• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Mar 15, 2011; 108(Suppl 1): 4539–4546.
Published online Jun 25, 2010. doi:  10.1073/pnas.1000066107
PMCID: PMC3063586
Colloquium Paper

A metagenomic β-glucuronidase uncovers a core adaptive function of the human intestinal microbiome


In the human gastrointestinal tract, bacterial β-D-glucuronidases (BG; E.C. are involved both in xenobiotic metabolism and in some of the beneficial effects of dietary compounds. Despite their biological significance, investigations are hampered by the fact that only a few BGs have so far been studied. A functional metagenomic approach was therefore performed on intestinal metagenomic libraries using chromogenic glucuronides as probes. Using this strategy, 19 positive metagenomic clones were identified but only one exhibited strong β-D-glucuronidase activity when subcloned into an expression vector. The cloned gene encoded a β-D-glucuronidase (called H11G11-BG) that had distant amino acid sequence homologies and an additional C terminus domain compared with known β-D-glucuronidases. Fifteen homologs were identified in public bacterial genome databases (38–57% identity with H11G11-BG) in the Firmicutes phylum. The genomes identified derived from strains from Ruminococcaceae, Lachnospiraceae, and Clostridiaceae. The genetic context diversity, with closely related symporters and gene duplication, argued for functional diversity and contribution to adaptive mechanisms. In contrast to the previously known β-D-glucuronidases, this previously undescribed type was present in the published microbiome of each healthy adult/child investigated (n = 11) and was specific to the human gut ecosystem. In conclusion, our functional metagenomic approach revealed a class of BGs that may be part of a functional core specifically evolved to adapt to the human gut environment with major health implications. We propose consensus motifs for this unique Firmicutes β-D-glucuronidase subfamily and for the glycosyl hydrolase family 2.

Keywords: functional core, intestinal microbiota, functional metagenomics, glycosyl hydrolase, Firmicutes

Bacterial β-D-glucuronidases (BG; E.C. are part of the extensive human intestinal microbiome and are involved in the metabolism and bioavailability of food and drug compounds in the human body. They catalyze the hydrolysis of exogenous β-glucuronides naturally occurring in diet and drugs as well as endogenous β-glucuronides produced in the liver by glucuronosyltransferases, a major xenobiotic detoxication pathway. Indeed, numerous compounds including metabolites, vitamins, steroid hormones, xenobiotics, and drugs are excreted via the bile and the digestive tract after conversion to a more hydrophilic glucuronidated form. Nevertheless, secondary deglucuronidation, primarily due to intestinal bacteria (1), promotes recycling of the aglycone forms through the enterohepatic cycle, which prevents their removal from the human body. This phenomenon is particularly well described in the case of xenobiotics (2, 3) and is suggested for circulating hormones (4, 5). Despite their detrimental effects, bacterial BGs are also thought to have beneficial effects, notably on the bioavailability of active metabolites derived from dietary compounds, including lignans, flavonoids, sphingolipids, and glycyrrhizin (68).

Finally, observations argue in favor of the involvement of bacterial BGs in several pathologies. Indeed, a lower BG activity was detected in the feces of Crohn's disease patients compared to healthy subjects (9). Conversely, a high BG activity is recognized as a prognosis marker for colon cancer (10). Furthermore, the Clostridium leptum group (cluster IV) or Lachnospiraceae (cluster XIVa), major reservoirs of BG-positive bacteria (11), are dysbiosis markers in Crohn's disease (1215) and in intestinal carcinogenesis (16). Nevertheless, the relative contribution of bacterial strains to the global intestinal BG activity remains unclear, and the molecular bases are essentially unknown.

In the human digestive tract, bacterial BGs are known to be distributed among the Enterobacteriaceae family in some Firmicutes genera (Lactobacillus, Streptococcus, Clostridium, Ruminococcus, Roseburia, and Faecalibacterium) and in a specific species of Actinobacteria (Bifidobacterium dentium) (11, 17, 18). On the basis of their sequence, BGs are highly homologous to β-galactosidases and only a few of them have been investigated and clearly annotated as β-D-glucuronidases in sequence databases. The most notorious prokaryotic BGs are those encoded by uidA homologs (also named gusA) from Escherichia coli, Lactobacillus gasseri, and Ruminococcus gnavus (17, 19, 20).

In the present study, we report the identification of a BG from intestinal bacteria by a functional metagenomic approach. The identified enzyme is unrelated to the previous ones and expands the number of bacterial enzymes responsible for glucuronidase activity. Furthermore, its sequence and distribution among bacterial genomes and metagenomes highlights the relevance and the diversity of BGs within the intestinal microbiome.


High Frequency and Diversity of BG Activities in Metagenomic Libraries.

A strategy was developed to assay BG activity from metagenomic inserts in spite of the constitutive activity of the E. coli strain used to built the metagenomic libraries (E. coli DH10B) (Fig. 1). The method consisted of a two-step screen of clones containing large metagenomic inserts. Using p-nitrophenyl β-D-glucuronide as chromogenic substrate (PNP-G deglucuronidation test), the clones’ activity was first compared to the basal activity of the receiving strain. Fosmids born by positive clones were further transferred into an E. coli strain deprived of BG activity (E. coli L90 ΔuidA). The clones were then grown with 5-bromo-4-chloro-3-indolyl-β-D-glucuronide in the medium (X-GlcA deglucuronidation test) to confirm BG activity.

Fig. 1.
Screening of BG activity from metagenomic libraries. The metagenomic clones originated from different libraries derived from healthy and Crohn's disease subjects’ feces and from a distal ileum biopsy. Metagenomic libraries were screened for over-expressing ...

With this strategy, two libraries from human gut microbiota were screened: a metagenomic library of 4,608 clones combining uncharacterized inserts (Materials and Methods) and a sublibrary of 1,536 clones previously selected for the presence of 16S rRNA genes (14). An initial visual screen was performed and revealed that 1.79% of clones over-expressed BG activity in the uncharacterized library. The quantitative validation step confirmed an over-expression for 40% of them (Fig. 1B), and transfer of the corresponding metagenomic inserts into the ΔuidA strain allowed the determination of a hit rate of 0.41% of positive fosmids. Including the second screen performed on the 16S rRNA gene sublibrary, 19 positive fosmids were identified of a total of 6,144 metagenomic clones. BG activity in the positive clones ranged from 0.02 to 1.30 U (units) with, as expected, a lower global level of BG activity when metagenomic inserts were expressed in the ΔuidA E. coli strain (Fig. 2).

Fig. 2.
BG activity of positive metagenomic fosmids in E. coli DH10B and E. coli L90 ΔuidA. Activity was assayed using PNP-G as substrate and expressed in units (U) of BG activity (ΔOD 405 nm·min−1 · mg−1). Means ...

A Unique Class of BGs in the Firmicutes Phylum.

An initial attempt was made to sequence the potential BGs present in the 19 BG-positive clones on the basis of degenerate primers. The sequence of the primers used derived from the sequences of known prokaryotic BG genes (17, 19). As this approach proved to be unsuccessful, the BG-positive fosmids were subcloned and subsequently sequenced. The 40-kbp inserts were fully sequenced, and six candidate genes were identified as potentially encoding BG activity. The predicted proteins encoded by these genes were the following: β-galactosidase, antimicrobial peptide ABC transport system, serine/threonine kinase, sporulation protein, 2C-methyl-D-erythritol 2,4-cyclodiphosphate synthase, and aspartyl-tRNA synthetase (Table S1). The corresponding genes were subcloned into an expression plasmid (pRSF plasmid) and further transformed into E. coli BL21(DE3) or E. coli L90 T7+ (uidA+ and ΔuidA strains, respectively).

Only the gene encoding the putative β-galactosidase was able to promote BG over-expression in the uidA+ strain (5.3 versus 0.4 U for the control deprived of insert) and to induce a BG activity in the ΔuidA strain (4.9 U). This gene originated from a metagenomic insert (H11G11 insert) in which BG activity was in the range of the 19 fosmids (Fig. 2). This insert was from a pool of fecal samples from healthy individuals. Remarkably, the identified glycosidase had low homology with known BGs (<33% identity, Fig. 3) and signatures that diverged from the two consensus sequences of glycosyl hydrolase family 2 (82 and 93% similarity with Prosite patterns PS00719 and PS00608, respectively; Table S2). This glycosidase also had low homology with putative BGs recently identified in the gut dominant species Roseburia intestinalis and Bacteroides ovatus using degenerate PCR primers (27% and 33% identity, respectively) (11). However, the two glutamate residues involved in the catalytic mechanism of other members of family 2 BGs (21) were conserved in this BG, and a search in the Conserved Domains Database (22) revealed the highest matches to the TIM barrel domain of glycosyl hydrolase family 2 (pfam02836) and BGs (PRK10150) and lower matches to β-galactosidases (COG3250, PRK09525, PRK10340). An additional C-terminal domain compared to known BGs was identified, but no putative function could be proposed (expected value 0.01).

Fig. 3.
Amino acid sequence alignment of the unique BG (H11G11-BG) and known BGs from the gut microbiota. Alignment was performed using the ClustalW multiple sequence alignment program. Framed amino acids were conserved in at least three sequences.

A search of homologous proteins in the National Center for Biotechnology Information (NCBI) nr database revealed a set of 15 H11G11-BG homologs with more than 50% coverage and similarity (Fig. 4). On the basis of amino acid sequence alignment, all homologs constituted a biologically relevant group (COFFEE score: from 89 to 95) with 38–57% identity with H11G11-BG. These proteins were all identified in prokaryotic genomes under sequencing and were annotated as “uncharacterized protein” except the one from Paenibacillus sp. JDR-2, which was annotated as a β-galactosidase. The H11G11-BG homologs were identified only in gut bacteria from the Firmicutes phylum including Bacteroides capillosus, recently reassigned to Gram-positive bacteria (23). The only exception was Shuttleworthia satelles DSM 14600, which is a Firmicutes isolated from the oral cavity (24). The closest neighbors were functionally uncharacterized genes in gut Bacteroidales and represented a likely subgroup of Bacteroidetes BGs (Fig. 1B).

Fig. 4.
Identification of a unique class of β-D-glucuronidase, phylogenetic affiliation, and association with potential symporters. Detailed tree view of the H11G11BG-like homologs identified in GenBank genomic databases (Fig. S1). The unique BGs (>50% ...

The unique Firmicutes BGs had amino acid sequences distantly related to known BGs, even from the ones previously characterized in this phylum (Table S2 and Fig. S1), and had specific uncharacterized and additional domains (Fig. S2). Most of the identified genes were restricted to the Clostridiales order, including members of the Ruminococcaceae, Lachnospiraceae, and Clostridiaceae families (Fig. 4). Three strains were previously known to exhibit BG activity (i.e., Faecalibacterium prausnitzii M21/2, R. gnavus ATCC 29149, and Bacteroides capillosus ATCC 29799). Seven strains of Firmicutes and Bacteroidetes containing H11G11-BG homologs in their genomes were tested for their ability to hydrolyze PNP-G (i.e., B. ovatus ATCC 8483, Bryantella formatexigens DSM 14469, Clostridium bartlettii DSM 16795, Roseburia inulinivorans DSM 16841, Subdoligranulum variabile DSM 15176, Parabacteroides johnsonii DSM 18315, and Parabacteroides merdae ATCC 43184). Except for R. inulinivorans DSM 16841 (BG activity <0.025 U), all strains tested exhibited high BG activity. The search of known uidA homologs in their genomes and in the three Firmicutes strains with known BG activity led to different situations (Table S3). For most of the genomes tested (7/10), the homologs found were more closely related to the H11G11 protein sequence than to UidA. The presence of the two classes of BG (uidA and H11G11-like) within the same genome remained possible for F. prausnitzii M21/2 and B. formatexigens DSM 14469, and even a third class was suggested in B. ovatus ATCC 8483. Nevertheless, these data suggest that in most cases H11G11-BG was responsible for the observed BG activity. Furthermore, paralogs were found in the genome of S. variabile DSM 15176, B. formatexigens DSM 14469, and R. gnavus ATCC 29149.

Because the two existing Prosite consensus signatures (PS00719 and PS00608) for glycosyl hydrolase family 2 could not include these unique BGs (Table S2), we propose different consensus motifs and a signature for the unique Firmicutes BG group (Fig. 5 AB) and more broadly for the glycosyl hydrolase family 2 (Fig. 5C). Interestingly, a clear distinction inside the unique BG family was observed. Indeed, the invariant histidine/tyrosine/glutamine motif was conserved in the Bacteroidetes subgroup but was replaced by a histidine/tyrosine/proline motif in the Firmicutes subgroup (Fig. 5C and Fig. S3).

Fig. 5.
Conserved motifs within the H11G11-BG group. (A) Conserved motifs within the 17 proteins identified in this study (>50% coverage and similarity with H11G11-BG). (B) Signatures proposed for the unique Firmicutes BGs. (C) Previously undescribed ...

Is Functional Diversity of Unique BGs Based on Associated Symporters?

Insert sequence alignments and genetic environment exploration of 15 of the 17 Firmicutes BGs identified revealed that the conserved genetic environments close to the BG genes were restricted to symporters. Fifteen BGs were associated with putative symporters that, by sequence homology, clustered into four types (Fig. S4). The symporter homologies correlate well to the BG subclades (Fig. 4). Despite the diversity of the genetic organizations observed (direct vicinity, intermediate sequences, lengths) (Figs. S4 and S5), most of these putative symporters had best hits for conserved domains corresponding to Na+/melibiose symporter and related transporters (COG2211) or the sugar transporter Gph (TIGR00792), indicating that these BGs are involved in the metabolism of extracellular glucuronides.

The Unique BG Within the Phylogenetic Core.

Among healthy adults the human intestinal microbiota phylogenetic core consists of a restricted set of dominant species (25). To determine the presence of the H11G11-BG–positive species within this phylogenetic core, sequence alignments between the 16S rRNA of H11G11-BG-positive strains and the 66 prevalent operational taxonomic units (OTUs) (25) were performed. Almost half of the BG-positive strains (i.e., F. prausnitzii M21/2, R. inulinivorans DSM 16841, S. variabile DSM 15176) belonged to the phylogenetic core (>99% 16S rRNA identity). Because the phylogenetic core represents only a very few highly prevalent and dominant OTUs, this result underlined the importance of this BG function in the human gut.

A BG Highly Prevalent in and Specific to the Human Gut.

The frequency of the unique Firmicutes or Bacteroidetes BGs reached an average of 9.5 ± 3.0 × 10−8 hit/bp within the human gut metagenomes (Fig. 6). At least one homolog can therefore be found every 107 bp, which is approximately equivalent to 104 bacterial genes. Previously known BG genes (uidA homologs from Clostridiales, Lactobacillales, and E. coli K12) were at least three times less frequently represented in the human gut metagenomes. Remarkably, the H11G11-BGs subgroup was almost not represented or was totally absent in microbial metagenomes arising from other environments whereas the uidA BGs subgroup was more systematically found in soil ecosystems (Waseca County Farm Soil Metagenome ID 13699).

Fig. 6.
Frequency of the unique β-glucuronidases and known β-glucuronidases in environmental and human gut metagenomes. Homologs of unique and known BG proteins were searched in NCBI metagenomic project datasets (Materials and Methods). The hit ...

To determine the distribution among the human population, homologs of H11G11-BGs were searched in gut metagenomic databases from different healthy individuals (26) (Fig. 7). Homologs of the unique Firmicutes or Bacteroidetes BGs were identified in all 11 adults and children but only in one infant among the four explored (Fig. 7 and Fig. S6). In contrast, uidA BGs were less abundant among the “adults/children” group (6/11, 8/11, and 7/11 for E. coli, Clostridiales, and Lactobacillales BGs, respectively) but were present in all infants. Furthermore, in adults and children, the H11G11-BG frequencies were always higher than those of the gusA/uidA BGs (P ≤ 0.01). They ranged from 2.3 × 10−8 to 2.2 × 10−7 hit/bp, and the frequency in each individual was not correlated to the previously determined level of COG3250 (β-galactosidase/β-glucuronidase) (26).

Fig. 7.
Distribution of the unique and known BGs among adult, child, and infant gut metagenomes. Homologs of unique and known BG proteins were searched in the two projects Human Gut Metagenome (13 healthy individuals) (ID 28117) and Human Distal Gut Biome (ID16729) ...


Using a functional metagenomic approach, we revealed a high frequency of BG activity (0.41%) in the human gut microbiota. In the high range of hits usually obtained during functional screens of glycosyl hydrolases from environmental or gut metagenomic libraries (from 0.00002% to 0.8%) (27), this result is coherent with the particular enrichment of the gut microbiome with β-galactosidase/β-glucuronidase family (28). This result probably reflects the abundance and the diversity of glucuronide compounds that reach the intestine and the critical role of BGs in several bacterial metabolic pathways. However, this high score could also be due to the secondary deglucuronidation potentialities of proteins that exhibit no obvious sequence homologies with BGs as suggested in this study. We were unable to find evidence of BG activity among these cloned genes. Further investigations of these genes in their genetic context will be required to definitely demonstrate their functionality.

Our functional screen led to the identification of a unique β-D-glucuronidase from the human gut microbiota that possesses several key features, including an additional C terminus domain and distant sequence homologies with known BGs. Mining of genome sequence databases demonstrated that this unique BG is specifically present into several Firmicutes species of the gastrointestinal tract or the oral cavity including C. bartlettii, S. variabile, B. formatexigens, R. gnavus, Penibacillus sp., R. inulinivorans, S. satelles, and F. prausnitzii. This group includes bacterial strains already reported to exhibit BG activity (i.e., F. prausnitzii M21/2, R. gnavus ATCC 29149, and B. capillosus ATCC 29799) (11, 17, 29) and strains that we demonstrated here as BG positive (i.e., S. variabile DSM 15176, B. formatexigens DSM 14469, C. bartlettii DSM 16795). In the same way, BG-positive strains from the Bacteroidetes phylum were evidenced (i.e., B. ovatus ATCC 8483, P. merdae ATCC 43184, and P. johnsonii DSM 18315). Thus this class of BGs, which is present in many dominant Firmicutes and Bacteroidetes species, might represent a major deglucuronidation pathway in the human gut. The unique BGs exhibit conserved domains that diverged from the Prosite consensus sequences (PS00719 and PS00608) of glycosyl hydrolase (GHase) family 2. We thus proposed consensus signatures to include this unique class of BGs. Interestingly, the invariant “HYP” amino acid motif found in the PS00719 signature was maintained in Bacteroidetes BGs but specifically replaced by a “HYQ” motif in Firmicutes BGs. These Firmicutes BGs may therefore have originated from a common ancestor and evolved in the context of the intestinal environment. Considering the importance of GHase family 2, which is the third most abundant family of GHases in the Human gut (27), the functional significance of these sequence specificities deserves further investigation.

Our observations support an important physiological role for these unique BGs. Indeed, in contrast to previously identified BGs (UidA/GusA homologs), these BGs are found with high frequency and high specificity in the human gut microbiome and are well represented and distributed among adults and children but absent in most of the very young infants tested despite the relative known abundance of Firmicutes in their microbiota (26, 30). The absence of correlation between frequencies of H11G11-BG homologs and the galactosidase/glucuronidase COG family within human gut microbiomes demonstrates the necessity of improving this family classification and annotation. Furthermore, among the bacterial species harboring this enzyme family, three Firmicutes species belong to the restricted phylogenetic core of the human intestinal microbiota (25), suggesting an ecological drive that ensures the presence of the activity in spite of microbiota variability among humans. This unique class of BGs might therefore play a critical role in the establishment of a definitive and stabilized adult microbiota (31). Results also revealed duplication of unique BG-encoding genes in several strains and association with diverse permeases. These organizations are biologically highly relevant because they are important in generating genetic variability and in facilitating adaptive evolution and functional diversity. Interestingly, most of the species harboring the unique BG were from the C. leptum group (cluster IV) or Lachnospiraceae (cluster XIVa), both subject to microbiota disturbances in the physiopathological contexts of Crohn's disease (12, 3234), ulcerative colitis (35), familial Mediterranean fever (36), and intestinal carcinogenesis in Apc (Min) mice (16). Recent studies revealed a functional link between xenobiotic neutralization, steroid metabolism, and inflammation via mutual repression of xenobiotic and hormone nuclear receptors and the NF-κB–signalling pathways in the intestine (3739). Determination of potential xenobiotics or hormone substrates for BGs may be critical to determining their physiological function(s) and potential contribution to the modulation of eukaryotic regulatory networks.

In conclusion, the functional metagenomic approach reported here allows the identification of a unique BG type, which is predominant in the gut microbiota of human adults and children. This BG family, identified in several Firmicutes, may be part of the functional core of the human microbiome recently proposed for the GIT ecosystem (40) and may represent an adaptive evolution. Further investigations should address the impact of functional BGs in individuals‘ gut physiology and their involvement in the crosstalk with host cells, especially with respect to immune homeostasis.

Materials and Methods

Metagenomic Libraries.

Fosmidic metagenomic libraries were constructed as previously described using the EpiFOS Library production kit (Epicentre Technologies) and E. coli DH10B as a host (14). Three fosmidic metagenomic libraries were used in this study; namely the “Healthy,” “Crohn,” and “Ileum” libraries. The first two libraries were previously constructed from bacterial DNA extracted from pools of feces of six healthy individuals and six inflammatory bowel disease patients, respectively (14), and the third from a colorectal cancer patient biopsy obtained from the healthy distal part of the ileum. From these we used a set of 4,600 clones that contained random genomic bacterial DNA and a set of 1,536 clones selected for the presence of 16S rRNA genes by DNA hybridization (14).

Screening of BG-Positive Fosmid Clones.

Because the E. coli strain used to build the metagenomic libraries (E. coli DH10B) constitutively expressed a β-glucuronidase encoded by the uidA gene, an initial visual screening was performed to detect metagenomic clones over-expressing BG activity. Metagenomic clones were grown overnight in 2YT medium with chloramphenicol (12.5 μg · ml−1), and the activity was measured on permeabilized cells after toluene treatment (10%, 10 min at 4 °C). BG activity was measured by adding 2 mM of PNP-G (Sigma-Aldrich) (41). Positive clones were submitted to a second quantitative screen using a kinetic measurement with E. coli DH10B transformed with an empty fosmid as control. Eight replicates were performed for each clone. Activity and bacterial density (E. coli dry weight) were estimated by measuring ΔOD405nm and OD600nm, respectively. The BG activity was expressed in units of ΔOD405nm.min−1·mg−1 of dry weight.

As E. coli DH10B possesses a β-glucuronidase activity encoded by the uidA gene, the ΔuidA E.coli L90 strain was used to unambiguously identify fosmids expressing β-glucuronidase activity. This strain was derived from the TG1 ΔuidA::Kanr strain that was previously constructed by gene homologous recombination (3). The kanamycine resistance cassette was removed using FLP recombinase borne by the pCP20 plasmid. Purified fosmids (NucleoSpin 96 Flash, extraction kit, Marchery Nagel) were then electro-transferred (25 μFd, 200 Ω, and 2.5 Kv) in the E. coli L90 strain. Transformants were grown on LB agar containing the appropriate antibiotic with 5-bromo-4-chloro-3-indolyl-β-D-glucuronide (25 μg·ml−1) (X-gluc, VWR) to reveal BG activity.

Fosmids Subcloning and Sequence Analysis.

Metagenomic fosmids were purified using the NucleoSpin 96 Flash extraction kit (Macherey-Nagel). Insert size was estimated by HindIII digestion followed by pulsed-field gel electrophoresis. Fosmid DNA was subjected to mechanical shearing, and 3-kbp fragments were inserted into the high-copy plasmid pcDNA2.1. A total of 384 subclones originating from each metagenomic BG-positive clone were screened for BG activity, and positive subclones were further sequenced (MWG Biotech). A homolog search was made using the Blastp algorithm available on the NCBI web site (http://www.ncbi.nlm.nih.gov/BLAST/). Domain architectures, potential functions, and associated patterns were analyzed using the InterProScan (http://www.ebi.ac.uk/InterProScan/), MyHits (http://www.isb-sib.ch/), PROSITE (http://www.expasy.org/prosite/), SMART (http://smart.embl-heidelberg.de/), PFAM (http://www.sanger.ac.uk/Software/Pfam/), Motif Scan (http://myhits.isb-sib.ch/cgi-bin/motif_scan), or SUPERFAMILY (http://supfam.org/SUPERFAMILY/) softwares.

β-Glucuronidase Candidates, Cloning, and Expression.

Genes encoding putative BG were amplified by PCR with metagenomic fosmid DNA as templates and appropriate primers (Table S4). PCR products were purified by agarose gel electrophoresis and the Wizard SV Gel and PCR clean up system kit (Promega). PCR products were subcloned into pGEM-T plasmid (Promega) and fully sequenced. pGEM-T plasmids with exact insert sequence were digested by BamHI and PstI restriction enzymes (except for the putative serine/threonine protein kinase cloned with the EcoRI and PstI restriction site). Inserts were further purified by agarose gel electrophoresis and the Wizard SV Gel and PCR clean up system kit (Promega). Purified inserts were ligated into pRSFDuet-1 plasmid (Novagen) previously digested with the same restriction enzymes. All cloned genes were sequenced to ensure that no errors were introduced. The plasmids obtained were transformed into both E. coli BL21(DE3) (Stratagene) and the L90 T7+ E. coli strain. The L90 T7+ E. coli strain was obtained by transforming the L90 ΔuidA E. coli strain with the pAR1219 plasmid, which expresses the T7 polymerase under the control of an isopropyl 1-thio-β-D-galactopyranoside (IPTG)-inducible promoter. The resulting clones were analyzed for PNP-G deglucuronidation as described above, after gene expression induction with IPTG.

Sequence Analysis of the Unique BG Group and Associated Symporters.

Homologs of H11G11-BG were searched on May 2009 by Blastp method (limit expect E-value ≤2e-45) on the NCBI nr database, including all nonredundant GenBank coding sequences. Radial tree of Blastp results was computed by the fast minimum evolution method (threshold of maximum sequence difference: 0.85). Definitive presentation and phylogenetic symbols were realized using Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0 after conversion in Newick file format. Proteins showing at least 50% length coverage and 50% similarity of aligned sequences were retained to constitute the group of H11G11-BG homologs. Sequence similarities with the current conserved patterns of the glycosyl hydrolases family 2 (BG and galactosidases, PROSITE documentation PDOC00531, http://www.expasy.org/prosite/) were determined using PATTINPROT (http://npsa-pbil.ibcp.fr/). Conserved motifs were determined by using the motif-based sequence analysis tool MEME (http://meme.nbcr.net/meme4_1/). The patterns proposed were obtained by using the discover patterns tool PRATT (http://www.expasy.ch/tools/pratt/). Comparative analysis of codon usage was performed with known Firmicutes BGs and H11G11-like BG proteins identified by using the Blastp software.

The genetic environments of the H11G11 and C7D2 BGs were analyzed using GeneMark (Version 2.4) for Prokaryotes (http://opal.biology.gatech.edu/GeneMark/gmhmm2_prok.cgi). The different types of putative symporters and their identities/similarities within types were determined after homology search and ClustalW multiple alignment.

Phylogenetic Assignment.

The phyla and family assignments of H11G11-BG homologs from genomic databases were from the taxonomy reports automatically generated by NCBI (completed with the classification of R. gnavus ATCC 29149 using the Ribosomal Database Project).

The C7D2 insert was isolated from the 16S rDNA sublibrary and thus harbored a 16S rRNA gene (NCBI accession no. AY850499) allowing phylogenetic assignment by the Ribosomal Database Project classifier (http://rdp.cme.msu.edu/). However, for C7D2 and H11G11 inserts, potential ORFs were determined using MetaGene (http://metagene.cb.k.u-tokyo.ac.jp/), and sequences were further phylogenetically assigned using the best blast hits (Blastn or Blastx software) on NCBI databases. The final phylogenetic assignment was expressed in percentage of insert length coverage.

Phylogenetic relationships with the human gut bacterial phylogenetic core (25) were analyzed in retrieving the 16s rRNA gene sequences of the H11G11-BG–containing strains and analyzing them with the RapidOTU pipeline (25) (0.02% identity cut-off, Kimura-2 parameter distance method). The percentage of identity between the 16s rRNA sequences of BG-positive strains and their closest phylogenetic core homologs was determined using pairwise alignment with Jalview 2.4.0b2.

BG Activity Assay in Gut Bacteria.

Strains encoding a H11G11 homolog in their genomes were grown for 24 h at 37 °C under anaerobic conditions using the Hungate method in appropriate media. The Cooked Meat Medium (Difco) supplemented with yeast extract (5 g/L, Sigma), KH2PO4 (5 g/L, Sigma), and cysteine (0.5 g/L, Sigma) was used for C. bartlettii DSM 16795. B. ovatus ATCC 8483 and R. gnavus ATCC 29149 were grown in LYHBHI medium [brain–heart infusion medium with 0.5% yeast extract (Difco) and 5 mg/L hemin] supplemented with cysteine (0.25 mg/mL; Sigma). The same medium was supplemented with cellobiose (0.5 mg/mL; Sigma–Aldrich) and maltose (0.5 mg/mL; Sigma) for S. variabile DSM 15176 and F. prausnitzii M21/2. Soluble starch (0.5 mg/mL; Sigma) was added to this medium to grow P. merdae ATCC 43184 and P. johnsonii DSM 18315, and 0.4% of clarified rumen juice (42) was also necessary to grow B. formatexigens DSM 14469 and R. inulinivorans DSM 16841. PNP-G activity was tested as described above for E. coli.

Comparative Metagenomic Analysis.

Comparative metagenomic analysis was performed on May 2009 using the tBlastn software to search for H11G11-BG homologous proteins in the following NCBI metagenomic project datasets (http://metasystems.riken.jp/metabiome/metagenome.php): Human Gut Metagenome (13 healthy individuals) (ID 28117), Human Distal Gut Biome (ID16729), Combined Gut Metagenome from Obese and Lean Mice (ID17401), Termite Gut Metagenome (ID19107), Fish Metagenome (ID28955), Whale Fall Metagenome (ID13700), Microbial Mat Isolate (ID29795), Waseca County Farm Soil Metagenome (ID13699), and Global Ocean Sampling (ID19733). The same approach was performed with representative known or putative BGs (UniProtKB databases, http://www.uniprot.org/, June 2009). The similarity hit threshold was at least 50% similarity with 50% sequence coverage. To compare frequencies between databases, results are expressed as hits per base-pair unit.

Supplementary Material

Supporting Information:


We thank Dr. J. Tap for helpful discussions, M. Serezat for technical assistance, and C. Bridonneau for advice about gut bacteria cultivation. This research was supported by the French Ministry of Research under project GenoTube, and insert sequences were performed by the French National Sequencing Center (Genoscope, Evry, France).


This paper results from the Arthur M. Sackler Colloquium of the National Academy of Sciences, “Microbes and Health,” held November 2–3, 2009, at the Arnold and Mabel Beckman Center of the National Academies of Sciences and Engineering in Irvine, CA. The complete program and audio files of most presentations are available on the NAS Web site at http://www.nasonline.org/SACKLER_Microbes_and_Health.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data depostion: BG sequences from C7D2 and H11G11 metagenomic inserts have been deposited at EMBL/GenBank/DDBJ databases under accession nos. FN666674 and FN666673, respectively.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1000066107/-/DCSupplemental.


1. Rod TO, Midtvedt T. Origin of intestinal beta-glucuronidase in germfree, monocontaminated and conventional rats. Acta Pathol Microbiol Scand [B] 1977;85:271–276. [PubMed]
2. Knasmuller S, et al. Impact of bacteria in dairy products and of the intestinal microflora on the genotoxic and carcinogenic effects of heterocyclic aromatic amines. Mutat Res. 2001;480–481:129–138. [PubMed]
3. Humblot C, et al. Beta-glucuronidase in human intestinal microbiota is necessary for the colonic genotoxicity of the food-borne carcinogen 2-amino-3-methylimidazo[4,5-f]quinoline in rats. Carcinogenesis. 2007;28:2419–2425. [PubMed]
4. de Herder WW, Hazenberg MP, Pennock-Schroder AM, Hennemann G, Visser TJ. Rapid and bacteria-dependent in vitro hydrolysis of iodothyronine-conjugates by intestinal contents of humans and rats. Med Biol. 1986;64:31–35. [PubMed]
5. Graef V, Furuya E, Nishikaze O. Hydrolysis of steroid glucuronides with beta-glucuronidase preparations from bovine liver, Helix pomatia, and E. coli. Clin Chem. 1977;23:532–535. [PubMed]
6. Kim DH, et al. Intestinal bacterial metabolism of flavonoids and its relation to some biological activities. Arch Pharm Res. 1998;21:17–23. [PubMed]
7. Kim DH, et al. Biotransformation of glycyrrhizin by human intestinal bacteria and its relation to biological activities. Arch Pharm Res. 2000;23:172–177. [PubMed]
8. Schmelz EM, et al. Ceramide-beta-D-glucuronide: Synthesis, digestion, and suppression of early markers of colon carcinogenesis. Cancer Res. 1999;59:5768–5772. [PubMed]
9. Carrette O, et al. Bacterial enzymes used for colon-specific drug delivery are decreased in active Crohn's disease. Dig Dis Sci. 1995;40:2641–2646. [PubMed]
10. Geier MS, Butler RN, Howarth GS. Probiotics, prebiotics and synbiotics: A role in chemoprevention for colorectal cancer? Cancer Biol Ther. 2006;5:1265–1269. [PubMed]
11. Dabek M, McCrae SI, Stevens VJ, Duncan SH, Louis P. Distribution of beta-glucosidase and beta-glucuronidase activity and of beta-glucuronidase gene gus in human colonic bacteria. FEMS Microbiol Ecol. 2008;66:487–495. [PubMed]
12. Baumgart M, et al. Culture independent analysis of ileal mucosa reveals a selective increase in invasive Escherichia coli of novel phylogeny relative to depletion of Clostridiales in Crohn's disease involving the ileum. ISME J. 2007;1:403–418. [PubMed]
13. Frank DN, et al. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci USA. 2007;104:13780–13785. [PMC free article] [PubMed]
14. Manichanh C, et al. Reduced diversity of faecal microbiota in Crohn's disease revealed by a metagenomic approach. Gut. 2006;55:205–211. [PMC free article] [PubMed]
15. Scanlan PD, Shanahan F, O'Mahony C, Marchesi JR. Culture-independent analyses of temporal variation of the dominant fecal microbiota and targeted bacterial subgroups in Crohn's disease. J Clin Microbiol. 2006;44:3980–3988. [PMC free article] [PubMed]
16. Mai V, Colbert LH, Perkins SN, Schatzkin A, Hursting SD. Intestinal microbiota: A potential diet-responsive prevention target in ApcMin mice. Mol Carcinog. 2007;46:42–48. [PubMed]
17. Beaud D, Tailliez P, Anba-Mondoloni J. Genetic characterization of the beta-glucuronidase enzyme from a human intestinal bacterium, Ruminococcus gnavus. Microbiology. 2005;151:2323–2330. [PubMed]
18. Roy D, Ward P. Rapid detection of Bifidobacterium dentium by enzymatic hydrolysis of beta-glucoronide substrates. J Food Prot. 1992;55:291–295.
19. Russell WM, Klaenhammer TR. Identification and cloning of gusA, encoding a new beta-glucuronidase from Lactobacillus gasseri ADH. Appl Environ Microbiol. 2001;67:1253–1261. [PMC free article] [PubMed]
20. Blanco C. Transcriptional and translational signals of the uidA gene in Escherichia coli K12. Mol Gen Genet. 1987;208:490–498. [PubMed]
21. Salleh HM, et al. Cloning and characterization of Thermotoga maritima beta-glucuronidase. Carbohydr Res. 2006;341:49–59. [PubMed]
22. Marchler-Bauer A, Bryant SH. CD-Search: Protein domain annotations on the fly. Nucleic Acids Res. 2004;32(Web Server issue):W327–W331. [PMC free article] [PubMed]
23. Carlier JP, Bedora-Faure M, K'Ouas G, Alauzet C, Mory F. Proposal to unify Clostridium orbiscindens (Winter et al. 1991) and Eubacterium plautii (Seguin 1928; Hofstad and Aasjord 1982) with description of Flavonifractor plautii gen. nov., comb. nov. and reassignment of Bacteroides capillosus to Pseudoflavonifractor capillosus gen. nov., comb. nov. Int J Syst Evol Microbiol. 2010;60:585–590. [PubMed]
24. Downes J, Munson MA, Radford DR, Spratt DA, Wade WG. Shuttleworthia satelles gen. nov., sp. nov., isolated from the human oral cavity. Int J Syst Evol Microbiol. 2002;52:1469–1475. [PubMed]
25. Tap J, et al. Towards the human intestinal microbiota phylogenetic core. Environ Microbiol. 2009;11:2574–2584. [PubMed]
26. Kurokawa K, et al. Comparative metagenomics revealed commonly enriched gene sets in human gut microbiomes. DNA Res. 2007;14:169–181. [PMC free article] [PubMed]
27. Li LL, McCorkle SR, Monchy S, Taghavi S, van der Lelie D. Bioprospecting metagenomes: Glycosyl hydrolases for converting biomass. Biotechnol Biofuels. 2009;2:10. [PMC free article] [PubMed]
28. Gill SR, et al. Metagenomic analysis of the human distal gut microbiome. Science. 2006;312:1355–1359. [PMC free article] [PubMed]
29. Knivett VA, Shah HN, McKee AS, Hardie JM. Numerical taxonomy of some non-saccharolytic and saccharolytic Bacteroides species. J Appl Bacteriol. 1983;55:71–80. [PubMed]
30. Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO. Development of the human infant intestinal microbiota. PLoS Biol. 2007;5:e177. [PMC free article] [PubMed]
31. Paliy O, Kenche H, Abernathy F, Michail S. High-throughput quantitative analysis of the human intestinal microbiota with a phylogenetic microarray. Appl Environ Microbiol. 2009;75:3572–3579. [PMC free article] [PubMed]
32. Jansson J, et al. Metabolomics reveals metabolic biomarkers of Crohn's disease. PLoS One. 2009;4:e6386. [PMC free article] [PubMed]
33. Sokol H, et al. Low counts of Faecalibacterium prausnitzii in colitis microbiota. Inflamm Bowel Dis. 2009;15:1183–1189. [PubMed]
34. Willing B, et al. Twin studies reveal specific imbalances in the mucosa-associated microbiota of patients with ileal Crohn's disease. Inflamm Bowel Dis. 2009;15:653–660. [PubMed]
35. Zhang M, et al. Structural shifts of mucosa-associated lactobacilli and Clostridium leptum subgroup in patients with ulcerative colitis. J Clin Microbiol. 2007;45:496–500. [PMC free article] [PubMed]
36. Khachatryan ZA, et al. Predominant role of host genetics in controlling the composition of gut microbiota. PLoS One. 2008;3:e3064. [PMC free article] [PubMed]
37. Wahli W. A gut feeling of the PXR, PPAR and NF-kappaB connection. J Intern Med. 2008;263:613–619. [PubMed]
38. Zhou C, et al. Mutual repression between steroid and xenobiotic receptor and NF-kappaB signaling pathways links xenobiotic metabolism and inflammation. J Clin Invest. 2006;116:2280–2289. [PMC free article] [PubMed]
39. Zhou C, Verma S, Blumberg B. The steroid and xenobiotic receptor (SXR), beyond xenobiotic metabolism. Nucl Recept Signal. 2009;7:e001. [PMC free article] [PubMed]
40. Turnbaugh PJ, et al. A core gut microbiome in obese and lean twins. Nature. 2009;457:480–484. [PMC free article] [PubMed]
41. Bardonnet N, Blanco C. uidA-antibiotic-resistance cassettes for insertion mutagenesis, gene fusions and genetic constructions. FEMS Microbiol Lett. 1992;72:243–247. [PubMed]
42. Leedle JA, Hespell RB. Differential carbohydrate media and anaerobic replica plating techniques in delineating carbohydrate-utilizing subgroups in rumen bacterial populations. Appl Environ Microbiol. 1980;39:709–719. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...