Binding of enterotoxigenic Escherichia coli to isolated enterocytes and intestinal mucus

Binding of human enterotoxigenic Escherichia coli (ETEC) to the small intestine is a prerequisite for colonization and is mediated by colonization factor (CF) antigens. Coli surface antigen 6 (CS6) is considered a CF but binding to isolated enterocytes has not been established. In this study bacteria expressing CS6 were analysed for binding to enterocytes from human and rabbit small intestine, isolated using either an EDTA-containing buffer or a buffer devoid of EDTA. We found that the bacteria bound to enterocytes from rabbit ileum and human duodenum, but only when the cells had been isolated in the absence of EDTA. Pretreatment of rabbit enterocytes with meta-periodate resulted in a decreased proportion of cells with bound bacteria. Purified CS6, and for comparison other ETEC CFs, were also tested for binding to different human and rabbit mucus fractions. These analyses showed that purified CS6 bound to mucus from rabbit duodenum and ileum as well as from human duodenum, jejunum and ileum and that this binding was abolished by pretreatment of the mucus material with meta-periodate or Proteinase K. CFA/I, CS1 to CS5, CS7, CS17, putative CF (PCF) O159 (CS12), PCFO166 (CS14), and CFA/III (CS8) also bound to the rabbit mucus material although with different patterns; the binding of CS2 and CS5 was abolished by meta-periodate treatment. Thus, ETEC bacteria expressing CS6 might bind to carbohydrate-containing structure(s) in the apical membrane of isolated rabbit ileal and human duodenal enterocytes that could probably be released by EDTA treatment. In addition, CS6 and other ETEC CFs bind to component(s), in some instances protein-associated carbohydrate structures, in mucus fractions from small intestine.     

Isolation of enterotoxigenic Escherichia coli from British troops in Saudi Arabia

Specimens from 181 patients with diarrhoea were examined by a Military General Hospital in a 3-month period during deployment of troops to Saudi Arabia in 1990/1. DNA probes for heat labile (LT) and heat stable (ST) enterotoxin genes identified enterotoxigenic Escherichia coli (ETEC) in 47 of the specimens (26%) and 49 ETEC strains were isolated. The majority (55%) belonged to a novel ETEC serotype having the O-antigen 159 and a flagellar antigen designated as a provisional new type. They produced ST and the coli surface associated antigen (CS)6. Strains of serotype O6:H16 represented 22% of the ETEC examined. They produced ST, LT and CS3 together with either CS1 or CS2. The remaining ETEC belonged to seven O:H serotypes. Overall, ST was the only enterotoxin gene identified in 73% of the ETEC and 67% of the strains expressed CS6 in the absence of other colonization antigens. Resistance to three or more antibiotics was observed in 53% of the ETEC, including most of the O159 strains.     

Development and use of a multiplex PCR system for the rapid screening of heat labile toxin I, heat stable toxin II and shiga-like toxin I and II genes of Escherichia coli in water

Enterotoxigenic Escherichia coli (ETEC) may produce heat-labile toxin (LT) I and LTII and heat-stable toxin (ST) I and STII, while shiga toxin producing E. coli (STEC) strains, including enterohaemorrhagic E. coli (EHEC), may produce shiga-like toxin (SLT) I and/or SLTII. Both ETEC and STEC are pathogenic to humans, pigs and cattle. As contamination of environmental water by any of these pathogenic E. coli cells is possible, a multiplex polymerase chain reaction (PCR) system for the rapid screening of LTI, STII, and SLTI and SLTII genes of E. coli was developed. The PCR primers used were the SLTI and SLTII genes specific primers developed by the present authors and the LTI and STII genes specific primers reported by other laboratories. The detection specificity of this multiplex PCR system was confirmed by PCR assay of ETEC, STEC and other E. coli cells as well as non-E. coli bacteria. Its detection limit was 10(2)-10(3) cfu each of the target cells per assay. When this multiplex PCR system was used for the rapid screening of LTI, STII ETEC and STEC in water samples such as tap, underground and lake waters, it was found that after the enrichment step, as few as 10(0) cells 100 ml-1 of the water sample could be detected. Therefore, this PCR system could be used for the rapid monitoring of ETEC and/or STEC cells contaminating water samples.     

Antibody responses in humans against coli surface antigen 6 of enterotoxigenic Escherichia coli

Enterotoxigenic Escherichia coli (ETEC) strains expressing only coli surface antigen 6 (CS6) have previously been isolated from patients with diarrhea, but the immunogenicity of CS6 has not been established in humans. We have detected CS6-specific immunoglobulin A responses in the feces and blood of patients convalescing from natural ETEC disease and of volunteers given an oral ETEC vaccine.     

An outbreak of foodborne illness caused by Escherichia coli O39:NM, an agent not fitting into the existing scheme for classifying diarrheogenic E. coli

An outbreak of gastrointestinal illness with clinical and epidemiologic features of enterotoxigenic Escherichia coli (ETEC) occurred among patrons of a restaurant during April 1991. Illnesses among several groups of patrons were characterized by diarrhea (100%) and cramps (79%-88%) lasting a median of 3-5 days. Median incubation periods ranged from 50 to 56 h. A nonmotile strain of E. coli (E. coli O39), which was negative for heat-labile (LT) and heat-stable (STa, STb) ETEC toxins, was isolated only from ill patrons. This organism produced enteroaggregative E. coli heat-stable enterotoxin 1 and contained the enteropathogenic E. coli gene locus for enterocyte effacement; it did not display mannose-resistant adherence, but produced attaching and effacing lesions in the absence of mannose on cultured HEp-2 cells. E. coli that are not part of highly characterized but narrowly defined groups may be important causes of foodborne illness.     

Evolutionary relationships among pathogenic and nonpathogenic Escherichia coli strains inferred from multilocus enzyme electrophoresis and mdh sequence studies

Within the species Escherichia coli, there are commensal strains and a variety of pathogenic strains, including enteropathogenic E. coli (EPEC), enterohemorrhagic E. coli (EHEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), and urinary tract infection (UTI) strains. The pathogenic strains are identified by serotype and by possession of specific virulence determinants (toxins and adhesions, etc.) encoded by either monocistronic genes, plasmids, or pathogenicity islands. Although there are studies on the relationships between selected pathogenic strains, the relatedness among the majority of the pathogenic forms to each other, to commensal E. coli, and to the genus Shigella (which has often been suggested to be part of E. coli) has not been determined. We used multilocus enzyme electrophoresis (MLEE) at 10 enzyme loci and the sequence of the mdh housekeeping gene to study the genetic relationships of pathogenic E. coli strains (including Shigella clones), namely, 5 EPEC strains (serotypes O111 and O55), 3 EHEC strains (serotype O157), 6 ETEC strains (serotypes O78, O159, and O148), 5 EIEC strains (serotypes O124, O28, and O112), and 13 Shigella strains representing clones Flexneri, Dysenteriae, Boydii, and Sonnei, to commensal E. coli strains. Both the MLEE and mdh sequence trees reveal that EPEC, EHEC, ETEC, EIEC, and UTI strains are distributed among the ECOR set groups, with no overall clustering of EPEC, ETEC, EIEC, or UTI strains. The genus Shigella is shown to comprise a group of closely related pathogenic E. coli strains. Six pathogenic strains, i.e., M502 (EIEC; O112ac:NM), M503 (EPEC; O111:H12), M526 (ETEC; O159:H4), M522 (EPEC; O111ac:H12), M524 (ETEC; O78:H11), and M506 (ETEC; O78:H11), were found to have mdh sequences identical to those of five ECOR group A strains (ECOR5, ECOR10, ECOR14, ECOR6, and K-12). All 11 strains are closely related by MLEE. The results indicate that pathogenic strains of E. coli do not have a single evolutionary origin within E. coli but have arisen many times. The results also suggest the possibility that any E. coli strain acquiring the appropriate virulence factors may give rise to a pathogenic form.     

Adhesion of human bifidobacterial strains to cultured human intestinal epithelial cells and inhibition of enteropathogen-cell interactions

Thirteen human bifidobacterial strains were tested for their abilities to adhere to human enterocyte-like Caco-2 cells in culture. The adhering strains were also tested for binding to the mucus produced by the human mucus-secreting HT29-MTX cell line in culture. A high level of calcium-independent adherence was observed for Bifidobacterium breve 4, for Bifidobacterium infantis 1, and for three fresh human isolates from adults. As observed by scanning electron microscopy, adhesion occurs to the apical brush border of the enterocytic Caco-2 cells and to the mucus secreted by the HT29-MTX mucus-secreting cells. The bacteria interacted with the well-defined apical microvilli of Caco-2 cells without cell damage. The adhesion to Caco-2 cells of bifidobacteria did not require calcium and was mediated by a proteinaceous adhesion-promoting factor which was present both in the bacterial whole cells and in the spent supernatant of bifidobacterium culture. This adhesion-promoting factor appeared species specific, as are the adhesion-promoting factors of lactobacilli. We investigated the inhibitory effect of adhering human bifidobacterial strains against intestinal cell monolayer colonization by a variety of diarrheagenic bacteria. B. breve 4, B. infantis 1, and fresh human isolates were shown to inhibit cell association of enterotoxigenic, enteropathogenic, diffusely adhering Escherichia coli and Salmonella typhimurium strains to enterocytic Caco-2 cells in a concentration-dependent manner. Moreover, B. breve 4 and B. infantis 1 strains inhibited, dose dependently, Caco-2 cell invasion by enteropathogenic E. coli, Yersinia pseudotuberculosis, and S. typhimurium strains.     

Hereditary isolated renal magnesium loss maps to chromosome 11q23

Dendritic cells (DC) expanded in the presence of GM-CSF from the bone marrow of C57BL/6 mice process Gram-negative bacteria expressing the model antigen Crl-OVA for peptide presentation on MHC class I molecules. Here we show that presentation of OVA(257-264) processed by DC co-incubated with E. coli expressing Crl-OVA, which contains the Kb-binding OVA(257-264) epitope, occurs by a cytosolic MHC-I presentation pathway. First, we demonstrate the requirement for the transporter associated with antigen processing (TAP) by showing that DC from TAP1-/- mice co-incubated with E. coli expressing Crl-OVA did not result in Kb presentation of OVA(257-264). Second, the proteasome inhibitor MG132 abrogated presentation of OVA(257-264) on Kb when C57BL/6 DC phagocytosed and processed E. coli expressing Crl-OVA. Third, inhibiting protein synthesis using cycloheximide or blocking exocytosis of newly synthesized proteins from the endoplasmic reticulum using brefeldin A abrogated presentation of OVA(257-264) processed from bacteria expressing Crl-OVA by C57BL/6 DC. Finally, peptide regurgitation and loading of OVA(257-264) on neighboring bystander Kb-expressing antigen-presenting cells after BALB/c (H-2d) DC phagocytosed E. coli expressing Crl-OVA could not be detected. Together, these data support a cytosolic MHC-I presentation pathway for OVA(257-264) processed from E. coli expressing Crl-OVA by bone marrow-derived DC.     

Clinical features of infections due to Escherichia coli producing heat-stable toxin during an outbreak in Wisconsin: a rarely suspected cause of diarrhea in the United States

In September 1994, a foodborne outbreak of enterotoxigenic Escherichia coli (ETEC) infection occurred in attendees of a banquet in Milwaukee. E. coli was isolated from stool specimens from 13 patients that were comprehensively tested; isolates from five patients were positive for E. coli producing heat-stable toxin, were biochemically identified and serotyped as E. coli O153:H45, and were all resistant to tetracycline, ampicillin, sulfisoxazole, and streptomycin. Diarrhea (100%) and abdominal cramps (83%) were the most prevalent symptoms in 205 cases; vomiting (13%) and fever (19%) were less common. The median duration of diarrhea and abdominal cramps was 6 days and 5 days, respectively. In the United States, health care providers rarely consider ETEC as a possible cause of diarrhea in their patients, and few laboratories offer testing to identify ETEC. Hence, outbreaks of ETEC infection may be underdiagnosed and underreported. As in this outbreak, the relatively high prevalence of diarrhea and cramps lasting > or = 4 days and the low prevalence of vomiting and fever can help distinguish ETEC infection from Norwalk-like virus infection and gastroenteritis due to other causes with incubation times of > or = 15 hours and can provide direction for confirmatory laboratory testing.     

Evaluation of three different STb assays and comparison of enterotoxin pattern over a five-year period in Swedish porcine Escherichia coli

The pig intestinal loop (PIL) assay, inhibition enzyme-linked immunosorbent assay (ELISA), and DNA hybridization assay were compared for analysis of Escherichia coli heat-stable enterotoxin b (STb) on 201 porcine E. coli strains. The DNA hybridization had a 95% correlation with the STb ELISA and was therefore chosen as the method for subsequent screening of enterotoxin genes: heat labile (LT), heat-stable a (STa), and/or STb. In contrast to the PIL assay, both the STb ELISA and DNA hybridization assays were more sensitive, reliable, reproducible, and showed good correlation with each other. Consequently, the STb ELISA is preferable for analysis of toxin preparations and screening of E. coli, whereas the DNA hybridization is better for large-scale epidemiologic screening. Escherichia coli strains (n = 437) associated with porcine diarrhea isolated in Sweden during 1989 were investigated. Of the strains, 135 (31%) were positive for at least one of these toxins and, therefore, designated enterotoxigenic E. coli (ETEC). Our results were compared with the enterotoxin pattern found in earlier studies of Swedish porcine strains. The only change in occurrence of toxins was found in strains isolated from piglets less than 1 week of age. LT- and STb-producing ETEC had decreased, and STa-producing ETEC had increased in prevalence. The occurrence of STb among ETEC of weaned pigs was 93%. This toxin was also found to be more common than STa when strains from all age groups were considered.     

Localization of a domain in the FimH adhesin of Escherichia coli type 1 fimbriae capable of receptor recognition and use of a domain-specific antibody to confer protection against experimental urinary tract infection

The FimH subunit of type 1-fimbriated Escherichia coli has been implicated as an important determinant of bacterial adherence and colonization of the urinary tract. Here, we sought to localize the functionally important domain(s) within the FimH molecule and to determine if antibodies against this domain would block adherence of type 1-fimbriated E. coli to the bladder mucosa in situ and in vivo in an established mouse model of cystitis. We generated translational fusion proteins of disparate regions of the FimH molecule with an affinity tag MalE, and tested each of the fusion products in vitro for functional activity. The minimum region responsible for binding mouse bladder epithelial cells and a soluble mannoprotein, horseradish peroxidase, was contained within residues 1-100 of the FimH molecule. We validated and extended these findings by demonstrating that antibodies directed at the putative binding region of FimH or at synthetic peptides corresponding to epitopes within the binding domain could specifically block type 1 fimbriae-mediated bacterial adherence to bladder epithelial cells in situ and yeast cells in vitro. Next, we compared the ability of mice passively immunized intraperitoneally with antisera raised against residues 1-25 and 253-264 of FimH or 1-13 of FimA to resist bladder colonization in vivo after intravesicular challenge with type 1-fimbriated E. coli. Only the antibody directed at the putative binding region of FimH (anti- s-FimH1-25) significantly reduced E. coli bladder infections in the experimental mouse model of urinary tract infections. Similar results were obtained when the mice were actively immunized with synthetic peptides corresponding to residues 1-25 and 253-264 of FimH or 1-13 of FimA. The mechanism of protection was attributed, at least in part, to inhibition of bacterial adherence to the bladder surface by s-FimH1-25-specific antibody molecules that had filtered through the kidneys into the urine. The level of FimH antibodies entering the bladder from the circulatory system of the immunized mice was found to be markedly enhanced upon bacterial challenge. The potential broad spectrum activity of the protective FimH antibody was indicated from its serologic cross-reactivity with various urinary tract bacterial isolates bearing type 1 fimbriae. These findings could be relevant in the design of an efficacious and broadly reactive FimH vaccine against urinary tract infections.     

Digoxigenin-labeled deoxyribonucleic acid probes for the enumeration of bifidobacteria in fecal samples

The numbers of bifidobacteria in fecal samples were specifically determined by colony hybridization with the mixture of digoxigenin-labeled DNA probes that were prepared from whole chromosomal DNA of Bifidobacterium longum 6001 and Bifidobacterium adolescentis 6003. These DNA probes strongly hybridized with DNA of B. longum, B. adolescentis, Bifidobacterium breve, Bifidobacterium suis, Bifidobacterium infantis, Bifidobacterium bifidum, Bifidobacterium angulatum, and Bifidobacterium animalis. Detectable positive signals with DNA of Bifidobacterium pseudolongum ssp. pseudolongum, Bifidobacterium catenulatum, and Bifidobacterium thermophilum were also found after hybridization. When dot-blot hybridization was performed with whole cells of 47 reference strains containing 11 species (16 strains) of bifidobacteria, all of the bifidobacteria tested could be specifically detected by using these DNA probes; Lactobacillus fermentum JCM 1173, however, showed a slight nonspecific signal. The counts of bifidobacteria by colony hybridization in the fecal samples of four of the five subjects were the same as the counts that were obtained by the conventional method using BL agar medium. Furthermore, no significant difference existed in the number of bifidobacteria that were determined by either method.     

Epitope analysis of the CS3 fimbrial subunit of human enterotoxigenic Escherichia coli and the construction of novel CS3::ST and CS3::LT-B immunogens

Enterotoxigenic E. coli (ETEC) are the major cause of traveler's diarrhoea and the CS3 fimbriae/fibrillae are expressed by most strains bearing the colonization factor CFA/II. The cstAH gene cluster determining CS3 biosynthesis has been previously cloned and sequenced and it has been shown that cstH encodes the major fimbrial subunit and cstA-G encode an assembly cassette. In the work described here we have sought to define the surface exposed domains on CS3 and to manipulate them so that CS3 can be used as a means of expressing foreign antigenic determinants on the bacterial surface. Using a panel of 21 monoclonal antibodies, which we have used in western blotting, immunofluorescence microscopy and colony blotting, together with computer predictions, we have identified three domains within CstH. Two of these sites were permissive for insertion and we have introduced, in-frame, either an epitope from the B subunit of LT (heat labile toxin) or the entire coding sequence of mature ST (heat stable toxin) to construct hybrid proteins. These proteins could be assembled into hybrid fimbriae which could be recognized by antibodies to both CS3 and the foreign epitope as shown by immunofluorescence microscopy and colony blotting. The immunogenicity of the constructs has been evaluated following both oral and intraperitoneal immunization of mice with the attenuated Salmonella typhimurium strain G30 harbouring the hybrid cst operons. Although plasmid stability is currently a problem, these experiments showed that antibodies to both the carrier and the foreign epitope were generated.     

In vivo signal transduction of tetrodotoxin-sensitive nociceptive responses by substance P given into the planta of the mouse hind limb

Diarrhea, sudden death after short duration of diarrhea and sudden death without apparent signs were observed in a herd of breeder pigs. Five pigs that died suddenly with diarrhea (SDD pigs) and 6 pigs that died suddenly without signs (SD pigs) were examined. The average age of the pigs was about 28 days. Twelve pigs of age 10 to 14 days old showing diarrhea (D pigs) were also examined. Eleven of them recovered. Large numbers of Escherichia coli were detected in all organs of every SDD and SD pig and in feces of D pigs. All of the isolates were identified as enterotoxigenic E. coli (ETEC) by the polymerase chain reaction (PCR). Porcine reproductive and respiratory syndrome (PRRS) virus cDNA was also detected from the lung of every SD and SDD pig by the RT-PCR. High and low titers of antibodies to PRRS virus were found in 10-day-old and 1-month-old pigs, respectively. In an experiment, 3 ETEC were isolated from 9 healthy weaning pigs during the quiescent stage in the herd. These data showed that growth of the ETEC was not active in healthy weaning pigs; however, following infection with PRRS virus ETEC infection became systemic and caused peracute death in the weaning pigs. It suggested also that infection with PRRS virus in 10-day-old pigs were protected by the colostral antibodies, and fatal infection by ETEC did not occur as a result.     

Oral immunization with a Salmonella typhimurium vaccine vector expressing recombinant enterotoxigenic Escherichia coli K99 fimbriae elicits elevated antibody titers for protective immunity

Bovine enterotoxigenic Escherichia coli (ETEC) continues to cause mortality in piglets and newborn calves. In an effort to develop a safe and effective vaccine for the prevention of F5(+) ETEC infections, a balanced lethal asd+ plasmid carrying the complete K99 operon was constructed and designated pMAK99-asd+. Introduction of this plasmid into an attenuated Salmonella typhimurium Deltaaro Deltaasd strain, H683, resulted in strain AP112, which stably expresses E. coli K99 fimbriae. A single oral immunization of BALB/c and CD-1 mice with strain AP112 elicited significant mucosal immunoglobulin A (IgA) titers that remained elevated for >11 weeks. IgA and IgG responses in serum specific for K99 fimbriae were also induced, with a prominent IgG1, as well as IgG2a and IgG2b, titer. To assess the derivation of these antibodies, a K99 isotype-specific B-cell ELISPOT analysis was conducted by using mononuclear cells from the lamina propria of the small intestines (LP), Peyer's patches (PP), and spleens of vaccinated and control BALB/c mice. This analysis revealed elevated numbers of K99 fimbria-specific IgA-producing cells in the LP, PP, and spleen, whereas elevated K99 fimbria-specific IgG-producing cells were detected only in the PP and spleen. These antibodies were important for protective immunity. One-day-old neonates from dams orally immunized with AP112 were provided passive protection against oral challenge with wild-type ETEC, in contrast to challenged neonates from unvaccinated dams or from dams vaccinated with a control Salmonella vector. These results confirm that oral Salmonella vaccine vectors effectively deliver K99 fimbriae to mucosal inductive sites for sustained elevation of IgA and IgG antibodies and for eliciting protective immunity.     

Mast cell chymase-like protease(s) modulates Escherichia coli lipopolysaccharide-induced vasomotor dysfunction in skeletal muscle in vivo

This study investigated whether short-term exposure to Escherichia coli lipopolysaccharide (LPS) elicits vasomotor dysfunction in skeletal muscle in vivo and, if so, whether perivascular mast cell proteases partly modulate this response. With intravital microscopy, we found that suffusion of E. coli LPS on the in situ hamster spinotrapezius muscle for 60 min elicits immediate vasoconstriction followed by vasodilation. Vasoconstriction is abrogated by SK&F 108566, a selective, nonpeptide angiotensin II (AT II) subtype 1 receptor antagonist, chymostatin and soybean trypsin inhibitor. These compounds also attenuate E. coli LPS-induced vasodilation. By contrast, superoxide dismutase, catalase and indomethacin attenuate only E. coli LPS-induced vasodilation. Endothelin receptor antagonists, lisinopril, leupeptin, Bestatin and DL-2-mercaptomethyl-3-guanidinoethylthiopropanoic acid are ineffective. Histochemical analysis of the spinotrapezius muscle reveals abundant perivascular mast cells with chymostatin-inhibitable chymase-like activity. Pretreatment of hamsters with compound 48/80 for 4 days curtails E. coli LPS-induced vasoconstriction and converts vasodilation to vasoconstriction. On balance, these data indicate that E. coli LPS stimulates perivascular mast cells in the in situ hamster spinotrapezius muscle to release an AT II-producing chymase-like protease(s). AT II thus produced elicits local vasoconstriction and elaborates reactive oxygen species which, in turn, generate vasodilator prostaglandins.     

Quaternary structure, Mg2+ interactions, and some kinetic properties of the beta-galactosidase from Thermoanaerobacterium thermosulfurigenes EM1

The beta-galactosidase from Thermoanaerobacterium thermosulfurigenes EM1 was found to be a dimer with a monomer molecular weight of about 85,000. It lacks the alpha-peptide and an important alpha-helix that are both needed for dimer-dimer interaction and there is no homology in other important dimer-dimer interaction areas. These differences in structure probably account for the dimeric (rather than tetrameric) structure. Only 0.19 Mg2+ bound per monomer and Mg2+ had only small effects on the activity and heat stability. The absence of residues equivalent to Glu-416 and His-418 (two of the three ligands to Mg2+ in the beta-galactosidase from Escherichia coli) probably accounts for the low level of Mg2+ binding and the consequent lack of response to Mg2+. Both Na+ and K+ also had no effect on the activity. The enzyme activity with o-nitrophenyl-beta-D-galactopyanoside (ONPG) was very similar to that with p-nitrophenyl-beta-D-beta-D-galactopyranoside (PNPG) and the ONPG pH profile was very similar to the PNPG pH profile. These differences are in contrast to the E.coli beta-galactosidase, which dramatically discriminates between these two substrates. The lack of discrimination by the T. thermosulfurigenes beta-galactosidase could be due to the absence of the sequence equivalent to residues 910-1023 of the E. coli beta-galactosidase. Trp-999 is probably of the most importance. Trp-999 of the E. coli beta-galactosidase is important for aglycone binding and ONPG and PNPG differ only in their aglycones. The suggestion that the aglycone site of the T. thermosulfurigenes beta-galactosidase is different was strengthened by competitive inhibition studies. Compared to E. coli beta-galactosidase, D-galactonolactone was a very good inhibitor of the T. thermosulfurigenes enzyme, while L-ribose inhibited poorly. These are transition-state analogs and the results indicate that T. thermosulfurigenes beta-galactosidase binds the transition state differently than does E. coli beta-galactosidase. Methanol and glucose were good acceptors of galactose, and allolactose was formed when glucose was the acceptor. Allolactose could not, however, be detected by TLC when lactose was the substrate. The differences noted may be due to the thermophilic nature of T. thermosulfurigenes.     

Experimental investigation of the distribution of residual strains in the artery wall

Diarrheal diseases often result from ingestion of contaminated water or food. The population of La Paz, Bolivia is directly or indirectly exposed to the sewage-contaminated La Paz River. We conducted a bacteriologic survey of the La Paz River to quantify the level of bacterial contamination, with particular reference to enteropathogens. A total bacterial count exceeding 10(6) colony-forming units (CFU)/ml, including lactose fermenting and nonfermenting, gram-negative bacilli of approximately 10(5) CFU/ml, respectively, were detected in river water samples collected near two densely populated areas. A total bacterial count of 10(5) CFU/ml was also detected at the most downstream area of the river near a sparsely populated area. At four sampling locations, several enteropathogens were detected, including five enterotoxigenic Escherichia coli (ETEC) (serotype O6, O15, and O159), two enteropathogenic E. coli (EPEC) (serotype O44), two enteroinvasive E. coli (EIEC) (serotype O29), and three Salmonella O4 group isolates. The heat-labile enterotoxin gene and the invasive toxin gene were detected in all ETEC and EIEC isolates by polymerase chain reaction analysis. Nine isolates of E. coli were found by the agar dilution method to be susceptible to ampicillin, kanamycin, nalidixic acid, tetracycline, and chloramphenicol, and ampicillin resistance was found in only two isolates of EIEC 7-4 (serotype O29) and EPEC 7-5 (serotype O44). Ampicillin resistance was coded on plasmids and transferred conjugatively to E. coli chi1037 at a frequency of 10(-5) CFU/donor by the broth mating method. Strains of Aeromonas caviae, which can cause diarrheal disease in infants, were detected in vegetables grown in fields irrigated by water from the La Paz River. The survival of nine isolates of E. coli in filtered river water was compared with that of laboratory strains (E. coli chi1037, W3110, and ATCC29577). The survival time of seven isolates, excluding two ampicillin-resistant isolates, was markedly longer than that of the laboratory strains. Our results show a high bacterial contamination of the La Paz river and suggest that such levels may contribute to the high incidence of diarrheal disease in the city of La Paz.     

Determination of the amount of information in nucleotide sequences

Some strains of enterotoxigenic Escherichia coli associated with human diarrhoeal disease produce a class of pili represented by those called CS1. For the assembly of the major-pilin subunit, CooA, into pili, each of four linked genes, cooB, A, C, and D, is required. In this study, we have determined the subcellular localization of CooB, C and D, and investigated the molecular interactions of these proteins using specific antisera. CooD appears to be an integral pilus protein because it co-purifies with, and is strongly associated with, CS1 pili. In keeping with its role as an assembly protein, the CooD minor pilin (when overexpressed in CS1-piliated strains) was detected in periplasmic intermolecular complexes with the major-pilin subunit CooA. CooB is an assembly protein found exclusively in the periplasm of CS1-piliated strains. CooB also forms periplasmic intermolecular complexes with CooA, but does not constitute part of the final pilus structure. Immunoblot analysis of cell fractions showed that CooC is an outer membrane protein of CS1-piliated E. coli. Based on this information, we have proposed a model for CS1-pilus assembly which is very similar to the model for polymerization of the PapA pilin of uropathogenic E. coli. As the assembly proteins of Pap and CS1 pili are structurally unrelated, this may represent a case of convergent evolution.