Taxonomy of the Bacteroides: II. Correlation of Phenotypic Characteristics with Deoxyribonucleic Acid Homology Groupings for Bacteroides fragilis and Other Saccharolytic Bacteroides Species

Publisher Website
Google Scholar
Abstract
Two hundred ninety strains of Bacteroides, each belonging to B. fragilis, B. thetaiotaomicron, B. ovatus, B. uniformis, B. vulgatus, B. distasonis, or B. eggerthii, or to one of three unnamed taxa designated “B. fragilis subsp. a,” “3452-A,” and “T4-1,” were used to compare the patterns of 34 phenotypic traits with deoxyribonucleic acid (DNA) homology groupings. The responses for all but 13 of the traits were very similar for all of the DNA homology groups. The responses for the 13 traits (the production of indole, catalase, and hydrogen, and acid production from arabinose, cellobiose, melibiose, melezitose, raffinose, rhamnose, ribose, salicin, sucrose, and trehalose) varied between the various homology groups. Similarity coefficients, for which all of the strains were compared with the DNA reference strains, were calculated. Similarity coefficient averages of strains within a given DNA homology group with the DNA reference strains of that group ranged form 83 to 94%. Organisms having about 60 to 65% DNA homology to each other were in most cases also phenotypically very similar. In a couple of instances, however, homology groups with strains having 72 to 95% intrastrain homology were quite variable with respect to certain phenotypic traits. There was no general pattern between levels of intergroup homologies and phenotypic properties. For example, B. eggerthii and the “subsp. a” homology groups, having about 50% intergroup homology, had average intergroup phenotypic similarity values of 65 and 69%, whereas B. ovatus and B. thetaiotaomicron homology groups, having only about 35% intergroup DNA homology, had intergroup similarity values that ranged from 86 to 93%. Similarity coefficients between homology groups having low levels of DNA homology ranged from 38 to 78%. Response probability values for the phenotypic traits were estimated for strains belonging to each of the DNA homology groups. The probability values obtained with 20 to 40 strains were not altered significantly by including more strains. A test identification matrix was constructed using the probability values of the 13 phenotypic traits listed above. On the basis of these traits, 93% of the strains could be assigned to the correct homology group if groups having 60% or greater intragroup homology were considered to be phenotypically the same.