Detection and characterization of toxinogenic Clostridium difficile by non-genetic and genetic techniques.

Abstract

A semi-quantitative, polymerase chain reaction-enzyme immunoassay (PCR-EIA) was evaluated for the detection of toxin A gene sequences from Clostridium difficile. Amplified DNA target sequences were hybridized in solution to a biotinylated RNA probe, and DNA:RNA hybrids were measured in fluorescence units (fu) using an immunoassay. PCR-EIA detected 0.9 bacterial cells from spiked stool specimens (16 {dollar}\pm{dollar} 0 fu) and 72 fg DNA or 2 genome equivalents (ge) (45 {dollar}\pm{dollar} 3 fu) from amplified DNA dilutions. By gel analysis, 290 fg or 9 ge were detected. Overall test variation was 21.9%. PCR-EIA was positive for 16 tissue culture cytotoxic (TC+) isolates (range, 261-1400 fu). The predicted 410 bp band was present by gel analysis, and the band hybridized with a radiolabeled RNA probe using Southern blot analysis. Seventeen TC+ C. difficile isolates were positive by PCR-EIA with a range of 261-1400 fu. Ten TC{dollar}-{dollar} C. difficile isolates and 10 other clostridial isolates were negative by gel and by PCR-EIA with mean fu = {dollar}-{dollar}6 and {dollar}-{dollar}8, respectively. Nine TC+ C. difficile and 2 sporulation mutants were compared by semi-quantitative methods for toxin A protein and toxin A gene using enzyme-immunoassay (EIA-A) and PCR-EIA, respectively. Strains giving low EIA-A values for toxin A protein had PCR-EIA values equivalent to high toxin producers for the amplified area of the gene. Twenty-nine stool specimens from nursing home and acute-care patients were tested by TC, bacterial culture (BC) and PCR-EIA. For groups BC{dollar}-{dollar}/TC{dollar}-{dollar}, BC+/TC{dollar}-{dollar}, and TC+, PCR-EIA detected C. difficile in 0% (10/10, mean fu = {dollar}-{dollar}5), 83.3% (5/6, mean fu = 135), and 84.6% (11/13, mean fu = 452), respectively. Inhibitors to PCR were found in one stool specimen. The epidemiology of a diarrhea outbreak in a nursing home caused by C. difficile was investigated by microbiological means. Standard methods for typing isolates were compared to genetic methods. Standard methods included antibiotic and phage susceptibility, ({dollar}\sp{lcub}35{rcub}{dollar}S) methionine protein autoradiography and serogrouping. Genetic methods were plasmid profiles, Hinf I chromosomal restriction length polymorphism-RFLP, and toxin A PCR-RFLP using Alu I. Hinf I was the most discriminating method for epidemiological purposes indicating that 8 diarrhea types were present during the outbreak. Alu I PCR-RFLP for a segment of the toxin A gene indicated variability in the toxin A genome, and sequence analysis supported this finding.