Search

While searching for Escherichia coli O157 in the aquatic environment of Calcutta using an immunodetection procedure, we fortuitously detected five strains of Citrobacter braakii, which cross-reacted with the commercially available O157 polyvalent antiserum. The five C. braakii isolates gave positive results when a sensitive dot-ELISA was performed with E. coli O157 monoclonal antibody. Further, the O157 monoclonal antibody recognized the bands of proteinase K treated whole cells of lipopolysaccharide of all the C. braakii isolates. Apart from weak reactions with two or three of the DNA probes, all the C. braakii strains did not hybridize with the other probes spanning the minimum region required for O157 O-antigen biosynthesis. These strains did not possess any of the virulence genes that are commonly found in the Shiga toxin-producing E. coli (STEC) specially the serotype O157∶H7. Therefore, it appears that the serological cross-reaction between C. braakii and E. coli O157 antiserum is based on structural mimicry between the O-polysaccharide of C. braakii and E. coli O157.

Active surveillance of Vibrio parahaemolyticus infection among hospitalized patients in Calcutta, India, showed the appearance of the O4∶K68 serovar for the first time in March 1998 alongside the continued predominant incidence of the O3∶K6 serovar. Strains belonging to both these serovars have been reported to possess pandemic potential. The genomes of O3∶K6 and O4∶K68 strains and for comparison, non-O3∶K6 and non-O4∶K68 strains isolated from two different countries, India and Thailand, were examined by different molecular techniques to determine their relatedness. The O3∶K6 and O4∶K68 strains from Calcutta and Bangkok carried the tdh gene but not the trh gene. Characterization of representative strains of these two serovars by ribotyping and by arbitrarily primed-polymerase chain reaction (AP-PCR) showed that the isolates had identical ribotype and DNA fingerprint. Pulsed-field gel electrophoresis (PFGE) performed with the same set of strains yielded nearly similar restriction fragment length polymorphism (RFLP) patterns for the O3∶K6 and O4∶K68 isolates from Calcutta and Thailand. Phylogenetic analysis of the NotI RFLP showed that the O3∶K6 and O4∶K68 strains formed a cluster with 78–91% similarity thus indicating close genetic relationship between the two different serovars isolated during the same time-frame but from widely separated geographical regions. The non-O3∶K6 and non-O4∶K68, in contrast, showed different ribotype, AP-PCR and PFGE patterns.

This study was initiated to search for a homologue of the Vibrio cholerae zotgene in uropathogenic Escherichia coli(UPEC) using a specific DNA probe. The faint signal obtained at low stringency with some UPEC strains associated with prostatitis cases prompted us to examine UPEC strains by PCR using primers designed from the conserved regions of the proteins of the Zot group of putative NTPases containing the classical NTP binding motif. This led to the discovery of a DNA fragment in UPEC strains which hybridized with a probe designed from the PCR. Further analysis of this DNA fragment revealed an ORF which was designated as uropathogenic specific protein (Usp). The gene encoding Usp was 1038 bp long and codes for 346 amino acids with an appropriate SD sequence. Upstream and downstream analysis of usprevealed motifs of prokaryotic consensus promoters and three small ORFs with SDs and ribosome binding sites transcribed in the same direction of usp. The proximity of these set of genes in a specific area of the bacterial chromosome resembling a block of genes preferentially associated with UPEC coupled with the presence of a motif matching that of a Tn3 transposon family lead us to believe that this could be an hitherto unknown pathogenicity island.

The DNA hybridization assay for genes encoding for Escherichia coli enterotoxins was used to examine water specimens in Thailand. In a reconstruction experiment, the DNA hybridization assay was 10(4) times more sensitive than testing random E. coli in the Y-1 adrenal and suckling mouse assays in identifying enterotoxigenic E. coli (ETEC) in water. Drinking and bathing water collected from 2 of 10 different homes of individuals with ETEC-associated diarrhea and 6% (5 of 78) and 11% (11 of 78) of drinking and bathing water samples collected from homes of individuals with diarrhea without ETEC infections, as well as 6% (5 of 77) and 8% (6 of 77) of drinking and bathing water collected from homes in which no inhabitants had diarrhea, were homologous with the DNA probes. Ten E. coli from each of the 31 water specimens which contained bacteria which were homologous with the DNA probes were tested in the Y-1 adrenal and suckling mouse assay. In only 2 of these 31 specimens could ETEC be identified with the standard assays. The DNA hybridization assay is a much more sensitive means of detecting organisms carrying genes coding for enterotoxin production than testing 10 individual colonies in the Y-1 adrenal and suckling mouse assays. This novel application of recombinant DNA technology provides a sensitive method of detecting organisms carrying genes coding for enterotoxin, and this method will be useful in defining the epidemiology of ETEC.

Monoclonal antibodies were raised against Barber antigen (Ba) of Salmonella typhi 0901. Antibodies produced to antigen 9 of group D salmonellae were used in double- and triple-sandwich antibody enzyme-linked immunosorbent assays (ELISAs) for detecting antigen 9 in urine and plasma specimens from three groups of patients and 49 controls. The triple-antibody ELISA detected the antigen in urine samples from 11 of 18 (65%) patients with hemoculture-proven typhoid (group 1) and 12 of 39 (31%) patients with clinical features compatible with typhoid but whose hemocultures were negative (group 2). This ELISA was negative in three patients from whom Salmonella paratyphi A, Escherichia coli, and Klebsiella pneumoniae (group 3) were isolated by hemoculture and in all healthy controls. The double-antibody sandwich ELISA was positive in 41 and 15% of urine samples from patients in groups 1 and 2, respectively, and was negative with samples from two patients from group 3 and all controls. The sensitivity and specificity compared with those for healthy controls were 65 and 100%, respectively, for the triple-antibody ELISA. Although as little as 7.8 ng of homologous lipopolysaccharide could be detected, background in clinical specimens prevented accurate interpretation of the detection of this antigen in serum. Results were best with urine specimens.