Cultivable bacterial diversity from the human colon


S.H. Duncan, Rowett Research Institute, Microbial Ecology Group, Greenburn Road, Bucksburn, Aberdeen, AB21 9SB, UK. E-mail:


Knowledge of the composition of the colonic microbiota is important for our understanding of how the balance of these microbes is influenced by diet and the environment, and which bacterial groups are important in maintaining gut health or promoting disease. Molecular methodologies have advanced our understanding of the composition and diversity of the colonic microbiota. Importantly, however, it is the continued isolation of bacterial representatives of key groups that offers the best opportunity to conduct detailed metabolic and functional studies. This also permits bacterial genome sequencing which will accelerate the linkage to functionality. Obtaining new human colonic bacterial isolates can be challenging, because most of these are strict anaerobes and many have rather exact nutritional and physical requirements. Despite this many new species are being isolated and described that occupy distinct niches in the colonic microbial community. This review focuses on these under-studied yet important gut anaerobes.


The large intestine is a complex microbial ecosystem harbouring more than 500 different bacterial species and around 75% of these remain uncultured (Eckburg et al. 2005; Flint 2006). There are clearly compositional changes in the gut microbiota related to diet, age and disease (Manichanh et al. 2006). The human gut microbial community is considered to play a crucial role in gut health and disease and it is vital therefore to gain a clear and detailed understanding of its microbiology. In particular there is much interest in modulating microbial metabolic activity in the colon through diet.

Intestinal bacteria, particularly in the large intestine, have a fundamental role in supplying energy to the host through anaerobic fermentative processes. Dietary material that escapes digestion in the upper gastrointestinal tract and host secretions are available for growth of the colonic microbiota. These complex substrates are colonized primarily by rather specialized bacterial groups that possess elaborate enzyme systems for their hydrolysis (Flint 2006) with each chemically defined component of the substrate giving rise to a specialized bacterial consortium. A myriad of other bacterial groups are then likely to compete for released substrate breakdown products. In addition, a second tier of cross feeding involves utilisation of some of the fermentation products, such as succinate and lactate (Bourriaud et al. 2005; Belenguer et al. 2006; Morrison et al. 2006), which in healthy individuals are only detected at low concentrations, if at all. The main short chain fatty acids (SCFA) detected in faeces, resulting from microbial fermentation, are acetate, propionate and butyrate, which in combination reach concentrations of around 100 mmol l−1 (Cummings et al. 1987; Cummings 1995). These SCFA can supply approximately 10% of the energy requirements of the host (McNeil 1984). Gases are also produced with hydrogen providing a route for disposal of reducing equivalents but high partial pressures of hydrogen in anaerobic ecosystems are likely to lower the fermentation efficiency (Wolin 1981).

Advances in the use of molecular techniques have revolutionized our understanding of the identity and abundance of the major groups of bacteria that inhabit the colon (Eckburg et al. 2005; Flint 2006). Enriching culture collections with new isolates is crucial to understanding their impact on key metabolic processes and on gut health. The availability of key representative isolates will allow detailed analyses to be conducted on the physiology and metabolism of important bacterial groups. In the future this knowledge base will allow researchers to gain a much clearer understanding of key microbial processes in the large intestine and help to shape future strategies for promoting gut health.

Molecular analysis of the composition of human colonic microbiota

Molecular analysis, using PCR-based methods and fluorescent in situ hybridisation (FISH), (Suau et al. 1999; Hayashi et al. 2002; Hold et al. 2002; Eckburg et al. 2005; Flint 2006) has revealed that the predominant bacterial species present in the human colon and in faeces belong to two phyla, these are the Gram-negative Cytophaga-Flavobacterium-Bacteroides (CFB) and the low G + C Gram-positive Firmicutes with the latter comprising several clostridial clusters (Collins et al. 1994). These analyses suggest that around 60–80% of colonic or faecal bacteria fall into these phyla. The Bacteroidetes appear to comprise around one-quarter of total bacteria. Bifidobacteria that belong to the high G + C Gram-positive Actinomycetes form on average around 3–5% of the colonic microbiota (Flint 2006). In addition, however, members of some of the less abundant phyla in the human gastrointestinal tract can have a disproportionately large influence on health maintenance and disease progression. The proteobacteria usually form a minor component of the colonic microbiota. These include pathogens such as Escherichia coli and Salmonella as well as sulphate reducing bacteria including Desulfovibrio sp. that have been implicated as an aetiological agent for ulcerative colitis (Gibson 1990). This review will focus on the less studied but numerically dominant groups of anaerobic gut bacteria, particularly on new isolates that have been described during the course of the last 5 years.

Isolation and cultivation of new human colonic bacterial species

Gut microorganisms are adapted to an environment with low partial oxygen pressures. Many of these bacteria lack electron transport chains found in facultatively anaerobic bacteria to regenerate the reduced co-factors (usually NADH2) and therefore do not gain further energy by electron transport level phosphorylation. Instead, metabolic intermediates are reduced to regenerate NAD+, which results mainly in the formation of acidic fermentation products. As a result of the low energy gain, the turnover of substrate, and consequently production of fermentation acids, is usually high in these organisms. Some gut bacteria perform anaerobic respiration involving electron transport chains by using electron acceptors such as sulphate or carbon dioxide.

To avoid oxygen toxicity when culturing anaerobes the medium is prepared under an O2-free gas phase and a reducing agent added to remove traces of residual oxygen. Moore and Moore (1995) isolated many representative bacterial groups and species following these general principles, although many of these isolates have subsequently been lost from collections. Despite the recent increase in the use of molecular methods for analysing complex microbial ecosystems, there has been renewed interest in isolating new bacterial species from this habitat (see Table 1).

Table 1.   Newly described human intestinal and faecal isolates
New speciesType strain designationGroupKey substrate(s) utilised*Fermentation products (major acidic)Other traitsReferences
  1. *Only selected substrates that are utilised by the strains are shown here although most can ferment a wide range of soluble sugars.

  2. †Fermentation products are those derived from glucose unless stated otherwise.

  3. ‡CFB = Cytophaga-Flavobacterium-Bacteroides phylum.

Gram + ve bacteria
Anaerotruncus colihominisCCUG 45055TCluster IVSaccharolyticAcetate, butyrateNo growth on starchLawson et al. (2004)
Subdoligranulum variabileCCUG 47106TCluster IVSaccharolyticButyrate, lactateNo growth on starch; utilises fucoseHolmstrøm et al. (2004)
Clostridium bartlettiiCCUG 48940TCluster XISaccharolyticIso-butyrate, iso-valerate, phenyl-acetic acidSpore formerSong et al. (2004)
Roseburia intestinalisDSM 14610TCluster XIVaStarch, xylanButyrateAcetate utiliserDuncan et al. (2002a)
Roseburia hominisDSM 16839TCluster XIVaSaccharolyticButyrateAcetate utiliserDuncan et al. (2006)
Roseburia faecisDSM 16840TCluster XIVaStarch and weak growth on xylan and inulinButyrateAcetate utiliserDuncan et al. (2006)
Roseburia inulinivoransDSM 16841TCluster XIVaStarch, inulin, fucoseButyrate, propionate (depending on substrate)Acetate utiliserDuncan et al. (2006); Scott et al. (2006)
Anaerostipes caccaeDSM 14662TCluster XIVaStarch utiliser; SaccharolyticButyrateAcetate and lactate utiliserSchwiertz et al. (2002); Duncan et al. (2004)
Ruminococcus lutiDSM 14534TCluster XIVaSaccharolytic, ferments starch, inulin and fucoseAcetate, succinate, lactateGrows on peptoneSimmering et al. (2002)
Dorea longicatenaDSM 13814TCluster XIVaSaccharolyticFormate, acetateForms ethanolTaras et al. (2002)
Clostridium hathewayiDSM 13479TCluster XIVaSaccharolyticAcetateSpore formerSteer et al. (2001)
Clostridium asparagiformeDSM 15981TCluster XIVaSaccharolyticAcetate, lactateForms ethanolMohan et al. (2006)
Anaerofustis stercorihominisCCUG 47767TCluster XVSaccharolyticAcetate, butyrateBile resistantFinegold et al. (2004)
Gram negative bacteria
Cetobacterium someraeCCUG 46254TCFB phylumPeptides and carbohydratesAcetateBile resistantFinegold et al. (2003)
Alistipes finegoldiiCCUG 46020TCFB phylumFermentation substrate difficult to demonstrateSuccinateBile resistant, pigmented cells, proteolytic, weakly β-hemolyticRautio et al. (2003)
Bacteroides goldsteiniiCCUG 48944TCFB phylumSaccharolyticAcetate, succinateBile resistant, not pigmented, possesses β-glucuronidase activitySong et al. (2005)
Akkermansia mucinophilaCIP 107961TVerrucomicrobiaMucin degrader, No growth on glucoseForms acetate and propionate from mucin Derrien et al. (2004)
Victivallis vadensisDSM 14823TVerrucomicrobiaFerments cellobioseAcetateUnable to grow on normal agar plates; does not use mucusZoetendal et al. (2003)
Methanosphaera stadtmanaeDSM 3091ArchaeonMethanol Methane producerFricke et al. (2006)

Low mol % G + C Gram-positive bacteria (Firmicutes)

The Firmicutes probably make up around 40–65% of the colonic or faecal microbiota (Flint 2006). The most abundant phylotypes fall into three main clostridial clusters (IV, IX and XIV) as defined by 16S rRNA sequencing (Collins et al. 1994) with a lower abundance of several other clusters including I, II, III, XI, XV and XVI.

Clostridial cluster IV This cluster is also referred to as the Clostridium leptum group and comprises around 25% of the colonic microbiota (Flint 2006). Faecalibacterium prausnitzii is a predominant species within this cluster, which was previously assigned to the genus Fusobacterium. F. prausnitzii can metabolize a wide range of carbohydrate substrates, including starch and inulin, to form butyrate and D-lactate. It has an absolute requirement for acetate in the growth medium (Duncan et al. 2002b). Subdoligranulum variabile, a recently described isolate (Holmstrøm et al. 2004), is most closely related to F. prausnitzii (Fig. 1) and can grow on fucose but not starch with butyrate and lactate as fermentation end products (Table 1). Another recently described isolate belonging to this group is Anaerotruncus colihominis (Lawson et al. 2004), which also does not utilize starch and forms butyrate and acetate (Table 1).

Figure 1.

 Phylogenetic tree of low mol% G + C Gram-positive bacteria based on 16S rRNA sequence corresponding to positions 122 to 1426 of the Bifidobacterium bifidum sequence S83624, which was used as the outgroup. Newly isolated strains (ie since 2001) are shown in boldface. Accession numbers for sequences are given in brackets. Bootstrap values greater than 95 (per 100 replications) are shown at branch points. The scale bar represents genetic distance (10 substitutions per 100 nucleotides). Clostridial clusters are indicated by roman numerals.

Many of the ruminococci that belong to cluster IV possess the ability to hydrolyse complex carbohydrates and form acetate as a major fermentation product. Ruminococcus bromii strains can degrade starch whilst others can metabolize complex plant cell wall material. Ruminococcus albus and Ruminococcus flavefaciens relatives have been shown to degrade recalcitrant substrates such as cellulose and R. flavefaciens carries sophisticated enzyme systems organized as cellulosomes on the cell surface (Flint 2006). Most of these have been isolated from the rumen ecosystem but more recently Robert and Bernalier-Donadille (2003) described five novel Ruminococcus isolates from human faeces that could degrade microcrystalline cellulose. These cellulose-degrading bacteria, that produce hydrogen, were only detected in subjects that were methane producers.

Clostridial cluster IX Little detailed work has been carried out on the species composition or functions of this group of bacteria in the human colon, which have been estimated to be around 7% in faecal samples from healthy donors using FISH (Walker et al. 2005; Flint 2006). In other gut ecosystems, such as in ruminants and pigs, many of the species belong to the genera Megasphaera, Veillonella, Selenomonas and Megamonas. Species that have been identified within this cluster are predominantly saccharolytic and some employ the acrylate pathway for propionate formation. Megasphaera elsdenii, a lactate utilizer, can form both butyrate and propionate.

Clostridial cluster XIVa This cluster is made up of a disparate collection of bacterial genera and species and has been estimated to make up around 25% of bacteria found in the colon (Flint 2006). Species in this cluster belong to a number of different genera including Anaerostipes, Clostridium, Coprococcus, Eubacterium, Roseburia and Ruminococcus (Fig. 1). The ruminococci from humans fall into two clusters, namely IV and XIVa, and species belonging to this latter cluster are not all closely related. The Ruminococcus obeum group appears to be abundant in the colon (Suau et al. 1999). They are saccharolytic bacteria that form mainly acetate and include the new related species, Ruminococcus luti (Simmering et al. 2002) (Table 1; Fig. 1). Two other recently isolated saccharolytic bacteria belonging to cluster XIVa, namely Dorea longicatena and Clostridium asparagiforme, can form ethanol (Table 1).

Roseburia spp. and Eubacterium rectale are a major component of this cluster and make up around 7% of the faecal microbiota (Aminov et al. 2006). The Roseburia/E. rectale group along with others including F. prausnitzii (cluster IV), employ the butyryl CoA:acetate CoA transferase route for butyrate formation (Louis et al. 2004). This may be the most important route for butyrate formation in the large intestine although other bacterial species, including isolates related to Clostridium nexile and Coprococcus sp. employ the butyrate kinase route (Louis et al. 2004). In addition Scott et al. (2006), have shown that Roseburia inulinivorans can modulate its metabolic activities and in addition to butyrate can form propionate and propanol when grown on fucose. Butyrate is considered beneficial to colonic health (Csordas 1996; Archer et al. 1998; Avivi-Green et al. 2000) whilst lactate accumulation, in particular the D-isomer, is detrimental. It is interesting therefore that within this cluster a number of bacterial species, including Eubacterium hallii and Anaerostipes caccae can metabolize lactate to form butyrate (Duncan et al. 2004).

Many of the species belonging to this cluster may be difficult to isolate as some are extremely sensitive to exposure to air. Recently, five Roseburia strains belonging to three species were screened for their ability to survive exposure to air on the surface of agar plates and all failed to grow following two minutes exposure (Duncan et al. 2006). Despite this, recently cultured strains (Table 1) apparently cover most of the diversity described by direct 16S rRNA sequencing in this group (Aminov et al. 2006).

Other low G + C clostridial clustersClostridium bartlettii belongs to cluster XI (Song et al. 2004) and is one of the relatively few gut commensal organisms in which spore formation has been readily identified. This organism forms phenylacetic acid (Table 1). Some human gut bacterial species fall into other less abundant clostridial clusters, including XI, XV and XVI. Anaerofustis stercorihominis belongs to clostridial cluster XV and is another butyrate producer that is tolerant of relatively high concentrations of bile (Finegold et al. 2004).

Gram-negative bacteria

Bacteroidetes (Cytophaga-Flavobacterium-Bacteroides) phylum These bacteria make up around 25% of the human colonic microbiota and are Gram-negative, anaerobic rods. Most can metabolize carbohydrates (Salyers et al. 1977), peptones, and/or metabolic intermediates. Saccharolytic species form succinate, acetate, lactate, formate, or propionate as major products. In addition to the low G + C clostridial cluster IX, Bacteroides species are possibly the other main contributors to propionate formation in the colon. Bacteroides fragilis is the type species and it has been suggested that several species may have a role in human disease. Bacteroides spp. may also be particularly dominant in mucosal biofilms from ulcerative colitis patients (Swidsinski et al. 2005). The new genera and species that have been described are proteolytic and tolerant of high concentrations of bile (Table 1).

Verrucomicrobia Two new species within this group, which make up around 1–2% of the colonic microbiota (Zoetendal et al. 2006), are Akkermansia mucinophila (Derrien et al. 2004) and Victivallis vadensis (Zoetendal et al. 2003). A. mucinophila is a mucin degrader whilst V. vadensis cannot grow on mucus and curiously does not grow on the surface of agar plates.


Methanogens are a specialized group of Archaea and approximately one-third of adults are methanogenic (Florin et al. 2000). Only two methanogens from the human gut have been described, namely Methanobrevibacter smithii and Methanosphaera stadtmanae. Complete genome sequencing of M. stadtmanae (Fricke et al. 2006) has revealed that it has the most restricted metabolism of all methanogenic archaea. M. smithii uses H2 and CO2 or formate to form methane whilst M. stadtmanae uses methanol which is likely to arise from pectin degradation by anaerobes including Bacteroides species (Jensen and Canale-Parola 1986).


The colon harbours a highly complex microbial ecosystem. Molecular techniques, such as real-time PCR, clone libraries and PCR independent techniques have allowed researchers to both identify and estimate the abundance of the bacterial groups and species particularly in faecal samples. These diversity studies have begun to reveal where key cultured representatives from the human colon are missing. Examples of deficient groups fall within the low G + C clusters, including clusters IX and XIVa. More recently these clusters have been enriched through careful isolation work using strictly anaerobic techniques. Detailed profiling of these isolates, with respect to substrate utilization, together with in vivo microbial ecology studies will indicate which bacterial species are promoted on different human diets. Currently, limited numbers of complete genome sequences from these bacteria are available but hopefully this deficiency will be corrected in the near future, yielding much valuable information on the colonic microbial community.

Future prospects

Understanding of the microbial ecosystem of the human colon will continue to benefit from a variety of approaches. These include studies in pure culture on the physiology, metabolism and genomics of new and existing bacterial isolates, but preferably with improved ability to create mutants to investigate the functionality of particular genes. Additionally, studies need to be conducted on defined mixed culture systems in vitro in an effort to understand microbe–microbe interactions and in animal studies to understand host-microbe interactions. Most important, however, will be the further development of methodologies for studying the mixed community in vivo using molecular approaches and isotopic tracers to track changes in bacterial populations and their metabolic activities in response to diet and disease states.


The Rowett Research Institute receives financial support from the Scottish Executive Environmental and Rural Affairs Department. We thank Karen Scott for critical reading of the manuscript.