Enteropathogenic Escherichia coli (EPEC) is a Gram-negative bacterial pathogen that adheres to human intestinal epithelial cells, resulting in watery, persistent diarrhoea. Despite the advances made in understanding EPEC–host cell interactions, the molecular mechanisms underlying watery diarrhoea have not been understood fully. Loss of transepithelial resistance and increased monolayer permeability by disruption of tight junctions has been implicated in this process. Apart from disruption of tight junctions, an important factor known to regulate monolayer permeability is E-cadherin and its interaction with β-catenin, both of which constitute the adherens junctions. Our previous studies using HEp-2 cells demonstrated the morphological and cytoskeletal changes caused by cell-free outer membrane preparations (OMPs) of EPEC. In this study, we have shown that EPEC and its OMP induce significant changes in the adherens junctions of Caco-2 monolayers. We also observed significant phosphorylation of protein kinase Cα (PKCα) in cells treated with either whole EPEC or its OMP. Immunoprecipitation of cell lysates with anti-E-cadherin and probing with phospho-PKCα monoclonal antibodies and anti-β-catenins revealed that in these cells, phosphorylated PKCα is associated with cadherins, leading to the dissociation of the cadherin/β-catenin complex. Immunofluorescence showed β-catenins dissociated from the membrane-bound cadherins and redistributed into the cytoplasm. Expression of dominant negative PKCα reversed these effects caused by either whole EPEC or its OMP and also reduced the associated increase in monolayer permeability. It is possible that this mechanism may complement the earlier known pathways for loss of barrier function involving myosin light chain kinase activation and also may play a role in causing host cell death by apoptosis.

Enteropathogenic Escherichia coli (EPEC) causes persistent infantile diarrhoea. This nontoxigenic E. coli exhibits a complicated pathogenic mechanism in which its outer membrane proteins and type III secretory proteins damage intestinal epithelium and cause diarrhoea. In accordance with this, our previous study using HEp-2 cells demonstrated cytopathic effects caused by cell-free outer membrane preparations of EPEC. In this study, we report the extrusion of actin-positive strands from HEp-2 and Int 407 cells when treated with outer membrane preparations. An interesting observation of this work, perhaps relevant to the characteristic localized three-dimensional colony formation of EPEC, is the attachment of a wild type EPEC strain to these actin-positive strands.

The ability of the siderophores aerobactin and enterochelin to remove iron from transferrin is reported. Aerobactin removes iron from both high‐affinity sites on the transferrin molecule, but shows a marked preference for the C‐terminal site. This preference is different to that of many iron chelators. Enterochelin removes iron perferentially from the N‐terminal site. No evidence for synergism between aerobactin and bidentate ligands could be detected.

To assess the importance of aerobactin-mediated iron uptake as a bacterial virulence determinant in animal infections, a total of 576 strains of Escherichia coli isolated from cattle, chickens, sheep and pigs were screened by colony hybridization to determine the presence of the aerobactin genetic determinants, and by a bioassay to detect aerobactin secretion in iron-limited conditions. Results obtained by the two complementary methods correlated well. The incidence of the aerobactin system was very high among septicaemia isolates, particularly those from cattle and chickens, an observation that strongly suggests an important role for this mechanism of iron assimilation in pathogenesis. On the other hand, the incidence of the aerobactin system among mastitis strains was not significantly higher than among faecal isolates from healthy animals. No classical enterotoxigenic E. coli strains tested carried the aerobactin genetic determinants. Although most strains that produced aerobactin were also able to make colicin V, the fact that the two characteristics existed separately in a significant minority of isolates suggested that colicin testing alone could not be reliably used to determine the presence of the aerobactin system.

Single-gene mutants of Escherichia coli defective in aerobactin biosynthesis were incubated under non-growing conditions for 2 h with radiolabelled lysine. Analysis of the intermediates produced suggested that acetylation of lysine may be the first step in aerobactin production.

Plasmid pABN1 contains the genetic determinants for the aerobactin iron uptake system of plasmid ColV-K30. Transposon Tn1000 mutants of pABN1 defective in synthesis of a 50,000-dalton polypeptide were found neither to secrete nor to accumulate aerobactin, but were not impaired in iron transport functions, clearly indicating a role for this polypeptide in aerobactin biosynthesis. RNA-DNA hybridization studies with probes spanning the entire aerobactin gene cluster showed that the system is regulated at the transcriptional level by the availability of iron in the external medium. When induced by low-iron stress, all five genes of the cluster were transcribed at a uniformly high level. When repressed by excess iron, transcripts of the four biosynthesis genes were some 30-fold less abundant in the case of the parental ColV-K30 plasmid and 10-fold less for the recombinant plasmid pABN1, whereas the receptor gene in either plasmid was transcribed at only about a third of the induced level.

Preparations of ColEl plasmid DNA synthesized in the presence of chloramphenicol were separated into samples having gaps resulting from removal of ribonucleotides in one or the other of the complementary DNA strands. These samples were used as templates for repair resynthesis reactions using DNA polymerase of Rous sarcoma virus and α-32P-labeled deoxyribonucleoside 5′-triphosphates. Reactions involved the incorporation of each labeled nucleotide in the presence of three unlabeled nucleotides, and also the incorporation of all four labeled nucleotides followed by complete digestion and electrophoretic separation of the products. By these two methods the RNA integrated in the light strand of ColEl DNA was found to comprise an average of 38 ribonucleotides with a base composition of 17G, 5A, 8C, and 8U. The RNA segment in the heavy strand consists of an average of 15 ribonucleotides with a base composition of 5G, 2A, 4C, and 4U.

Supercoiled plasmid molecules sensitive to nicking by RNase or alkali have been shown to accumulate during replication of colicinogenic factor E1 (ColE1) in Escherichia coli in the presence of chloramphenicol. The possibility that this sensitivity is due to the covalent integration of RNA molecules during the synthesis of plasmid DNA is supported by the demonstration that (a) strands of supercoiled ColE1 newly replicated in the presence of chloramphenicol exhibit sensitivity to RNase and alkali treatment, while (b) RNase- and alkali-resistant circular strands of plasmid DNA synthesized either before or after the addition of chloramphenicol remain resistant during subsequent replication of the plasmid in the presence of chloramphenicol. Furthermore, newly made plasmid DNA strands cannot act as templates for further rounds of replication if they possess an RNA segment. The existence of a repair mechanism for the removal of the RNA segment from supercoiled ColE1 DNA molecules was demonstrated by pulse-chase experiments. It was observed that the proportion of RNase-sensitive molecules is considerably higher in pulse-labeled as compared to continuously labeled ColE1 DNA synthesized in the presence of chloramphenicol, and the proportion of pulse-labeled ColE1 DNA that is RNase sensitive is greatly reduced during a chase period. Removal of the RNA segment is also carried out effectively at the restrictive temperature in temperature-sensitive DNA polymerase I mutants. In a survey of other bacterial mutants defective in the repair of damaged DNA, a substantial increase in the rate of accumulation of RNase-and alkali-sensitive supercoiled ColE1 DNA in the presence of chloramphenicol was observed in recBC and uvrA mutants in comparison with the wild-type strains.

Pulse-labeled ColEl DNA molecules, undergoing replication in Escherichia coli cells either in the absence or presence of chloramphenicol, were extracted and purified by neutral sucrose density gradient sedimentation and equilibrium centrifugation in an ethidium bromide-cesium chloride gradient. In the dye-buoyant density gradient, the replicating molecules were found in regions between the supercoiled and open-circular nonreplicating plasmid DNA, as well as in the open-circular region. In a neutral sucrose gradient, peaks of pulse label were found in the region of 26 to 38 S as well as at the 23 and 17 S positions corresponding to the positions of supercoiled and open-circular ColEl DNA. In alkaline sucrose gradient, nascent ColEl DNA was found to sediment as discrete peaks corresponding to 5-6, 7-9, and 14-16 S, indicating that at least one growing strand of the replicating molecule is produced discontinuously. In the electron microscope, many of the molecules appeared as partially supercoiled structures containing two open-circular branches of equal length, of less than 20% to more than 90% replicated. Branched open-circular molecules were not observed to any significant extent without prior treatment to induce single-strand scissions. The parental strands of the replicating molecules were determined to be covalently closed, but the superhelical density of the DNA was shown to be progressively decreased as replication proceeded.

Mutants of the dnaA, dnaC, dnaD, polC, dnaF and dnaG gene loci were tested for their capacity for colicinogenic plasmid E1 (ColE1) replication at a non-permissive temperature. It was found that ColE1 replication was independent of the dnaA gene function and dependent on dnaC, D, F and G. ColE1 replication in the polC mutant E486 continued for several hours but at a greatly reduced rate. No effect was found of the dnaG mutation on thymine-deprivation-induced “priming” of ColE1 replication at the non-permissive temperature. The mutants also were tested for aberrant replication intermediates of plasmid DNA as well as a temperature sensitive supercoiled DNA-protein relaxation complex. RNA-containing supercoils were found to accumulated in a polC mutant also blocked for protein synthesis.

In Bacillus subtilis 168, at least three enzymes of arginine biosynthesis are repressed, and the enzymes arginase and ornithine transaminase (OTAse) are induced, by arginine.The strain is inhibited by the analogue arginine hydroxamate; Kisumi et al. (Appl.Microbial., 22, 987, 1971) showed that some resistant mutants excrete arginine.We obtain spontaneous arginine hydroxamate-resistant mutants at frequencies of the order of 3 x lO-.Although some isolates are unaltered in the activity of enzymes of the arginine pathways, many show reduced levels of arginase and OTAse.We have so far identified seven such phenotypic classes.In all of these, arginase and OTAse are both affected but to different extents, and we therefore tentatively propose that they are control mutants.In at least three mutant classes, one or more of the biosynthetic enzymes is synthesised at a high partially or completely constitutive level also.We conclude that the control systems for repression of the biosynthetic enzymes and induction of the catabolic enzymes share at least one component.It is noteworthy that these pathways in the yeast Saccharomces cereoisiae, which are biochemically similar to those in B. subtilis, show (at least formally) the same linked control (Wiame, Current Topics in Cellular Regulation, 4, 1, 1971).Representatives of six classes have been roughly mapped by transduction with phage PBS I.They are c.15 per cent co-transducible with csA.