Strain differentiation in Bacteroides fragilis by RAPD and Dendron computer-assisted gel analysis

Citation
Er. Eribe et I. Olsen, Strain differentiation in Bacteroides fragilis by RAPD and Dendron computer-assisted gel analysis, APMIS, 108(10), 2000, pp. 676-684
Citations number
19
Categorie Soggetti
Medical Research General Topics
Journal title
APMIS
ISSN journal
09034641 → ACNP
Volume
108
Issue
10
Year of publication
2000
Pages
676 - 684
Database
ISI
SICI code
0903-4641(200010)108:10<676:SDIBFB>2.0.ZU;2-N
Abstract
Bacteroides fragilis is the anaerobe most frequently isolated from human in fections. Strains of this species are not easily distinguished by phenotypi c tests. It is important to make this distinction because virulence may var y between strains and because B. fragilis seems to be a heterogeneous speci es. The aim of this study was to assess the utility of randomly amplified p olymorphic DNA (RAPD) analysis for differentiation of 46 strains of B. frag ilis. Twenty-seven of the strains belonged to Johnson's DNA homology group I and 8 to group II, while II strains had not been assigned to any of these groups (NI group). The primers OPA16 and 18 were chosen among 30 primers t ested for optimal RAPD analysis. OPA18 gave best discrimination, revealing a total of 15 genotypes while OPA16 gave 13. The gels obtained after RAPD a nalysis were evaluated with the Dendron computer-assisted program. Most str ains showed similarity levels (S-AB) within 70%. Strain clusters thus estab lished were not always in agreement with DNA homology since strains from bo th homology groups fell in the same cluster. Similarly, strains of the NI g roup fell among the group I and II homology strains. RAPD was useful for di fferentiation of B. fragilis strains and thus probably suitable for epidemi ological studies. On the other hand, DNA-DNA homology, comparing the entire genome of strains rather than a few random priming sites, would be more re liable for taxonomy. Computer-assisted gel analysis made it possible to obj ectively evaluate multiple banding patterns, thereby increasing the reliabi lity of the RAPD analysis.