B.H. Lee, Department of Food Science and Agricultural Chemistry, McGill University, Montreal, Quebec, Canada H9X 3V9. E-mail: email@example.com
Aims: To clone, sequence and characterize a new bile salt hydrolase from a bile tolerant strain of Bifidobacterium animalis ssp. lactis KL612, and further analysis of the bsh promoter and an operon-like structure containing the bsh gene in the genus Bifidobacterium.
Methods and Results: A new type of bile salt hydrolase from a bile tolerant strain of Bifidobacterium was cloned, completely sequenced and characterized. The putative bsh promoter sequence was analysed by primer extension to determine the transcriptional start point by applying the genomic walking-PCR, an operon-like structure containing the bsh gene and two more open reading frames located within a complete set ranging from a promoter to a transcription terminator sequence is reported for the first time in the genus Bifidobacterium. The polycistronic bsh transcript was revealed by reverse transcriptase-PCR (RT-PCR) as well as by Northern hybridization.
Conclusions: Most of bile tolerant strains of bifidobacteria showed a similar genetic organization around the bsh gene. This finding suggests that bile tolerance of those strains is possibly because of the bile salt hydrolase and some transporter proteins, which are functionally related to each other to respond efficiently to the stress from bile salts.
Significance and Impact of the Study: Knowledge gained through BSH research would provide further insight into the survival of probiotics in the gastrointestinal tract and some physiological functions of this enzyme in relation to the host as well as the enzyme-producing bacteria.
The amphipathic nature of bile acids enables them to play an important role in the emulsification of dietary lipids by increasing the contact between lipases and lipid substrates. In addition to their function as natural emulsifiers in the intestine, bile salts exhibit some detergent-like antimicrobial properties. Gastrointestinal (GI) microbiota and enteric pathogens have developed mechanisms to resist antimicrobial activity of bile and this trait has been used for years in clinical and research environments to selectively enrich bacterial samples for enteric organisms (e.g. MacConkey agar, Salmonella-Shigella agar). Several mechanisms evolved by micro-organisms to counteract bile acid toxicity have been proposed (Gunn 2000; Begley et al. 2005). In general, micro-organisms can resist toxic compounds by various responses, which are activated upon exposure to stress (Barthlmebs et al. 2000). Most of the time, detoxification involves either efflux pump of the toxic compound from the cell by highly specific systems (George 1996; Alekshun and Levy 1999) or enzymatic conversion of the toxic compound into a less toxic form (Pereira and Gibson 2002).
Bile salt hydrolase (BSH, EC 184.108.40.206), which is an enzyme that catalyses the hydrolysis of glycine- and/or taurine-conjugated bile salts into amino acid residues and free bile acids, is widely spread not only in many enteric commensals but also in enteric pathogens. BSH activity in pathogenic strains was reported as a novel virulence factor that contributes to bacterial persistence in the host and consequently to dissemination (Dussurget et al. 2002; Shankar et al. 2002). On the contrary, BSH activity in commensals and probiotic strains was proposed as a detoxification mechanism (De Smet et al. 1995; De Boever and Verstraete 1999), which is an important feature for their survival during host residence. Tolerance to the deleterious actions of bile is necessary for commensal and pathogenic micro-organisms to survive in the human GI tract. In probiotic research, bile tolerance is considered of primary importance in the selection of strains (Salminen et al. 1998) because it is believed that bile tolerance enables the bacteria to survive its transit along the duodenum and subsequently to grow and colonize the gut epithelia by adhesion to enterocytes (Kociubinski et al. 1999). Considering the wide distribution and high activity of the BSH in bifidobacteria (Grill et al. 1995; Tanaka et al. 1999, 2000) and the hydrolytic products of the BSH are more toxic to the bacterial cells than the substrates, bile tolerant strains of bifidobacteria appear to have another defense mechanism to protect themselves from the toxic nature of the deconjugated bile acids. Tannock (2000) proposed that there are some differences in the efficiency of cell membrane-associated transport mechanisms between strains that can be used to maintain intracellular levels of cholic acid at nontoxic concentration. In our previous work (Kim et al. 2004a), it has been revealed that many strains of Bifidobacterium from dairy products possessed the same type of BSH enzyme and they exhibited bile tolerance in spite of high BSH activity. To improve our understanding of bile salt metabolism and bile tolerance in bifidobacteria, we have screened bifidobacteria strains that possess high BSH activity as well as bile tolerance.
In this study, we describe the molecular cloning, sequencing and characterization of a bile salt hydrolase from a bile tolerant strain of Bifidobacterium. The putative bsh promoter sequence was analysed by primer extension to determine the transcriptional start point. An operon-like structure containing the bsh gene and two more open reading frames (ORFs), located within a complete set ranging from a promoter to a transcription terminator sequence is reported in this study for the first time in the genus Bifidobacterium.
Materials and methods
Bacterial strains, plasmids and growth conditions
The bacterial strains and plasmids used in this study are listed in Table 1. The species of bifidobacterial strains were verified by 16S rRNA gene sequence typing and its 16S–23S interspace region according to the method of Ventura et al. (2004b). Bifidobacteria were propagated anaerobically at 37°C in deMan, Rogosa, Sharpe broth (Difco, Detroit, MI, USA) supplemented with 0·05% (w/v) cysteine HCl (MRS-Cys). Escherichia coli cells were propagated at 37°C in Luria–Bertani (LB) broth with vigorous shaking or on LB medium solidified with 1·5% agar. When appropriate, ampicillin (200 μg ml−1) and kanamycin (40 μg ml−1) were added. Tolerance of the strains against bile was tested according to Pereira and Gibson (2002). Briefly, overnight cultures were inoculated (1%) into MRS-Cys broth without and with 0·3% oxgall (Difco) and incubated at 37°C. The number of viable bifidobacteria was determined at 0 h and after 24 h of incubation on MRS agar plates.
Table 1. Bacterial strains and plasmids used in this study
pUC19 with 1·3-kb B. infantis KL412 chromosomal insert; BSH+
pUC19 with 4·2-kb B. infantis KL412 chromosomal insert; BSH+
pET36b(+) with 950-bp NdeI & HindIII insert from pDR095; BSH+
BSH activity in cell-free extracts (CFEs) was measured by the hydrolysis of sodium taurocholate (TC) and/or sodium glycocholate (GC) (Sigma, St Louis, MO, USA) at 37°C in sodium phosphate buffer (0·1 mol l−1, pH 6·5). The amounts of the amino acids released from conjugated bile salts were measured by the ninhydrin assay (Kim et al. 2004a). One unit of BSH activity was defined as the amount of enzyme that liberated 1 μmol of amino acids from the substrate per minute. Specific activity was defined as units per milligram of protein. Protein concentrations were determined by the Bio-Rad protein assay using bovine serum albumin as a standard.
DNA isolation, manipulation and transformation
Genomic DNA of Bifidobacterium was isolated according to the method of Kim et al. (2004b). Routine DNA manipulations were performed according to standard procedures as described elsewhere (Sambrook et al. 1989). T4 DNA ligase and other DNA-modifying enzymes were purchased from New England Biolabs Inc. or Invitrogen Life Technologiesand used according to the manufacturers’ specifications. Electroporation was performed with a Gene-Pulser II electroporation apparatus (Bio-Rad) according to the manufacturer’s specifications.
Construction and screening of Bifidobacterium lactis BSH genomic library
The isolated chromosomal DNA from Bifidobacterium lactis KL 612 was partially digested with the restriction enzyme EcoRI and analysed by DNA electrophoresis. DNA fragments ranging from 2·0 to 6·0 kb were isolated using QIAquick Gel Extraction kit (Qiagen) and then ligated to the pUC19 vector that had been restricted with EcoRI and dephosphorylated. The ligation mixture was transformed into E. coli DH5α competent cells to create a plasmid library. The transformed E. coli was plated on LBGCT medium (LB agar plates containing 1% glucose, 0·035% calcium chloride and 3 mmol l−1 taurodeoxycholic acid) and BSH-positive clones were detected based on the formation of deoxycholate precipitate around the colony (Coleman and Hudson 1995).
DNA sequencing and sequence analysis
The nucleotide sequences were determined by AmpliTaq FS DNA polymerase fluorescent dye terminator reactions using an Applied Biosystems 373 stretch automated sequencer (Applied Biosystems Inc., Foster City, CA, USA). ORF prediction was carried out by the National Center for Biotechnology Information ORF finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). The predicted ORFs were translated into putative proteins that were submitted to BlastP analysis and protein homology searches were performed with the Blast at the website of the NCBI (http://www.ncbi.nlm.nih.gov). The transmembrane spanning domain and topology of the membrane protein were predicted using the tmap program (Persson and Argos 1994, 1996). The secondary structure of the putative transcription termination sequence was predicted using the Mfold web server (Zuker 2003) and the free energy of this structure was determined using the web server (http://www.bioinfo.rpi.edu/applications/mfold/old/rna/energy/form1.cgi) (Mathews et al. 1999).
Extension of the genomic DNA sequence
Genomic walking-PCR was used to amplify and extend the genomic DNA sequence 5′ of the cloned fragment (Fig. 1). A nested set of four sequence-specific primers (SSPs), GW1, GW2, GW3 and GW4, to the 5′ end of the clone was designed (Table 2). Genomic walking-PCR was performed on genomic DNA with nested SSPs and degenerate random tagging primers (DRTs) along with the APAgene Genomic Walking Kit (Bio S&T, Montreal, Quebec, Canada). Three consecutive PCR amplifications were carried out with a genomic DNA template from B. lactis KL612. Each successive PCR used a different nested SSP, the same DRT primer and the template from the previous reaction according to the supplier’s instruction. The DNA amplicon from the third reaction was purified (PCR purification kit; Qiagen) and sequenced using GW4 as a sequencing primer. The sequence was reconfirmed by directly amplifying genomic DNA with GW1 and the primer UP1 designed on the basis of the upstream end of the new sequence.
Table 2. Oligonucleotide primers used in this study
Nucleotide sequence (5′-3′)*
*Engineered NdeI and HindIII sites are underlined. Changed nucleotide sequences for the site-directed mutagenesis are shaded.
† Based on the ATG start codon as a starting point (#1).
−1205 to −1185
−550 to −530
Northern blot analysis
Northern blot analysis
Northern blot analysis
Northern blot analysis
−550 to −530
Primer extension sequencing
RNA was isolated from B. lactis with the RNeasy kit (Qiagen), starting with digestion of the bacterial cell wall in DEPC-treated TE buffer containing 15 mg lysozyme and 1 kU mutanolysin per ml for 15 min at 37°C. The purified RNA was treated with RNase-free DNase according to the manufacturer’s instruction (Qiagen) and resuspended in RNase-free water. The absence of any significant DNA contamination of the RNA was confirmed by PCR. The integrity of purified RNA was confirmed by a formaldehyde agarose gel electrophoresis according to the protocol described previously (Sambrook et al. 1989). After separation on the gel, 16S rRNA and 23 rRNA bands were visualized with SYRB Gold (Molecular Probes Inc., Eugene, OR, USA). The RNA was quantified by spectrophotometric measurement of OD at 260 nm and stored at −80°C.
PCR, multiplex PCR and RT-PCR
The PCRs were performed using genomic DNA or plasmid DNA from BSH positive clones as templates for amplifying the target genes. When appropriate, restriction sites were designed in the 5′ end of the primers to facilitate future cloning steps. Template DNA and primers were added to 50 μl of PCR mixture containing 200 μmol l−1 of each dNTP, PCR buffer and 2·5 units of HotStarTaq DNA polymerase (Qiagen). The PCR was conducted in a Perkin Elmer GeneAmp under conditions described by Kim et al. (2004b).
For the simultaneous amplification of three loci in the bsh operon, a multiplex PCR was performed with a Multiplex PCR kit (Qiagen) according to the supplier’s instruction.
For the RT-PCR, the first strand cDNA was synthesized by using the Invitrogen RT-PCR kit as recommended by the manufacturer by using either primer F or primer H (Table 2 and Fig. 3). cDNA products were subsequently amplified by PCR using primers internal to genes of interest. PCR products were analysed by electrophoresis in agarose gels containing EtBr (1 μg ml−1) and, if necessary, purified using the QIAquick PCR purification kit (Qiagen).
Identification of the bsh promoter sequence
Promoter sequences were predicted by using a Neural Network Promoter Prediction program that is available on the Internet (http://www.fruitfly.org/seq_tools/promoter.html). To verify the predicted promoter sequence and to determine the transcription start point of the mRNA, primer extension analysis was performed under conditions described by Kim et al. (2004b) using the reverse primers PEA-1 and PEA-2 (Table 2). The same primers and a PCR product corresponding to the region from the bsh gene, containing the putative transcriptional start point, were used to generate sequencing ladders by dideoxychain termination method with a T7 Sequencing kit (USB Corp., Cleveland, OH, USA) according to the manufacturer’s description. The resultant gel was dried and exposed to BioMax MS Film (Eastman Kodak, Rochester, NY, USA).
RNA (20 μg) was separated on formaldehyde agarose gels prepared, and after the transfer, ribonucleic acids were cross-linked to the nylon membrane by UV irradiation. For the Northern hybridization, two probes (probe BSH and probe 2&3) were amplified with two pairs of primers, BIF-F and BIF-R, G and H, respectively (Table 2). PCR products were purified by agarose gel electrophoresis and subsequently extracted with the QIAquick kit (Qiagen). Nick translation of both strands of purified probes was carried out with [α-32P]-dCTP (Amersham) in a Klenow reaction and random primers (NEB). DNA–RNA hybridization was performed as described by Miyamoto et al. (1985). The sizes of the transcripts were estimated by direct comparison with a molecular RNA ladder (Invitrogen).
Nucleotide sequence accession number
The sequence of the 3170-bp DNA fragment containing the bsh operon has been deposited in the GenBank database under accession number AY530821.
Construction of a gene bank for B. lactis KL612 and isolation of the bsh gene
Among many strains of bifidobacteria, B. lactis KL612 was selected to be the subject of the cloning experiments because this strain exhibited the high level of BSH activity as well as in vitro bile salt tolerance property (Table 3).
Table 3. Bile tolerance and BSH activity of the Bifidobacterium strains tested
Bile tolerance (Delta log G*
*Delta log G = [log (CFU ml−1) in MRS broth]−[log (CFU ml−1) in MRS broth + 0·3% oxgall]; cell growth was measured as CFU ml−1 after 24 h incubation at 37°C.
†No bacterial growth was observed in the presence of 0·3% oxgall.
‡Specific activity (U mg−1) of the bile salt hydrolase measured with cell free extracts from each strains; one unit of enzyme activity is defined as the amount of enzyme that liberated 1 μmol of amino acids from sodium glycocholate per minute.
B. longum KL 515
B. bifidum ATCC 11863
B. lactis KL 612
Bifidobacterium KL 701
Bifidobacterium KL 714
Bifidobacterium KL 725
To identify the gene(s) encoding the BSH enzyme, an EcoRI-digested genomic library for B. lactis KL612 was created in pUC19 and screened in E. coli DH5α cells. Two positive clones were identified by the formation of white precipitation around the colonies on the LBGCT medium supplemented with 200 μg ml−1 ampicillin, and were designated pBSH13 and pBSH42. Preliminary restriction enzyme analysis revealed that pBSH13 contained a 1·3-kb insert and pBSH42, a 4·2-kb insert (data not shown).
Sequencing and extension of the 4146-bp insert in pBSH42
Sequence analysis of the 1·3-kb insert revealed one complete ORF encoding a 314-amino-acid protein. The N-terminal amino acid sequence of the deduced protein was identical to that of the previously purified BSH enzyme (Kim et al. 2004a). The deduced protein had a theoretical molecular mass of 35 037 Da and a pI of 4·70, which were in good agreement with the measured biochemical data obtained from the purified enzyme. Sequencing analysis of the 4·2-kb insert revealed that 5′-end of the sequence exactly matched that of the 1·3-kb insert. In both inserts, the EcoRI site at the 5′-end was identified 72 nucleotides upstream of the ATG start codon of the ORF. To identify the promoter sequence and to better characterize the genetic organization, upstream DNA sequence was extended. A DNA product of approximately 1·2kb was amplified by genomic walking (Fig. 1). Once it was sequenced by using GW-4 as the primer, the amplified fragment extended the known genomic sequence 1133 nucleotides 5′ of the EcoRI-cloned DNA.
Analysis of ORFs and homology studies with the BSH
In all, a total of 5279 bp was sequenced from plasmid pBSH42 (4146 bp) and the amplified fragment (1133 bp) by genomic walking-PCR to reveal the presence of three complete and two partial ORFs (Figs 1 and 2). Comparison of the predicted amino acid sequence of the first complete ORF to the database by Blast analysis revealed the highest similarities to the BSHs of Bifidobacterium adolescentis, Bifidobacterium bifidum and Bifidobacterium longum (53%, 52% and 50% identity, respectively) (Tanaka et al. 2000; Kim et al. 2004b, 2005). High similarities were also found to the BSH enzymes of several Lactobacillus strains (31–42% identity) (GenBank accession numbers AF054971, AF305888, AF091248, AF297873, M96175) as well as Enterococcus strains (31–33%) (GenBank accession numbers AY032999, AY260046), Listeria monocytogenes (31%) (Dussurget et al. 2002), Clostridium perfringens (30%) (Coleman and Hudson 1995) and penicillin V amidase (PVA) of Bacillus sphaericus (23%) (GenBank accession number M15660).
A second ORF, designated ORF-2, begins on the same strand, 121 nucleotides downstream of bsh, with an ATG start codon at nucleotide 2271, and ends with a TGA codon at nucleotide 3044 (Fig. 2).
Blast analysis of the 257 amino acids encoded by ORF-2 revealed similarities to a number of 4-carboxymuconolactone decarboxylase. The strongest of these similarities was to the gammacarboxymuconolactone decarboxylase subunit of Methanosarcina barkeri (36% identity and 56% similarity over 244 residues) (GenBank accession number NP_615377).
The third complete ORF encoding a protein of 354 amino acids, is located 51 nucleotides downstream of the ORF-2 on the same strand and is designated ORF-3. The predicted amino acid sequence encoded by ORF-3 showed high similarities to the transporter proteins. The highest similarities were to the putative transport proteins of Streptomyces avermitilis MA-4680 (35% identity and 59% similarity over 313 residues) (GenBank accession number NP_826229; Omura et al. 2001) and of Streptomyces coelicolor A3(2) (34% identity over 319 residues) (GenBank accession number NP_627246; Redenbach et al. 1996). The predicted protein encoded by ORF-3 was identified as a membrane protein anchored through a series of transmembrane segments, which are composed of several hydrophobic residues. Five transmembrane segments were detected by using a TMAP program (data not shown), and the topology of this transmembrane protein was predicted as Nout (the N-terminal portion of the protein is likely to be located in the extracellular space) (Persson and Argos 1996).
Two incomplete ORFs were also identified on the fragment, both in the opposite orientation from that of bsh (Fig. 1). The first, ORF-4, contains 335 codons for the C-terminal end of a protein whose amino acid sequence shows the highest similarities to the ATP binding protein of ABC transporter of B. longum NCC2705 (92% identity and 94% similarity over 296 residues) (GenBank accession number NP_695873; Schell et al. 2002).
The ORF5, located upstream of bsh, begins with an ATG start codon at nucleotide 996 and contains 331 codons. The deduced amino acid sequence shows the highest similarities to the conserved hypothetical protein of B. longum NCC2705 (58% identity and 74% similarity over 322 residues) (GenBank accession number NP_695963; Schell et al. 2002).
Characterization of the bsh transcript
The bsh transcription start point was determined by primer extension using total RNA isolated from log-phase B. lactis KL612 cells. A putative bsh promoter (Pbsh) sequence was identified upstream of the mRNA 5′ end at nucleotides 1159/1160 (Fig. 2), which is in accordance with the promoter sequence predicted by NNPP (Fig. 3). Potential −35 (TCGACA) and −10 (TAGAAT) regions of the Pbsh, which share 5 bp each with the consensus hexamers, are separated by 19 bp, and the −10 sequence precedes the dual mRNA start site at a distance of 5 bp. A putative rho-independent type transcription terminator sequence (ΔG = −35·2 Kcal mol−1 at 25°C) was recognized 15 nucleotides downstream of the stop codon of the ORF-3 (Fig. 2), suggesting that the bsh gene could be transcriptionally coupled with ORF-2 and ORF-3.
To examine if the bsh gene is transcribed polycistronically, RT-PCR was carried out with several primer pairs that probed four different regions (Fig. 4). Primers B and E were used for an internal control, primers D and F were used to detect transcripts running from bsh to ORF-2, primers G and H for transcripts running from ORF-2 to ORF-3, and primers A and C were used to examine whether mRNA extended upstream beyond the possible transcription start point of bsh. First-strand cDNA was synthesized both with primers F and H, and the similar results were obtained with both cDNAs (Fig. 4b,c). In addition to the positive DNA control, clear signals were obtained with primers D and F, and primers G and H. However, no RT-PCR product was obtained with primers A and C. This indicated that three genes were present on the same transcript, but the transcript did not extend upstream of bsh. The opposite orientation of ORF-5 to the bsh gene (Fig. 4) and the presence of another promoter sequence (Fig. 2) directly upstream of ORF-5, which was also predicted by using NNPP, demonstrated that this gene is most likely transcribed independently from the bsh transcript as a function of its own promoter.
In Northern hybridizations, two bands of 3·1 kb and 1·1 kb with probe BSH, and two bands of 3·1 kb and 2·0 kb with probe 2&3 were observed (Fig. 5). A small stem-loop-like structure within the intergenic spacer downstream of bsh may stop readthrough transcription, resulting in single transcription of bsh (or in cotranscription with ORF2 and ORF3).
Screening of Bifidobacterium strains for the bsh locus
To examine the distribution of the bsh locus, genomic DNA from each strain was screened using standard PCR with two primers, BSH-F and BSH-R. Four strains of dairy origin produced the same size (950 bp) of amplicon (data not shown). To simultaneously obtain the PCR products for three genes encoding BSH, ORF-2 and ORF-3, a multiplex PCR was performed with genomic DNA isolated from Bifidobacterium strains (Table 3). As shown in Fig. 6, all the strains showing bile tolerance property produced the same patterns of amplicons (1479 bp, 1071 bp and 771 bp), suggesting that those strains possessed the same genetic organization around the bsh locus.
In this study, we report the cloning and transcriptional analysis of a new bile salt hydrolase gene from B. lactis KL612. In our previous reports, two bsh genes have been cloned and characterized, including the bsh genes from B. adolescentis ATCC 15705 (Kim et al. 2005) and B. bifidum ATCC 11863 (Kim et al. 2004b). This study presents another type of bsh gene from a bile tolerant strain of Bifidobacterium isolated from a dairy product.
Two bsh positive clones obtained from screening of the EcoRI-cloned genomic library contained only 72 nucleotides of the 5′-end flanking region, which was not long enough to include all the components of the transcription machinery. By applying the genomic walking-PCR technique to the 5′-end of the positive clone, we were able to amplify one fragment of 1·2 kb that was identical in sequence to genomic DNA from the upstream end of the EcoRI-cloned DNA. As the original method was developed for screening yeast artificial chromosome (Liu and Whittier 1995), this semi-nested PCR technique has also been used for the recovery of DNA fragments adjacent to known sequences of genes from humans and plants (Hahn et al. 1996; Nakayama et al. 2000, 2001). The versatile applicability of this technique has been proposed and attempts have been made to apply this technique to the bacterial artificial chromosome (Liu and Huang 1998). A similar approach was made to amplify and extend the genomic DNA sequence 3′ of the bsh positive clone from Lactobacillus johnsonii 100-100 (Elkins and Savage 1998).
From the homology comparison of the deduced protein to previously known sequences, B. lactis BSH shared the highest amino acid sequence similarity with the BSH from B. bifidum ATCC 11863 and B. longum strains (Schell et al. 2002). The predicted ORF of 945 bp displayed 63% and 62% DNA sequence identity, and 66% and 64% amino acid sequence similarity to BSHs of B. bifidum ATCC 11863 and B. longum NCC 2705, respectively. However, this sequence homology is of much lower value than that of observed between B. bifidum and B. longum BSHs.
Partial ORFs identified flanking the bsh operon of B. lactis showed high level of sequence similarities to those found in other Bifidobacterium genomes; ORF 4 shared 94% similarity over 296 residues to NBD of ABC transporter and ORF 5 shared 74% similarity over 322 residues to a conserved protein with unknown function, both of which were reported from B. longum NCC 2705 genome. However, the genetic organization of the B. lactis bsh gene and the surrounding region were not conserved when compared with those of B. bifidum (Kim et al. 2004b) and B. longum (Tanaka et al. 2000). In particular, ORF-2 and ORF-3 identified downstream the bsh gene did not display the significant sequence similarity to any known genes of Bifidobacterium genome, indicating that they are either specific to B. lactis or yet to be determined in other genomes. This is the first report on a complete BSH operon in the genus Bifidobacterium in which three ORFs are arranged unidirectional possibly to be controlled by a Pbsh and a transcription termination sequence.
Using Neural Network Promoter Prediction (NNPP, version 2·2; http://www.fruitfly.org/seq_tools/promoter.html), a putative bsh promoter sequence was predicted with a higher score value than that of B. bifidum bsh (Kim et al. 2004b). This predicted promoter sequence showed similarities to the consensus promoter sequences with regards to the −35 (TTGACA) and −10 (TATAAT) consensus hexamers. However, the TG motif upstream of the −10 hexamer was absent in this promoter sequence, which was the case in the B. bifidum bsh promoter (Kim et al. 2004b). The spacer (19 nt) between −35 and −10 region was located between the spacer (20 nt) observed in the B. bifidum bsh promoter (Kim et al. 2004b) and the spacers of the consensus promoter sequences, which were reported as 17 ± 1 nt (Harley and Reynolds 1987; McCracken and Timms 1999). For the primer extension analysis, two primers, PEA-1 and PEA-2 (Table 2), targeting different positions of the 5′-end of the bsh gene, were used with the bacterial RNA. Two different extension products were obtained with the primer PEA-1, corresponding to 5′-ends at the adjacent T and G positions (Fig. 3).
The same T position was revealed from the primer extension with the primer PEA-2, that was in agreement with the position predicted by NNPP (data not shown). The putative promoter sequence contains a region of dyad symmetry centred on position −21 (−8·6 kcal mol−1) relative to the bsh TSP. Interestingly, a palindromic sequence was also identified in the region surrounding −35 box of the bsh promoters from B. bifidum (Kim et al. 2004b) as well as from L. monocytogenes (Dussurget et al. 2002). Recent studies (Ventura et al. 2004a,b; Ventura et al. 2005) revealed that inverted repeat (IR) sequences are commonly detected in the promoter regions of some stress related genes in the genus Bifidobacterium, including the GroEL and GroES chaperones (Ventura et al. 2004a,b) and the dnaK operon (Ventura et al. 2005). It has been proposed that IR sequences in the promoter region could serve as an operator site for the transcriptional regulators in other bacterial groups (den Hengst et al. 2005). However, no definitive consensus sequence can be deduced from the promoter sequences characterized so far from the genus Bifidobacterium (Rossi et al. 2000; MacConaill et al. 2003; Kim et al. 2004b; Ventura et al. 2004a,b; Ventura et al. 2005). Further studies on the transcriptional analysis would be of great importance in gaining a better understanding of the RNA polymerase recognition sites and eventually the transcriptional machinery in the genus Bifidobacterium.
The terminator sequence, found on 15 nucleotides downstream of the stop codon of the ORF-3, could be a putative rho-independent transcription terminator sequence. As this sequence was identified on both DNA strands, it might function as a bidirectional terminator.
Downstream of bsh, ORF-2 and ORF-3 were found on the same strand. They showed high similarities to 4-carboxymuconolactone decarboxylases and the putative transport proteins, respectively. In the genome sequences reported so far, L. monocytogenes EGD-e genome contains a similar genetic organization around the bsh gene to that of B. lactis. The L. monocytogenes bsh gene (lmo0446) is located upstream of a glutamate decarboxylase (gadA) and an antiporter (gadE) (Wemekamp-Kamphuis et al. 2004), and this decarboxylase (GAD) acid resistance system was reported as an important feature in survival and adaptation of L. monocytogenes in acidic conditions. However, their functional relationship in this genetic organization has not been elucidated yet with regard to the bile salt resistance. Even though the functionality of ORF-2 and ORF-3 has not been elucidated in this study, it can be speculated that they are not related with catabolic metabolisms of bile salt but rather they might have some functionalities as transporters based on the fact that Bifidobacterium strains are unable to chemically modify deconjugated bile acid molecules (Takahashi and Morotomi 1994; Kurdi et al. 2003).
RT-PCR experiments have shown that the three genes are transcribed together (Fig. 4), indicating that B. lactis bsh could be a part of an operon. Northern blotting demonstrated that the proposed bsh operon is transcribed as a single polycistronic message. This was confirmed by the 3·1 kb transcript with the probes BSH and 2&3 (Fig. 5), which matched the size of the bsh operon. A signal from a transcript of approximately 1·2kb obtained using the probe BSH, corresponding the bsh gene, was possibly because of the presence of an imperfect inverted repeat (Fig. 2), capable of forming a hairpin structure with a predicted free energy (ΔG) of −35·1 kcal mol−1 and providing an internal terminator sequence. A 2·0 kb signal with the probe 2&3 was too small to be the bsh operon, but rather could be another transcript encoding ORF-2 and ORF-3 or possibly a degradation product.
The combination of bsh and two more genes in one operon is rather surprising, as there is no obvious functional relationship among the three proteins. Further studies are required to investigate how these three proteins are functionally related with the transport mechanisms of bile salts and/or any products produced by the enzymatic reaction of BSH and whether this genetic organization is really responsible for the bile salt resistance property in the Bifidobacterium strains that have the same organization around the bsh gene.
Although we were not able to investigate the functionality of the BSH operon with regards to the bile tolerance property at the molecular level because of a paucity of genetic tools and relatively low transformation efficiency in the genus Bifidobacterium, we observed that most of the bile tolerant strains of bifidobacteria investigated in this study show a similar genetic organization around the bsh gene. This finding suggests that the bile tolerance of those strains is possibly because of the genetic organization around the bile salt hydrolase and some transporter proteins, which are functionally related to each other to respond efficiently to the stress from bile salts. However, further study in the future will be carried out to establish the validity of this assumption.
This work was supported by Natural Sciences and Engineering Research Council of Canada (NSERC-222026) and the ‘GRRC’ Project of Gyeonggi Provincial Government, Republic of Korea. We thank Ms Carol Miyamoto and Dr Edward Meighen of the Department of Biochemistry, McGill University for their contribution in some of the experiments.