Minimal requirements of amino acids and vitamins were determined in chemically defined medium for five strains of Clostridium difficile. Cysteine, isoleucine, leucine, proline, tryptophan and valine were essential amino acids for growth of C. difficile. Arginine, glycine, histidine, methionine and threonine enhanced growth. Biotin, pantothenate and pyridoxine were essential vitamins. A defined medium containing the minimal requirements of amino acids and vitamins produced a rapid and heavy growth which was comparable to that in modified brain heart infusion, a complex medium. Adenine was able to substitute for glycine and threonine, suggesting that the two amino acids may be utilized as precursors of purine nucleotides. The defined medium developed here will assist physiological and biochemical studies on C. difficile.

Attempts were made to isolate Clostridium difJicile from a total of 431 fecal specimens from 149 young and 213 elderly healthy adults, and 69 elderly adults with cerebrovascular disease but no gastrointestinal disease. C. difJicile was isolated from 49 specimens, and the frequency of isolation was 15.4% in healthy young adults, 7.0% in healthy elderly adults, and 15.9% in elderly adults with cerebrovascular disease. Thirty-four (about 70%) of the 49 C. difJicile strains isolated produced cytotoxin which was neutralized by Clostridium sordellii antitoxin in vitro; in both young and elderly adults approximately 30% of the C. difJicile isolates were nontoxigenic. The mean concentration of C. difJicile in feces was 10 4 • 1/g in young adults and 104.6/g in elderly adults, with a range of 10 2 • 0 to 10 6 • 9/g. Antibody against C. difJicile toxin was found in most of the sera obtained from young adults carrying toxigenic C. difJicile, but not in sera of elderly adults, no matter how abundant was toxigenic C. difJicile in the feces.Toxigenic strains of Clostridium dijfu;ile have been shown to be a major cause of antibiotic-associated pseudomembranous colitis (PMC) of man (1,3,7,8,13). Feces from the vast majority of PMC patients contain a cytopathic toxin which is neutralized by Clostridium sordellii antitoxin, and toxigenic C. difficile in high concentration. Toxigenic strains of this organism have been isolated also at a high rate from feces of healthy neonates and infants (II, 15, 18) but rarely from those of healthy adults (6, 15). Kobayashi et al (14) isolated C. difficile at a high rate from the feces of healthy adults by using a selective medium devised by George et al (5), although there were fewer than 100 of the organisms per gram in most specimens. Recently, Willey and Bartlett (19) also developed a selective medium for the isolation of C. dijfu;ile, but failed to isolate the organism from the feces of healthy adults.Considering that antibiotic-associated PMC or diarrhea occurs more frequently in elderly adults (2, 10), we attempted to determine the frequency of isolation of C. difficile from fecal specimens, the toxigenicity of the isolates, and the presence of antibody against C. dijfu;ile toxin in the sera of young and elderly adults.

Type E botulinum toxin (BoNT/E)-producing Clostridium butyricum strains isolated from botulism cases or soil specimens in Italy and China were analyzed by using nucleotide sequencing of the bont/E gene, random amplified polymorphic DNA (RAPD) assay, pulsed-field gel electrophoresis (PFGE), and Southern blot hybridization for the bont/E gene. Nucleotide sequences of the bont/E genes of 11 Chinese isolates and of the Italian strain BL 6340 were determined. The nucleotide sequences of the bont/E genes of 11 C. butyricum isolates from China were identical. The deduced amino acid sequence of BoNT/E from the Chinese isolates showed 95.0 and 96.9% identity with those of BoNT/E from C. butyricum BL 6340 and Clostridium botulinum type E, respectively. The BoNT/E-producing C. butyricum strains were divided into the following three clusters based on the results of RAPD assay, PFGE profiles of genomic DNA digested with SmaI or XhoI, and Southern blot hybridization: strains associated with infant botulism in Italy, strains associated with food-borne botulism in China, and isolates from soil specimens of the Weishan lake area in China. A DNA probe for the bont/E gene hybridized with the nondigested chromosomal DNA of all toxigenic strains tested, indicating chromosomal localization of the bont/E gene in C. butyricum. The present results suggest that BoNT/E-producing C. butyricum is clonally distributed over a vast area.

The effect of biotin on toxin production by Clostridium diffile was examined in a defined medium. When toxin production by strain KZ 1647, which was isolated from a healthy adult, was examined in relation to its biotin requirement, it was found that with decreasing concentrations of biotin, bacterial growth was decreased, but production of both toxins A and B were remarkably increased, particularly with 0.05 nM biotin. The time course of production of both toxins in biotin-limited conditions was similar to that in biotin-enriched conditions. The biotin effect on toxin production was also observed in 15 other strains, suggesting that the effect occurs frequently amongst toxigenic C. difficile strains. The biotin effect is discussed in relation to the pathogenesis of C. diffile colitis.

Soil specimens collected from a site around the home of patients with food-borne type E botulism probably caused by neurotoxigenic Clostridium butyricum in Guanyun, Jiangsu province, China, were examined for the presence of neurotoxigenic C. butyricum. Five lakeside sites of Weishan lake, in an area near to the sites where the type E botulism outbreaks caused by neurotoxigenic C. butyricum occurred were also surveyed. Type E toxin-producing C. butyricum was isolated from soil from four sites including the site in Guanyun. Polymerase chain reaction assay demonstrated the presence of the type E toxin gene in all the toxigenic isolates. The biochemical properties of the isolates from the Guanyun soil and the lakeside soil were identical except for inulin fermentation and starch hydrolysis properties. These results indicate that neurotoxigenic C. butyricum has its principal habitat in soil.

Biochemical properties of Clostridium difficile were reinvestigated for the practical identification of the organism in clinical laboratories. Bacterial growth in 2% proteose peptone medium supplemented with 0.01 % r-cysteine Hel and 0.1 % agar supported sufficient growth to read the fermentation results just as well as did pre-reduced anaerobically sterilized medium. Incubation for 2 days was long enough for determining the ability to ferment fructose, glucose, mannitol, mannose, melezitose, and sorbitol. All of the 82 strains liquefied 2% but not 10% gelatin. The significance of mannitol fermentation and gelatin liquefaction is stressed since C. difficile is the only species fermenting mannitol among the gelatin-liquefying species of clostridia having subterminal spores.It is now recognized that Clostridium dijJicile is a major cause of antibiotic-associated pseudomembranous colitis of man (l). The carbohydrate-fermenting properties of C. dijJicile were described by Hall and O'Toole when they first isolated this microbe in 1935 (4). Recently, further detailed information about the carbohydrate-fermenting properties of this microbe has been reported by several investigators. Hafiz and Oakley (3) used proteose peptone broth as a basal medium for the carbohydrate fermentation test and incubated cultures anaerobically at 37 C for a maximun of 40 days, which is too long for practical purposes, while Holdeman et al (5) determined the properties by incubating the cultures in prereduced anaerobically sterilized (PRAS) medium at 37 C for 24 to 72 hr. The present paper describes the biochemical properties of C. dijJicile determined by a simple method for practical purposes, disclosing that PRAS medium is not necessary for the test. MATERIALS AND METHODSBacterial strains. A total of 82 C. dijJicile strains, 56 toxigenic and 26 nontoxigenic, were employed. Most of these strains were isolated from healthy adults and

Summary. Basal defined medium (BDM) containing vitamins, minerals and seven amino acids-(/L) tryptophan 0.1 g, methionine 0-2 g, valine 0.3 g, isoleucine 0.3 g, proline 0.3 g, leucine 0.4 g and cysteine 0.5 g-which appeared to be essential for good growth of Clostridium dzficile was prepared. Addition of glycine 0.2 g/L and threonine 0-4 g/L to BDM produced better growth of strain VPI 10463, and this defined medium was designated minimum amino acid-defined medium (MADM). Production of toxins A and B by strain VPI 10463 in 6 x MADM containing (/L) tryptophan 0.6 g, methionine 1.2 g, valine 1.8 g, isoleucine 1.8 g, proline 1.8 g, leucine 2.4 g, cysteine 0.5 g, glycine 0.2 g and threonine 0.4 g, was much greater than in MADM. Toxin production by 20 C. dzficile strains was examined in two defined media-6 x MADM and complete amino acid-defined medium (CADM) containing 18 amino acids-and one complex medium, modified brain heart infusion medium (m-BHI). Simultaneous production of toxins A and B by all test strains was demonstrated in m-BHI and the two defined media. It was also shown that 6 x MADM was generally better than CADM and as effective as m-BHI for stimulating toxin production by 13 strains. This defined medium would be useful for studies on the physiology, metabolism and pathogenicity of C. dzficile.

Accumulating evidence implicates tumor necrosis factor (TNF) and Fas systems in liver injury, although the interaction between these two systems remains to be investigated. In this study, we examined Propionibacterium acnes-primed TNF receptor p55-deficient (TNFRp55 ؊/؊) or Fas-deficient MRL/MpJ Lpr/Lpr mice challenged with lipopolysaccharide (LPS). Priming with P. acnes caused mononuclear cell infiltration into the hepatic lobules and granuloma formation in the livers of TNFRp55 wild-type mice. Subsequent LPS challenge caused massive liver injury and a marked increase in transaminase levels, leading to acute lethality in control wild-type mice. In contrast, the same treatment caused few pathological changes in livers of TNFRp55 ؊/؊ mice, and all animals survived. P. acnes and subsequent LPS challenge induced granuloma formation and apoptotic changes, respectively, in livers of MRL/MpJ Lpr/Lpr mice. However, liver injury was 50% of that in control MRL/MpJ ؉/؉ mice, suggesting some role of the Fas-Fas ligand system in this liver injury model. On the other hand, an agonistic anti-Fas antibody caused massive apoptosis and hemorrhagic changes of the liver without any priming with P. acnes, leading to death in both TNFRp55 ؊/؊ and control wild-type mice. These results suggest that TNFRp55 but not Fas was involved in P. acnes-induced granuloma formation as well as subsequent LPS-induced liver injury and that TNFRp55 and Fas independently induced apoptosis of hepatocytes in vivo. MATERIALS AND METHODS Experimental reagents. LPS (Escherichia coli serotype O55:B5) was obtained from Difco Laboratories (Detroit, Mich.). P. acnes (ATCC 11828) was heat killed and lyophilized as previously described (19). A purified hamster monoclonal agonistic anti-Fas antibody (Jo2) was purchased from Pharmingen (San Diego, Calif.). Animals. TNFRp55 Ϫ/Ϫ mice ([C57BL/6 ϫ 129]F 1) were generated as previously described (32) and bred under pathogen-free conditions at the Animal Research Center of Kanazawa University. MRL/MpJ Lpr/Lpr (hereafter called Lpr/Lpr) and MRL/MpJ ϩ/ϩ (hereafter called ϩ/ϩ) mice were obtained from Sankyo Laboratories (Tokyo, Japan) and CLEA Japan Inc. (Tokyo, Japan), respectively. Eight-to 16-week-old TNFRp55 ϩ/ϩ (hereafter called the wild type) and homozygous TNFRp55 Ϫ/Ϫ mice or 6-to 7-week-old ϩ/ϩ and Lpr/Lpr mice were used for the experiments. All animal procedures in this study complied with the standards set forth in the guidelines for the care and use of laboratory animals on the Takara-machi Campus of Kanazawa University.