Search

On the basis of DNA-DNA hybridization data, two main genomic relatedness groups (I and II) have been reported for a geographically varied collection of 52 strains of Flavobacterium meningosepticum. Herein, we have shown that genomic group II can be further divided into four subgroups (II:1 to II:4). To examine the taxonomic relevance of the ribosomal patterns of the 52 F. meningosepticum strains, the patterns were compared with existing DNA-DNA hybridization data with restriction enzymes PstI and HindIII. Ribotyping of the 52 F. meningosepticum strains showed banding patterns that could identify them correctly to one of the five genomic groups or subgroups. To assess the value of ribotyping for the interpretation of epidemiological data, the discriminatory power of the method was investigated for the 52 F. meningosepticum strains. With one to four restriction enzymes (PstI, HindIII, ClaI, EcoRI), a discriminatory index of 0.95 to 0.97 was found. The value of ribotyping in an epidemiological setting was assessed for three clinical isolates of F. meningosepticum from an outbreak of meningitis and bacteremia in the neonatal intensive care unit, Rigshospitalet, Copenhagen, Denmark. The three clinical isolates were shown to belong to the same ribotype, characteristic of genomic subgroup II:1. This ribotyping method will prove to be a useful tool for epidemiological studies concerning F. meningosepticum in the future.

DNA-DNA hybridization studies on 42 stains presumptively identified as members of Flavobacterium group IIb and Flavobacterium breve indicated pronounced genotypic heterogeneity within these taxa. Three large groups highly related to the type strains of F. gleum, F. indologenes and F. breve respectively, and eight small groups were found. The group containing the type strain of F. breve was phenotypically indistinguishable from another genomic group, and these two groups were significantly separated from the other flavobacteria studied. The other nine genomic groups, representing Flavobacterium group IIb, could not with certainty be differentiated from each other by phenotypic characteristics, and there is no evidence indicating that these genomic groups differ from each other with respect to pathogenicity or ecology. Thus, it is suggested that for the time being the name "Flavobacterium group IIb" rather than specific epithets continue to be used for these bacteria.

An attempt has been made to further characterize the Flavobacterium groups found in an earlier study by determining utilizable carbon-energy sources. As a preliminary, it was initially tried to define minimal growth factor requirements. It was found likely that most of the Flavobacterium groups have amino acid requirements, but a requirement for vitamins was also established for some strains. Testing Flavobacteria for carbon-energy sources by the replica plating method was found sufficiently technically difficult as to result in dubious readings at times. However, the overall results of the carbon-energy tests are in good agreement with the subdivision resulting from the previously performed biochemical tests, and the validity of the groups delimited in the earlier study is confirmed.

The genus Flavobacterium has undergone a number of changes during the last years resulting in a revised description of the genus in which restriction of Flavobacterium to certain low G + C content species is proposed. The purpose of the present study has been to see how a collection of 184 clinical and environmental strains of Flavobacterium-like organisms isolated in Denmark would fit into this latest revision. It was found that most of the different biochemical patterns represented gradually merged into each other, with certain patterns, however, occurring more frequently than others and corresponding to the already described taxa: F. meningosepticum, Flavobacterium group IIb, F. breve, F. odoratum, F. multivorum and F. spiritivorum. The lately proposed subdivision of the genus will need either revision or extension or both to cope satisfactorily with the situation revealed in this study.

By means of crossed immunoelectrophoresis and crossed-line immunoelectrophoresis, antigens were compared from four strains and five strains, respectively, of Flavobacterium meningosepticum DNA hybridization groups I and II. Two reference antisera raised against the pooled antigens of each of the two DNA groups were used. The two reference antigen-antibody systems produced 81 and 51 immunoprecipitates, respectively, and within each group nearly all antigens were common to all strains, giving matching coefficients of 0.98-1.00. Group specific antigens were not detected. When antigens from individual strains from each group were compared to individual strains of the other group, however, 5-15 of the antigens were not cross-reactive, giving matching coefficients of 0.83-0.94. The antigenic relatedness between strains belonging to the two different groups was, therefore, just below the level commonly found for strains belonging to the same species. The results thus confirm the previously published differences detected by a DNA hybridization technique.

Antimicrobial susceptibilities of 52 strains of Flavobacterium meningosepticum were determined by an agar plate dilution technique and by a tablet diffusion method. No useful susceptibility differences between the two main DNA relatedness groups within F. meningosepticum were found. All strains produced beta-lactamase and were resistant to most beta-lactam antibiotics. Ciprofloxacin, rifampicin and clindamycin appear to be the most promising for treating infections caused by F. meningosepticum.

DNA-DNA reassociation studies on 52 strains of Flavobacterium meningosepticum showed two main hybridization groups about 40-55% interrelated and comprising 4 and 48 strains respectively. The larger group could be further divided into four subgroups if differences in thermal stability of the reassociated duplexes were taken into consideration. Of a total of 20 strains isolated from cerebrospinal fluid, 18 were found in one subgroup of 28 strains, which also contained seven blood isolates. The results indicate that genetically defined subgroups within the species might differ with regard to pathogenic significance.

A total of 55 different antigens were demonstrated in pooled lysate of six Flavobacterium meningosepticum strains by crossed immunoelectrophoresis against homologous rabbit antibody. The six strains used represented the two DNA relatedness groups within this species. All 52 strains of F. meningosepticum investigated were found to cross-react with 53 to 55 of these F. meningosepticum reference antigens. Using this F. meningosepticum reference system, cross-reactions between the F. meningosepticum reference antigen and antigens from representatives of other Flavobacterium taxa were also investigated. A positive correlation was demonstrated between the number of cross-reacting antigens and the taxonomic relatedness of the investigated taxa as determined by other classification methods, including DNA-DNA hybridization. The crossed immunoelectrophoretic techniques were not useful in differentiating the two main DNA groups of F. meningosepticum.

Fifty-two strains found to belong to Flavobacterium meningosepticum on the basis of DNA-DNA hybridization analyses were characterized and found to be phenotypically homogeneous. All strains were oxidase, catalase, indole, gelatinase and beta-galactosidase positive, and produced acid from glucose, mannose, fructose, maltose and mannitol; nitrate was not reduced. F. meningosepticum could be differentiated from Flavobacterium group IIb by its ability to produce beta-galactosidase, and by the latter taxon's ability to produce a bright yellow pigment in contrast to the weak or non-existing pigmentation of F. meningosepticum. Phenotypic characteristics that could differentiate between the two main DNA relatedness groups of F. meningosepticum were not discovered, wherefore a subdivision of the present species into two species cannot be recommended. Strains from the DNA relatedness groups comprising 19 of 20 CSF isolates were found to be able to grow at 40 degrees C and to produce a weak yellow pigment; in contrast, strains from the three other DNA relatedness groups were unable to grow at 40 degrees C and produced no pigment.

An outbreak of infections due to Flavobacterium meningosepticum type C in a neonatal intensive care unit is described. During a period of two weeks, two infants developed meningitis and a third was colonized in the respiratory tract and had transient bacteremia. The two meningitis patients were treated with clindamycin, rifampicin and cefotaxime systemically, plus rifampicin intraventricularly. Bacteriological eradication was achieved within 48 h, and both infants recovered from the meningitis without apparent neurological sequelae; however, one infant died two months later of unrelated causes. Environmental surveillance cultures failed to demonstrate a reservoir for the epidemic strain, but other Flavobacterium strains were recovered. Two clinically healthy infants were found to be colonized in the nasopharynx with strains that were extremely difficult to differentiate phenotypically from the epidemic strain. Extensive characterization of strains is necessary in order to differentiate between strains and subsequently to determine a certain source of infection.