoxin (LT), and one coded for heat-stable enterotoxin (ST) and LT. The results of restriction enzyme digests and deoxyribonucleic acid reassociation experiments showed that all of the Ent plasmids were related, and supported the subdivision of the LT plasmids into three groups based on their genetic properties (M. M. McConnell et al., J. Bacteriol. 143: 158–167, 1980). Within group 1, two plasmids from South African strains were indistinguishable but differed in Eco RI and Hin dIII digests from the LT plasmid that originated from an Ethiopian strain. The three plasmids had >70% homology. The two non-autotransferring group 2 plasmids identified in O78.H11 strains from Bangladesh were indistinguishable. The group 3 plasmids were from strains belonging to serogroups O7 and O78 isolated in Bangladesh, India, and Thailand. They shared >95% homology but showed slight differences in fragment patterns when treated with Eco RI and Hin dIII. There was 60 to 70% homology between the plasmids of groups 1 and 3, and the group 2 plasmid had 40 to 50% homology with members of these two groups. The autotransferring Ent plasmids had up to 40% homology with R factors of incompatibility groups FI, FII, and FIV.

Summary: Enterotoxigenic strains of Escherichia coli, which belonged to serogroups other than 0 6 and produced colonization factor antigen 11, usually produced only coli surface antigen 3 (CS3) and gave weak mannose-resistant haemagglutination of bovine erythrocytes. A non-autotransfer-ring plasmid, NTP165, from a strain of E. coli 0168.H16 coded for heat-stable enterotoxin, heat-labile enterotoxin and CS antigens. The CS antigens expressed after acquisition of plasmid NTPl65 depended on the recipient strain : a biotype A strain of serotype 0 6. H 16 expressed CS1 and CS3; a biotype C strain of serotype 0 6. H 16 expressed CS2 and CS3; strain K 12 and strain El9446 of serotype 0139. H28 expressed only CS3. An exceptional wild-type strain, E24377, of serotype 0139. H28 produced CS1 and CS3 when isolated; a variant of E24377 which had lostthe plasmid coding for CS antigens produced both CSl and CS3 after the introduction of NTP165.

Enterotoxigenic Escherichia coli producing coli surface antigen 4 (CS4), CS5, and CS6 of colonization factor antigen IV were examined. This factor was originally called putative colonization factor 8775 (PCF8775). All of the coli surface antigens were plasmid coded and were usually carried on the same plasmid as the genes coding for heat-stable toxin (ST) or heat-labile toxin (LT); thus, CS5-CS6-ST, CS6-ST, and CS6-LT plasmids were found. In strains of serotype O25:H42, the genes coding for CS4 and CS6 were on a plasmid separate from that containing the genes coding for ST and LT. CS4 and CS5 were fimbrial antigens with a subunit molecular mass of about 17.0 and 21.0 kilodaltons (kDa), respectively. CS6 was found as a single polypeptide with a molecular mass of about 14.5 kDa in strains of serotypes O25:H42, O27:H7, and O27:H20 when heated extracts were run on sodium dodecyl sulfate-polyacrylamide gels. CS6-positive extracts of strains of serogroups O148, O159, and O167 showed two bands with molecular masses between 14.5 and 16.0 kDa.

EPEC adherence factor (EAF) plasmids from three strains of enteropathogenic Escherichia coli (EPEC) - E2347/69 (O127:H6), E20517 (O111:H2) and E24582 (O142:H6) - were examined. The EAF plasmids were all marked with ampicillin resistance by transposition of Tn801 to give pDEP1, pDEP2 and pDEP11, respectively. All three plasmids showed incompatibility with an FIme and an FIV plasmid and had some similarity in restriction enzyme digest patterns. Plasmid pDEP1 differed from pDEP2 and pDEP11 in being autotransferring and fertility-inhibition positive. An EAF probe consisting of a 1 kb BamHI-SalI restriction endonuclease fragment of the prototype EAF-associated plasmid pMAR2 hybridized to similar-sized SalI-BamHI fragments of pDEP1 and pDEP11 but to a different-sized fragment of plasmid pDEP2. Loss of the EAF plasmids from EPEC strains resulted in a marked reduction in the ability of these strains to adhere to HEp-2 cells. The EAF-plasmid-negative variants did not express a 94 kDa outer-membrane protein (OMP). When these EAF plasmids were reintroduced into EAF-plasmid-negative EPEC strains a high level of adherence equivalent to that of the parent EPEC strains was restored and a 94 kDa OMP was usually expressed. However, when EAF plasmids were transferred into E. coli K12 or non-EPEC E. coli the host strains either did not adhere or adhered poorly to the HEp-2 cells. These transconjugants did not express a 94 kDa OMP.

The role of the PCF8775 antigen and its antigenic subcomponents, in particular, the coli surface (CS) antigen CS6, as colonization factors and protective antigens was studied in the reversible intestinal tie adult rabbit diarrhea model. This was done by testing the abilities of different mutants which carried one or two of the CS components to colonize the intestine and to induce protective immunity against reinfection with PCF8775-positive enterotoxin-producing Escherichia coli. Infection with enterotoxigenic E. coli carrying CS4-CS6, CS5-CS6, or CS6 alone induced diarrhea in 75% or more of the rabbits, whereas the corresponding nonenterotoxigenic mutants, as well as enterotoxigenic but CS-negative strains, induced diarrhea in only a few cases. Mutants carrying CS6 alone colonized the intestine equally as well as strains carrying CS4-CS6 or CS5-CS6 did, whereas CS-negative mutants were excreted in the stool for a significantly shorter period. Rabbits previously infected with mutants carrying CS6 alone or CS6 in combination with CS4 or CS5 developed diarrhea with a significantly lower frequency after reinfection with a normally highly diarrheagenic dose of enterotoxigenic CS4-CS6-positive E. coli bacteria than did animals immunized with corresponding CS-negative mutants. Fecal excretion of the rechallenge strain was also of considerably shorter duration than that observed after initial infection with corresponding strains in 27 of the 30 animals (90%) immunized with strains carrying CS6 alone or in combination with CS4 or CS5. Such reduced shedding of the challenge strain was only seen in a few rabbits (3 of 12) initially infected with CS-negative bacteria. These results suggest that the CS6 component of PCF8775 is a colonization factor in rabbits and that it is also capable of inducing protective immunity.

Plasmids coding for colonization factor antigen I (CFA/I) and heat-stable enterotoxin (ST) were identified in 10 strains of human enterotoxigenic Escherichia coli. The strains, which belonged to serogroups O63, O114, O128, and O153, were isolated in Bangladesh, Latin America, Spain, and South Africa. Two strains produced heat-labile enterotoxin in addition to ST. CFA/I-ST plasmids were mobilized from two O128 strains into E. coli K-12 with the R factor R1-19K-. Like the prototype CFA/I-ST plasmid NTP113, mobilized previously from an E. coli O78 strain into K-12, these two plasmids were non-autotransferring. All 10 CFA/I-ST plasmids were incompatible with NTP113 and had molecular weights ranging from 59 X 10(6) to 72 X 10(6). The molecular properties of seven of these plasmids were compared with those of six CFA/I-ST plasmids previously mobilized from O78 strains from Ethiopia, South Africa, and Bangladesh and with those of one plasmid coding for CFA/I, ST and heat-labile enterotoxin from a South African strain of serogroup O63. Digestion with the restriction endonuclease HindIII showed that several plasmids had very similar fragment patterns and two were identical. Generally, a larger proportion of HindIII fragments were of common size in digests of plasmids identified in strains from related geographical areas, regardless of serogroup. However, all except one plasmid shared five or six HindIII fragments of the same size, one of which had been shown previously to be involved in CFA/I production. There was at least 90% DNA homology between CFA/I-ST plasmids with a molecular weight of about 58 X 10(6) from O78 strains from different sources. Most of the DNA sequences of these plasmids were present in a larger CFA/I-ST plasmid (72 X 10(6) from an O128 strain. The results of genetic and molecular studies suggest that CFA/I and ST production is determined by very similar plasmids in different serogroups of human enterotoxigenic E. coli from several sources.

Enterotoxigenic Escherichia coli (ETEC) of serogroup O166 gave mannose-resistant haemagglutination (MRHA) with bovine and human erythrocytes. The strains did not react with antisera prepared against the known colonization factors CFA/I, CFA/II, CFA/III, CFA/IV and PCFO159:H4. Strain E7476 of serotype O166:H27, which produced heat-stable enterotoxin (ST), was examined initially. It produced fimbriae about 7 nm in diameter. On SDS-PAGE two possible fimbrial polypeptides of molecular mass 15·5 and 17·0 kDa were seen. When variants of strain E7476 were isolated, loss of ST and MRHA together was associated with loss of a 98 MDa plasmid, while loss of ST alone correlated with plasmid deletion. An absorbed anti-strain E7476 antiserum reacted specifically with the 15·5 and 17·0 kDa polypeptides in Western immunoblotting and bound to the intact fimbriae by immuno-electron microscopy. When this antiserum was used in an ELISA to examine other strains of serogroup O166, a positive reaction was obtained with all the ST- and MRHA-positive strains. One strain of serotype O71:H27 and two strains of serotype O98:H- also reacted with the absorbed anti-strain E7476 antiserum. The antiserum did not react with ETEC carrying known colonization factors. E. coli K12 and a number of E. coli of different serotypes carrying a plasmid coding for ST transferred from strain E7476, all gave MRHA and reacted with the absorbed anti-strain E7476 antiserum. The term putative colonization factor O166 (PCFO166) is proposed to describe the adhesive factor(s) on ETEC of serogroup O166 because of the similarity of properties with those of known colonization factors.

Molecular cloning from a plasmid encoding colonization factor antigen I (CFA/I) and heat-stable enterotoxin isolated two regions, 1 and 2, that are required for the production of CFA/I fimbriae. The level of CFA/I synthesis measured by ELISA was similar in an Escherichia coli K12 strain carrying regions 1 and 2 cloned separately on compatible plasmid vectors to that in the same strain containing the parental plasmid. The structural gene for the CFA/I fimbrial subunit was within region 1. This region directed production in E. coli minicells of at least six independent polypeptides, of which the fimbrial subunit and at least three others appeared to be synthesized as precursor molecules that underwent processing. Cloned DNA containing CFA/I region 2 specified three polypeptides in minicells. Attempts to reduce the size of the cloned region 1 resulted in a derivative plasmid that carried the CFA/I structural gene but did not complement a region-2 recombinant plasmid to restore production of CFA/I fimbriae.

Two hundred forty-four specimens of Escherichia coli isolated in Bangladesh and Thailand and identified as enterotoxin producers were tested for the presence of adhesion antigens by mannose-resistant hemagglutination, immunodiffusion, and enzyme-linked immunosorbent assays (ELISAs). Specific antisera to the antigens colonization factor antigen (CFA)/I, CFA/II (consisting of coli surface antigens [CS] 1, 2, and 3), and putative colonization factor antigen (PCF) 8775 (consisting of CS4, 5, and 6) were used in immunodiffusion tests and ELISAs. The results showed that the antigens could be detected in more strains by ELISA than by immunodiffusion. Twenty-nine percent of specimens of E. coli from Thailand and 47% from Bangladesh carried an adhesion antigen. Many of the strains had lost the ability to produce enterotoxins. Forty percent of strains from Thailand and 64% from Bangladesh that were still enterotoxigenic carried adhesion factors. These antigens were found on strains with heat-stable or heat-stable and heat-labile enterotoxin but not on strains producing only heat-labile enterotoxin. PCF8775 antigens were associated mainly with strains from Bangladesh, where 10 strains that produced only CS6 were detected.

Nineteen enterotoxigenic Escherichia coli strains of serogroup O78, isolated in different geographical areas from humans with diarrheal diseases, were tested for their ability to transfer enterotoxin production. All of the strains originally produced heat-labile enterotoxin, and 16 also produced heat-stable enterotoxin and colonization factor antigen I. Plasmids coding for the production of heat-labile enterotoxin only were transferred from 13 strains. Some properties of these plasmids were compared. All were fi+, but they could be divided into three groups on the basis of their incompatibility reactions, ability to restrict E. coli K-12 phages, and size. The three heat-labile enterotoxin plasmids isolated from African strains all belonged to one enterotoxin plasmid group. The heat-labile enterotoxin plasmids from the Asian strains were divided into two groups, those from serotype O78.H11 differing from those from serotype O78.H12.

Entertoxigenic Escherichia coli (ETEC) strains of nineteen serogroups which produced colonization factors (coli-surface-associated antigens CS5, CS6, CS7 and CS17, colonization factor antigen CFA/III and putative colonization factors PCFO159:H4, PCFO166 and PCFO9) were tested for hybridization with a DNA probe containing the cfaD sequence that regulates expression of CFA/I. Strong colony hybridization, similar to that with the CFA/I-positive control strain H10407, occurred with ETEC strains of serogroups O27, O159 and O169 which produced CS6 antigen, and with all the strains which produced PCFO166 fimbriae. Weak colony hybridization, compared to the control strain, was found with ETEC producing CS5 fimbriae with CS6 antigen, CFA/III fimbriae with CS6 antigen, CS7 fimbriae or PCFO159:H4 fimbriae. CS6-antigen-positive strains of serogroups O79, O89 and O148 and all the CS17-antigen-positive and PCFO9-fimbriae-positive strains were negative in colony hybridization tests with the cfaD probe. Plasmid DNA of nine ETEC strains and their colonization-factor-negative derivatives was tested for hybridization with the cfaD probe and with ST and LT oligonucleotide probes. The sequences that hybridized with the cfaD probe were on the plasmids which coded for enterotoxin production. Fifteen strains were transformed with NTP513, a recombinant plasmid which contains the CFA/I region 1 fimbrial subunit operon but lacks a functional cfaD sequence, in order to determine whether DNA in any of these strains could substitute for the cfaD sequence in the regulation of production of CFA/I fimbriae.(ABSTRACT TRUNCATED AT 250 WORDS)

Twenty Escherichia coli strains of serogroup O167 were examined. They all produced the two surface antigens CS5 and CS6 of the putative colonization factor PCF8775, together with heat-stable enterotoxin and colicin Ia. A plasmid coding for CS5, CS6, heat-stable enterotoxin, and colicin Ia was demonstrated in each strain.