Characterization of some Clostridium species by gas-liquid chromatography using numerical analysis

We studied 42 strains of Clostridia belonging to 20 different species. All the strains were examined for morphological characters, biochemical reactions, and analyzed by means of gas-liquid chromatography (GLC) to determine metabolic patterns of short chain fatty acids and alcohols. To increase the number of criteria for the differentiation, specimens were grown on Peptone Yeast Extract medium (PY) with the addition of 13 different carbohydrates. The strains were compared using numerical taxonomic techniques based upon 20 unit qualitative and 224 quantitative characters. Data were examined using the simple matching coefficient (SSM) for qualitative characters, and degree of overlap between superimposed trace (So) for qualitative characters, and clustering was achieved using the unweighted pair group method with arithmetic averages (UPGMA) technique. DNA base composition was also determined. Numerical analysis showed remarkable difference between phenograms derived from SSM and So coefficients. The phenogram (SSM) is formed by 11 clusters and eight of these include strains from only one species. Only three clusters contained strains from different species. The cluster variability range of G + C base composition was never higher than 4 mol% except for one cluster where it reached 7 mol% G + C. In the phenogram (So) instead, there are 8 clusters and in only one case are strains from one species aggregated. In the remaining 7 clusters strains belonging to two or more species aggregated. The range values of base composition are over 4 mol% in three of the eight clusters.