- Volume 40, Issue 4, 1994

The effect of growth in iron-excess or iron-limitation conditions on the invasi veness for HeLa cells of Escherichia coli HB101 carrying plasmid pRI203 which bears the invasion gene of Yersinia pseudotuberculosis was examined. Iron-limitation reduced adhesion and the number of organisms internalised by HeLa cells by about 100-fold. The reduced adhesion of iron-starved bacteria correlated with reduced hydrophobicity and the reduced invasiveness appeared to depend on the plasmid copy number, which was 3·5-fold less than in bacteria grown in iron excess.

Shiga-like toxin-producing (SLT) Escherichia coli, particularly those belonging to serogroup O157, are responsible for haemorrhagic colitis, haemolytic uraemic syndrome and some cases of gastro-enteritis. The rapid and reliable diagnosis of all these infections is necessary for correct patient management and for epidemiological reasons, but is rarely possible with present methods. We compared the efficacy of two methods, (i) the culture of faeces in broth that contained mitomycin C followed by enzyme-linked immunosorbent assay (ELISA) for SLTs, and (ii) the culture of faeces on sorbitol MacConkey agar (SMA), in the detection of infections caused by SLT-producing E. coli. SLT-producing E. coli O157 strains were isolated on SMA from 42 of 475 faecal samples, but SLTs were detected by ELISA in culture supernates or lysates of 54 of 475 samples. SLT-producing E. coli strains were isolated subsequently from 11 of 12 ELISA-positive, SMA culture-negative samples by a colony blot technique. In four cases, SLT-producing E. coli of serogroups other than O157 were isolated and in seven cases E. coli O157 was isolated in small numbers. The ELISA is a rapid and sensitive technique for the diagnosis of SLT-producing E. coli infection, especially where low numbers of the organism are present in faeces and when the infection is caused by a serogroup other than O157.

Of 43 strains of Vibrio cholerae Ol biotype El Tor isolated over a span of almost three decades (1964–1990) from stools of children and adults with diarrhoea (25 isolates) and from sewage (three) and water from the river Ganges (15) examined for production of haemolysin and its correlation with enterotoxin production, 17 isolates showed haemolysis. The majority of isolates (26), including 68 % of diarrhoeal and 50 % of environmental origin, were non-haemolytic. The titre of haemolysin produced was 4-16 HU/ml, irrespective of the source of isolation. Haemolytic strains caused significantly more fluid accumulation than the non-haemolytic strains in the rabbit ileal loop (RIL) test. Twenty nine (67·4) V. cholerae biotype El Tor isolates—all the haemolytic and most (61·5 %) of the non-haemolytic isolates tested—caused fluid accumulation. The remaining non-haemolytic strains that caused little or no accumulation of fluid did so after one to four consecutive passage(s) through RIL without change in haemolytic character; these strains required more consecutive passage through rabbit gut to show haemolysis. All these strains reverted to their original non-haemolytic character on repeated subculture or on storage in the laboratory but continued to show enterotoxic activity. The present study indicated that El Tor haemolysin is not responsible for fluid accumulation in rabbit gut.

The effect of inoculum size on the results of agar dilution MIC tests was assessed for 20 Moraxella catarrhalis isolates with BRO-1 enzyme, 20 with BRO-2 enzyme and 15 isolates that did not produce β-lactamase. The compounds tested were ampicillin, coamoxiclav, cefaclor, cefixime and cefetamet, and the inocula were 104, 105, 106 and 107 cfu/spot. The MICs of ampicillin for BRO-1 and BRO-2 producers were consistently higher than those for non-producers at inocula of 107 cfu/spot but overlapped with those for non-producers at lower inocula. A small β-lactamase-related inoculum effect was seen with coamoxiclav; small inoculum effects also occurred with cefaclor and cefixime but were not related to enzyme presence or type. MICs of cefetamet were the least affected by the inoculum size. For all the compounds, the degree of correlation between MICs and the inhibition zones observed in disk diffusion tests was independent of the inoculum used in the MIC tests. These data suggest that high inocula should be used to determine MICs of ampicillin for M. catarrhalis but that this precaution is unnecessary with the cephalosporins tested or with coamoxiclav.

The usefulness of small-fragment restriction endonuclease digest analysis (SFREA) of group A streptococcal DNA with EcoRI, as a supplement to the more conventional T serotyping, was assessed for epidemiological characterisation. One hundred and thirty-five clinical isolates from 1988–1990 were examined. SF-REA provided characteristic fingerprints of all isolates, whereas eight isolates were non-typable by T serotyping. Generally, there was a striking correlation between the results obtained with the two techniques. Furthermore, SF-REA reliably classified the eight T-non-typable isolates and occasionally revealed subgroups within the T serotypes. In addition, SF-REA was useful for the clarification of discrepancies between serotyping results from two different reference laboratories. No obvious correlation was observed between the DNA fingerprints and the clinical manifestations of infection or the geographical origin of the group A streptococcal isolates. SF-REA is a valuable supplement to T typing in epidemiological studies and frequently appears to be a more efficient tool for strain differentiation.

The serum total IgG and IgG subclass and nasal wash IgA and IgG antibody responses of mice to influenza virus A/Hong Kong/68 (H3N2) subunit preparations administered parenterally as a single dose, incorporated either in immune stimulatory compounds (ISCOMs) or liposomes with Freund’s Complete Adjuvant, or as an aqueous material, as well as to live, infectious virus were measured by ELISA at 10 days and 3, 5, 7 and 22 weeks after immunisation. The protection of the upper and lower respiratory tracts provided by these preparations against homologous and heterologous challenge infection was assessed. Of the four variously-presented subunit preparations, influenza subunit ISCOMs induced relatively high and persisting levels of each of the different IgG subclasses, particularly IgG2a, throughout the study, and most nearly approached those observed after intranasal infection of mice with infectious virus. Furthermore, nasal wash IgA and IgG antibody levels, particularly at 5 or 7 weeks after immunisation, were also significantly greater in mice given the subunit ISCOM preparation than those induced by other subunit preparations with adjuvant or subunits given alone, and provided protection of both the upper and lower respiratory tracts against challenge as similar to that elicited by infectious virus.

The potential of a particle counting immunoassay (PACIA) for the direct detection of Clostridium difficile serogroup G specific antigen in faecal specimens was evaluated. F(ab′)2 fragments from a rabbit anti-serogroup G antiserum were covalently coupled to carboxylated latex beads. This reagent was mixed with acid extracts of faecal specimens and the reaction was assayed with an optical counter which discriminated unagglutinated from agglutinated latex particles. Culture for C. difficile, faecal cytotoxin detection, PACIA and serogrouping of C. difficile isolates were performed on 249 stools. Of the 71 culture-negative specimens, none gave a positive result in the cytotoxin assay or in PACIA. Faecal cytotoxin was detected in 100 of the 178 culture-positive specimens. PACIA was positive for 63 of the 71 faecal specimens that yielded serogroup G C. difficile on culture. PACIA gave negative results for all other culture-positive stools tested with one exception, from which a serogroup A7 C. difficile strain was isolated. PACIA detection of serogroup G antigen in faecal specimens showed a sensitivity of 88·7%, a specificity of 99·7%, a predictive value of a positive culture with a serogroup G strain of 98·4%, and a predictive value for specimens that were culture-negative for a serogroup G strain of 95·6%. The results indicate that PACIA with specific antiserum is a rapid and reliable method for detecting serogroup specific antigens of C. difficile in faecal specimens. Clinical applications of the method are discussed.

The genotypic diversity of 40 presumably epidemiologically unrelated strains of Pseudomonas aeruginosa belonging to nine different O-serotypes was analysed according to ribosomal DNA fingerprints. Ribotyping was performed with a digoxigenin-labelled DNA probe and four restriction endonucleases. Characteristic banding patterns of three to 12 bands were obtained with the different endonucleases. Among the 40 strains, eight, nine, 10 and 29 different ribotypes were differentiated with EcoRI, the combination EcoRI + HindIII, BamHI and PvuII, respectively. Poor correlations were noted between the results of serotyping and those of ribotyping. With the latter method, indices of discrimination were calculated for each enzyme from the data of the 40 unrelated strains: the values ranged from 0·678 for EcoRI to 0·979 for PvuII. Epidemiologically related samples were also tested; this enabled assessment of whether the method was able to cluster strains from a common origin with each of the enzymes tested. Ribotyping with PvuII endonuclease is proposed for screening large numbers of P. aeruginosa strains in epidemiological studies. Additional enzymes could be used to further increase the discrimination between isolates found to be indistinguishable with PvuII enzyme.

The surface structures of the cell envelopes of 16 clinical isolates of Pseudomonas aeruginosa were examined by electronmicroscopy with the new fixation technique of freezesubstitution. Two types of structures were observed among the organisms. In one group of strains, mostly isolated from blood, a dense fibrous layer c. 30 nm thick was found around the outer-membrane surface, whereas no such structure was observed in the other group of isolates, most of which were from sputum. Lipopolysaccharides extracted from the isolates with a dense fibrous layer were found by SDS-PAGE to have long O-polysaccharide chains, whereas strains without such a layer mostly had lipopolysaccharides that lacked high mol. wt. O-polysaccharide chains.

Enzyme-linked immunosorbent assay (ELISA) for IgA, IgG and IgM was evaluated with sera from 50 adult patients with pneumonia, selected on the basis of a positive complement fixation (CF) test for diagnosis of Mycoplasma pneumoniae infection and with sera from 105 healthy blood donors. The ELISA antigen for IgG and IgA was a sonicated suspension of M. pneumoniae solubilised by deoxycholate. For the IgM assay, the same antigen was directly conjugated to alkaline phosphatase and used in a μ-capture format. ELISA gave positive results with high or rising titres for one or several antibody classes in 47 (94%) patients. In two of the three ELISA-negative cases, the diagnosis of M. pneumoniae infection indicated by the CF test seemed unlikely on clinical grounds. Specific IgA antibodies was developed more regularly and more rapidly than IgM. IgA titres also started to decrease earlier than IgM or the late-peaking IgG response. Thus, the determination of IgA antibodies was found to be valuable for the early diagnosis of M. pneumoniae infection. The study also demonstrated that the determination of all three antibody classes is necessary to obtain an optimal level of serodiagnosis.

ELISA methods that measure IgG class antibodies to sonicated Borrelia burgdorferi may give false positive results. These errors could be traced to non-specific reactivity in subclass IgG2 in several instances. Sera were sampled randomly from two adult populations, which differed in having a high and low incidence of Lyme disease. If the binding of IgG2 subclass antibodies was left unrecorded in the test by the use of monoclonal reagent antibodies selective for IgGl and IgG3, the frequency of positivity in the ELISA test decreased in samples from the low risk group. Twenty-one samples were found to be positive in an immunoblot confirmatory test. Correct prediction of positivity was obtained for 15 sera by ELISA restricted to IgGl plus IgG3, for only four sera by ELISA restricted to IgG2 and for only six sera by IgG subclass non-restricted ELISA. A non-restricted ELISA with purified flagella of B. burgdorferi as the antigen predicted correctly 14 of the immunoblot-positive sera. The results of this ELISA correlated well with those of the IgGl plus IgG3 subclass restricted ELISA in the high risk population (r = 0·95, prevalence of seropositivity 12%), but was significantly worse for the low risk group (r = 0·47, prevalence 2·9 %). IgG subclass restriction also decreased cross-reactions of syphilitic sera in the ELISA with sonicated antigen.