The Genera Flavobacterium, Sphingobacterium and Weeksella

Abstract

The name Flavobacterium was proposed in 1923 for a genus of the family Bacteriaceae encompassing the rod-shaped, nonendosporeforming, chemoorganotrophic bacteria (Bergey et al., 1923). Most of the pigmented bacteria of the family were segregated in the tribe Chromobactereae, which contained four genera of aerobic bacteria separated from each other by differences in color. These genera were Chromobacterium, Flavobacterium, Pseudomonas, and Serratia, for the purple, yellow, green fluorescent, and red strains, respectively. This emphasis on pigmentation (a character shared by genetically diverse bacteria [Weeks, 1969]) for taxonomic assignment to Flavobacterium has given the genus a dubious reputation in the past (McMeekin et al., 1972; Weeks, 1969), and as a consequence the genus has served too frequently as a repository for pigmented bacteria that possess the general attributes of Flavobacterium but had not been subjected to detailed classification studies. Taxonomic heterogeneity and general uncertainty have characterized Flavobacterium from its inception, and its history is a record of proposals to achieve credibility for the genus.

Stanier (1947) recognized that the cytophagas have more than a casual phenetic resemblance to pigmented, Gram-negative eubacteria such as Flavobacterium and the relationship to the cytophagas has dominated the taxonomic consideration of Flavobacterium. Differentiation of flavobacteria from cytophagas has depended primarily upon demonstration of the gliding movement and colonial translocation characteristic of the latter bacteria, but absence of these features has not deterred assignment of flavobacterial species to Cytophaga (Mitchell et al., 1969). Freshly isolated cytophagas have the unusual ability to use a great variety of complex natural polymers, e.g., agar, alginates, cell walls, cellulose, chitin, DNA, keratins, lipids, pectin, porphyrins, proteins, RNA, and starch, as nutrients. This ability is not a general property of flavobacterial species, although strains of some species may hydrolyze casein, chitin, gelatin, or starch.

The problem of differentiating Flavobacterium from Cytophaga has been discussed extensively (Christensen, 1977a; McMeekin et al., 1972; Mitchell et al., 1969; Weeks, 1969). An outcome would be more nearly possible if two issues were resolved. These are the heterogeneity of Flavobacterium and the differentiation of nonmotile flavobacteria from cytophagas. A primary requirement for the resolution of both issues is an acceptable definition of Flavobacterium. The concept of Flavobacterium was hardly altered in successive editions of Bergey’s Manual of Determinative Bacteriology until the fifth (Bergey et al., 1939), which eliminated from the genus the least well-described species and the polarly flagellated strains. Those known to be Gram-positive were excluded in the seventh edition (Weeks and Breed, 1957). In the eight edition (Weeks, 1974) the genus remained heterogeneous, as evidenced by the two disparate reported ranges of GC content of the DNA: 30–42 and 63–70 mol%. Only in Bergey’s Manual of Systematic Bacteriology (Holmes et al., 1984a) did the genus at last become reasonably homogeneous by including only nonmotile low GC strains. Interspecies DNA-DNA hybridization studies of the latter organisms showed a background level of hybridization of about 20%, which may represent a common DNA complement (Callies and Mannheim, 1980; Owen and Holmes, 1981).

Despite sharing several features, the low GC flavobacteria nevertheless show some heterogeneity and fall into four natural groups (A–D, Table 1), as previously suggested by Holmes and Owen (1981). The strains of group D, so far found only on mammalian mucus membranes, are susceptible to penicillin and are nonpigmented; they now comprise the genus Weeksella. Strains of the remaining groups are free-living, show resistance to a wide range of antimicrobial agents and are yellow-pigmented (Table 1). These organisms constitute the genus Flavobacterium and they can be divided into groups A–C on the basis of differences in indole production, oxidation of carbohydrates, and proteolytic activity (Table 1). A detailed taxonomic study of these organisms by techniques such as DNA-rRNA hybridization is necessary to determine their phylogenetic relationships and therefore whether they all belong in the same genus. Pending such a study, it seems undesirable to divide the genus further or to admit to it taxa of uncertain affiliation, otherwise the genus is in danger once more of becoming heterogeneous. The limited data available from DNA-rRNA hybridizations (Bauwens and De Ley, 1981) indicate that F. breve, F. meningosepticum, and F. odoratum may not be closely related and that F. aquatile, the type species of the genus (and itself represented by a single strain), is not closely related to the above three species. For other reasons, F. aquatile had been considered unrepresentative of the genus. Consequently, Holmes and Owen (1979) requested that F. breve be made the type species instead. Their request was subsequently denied, and thus if the genus is further subdivided in the future the name Flavobacterium will remain with F. aquatile and the majority of organisms currently recognized as Flavobacterium will be transferred to one or more new genera. For the above reasons, F. aquatile is not considered in the tables in this paper (its characters are given by Holmes et al., 1984a). Perhaps prematurely, the genus Sphingobacterium has been proposed by Yabuuchi et al. (1983) for organisms in group C (Table 1) based on the fact that most of these organisms have been shown to possess novel sphingolipids in their cell walls. As this character is not easily determined routinely, and pending further study and possible revision of the genus as a whole, the group C strains are here retained in Flavobacterium. (Although all strains in group C have valid combinations of their specific epithets with Flavobacterium, there has as yet been no proposal to transfer F. thalpophilum or F. yabuuchiae to Sphingobacterium.)

Table 1. Differentiation of the flavobacteria.

With Flavobacterium better defined, it is now possible to reassess the differentiation of this genus from Cytophaga. It has been suggested that organisms of both these genera, which share a distinctive cellular fatty acid composition, a characteristic menaquinone system, and a low GC content, should all be included in Cytophaga (Oyaizu and Komagata, 1981). However, this view does not take into account the heterogeneity of the Flavobacterium-Cytophaga group as revealed by the preliminary DNA-rRNA hybridization results of Bauwens and De Ley (1981). More recent DNA-DNA hybridization studies (Bernardet, 1989) reveal no appreciable homology between Cytophaga johnsonae and Flavobacterium species of groups A, B, and C. Oligonucleotide cataloging (Paster et al., 1985) places C. johnsonae and F. aquatile in the same group while finding that F. breve is only peripherally related to the same group. Cataloging also places Cytophaga and Flavobacterium in the same eubacterial “phylum” as Bacteroides.

In summary, while Flavobacterium is now more homogeneous than it has ever been, further work is necessary to group the species into additional genera so as to reflect their phylogenetic relationships. Among the flavobacteria of clinical origin at least, these generic groupings may well be reflected by the groups A–C as defined in Table 1. While these organisms would appear to be distinct from Cytophaga, other flavobacteria, including F. aquatile, may be more closely related to Cytophaga. The latest arrangement of including the families Bacteroidaceae and Cytophagaceae plus a newly proposed family “Flavobacteriaceae” in the order Cytophagales (Reichenbach, 1989) is a good reflection of our present knowledge of the phylogenetic relationships of these organisms.