Phylogenetic placement of community members of human colonic biota.

and this genus is again a well-known major component of human The human colonic biota is a natural community of bacteria that colonic biota. are not only symbiotic but constitute an important host defense. Extensive work done two decades ago led to the conclusion that Phylogenetic cluster 3 consisted of the RDP group ‘‘Clostridium coccoides and relatives’’ (table 2), containing organisms the colonic ecosystem consists of phylogenetically diverse taxa including both gram-positive and gram-negative elements. Howfrom diverse taxonomic groups including Clostridium, Coprococcus, Eubacterium, Lachnospira, and Ruminococcus. Despite the ever, it has been possible only recently to accurately determine the phylogeny of these organisms and to study them without using diverse names, the organisms falling into this branch are highly similar to one another. For instance, the ruminococci falling in culture-based methods. In a recent study of the human colonic biota in which standard nonselective culture methods were comthis cluster are more similar phenotypically to the other organisms in the cluster than they are to the type species of their current pared with amplification and cloning of rDNA [1], both methods were found to be biased, but both detected organisms with similar genus. This point is illustrated by comparing Ruminococcus productus with Clostridium coccoides. Not only are their 16S rDNA phylogenetic distributions. Most of the organisms fell into only four clusters. Two of those clusters were the genera Bacteroides sequences practically identical, but they have the same G / C content, ferment the same sugars, and form the same end products and Bifidobacterium. The other two clusters fell into two separate branches of the phylogenetic tree containing organisms that have of fermentation. been classified in several different genera including Clostridium, The organisms in this cluster are likely to be important in carboEubacterium, Fusobacterium, Peptostreptococcus, and Ruminohydrate metabolism within the colon. Not only are they highly coccus. Studies of other bacterial communities have also generally stimulated by the presence of a fermentable carbon source, but shown limited phylogenetic diversity within an ecosystem. the only two organisms known to degrade gastrointestinal mucin, Ruminococcus gnavus and Ruminococcus torques, belong to the Herein, we more precisely map the organisms found in the above-mentioned rDNA-based study of one healthy volunteer and C. coccoides group. Unlike bacteroides, they are not known to transport and degrade complex carbohydrates. Because competicompare the results with the placement of organisms found in earlier investigations involving the nonselective isolation and idention for carbon sources is a major mechanism involved in bacterial nutrient competition, and most pathogens metabolize monosacchatification of human colonic biota. In earlier studies, Moore and Holdeman [2] and Holdeman et al. [3] described the numerically rides or disaccharides, the C. coccoides group is likely to be important in the phenomenon of ‘‘colonization resistance.’’ predominant biota of 20 Japanese Hawaiians and three astronauts. The standard for placement of all organisms in the present study Cluster 4 consisted of three groups as defined by RDP: Clostridwas their location in the phylogenetic tree generated in the May ium leptum and relatives, fusobacteria, and the Atopobium group. 1995 release of the Ribosomal Database Project (RDP) data base The C. leptum group contains the type species and three other [4]. The placement of organisms that belonged to named taxa members of Ruminococcus, as well as species presently assigned available in the RDP was established by determining their phylogeto Eubacterium and Clostridium (table 2). Many of these organisms netic location in that reference. Organisms not present in the RDP also tend to be highly stimulated by fermentable carbohydrates or but with available sequences were placed by using the Genetic to require these carbohydrates and thus are also likely to be inData Environment software package on a Sun Sparcstation 2 data volved in the normal control mechanisms of the biota. It is of base (Sun Microsystems, Mountain View, CA) to align their seinterest that the RDP C. leptum group contains ‘‘Fusobacterium’’ quences with those of known organisms. Aligned sequences were prausnitzii [4]. A close look at the phylogenetic tree based on analyzed phylogenetically in the same software package running rRNA sequences (figure 1) reveals this association to be more a parsimony analysis. harmonious than it seems at first glance. The RDP gram-positive tree essentially trifurcates, resulting in two ‘‘low G/ C’’ branches As shown in table 1, all the studies cited gave similar results in that most of the biomass found in the human colon fell into and one ‘‘high G / C’’ branch. C. leptum and relatives lie at the base of one ‘‘low G / C’’ branch; fusobacteria other than just four phylogenetic clusters. Cluster 1 consisted of the genus Bacteroides, well known as accounting for more than one-quarter F. prausnitzii lie at the base of the ‘‘high G / C’’ branch. Therefore, they are phylogenetically very close despite differing gramof all isolates [2, 3, 5]; this genus falls on two contiguous branches of the RDP tree. The genus Bifidobacterium was designated as stain characteristics. Eubacterium aerofaciens maps to the Atopobium group also at a position near the base of the ‘‘high G / C’’ cluster 2. Bifidobacterium falls into one RDP group of gram-positive bacteria with a high guanine plus cytosine (G / C) content, branch near fusobacteria. Thus, as far as the human colonic biota is concerned, all these organisms form one phylogenetic cluster. Their placement suggests that common ancestors may have colonized early animal forms before the main branches of the grampositive phylogenetic tree diverged. Comparative work should be Financial support: U.S. Department of Veterans Affairs. done in this area. Reprints or correspondence: Dr. Kenneth H. Wilson, Infectious Diseases The lack of variation in large pools of data from different popuSection, Veterans Affairs Medical Center, 508 Fulton Street, Durham, North lations (table 1) and the greater variation among three individuals Carolina 27705. (table 1) are in keeping with previous conclusions concerning the Clinical Infectious Diseases 1997;25(Suppl 2):S114–6 This article is in the public domain. variability of human biota. The culture data for the individual