Summary. Broth-culture filtrates of Campylobacterpylori induced non-lethal cytopathic effects in vitro in 7 of 9 mammalian cell lines tested. Transmission electronmicroscopy revealed that the response consisted of intracellular vacuolisation. Intestine 407 cells were among the most responsive and were used for routine assay. About 55% of isolates of C. pylori tested, originating from four geographic regions worldwide, produced cytotoxic activity. The activity was neutralisable by specific antisera to broth-culture filtrates or to sonicated bacteria but not by antisera to other bacterial preparations. Cytotoxic activity was heat-labile (70°C for 30 min), was proteasesensitive and ammonium-sulphate precipitable. It did not pass through an ultrafiltration membrane with a nominal mo1.-wt limit of 100 x lo3. It was concluded that C. pylori can produce a factor that alters cultured cells in vitro. The relevance of this factor to the pathogenesis of gastritis associated with C. pylori remains to be determined.

Summary. Non-motile variants of Helicobacterpylori (strain 26695) occurred with a frequency of 1.6 (SD 0.4) x lo-* variants/cell/division cycle, and reversion to the motile form occurred with a frequency of < lo-' variants/cell/division cycle. The two forms remained > 90 YO pure for up to 50 cell divisions and differed only in the presence or absence of motility and flagella. Bacteria were recovered from nine of 10 gnotobiotic piglets inoculated orally with motile H. pylori, but from only two of eight inoculated with the non-motile variant. The motile form survived for 21 days in infected piglets, but the non-motile variant survived for only 6 days. Bacteria recovered from piglets inoculated with the non-motile variant were non-motile. These data support the hypothesis that motility is a colonisation factor for H . pylori.

A mutant strain of Helicobacter pyloni with weak urease activity was created by using N-methyl-N'-nitro-N-nitrosoguanidine. The urease activity of the mutant (0.036 0.009 nmol of urea per ,ug of bacterial protein per min) was 0.4% of that of the parental strain (8.20 ± 2.30 nmol of urea per ,ug of bacterial protein per min). The mutant was otherwise indistinguishable from the parental strain. Both demonstrated prominent catalase and oxidase activities, and both produced vacuolating cytotoxin. Restriction endonuclease and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) patterns and ultrastructure were identical for the two strains. The mutant was fully motile, as evaluated by spreading in soft agar and by direct microscopic examination. Growth rate and colony size and morphology were identical for the mutant and parental strains. Seventeen gnotobiotic piglets were challenged with either the mutant or the parental strain and sacrificed 3 or 21 days after challenge. Gastric tissue was examined histologically and cultured for H. pyloni. Of seven piglets challenged with the parental strain, all became infected. H. pylori was not recovered from any of 10 piglets challenged with the urease-negative strain. Lymphofollicular gastritis was present in all seven piglets challenged with the parental strain but in none of the piglets challenged with the urease-negative strain. These results suggest that prominent urease activity is essential for colonization by H. pylori.

Sixteen cases of verotoxin producing Escherichia coli (VTEC) O 157:H7 Phage Type 49 infection were identified in the North West of England from 1 September to 1 November 1991, eight of whom lived in or around the same large town. Eleven of the cases were aged 10 years or less, and five of the affected children developed haemolytic uraemic syndrome. A case control study demonstrated a strong association between VTEC O 157:H7 PT 49 infection and the consumption of a locally produced live yoghurt. This is the first time that an outbreak of VTEC O 157 infection has been linked to the consumption of yoghurt and this vehicle of infection should be considered when investigating such outbreaks in future.