Phenotypes were compared in two different classes of mutants with defects in murein-lipoprotein (IkyD mutants of Salmonella typhimurium and an lpo mutant of Escherichia coli). Both mutations are associated with the same triad of phenotypic abnormalities, consisting of defective formation ofthe division septum, leakage of periplasmic proteins during growth, and increased sensitivity to several unrelated external toxic agents. The abnormality in septum formation consists of a defect in invagination of the outer membrane during formation of the nascent septum. The results suggest that formation of the murein-lipoprotein link plays an important role in differentiation of the division septum and perhaps also in maintaining the normal barrier function of the outer membrane.

Eighty-three serum specimens from 24 patients infected with Candida albicans were examined for circulating Candida protein antigens with the Candida Detection System (CAND-TEC; Ramco Laboratories, Inc., Houston, Tex.). The medical records of each patient were reviewed for clinical evidence of Candida colonization or disease, predisposing factors for infection, underlying illness, the presence of a contaminated indwelling venous catheter, intravenous amphotericin B therapy, and outcome. Forty-nine serum specimens with antigen titers of 1:2 or less were obtained either from colonized patients or at a time when disseminated disease was not yet clinically suspected. Except for five specimens from two colonized patients, one with a contaminated arterial line, the other specimens with titers of 1:8 or greater (n = 14) were obtained from patients who had been clinically diagnosed and treated for disseminated candidiasis. Serum specimens with titers of 1:4 were often from patients with deep-seated candidal infection but were not uniformly diagnostic; in this situation additional specimens should be tested for Candida antigen titers. Only 1 of 24 serum specimens from patients with no evidence of C. albicans infection had a Candida protein antigen titer of 1:8. With a 1:8 or greater titer as a criterion for dissemination, the sensitivity of the CAND-TEC system was 71%, with a specificity of 98%. If the 1:8 titer for the colonized patient with a contaminated arterial line is not considered a false-positive result, the CAND-TEC sensitivity was 83%. The latex agglutination assay appears to be a useful, rapid, and noninvasive means of laboratory diagnosis of systemic candidiasis. The recovery of C. albicans from at least three body sites may also be a useful predictor of disseminated disease.

A comparison of two commercially available kits for rapid herpes simplex virus (HSV) detection directly in patient specimens was performed. The immunofluorescence assay (IFA) utilized monoclonal antibodies to HSV, and the DNA probe assay utilized three HSV sequences cloned into pBR322. A sample of 243 specimens received in viral transport medium were inoculated into MRC-5 tissue cultures. The remainder of the specimen was centrifuged, and the cellular pellet was examined by IFA and DNA probes. One hundred and sixty-two (66.7%) specimens were considered satisfactory for IFA and DNA probe testingf based on a criterion of observing-2 intact cells per high-power field. Of the 162 specimens, 35 (21.6%) yielded HSV by culture. By IFA, the sensitivity of detecting HSV culture-positive specimens was 77.1%; specificity was 100%, positive predictive value was 100%, and negative predictive value was 93.3%. DNA probe sensitivity was 71.4%; specificity was 90.6%; positive predictive value was 67.6%; and negative predictive value was 92%. Forty-four (27.2%) of the 162 specimens exhibited nonspecific cytoplasmic staining with the DNA probe. IFA and DNA probe assays can be completed in 2 to 3 h, whereas the average time to culture positivity in this series was 2.2 days. Rapid HSV diagnosis can aid in timely and appropriate patient management.

Results of counterimmunoelectrophoresis (CIE) were compared with those of isolation of Clostridium difficile and assay for cytotoxicity in HeLa cells. On the basis of 471 stool specimens, CIE exhibited a sensitivity of 38% and a specificity of 88% as compared with the cytotoxin assay. The predictive value of a reactive CIE results is low (17%), whereas the predictive value of a nonreactive CIE result is significant (96%) and therefore warrants its use as a screening test. In addition, stool filtrates may nonspecifically precipitate with the C. difficile antitoxin in the CIE test. Such nonspecific reactions may be identified by simultaneous electrophoresis against nonimmune serum. Clostridium difficile and its toxin are central to the pathogenesis of antibiotic-associated diarrheal disease (AAD) (2, 3, 17). The laboratory diagnosis of AAD has been partially facilitated by the development of a selective and differential culture medium for isolating C. difficile from fecal specimens (8). In addition, demonstration of C. difficile toxin in fecal filtrates by a cell

Threshold concentrations of Streptococcus pneumoniae type 3, Haemophilus influenza type b, and Streptococcus sp. group B type Ib required for positive counterimmunoelectrophoresis reactions were determined in vivo and in vitro. Animals were infected intraperitoneally with various concentrations of microorganisms: adult mice with S. pneumoniae, suckling rats with H. influenza, and 3week-old mice with Streptococcus sp. group B. At 24 h after infection a minimum blood concentration of 103 colony-forming units (CFU)/ml was needed for S. pneumoniae or H. influenza before antigen was detected in the serum. A minimum concentration of 106 CFU/ml was needed for Streptococcus sp. group B at 10 h after infection. Larger threshold concentrations (104 CFU/ml for S. pneumoniae, 105 CFU/ml for H. influenza, and 107 CFU/ml for Streptococcus) were required in broth-grown cultures before cell-free antigens could be demonstrated by counterimmunoelectrophoresis in the medium. Marked levels of antigen release by group B streptococci were observed as the cultures entered early stationary phase. This study provides evidence of a long-accepted, though poorly substantiated, hypothesis that a threshold concentration of microorganism is necessary before counterimmunoelectrophoresis reactions become positive. Counterimmunoelectrophoresis results for clinical specimens should be interpreted cautiously in light of this evidence.