Abstract
Lactobacillus casei is a lactic acid bacterium commonly found in the gastrointestinal tract of animals, and some strains are used as probiotics. The ability of probiotic strains to survive the passage through the gastrointestinal tract is considered a key factor for their probiotic action. Therefore, tolerance to bile salts is a desirable feature for probiotic strains. In this study we have characterized the response of L. casei BL23 to bile by a transcriptomic and proteomic approach. The analysis revealed that exposure to bile induced changes in the abundance of 52 proteins and the transcript levels of 67 genes. The observed changes affected genes and proteins involved in the stress response, fatty acid and cell wall biosynthesis, metabolism of carbohydrates, transport of peptides, coenzyme levels, membrane H+-ATPase, and a number of uncharacterized genes and proteins. These data provide new insights into the mechanisms that enable L. casei BL23 to cope with bile stress.
-
The macroarray results discussed in this paper have been deposited in the NCBI Gene Expression Omnibus (GEO) database under series number GSE33047.
-
A supplementary figure and two supplementary tables are available with the online version of this paper.
-
Edited by: P. W. O’Toole
Introduction
Lactobacillus casei is a facultative heterofermentative lactic acid bacterium of interest to the food industry as a starter culture for milk fermentation and for maturation of some types of cheeses. Furthermore, some strains of L. casei have received considerable attention for their beneficial health effects as probiotics (de Vrese & Schrezenmeir, 2008). Probiotic micro-organisms are currently the focus of an intense research effort that aims to determine their possible health benefits and to identify the mechanisms through which they exert them (Oelschlaeger, 2010; Sánchez et al., 2010). Although there is evidence showing that dead probiotic cells can also confer some beneficial effects upon the host (Adams, 2010), it is generally agreed that probiotic micro-organisms must survive the transit through the gastrointestinal tract, where they will encounter a very acid environment in the stomach and a high concentration of bile salts in the upper small intestine (Corcoran et al., 2008).
Bile salts are one of the major components of bile. They are amphipathic molecules that play an important role in the emulsification of fats and absorption of hydrophobic vitamins. In addition, bile salts have antimicrobial activity against many micro-organisms, mainly by damaging their cell envelopes (Begley et al., 2005; Kurdi et al., 2006; Margolles & Yokota, 2011). Furthermore, several studies have indicated that bile salts can also damage DNA, since exposure to bile salts induces DNA repair systems, and strains defective in several DNA repair genes are more sensitive to bile than their parental strains (Begley et al., 2005; Margolles & Yokota, 2011; Merritt & Donaldson, 2009). Due to their amphipathic nature, bile salts may alter the conformation of proteins and they may also cause oxidative stress (Begley et al., 2005; Margolles & Yokota, 2011). Therefore, bile can act on several targets in the bacterial cell, and the defence mechanisms elicited by bacteria are likewise diverse (Begley et al., 2005; Margolles & Yokota, 2011; Merritt & Donaldson, 2009).
A number of studies have dealt with the response of lactobacilli to bile exposure (Bron et al., 2004; Burns et al., 2010; Hamon et al., 2011, 2012; Koskenniemi et al., 2011; Lee et al., 2008; Pfeiler et al., 2007; Pfeiler & Klaenhammer, 2009; Whitehead et al., 2008; Wu et al., 2010). These studies have spanned several species of lactobacilli and have detected considerable variability in the response, although it must be kept in mind that experimental conditions varied from one study to another. Furthermore, a comparative study conducted with six L. casei strains has shown significant differences between strains (Hamon et al., 2012). Therefore, extrapolation of results from other strains is of limited use in understanding the bile response of a certain strain. Notwithstanding this, some effects of bile on gene expression or protein content are usually observed. Induction of general stress proteins and a number of transport systems has been observed in most studies. In contrast, repression of proteins involved in the fatty acid biosynthetic pathway has been observed in Lactobacillus delbrueckii (Burns et al., 2010) and Lactobacillus rhamnosus (Koskenniemi et al., 2011), whereas it was not observed in Lactobacillus acidophilus (Pfeiler et al., 2007), Lactobacillus plantarum (Bron et al., 2004; Hamon et al., 2011) or Lactobacillus reuteri (Lee et al., 2008; Whitehead et al., 2008).
In this study we have used a transcriptomic and proteomic approach to understand the response of L. casei BL23 to bile. Earlier studies on the bile response of L. casei strains have not included strain BL23 and used exclusively a proteomic approach (Hamon et al., 2012; Wu et al., 2010). The study of strain BL23 is of special interest, since it has been used as a model strain for physiological studies (Monedero et al., 2007) and for its probiotic properties (Bäuerl et al., 2010). Our results show that the response of L. casei BL23 shares characteristics in common with other lactobacilli and displays others specific to this strain.
Methods
Bacterial strain, growth conditions, sample collection and microscopic observation.
L. casei BL23 was routinely grown in deMan Rogosa Sharpe (MRS) broth (Oxoid) at 37 °C without shaking. When required, agar was added at 1.8 % for plates. Cells were stored at −80 °C in MRS medium supplemented with 15 % (v/v) glycerol. For the assay, cells from the stock culture were inoculated on MRS agar plates. Single colonies were used to inoculate three 5 ml aliquots of MRS medium and the cultures were incubated overnight at 37 °C. Each culture was used to inoculate two 250 ml batches of pre-warmed MRS medium at OD595 0.05 for the treatment or control assays. Incubation was continued to OD595 0.5, after which 250 ml pre-warmed MRS with or without 0.4 % bovine bile (Sigma) was added so that a final 0.2 % concentration of bile was reached. Incubation was continued for 45 min. At this point, three 10 ml aliquots from each batch were taken for RNA isolation and the remaining culture was used for protein isolation. Samples for RNA isolation were centrifuged (5000 g, 10 min, 4 °C) and washed with 1 vol. 50 mM EDTA, pH 8.0, and the cell pellets were stored at −80 °C until use. For protein purification, cells were harvested by centrifugation (5000 g, 10 min, 4 °C) and washed twice with 1 vol. PN buffer (20 mM sodium phosphate buffer, pH 7.5, 140 mM NaCl), and the pellets were stored at −80 °C until use.
For the estimation of the average cell size, cells grown as outlined above were removed from cultures in MRS or MRS with 0.2 % bovine bile, washed with 0.9 % (w/v) NaCl, deposited on a microscope slide and air-dried. The preparations were subjected to negative staining with nigrosine and observed in a Nikon Eclipse 90i optical microscope at ×1500 magnification. Cell sizes were determined using the tools implemented in the NIS-Elements software (Nikon).
RNA isolation and analysis.
Total RNA was isolated from L. casei as described previously (Zúñiga et al., 2002). RNA samples were treated with the Ambion Turbo DNA-free kit (Applied Biosystems) using the routine DNase I treatment outlined by the supplier in order to remove contaminating DNA. The quality and concentration of the RNA samples were subsequently evaluated by using the Experion automated electrophoresis system (Bio-Rad). Samples with 23S : 16S ratios higher than 0.85 were used for subsequent analyses.
Transcriptomic analysis.
For transcriptomic analysis, first-strand cDNA was synthesized and radiolabelled by random priming using the Transcriptor First Strand cDNA Synthesis kit (Roche). The reaction mixture contained 45 µg sample RNA, 40 U RNase OUT, 24 µM random hexamers, dNTPs (16 mM dATP, dTTP and dGTP; 0.1 mM dCTP), 105 c.p.m. of [α-33P]dCTP (1.1×1014 Bq mmol−1; Hartmann Analytic) and 20 U reverse transcriptase. The retrotranscription reaction was performed as recommended by the manufacturer. The synthesized cDNA was purified using Illustra MicroSpin G50 columns (GE healthcare), quantified in a TriCarb 2800TR scintillation counter (Perkin Elmer) and stored at −20 °C until use.
A DNA macroarray based on the genome sequence of L. casei ATCC 334 (Makarova et al., 2006) was designed containing probes for 2390 out of 2906 annotated genes at the Sección de Chips de DNA-Servei Central de Suport a la Investigació Experimental (SCSIE) (University of Valencia). Details of the macroarray design and printing can be found at the Gene Expression Omnibus (GEO) database (platform GPL11074) and Alberola et al. (2004). Membranes were prehybridized in 0.5 M sodium phosphate (pH 7.0), 1 mM EDTA, 7 % SDS for 1 h at 65 °C. Hybridizations were performed in the same buffer containing 106 c.p.m. labelled cDNA at 65 °C for 24 h. After hybridization, the membranes were washed in 1× SSC, 0.1 % SDS for 10 min at room temperature and subsequently washed twice in 0.5× SSC, 0.1 % SDS for 30 min at 65 °C. The membranes were then transferred to sealable plastic bags and exposed for 10 days to an imaging plate (BAS-MP, FujiFilm).
Hybridization signals were detected at 50 µm resolution using an FLA-3000 fluorescent image analyser (Fujifilm). The image data obtained were imported into the software program ArrayVision 7.0 (Imaging Research) for spot detection and quantification of hybridization signals. Local backgrounds calculated from empty spots were subtracted to obtain raw signal intensities. Data were imported into ArrayStat 1.4 (Imaging Research), and raw cDNA hybridization signals were normalized between replicates by the median procedure and subsequently between conditions. The mean and sd were then calculated for each group. By using the mean signal intensity values for each strain, the logged (base 2) bile-treated : control group ratios were calculated. Statistical significance of ratios was assessed by a z-test implemented in the analysis package.
Reverse-transcription quantitative PCR (RT-qPCR).
RT-qPCR analyses and primer design were carried out essentially as previously described (Landete et al., 2010). Primers used are listed in Supplementary Table S1available with the online version of this paper. The lepA, ileS, pyrG and pcrA sequences were selected from a set of 10 reference genes (Landete et al., 2010) by using the geNorm application (Vandesompele et al., 2002). The relative expression based on the expression ratio between the target genes and reference genes was calculated using the software tool rest (Pfaffl et al., 2002). Linearity and amplification efficiency were determined for each primer pair. Every real-time PCR determination was performed at least six times.
Proteomic analysis.
Frozen pellets were thawed on ice and the cells resuspended in PN buffer to a final concentration of approximately 0.5 g ml−1. Cells were disrupted with 100 µm diameter glass beads (1 g ml−1) in a Mini-BeadBeater (Biospec). Unbroken cells and cell debris were removed by centrifugation (12 000 g, 5 min at 4 °C), and the supernatants were collected and centrifuged again (100 000 g, 60 min at 4 °C) to sediment membranes. The supernatant was collected (soluble fraction) and the membranes were washed with PN buffer, centrifuged again and finally resuspended in PN buffer (membrane fraction). The protein concentration of the cell-free fractions was measured using the Bradford Microassay (Bio-Rad). 2D gel electrophoresis and analysis were performed as described previously (Rivas-Sendra et al., 2011). Identification of selected spots was performed at the Proteomics Core Facility of the Centro de Investigación Príncipe Felipe or the Proteomic Unit of the Centro Nacional de Investigaciones Cardiovasculares (CNIC; Madrid, Spain) as described previously (Rivas-Sendra et al., 2011).
Results
Growth of L. casei BL23 in the presence of bile salts
The ability of L. casei BL23 to grow in the presence of bile salts was analysed in a previous study (Alcántara et al., 2011). In the present work, L. casei BL23 was exposed to a bile challenge at 0.2 %, since this concentration resulted in a moderate but significant decrease in growth rate (0.17 h−1 vs 0.28 h−1 for untreated cells; Fig. 1), indicating that this concentration affected growth, although it was not deleterious for this strain. Exposure to bile also led to a decrease in the mean cell size. After 1 h of bile exposure, mean cell size was 0.95 µm (sd 0.38), whereas the mean cell size for cells growing in MRS was 1.53 µm (sd 0.68). A significant difference of the means (P<0.0001) was determined by a two-tailed Student’s t test (n = 395) with Welch’s correction for unequal variances.
Growth of L. casei BL23 after twofold dilution with MRS or MRS supplemented with 0.4 % bile. The arrow indicates the point of dilution.
Proteomics of the L. casei BL23 response to bile
We compared the proteome of the L. casei BL23 cells with and without bile treatment. Since bile particularly affects the cell membrane, membrane and soluble fractions were separately analysed. Although integral membrane proteins cannot be detected by conventional 2D gel electrophoresis, proteins anchored to the membrane or associated with it can be resolved. Figs 2 and 3 show 2D gels of membrane and soluble protein extracts of L. casei BL23 grown in the absence or in the presence of bile salts. Comparison of protein patterns obtained under the two conditions revealed that 52 spots displayed different levels of expression in the conditions tested (34 in the membrane fraction and 18 in the soluble fraction; Table S2a, b). Among these proteins, nine proteins of the soluble fraction were significantly more abundant and nine were less abundant in bile-treated cells (Table S2b). In the membrane fraction, 23 proteins were significantly more abundant, whereas 11 were less abundant (Table S2a).
Silver-stained 2D electrophoresis gels of cell membrane fraction proteins extracted from L. casei BL23 cells untreated (a) and treated with 0.2 % bile (b). The figure shows one representative gel for each sample. Spot numbers indicate differentially expressed proteins.
Silver-stained 2D electrophoresis gels of cell soluble fraction proteins extracted from L. casei BL23 cells untreated (a) and treated with 0.2 % bile (b). The figure shows one representative gel for each sample. Spot numbers indicate differentially expressed proteins.
Ten spots in the soluble fraction and 29 spots in the membrane fraction could be identified by MS; the remaining spots did not render reliable mass spectra (Table S2a, b, and Tables 1 and 2). In some cases, isoforms of the same protein were identified: for example, three isoforms of the glucose phosphoenolpyruvate : phosphotransferase (PTS) transport subunit EIIAB (Yebra et al., 2006) (Table S2a). The comparison of the proteins identified in the soluble and membrane fractions showed that only one protein, glyceraldehyde-3-phosphate dehydrogenase, was detected in both fractions, thus indicating that two different subsets of the L. casei proteome were obtained by the fractionation procedure used in this study. Identified proteins were putatively involved in a wide variety of cell functions (Tables 1 and 2), thus indicating that bile induced global changes in L. casei BL23 physiology.
Symbols and abbreviations are as indicated in Table 1.
Transcriptional response to bile salts
In order to obtain a more comprehensive view of the L. casei BL23 response to bile, a DNA macroarray analysis was also carried out. The results obtained showed that errors increased as signal values decreased (Fig. S1). So, two restrictive criteria were introduced after the initial analysis of the results with ArrayStat in order to filter out low-confidence hits: first, only genes with ratios higher than 2.5 or lower than 0.4 were considered; second, the normalized signal had to be higher than 60 in at least one dataset. In this way, 67 genes were found to have significant expression changes out of 229 detected in the initial analysis. Obviously, under these restrictive criteria a number of genes actually responding to bile were missed, but those identified could be reliably considered to be involved in the bile response. Out of the total of 67 genes, 27 were upregulated, whereas 40 were downregulated (Tables 1 and 2).
Upregulated genes were classified in a variety of functional categories (Table 1), including four genes encoding chaperones and proteins involved in protein turnover, the stress-responsive transcriptional regulator CtsR and the glycinebetaine/lysine transporter GbuA. Furthermore, a number of transporter-encoding genes were induced in the presence of bile. Finally eight genes without a functional prediction were also induced. Among downregulated genes were genes involved in the fatty acid biosynthetic pathway, genes encoding an oligopeptide transport system and a putative glutamine ABC transporter (Table 2).
Validation of macroarray results by RT-qPCR
In order to assess the reliability of the results obtained by transcriptomic analysis, 16 genes were selected and changes in expression were determined by RT-qPCR. The results obtained showed a good correlation between the results obtained in the macroarray analysis and RT-qPCR (Table 3). Nine out of 16 tested genes showed similar fold changes in the macroarray and the RT-qPCR analysis. No significant differences in expression level were observed in the macroarray analysis for genes LCABL_08070, LCABL_15130 and LCABL_16980; however, RT-qPCR detected moderate but significant differences in the expression level of these genes, whereas for genes LCABL_08370, LCABL_16360, LCABL_16750 and LCABL_19410, although their fold changes were statistically significant, the values were below 1.5 (Table 3). Therefore, the macroarray analysis missed some genes responding to bile (false negatives), but we conclude on the basis of these results that the incidence of false positives is low.
Discussion
The combination of transcriptomic and proteomic analyses allows a more comprehensive view to be obtained of the response of L. casei BL23 to bile. Our results show that the bile response is a complex process that involves significant physiological changes in this bacterium. Earlier studies on the response of L. casei strains to bile made use only of a proteomic approach (Hamon et al., 2012; Wu et al., 2010). Furthermore, those studies focused on the differences observed between cells grown in the presence or absence of bile, whereas our experimental design aimed to determine the response of L. casei BL23 to a bile challenge. Therefore, differences in the results obtained are to be expected, especially considering the differences between strains observed by Hamon et al. (2012). The same argument applies to studies conducted with other species. For example, bile stress affects exopolysaccharide and bile salt hydrolase production in the closely related strain L. rhamnosus GG (Koskenniemi et al., 2011), but L. casei BL23 does not produce either exopolysaccharides or bile salt hydrolases (Mazé et al., 2010). For this reason, the Discussion will be mainly focused on coincidences with results obtained by other researchers rather than differences. In the following sections, genes and proteins differentially expressed and corresponding to different functional categories will be discussed.
Stress response
Exposure to bile induced a number of proteases and chaperones involved in the stress response (Table 1). Activation of stress proteins in response to bile has been observed in L. casei Zhang (Wu et al., 2010) and its close relative L. rhamnosus GG (Koskenniemi et al., 2011). In contrast, a clear activation of stress proteins was not observed by Hamon et al. (2012). A possible explanation for this difference is that Hamon et al. (2012) focused their analysis on a subset of proteins in order to identify biomarkers of bile tolerance. Upregulation of stress proteins in response to bile has also been observed in L. acidophilus (Pfeiler et al., 2007), L. delbrueckii subsp. lactis (Burns et al., 2010), L. plantarum (Hamon et al., 2011) and L. reuteri (Whitehead et al., 2008).
The activation of several proteolytic systems and chaperones is in agreement with the expected effect of protein denaturation of bile. Clp proteases play indispensable roles in cellular protein quality control systems (Frees et al., 2007). The role of HslUV in low-GC Gram-positive bacteria is poorly understood, but its role in misfolded protein degradation is well established in Escherichia coli (Sundar et al., 2010).
Two transcriptional regulators possibly involved in the stress response were upregulated in response to bile (Table 1). The transcriptional regulator CtsR controls the expression of ClpP in L. plantarum in response to various abiotic stress conditions (Fiocco et al., 2010). An increased abundance of CtsR and ClpP in response to p-coumaric acid has been observed in L. casei (Rivas-Sendra et al., 2011). In this study, a search of the genome sequence of L. casei BL23 with the CtsR binding site consensus sequence (Derré et al., 1999) revealed putative CtsR binding sites upstream of clpP (LCABL_10770), clpB (LCABL_15770) and clpE (LCABL_19810). In our transcriptomic analysis, clpE was upregulated (Table 1) and clpP was also detected as upregulated, although the ratio value (1.98) was below the established threshold, whereas no significant variation was detected for clpB.
As far as we know, the role of Rex in lactobacilli is unknown. However, Rex has been shown to be involved in the oxidative stress response in the related organism Streptococcus mutans (Bitoun et al., 2011). Rex might also play a similar role in lactobacilli.
Cell envelope-related functions
A strong increase in abundance of the subunit δ of the F0F1-ATPase in the membrane fraction was observed after exposure to bile (Table 1). However, at the transcriptional level, no significant differences were observed in any of the genes encoding the subunits of the ATPase. Wu et al. (2010) did not observe this effect, whereas Hamon et al. (2012) observed modest increments in the abundance of subunit δ in some L. casei strains. However, it cannot be ascertained whether this is due to different experimental designs or methodologies. Interestingly, Arikado et al. (1999) showed that the level of ATPase in Enterococcus hirae was regulated by pH at the assembly level and not at the transcriptional level. When the medium pH was lowered, the level of membrane-bound ATPase increased. This effect was explained by the assembly of ATPase subunits present in the cytoplasm in response to low pH. An increase of ATPase subunit α in the cell envelope fraction was observed in L. rhamnosus GG without any significant increment in the total proteome or transcriptome (Koskenniemi et al., 2011). These results suggest that the same regulatory mechanism of ATPase activity might operate in L. casei BL23 and L. rhamnosus GG. An increase in ATPase activity might be envisaged as a response to counteract the effect of dissipation of the proton motive force reported for bile acids in other lactic acid bacteria (Kurdi et al., 2006; Margolles & Yokota, 2011), although further evidence is obviously required to ascertain this point.
Two enzymes catalysing opposite reactions, glucosamine-6-phosphate deaminase (NagB) and glucosamine-fructose-6-phosphate aminotransferase (GlmS), were upregulated and downregulated, respectively (Tables 1 and 2). A strong upregulation of NagB and to a lesser extent of N-acetylglucosamine-6-phosphate deacetylase (NagA) has also been observed in L. rhamnosus GG (Koskenniemi et al., 2011), whereas GlmS showed a decrease in expression (although not significant). Koskenniemi et al. (2011) hypothesized that L. rhamnosus GG might utilize N-acetylglucosamine as an energy source under bile stress. Wu et al. also observed an increase in the abundance of NagA and NagB in response to low pH (Wu et al., 2011) and NagA in response to bile (Wu et al., 2010) in L. casei Zhang. In contrast, Hamon et al. (2012) observed lower abundance of NagB in most L. casei strains. In that study, three GlmS isoforms were detected whose abundance varied in response to bile from one strain to another. Again, data available are insufficient to explain this discrepancy. Glucosamine 6-phosphate occupies a key position between the glycolytic and the cell wall biosynthetic pathways. In Staphylococcus aureus, N-acetylglucosamine is almost exclusively used for cell wall synthesis in the presence of glucose, and the presence of this sugar inhibits the activity of GlmS (Komatsuzawa et al., 2004). Later studies have shown that GlmS expression is controlled by a riboswitch responding to the glucosamine 6-phosphate intracellular concentration in many Gram-positive organisms (Winkler et al., 2004). Therefore, the induction of NagB and repression of GlmS in L. casei BL23 strongly suggests that the cell wall biosynthetic activity is reduced in the presence of bile, and glucosamine, either from the growth medium or from cell wall hydrolysis, is delivered to the glycolytic pathway. The lower cell wall biosynthetic activity would also agree with the observed downregulation of phosphomevalonate kinase (Table 2), an enzyme of the biosynthetic pathway for terpenoids such as lipid II and lipid IV involved in peptidoglycan biosynthesis. Downregulation of GlmS and the genes encoding the three enzymes that lead from mevalonate to isopentenyl-pyrophosphate has also been observed in L. acidophilus (Pfeiler et al., 2007).
The microscopic observation of cells exposed to bile also indicates major changes in the cell shape and structural alterations, such as the formation of vesicle-like structures between the cell wall and the cell membrane (Bron et al., 2004; Ruiz et al., 2007; Taranto et al., 2006). The L. casei BL23 cells showed a significant decrease in cell size after exposure to bile. In relation to this, it is worth noting the lower abundance of LCABL_14850 detected in the membrane fraction (Table 2). This gene encodes an MreB homologue, a structural protein that forms filaments that spiral around the periphery of the cell (Jones et al., 2001). MreB interacts with membrane proteins and peptidoglycan synthases in a complex that directs cell wall growth (Young, 2010). A lower abundance of MreB also points to diminished cell wall-building activity.
Lipid metabolism
Most lactobacilli harbour a large gene cluster for fatty acid biosynthesis. In L. casei BL23, after addition of bile, most genes of this cluster were strongly downregulated, including the MarR family transcriptional regulator LCABL_23010 located in the same cluster (Table 2). Some of the proteins involved in fatty acid biosynthesis were also significantly less abundant in the proteomic analysis (Table 2). A similar result has been obtained for L. rhamnosus GG (Koskenniemi et al., 2011). Wu et al. (2010) also detected a significantly lower abundance of FabK in L. casei Zhang. Lower abundance of some fatty acid biosynthetic enzymes was also observed in L. delbrueckii subsp. lactis (Burns et al., 2010), and significantly lower levels of transcripts of some fatty acid biosynthetic enzymes were detected in L. reuteri cells grown with bile (Whitehead et al., 2008). Our results therefore suggest that L. casei and other lactobacilli lower their rates of fatty acid synthesis in response to bile stress. A number of studies have shown that exposure to bile leads to significant changes in the fatty acid composition of the membrane of different bacteria (Ruiz et al., 2007; Taranto et al., 2003, 2006). However, the data available are not sufficient to determine whether L. casei also changes the membrane fatty acid composition in response to bile stress.
Carbohydrate transport and metabolism
Bile also induced changes in carbohydrate transport and metabolism. Glyceraldehyde-3-phosphate dehydrogenase, as well as the genes LCABL_30330 and LCABL_30340, which encode the EIIC and EIIAB subunits of the L. casei BL23 major glucose transport system (Veyrat et al., 1994), were downregulated, suggesting a lower glycolytic flux after exposure to bile. Lower abundance of the EIIAB subunit was observed in most L. casei strains studied by Hamon et al. (2012) and in the surface-exposed proteome of L. rhamnosus GG (Koskenniemi et al., 2011). In addition, no changes in glycolytic enzymes were observed in L. rhamnosus GG. In contrast, a higher abundance of phosphofructokinase and phosphoglycerate mutase was observed in L. casei Zhang (Wu et al., 2010), whereas Hamon et al. (2012) observed changes in the abundance of a number of glycolytic enzymes, although these changes were strain-dependent in many cases. Differences in experimental conditions and sampling may account for these discrepancies.
Coenzyme metabolism
The expression or abundance of a number of proteins controlling the levels of key coenzymes was modified in the presence of bile. Gene LCABL_10160, encoding a putative NAD kinase, was upregulated in the presence of bile (Table 1). This is the only de novo pathway for the biosynthesis of NADP and plays a critical role in the maintenance of the NAD : NADP ratio (Shi et al., 2009). Recent research has shown that NADP(H+) is a key provider of reducing equivalents to maintain or regenerate the cellular detoxifying and antioxidative defence systems (Agledal et al., 2010). Therefore, the increase in NAD kinase expression may relate to oxidative stress caused by bile (Begley et al., 2005).
Adenylate kinase (ADK) is essential to equilibrate the levels of adenosine phosphates, since it catalyses the reversible phosphoryl transfer from ATP to AMP, yielding two ADP molecules. After exposure to bile, the abundance of ADK in the soluble protein fraction was significantly lower, although the transcript levels did not change significantly (Table 2). In contrast, Wu et al. (2010) detected a greater abundance of ADK after exposure to bile in L. casei Zhang. We do not have an explanation for this discrepancy, although differences in experimental design might account for it.
Signal transduction, transcription and translation
Bile also modified the abundance or the expression level of a number of proteins involved in signal transduction, transcriptional regulation or translation. A previous study showed that inactivation of response regulators RR01 (LCABL_02080), RR06 (LCABL_12050) and RR12 (LCABL_19600) resulted in severe growth defects in the presence of 0.5 % bile (Alcántara et al., 2011). In this study, we observed an increased abundance in the membrane fraction of the RR01 cognate histidine kinase LCABL_02090 (Table 1), although no significant differences in expression were detected at the transcriptomic level. Interestingly, an increased abundance of the corresponding response regulator LGG00252 was also detected in the cell surface proteome (surfome) fraction of L. rhamnosus GG (Koskenniemi et al., 2011). In that study, a significantly increased expression of systems LGG_00155–00156 (no orthologues in L. casei BL23), LGG_01003–01004 (orthologous to L. casei BL23 TC06) and LGG_01710–01711 (orthologous to TC10) was also detected. An increase of the expression of TC06 (LCABL_12050–12060) and TC10 (LCABL_18830–18840) components was detected in BL23 also, although below the established threshold (1.55 and 1.43 for TC06, and 1.60 and 1.63 for TC10). Inactivation of these two systems also led to increased sensitivity to bile (Alcántara et al., 2011). Taken together, these results indicate that these two-component systems may play a role in the response to bile of L. casei BL23 and L. rhamnosus GG. The involvement of a two-component system (Lba1430–1431) in the bile response has also been demonstrated in L. acidophilus (Pfeiler et al., 2007), although the L. casei BL23 genome does not encode any orthologous system (Zúñiga et al., 2011).
The expression level or abundance of a number of transcriptional regulators was also affected by bile. Some of them are discussed in other sections. Among them, gene LCABL_00610 was significantly induced in response to bile (Table 1). This gene constitutes an operon together with LCABL_00600 (ratio 1.40), which encodes an uncharacterized membrane protein. The expression of their corresponding orthologues lgg_00069–00068 in L. rhamnosus was also induced by bile (Koskenniemi et al., 2011).
One of the effects of bile stress most commonly observed is an alteration of the levels of ribosomal proteins (Bron et al., 2006; Hamon et al., 2011, 2012; Koskenniemi et al., 2011; Lee et al., 2008; Pfeiler et al., 2007; Whitehead et al., 2008). Similarly, levels of several ribosomal proteins or transcripts were altered by bile in L. casei BL23 (Tables 1 and 2). An increased abundance in the membrane fraction of protein S30AE (LCABL_10410) was observed in response to bile. Protein S30AE plays a role in the regulation of protein synthesis by reducing translation initiation under stress conditions (Vila-Sanjurjo et al., 2004). An increase in expression of protein S30AE has been observed in L. reuteri (Whitehead et al., 2008), whereas in an L. plantarum bile-sensitive strain, a decrease in the abundance of S30AE in response to bile was observed in a whole-cell protein extract (Hamon et al., 2011).
Peptide and amino acid transport
Amino acid and peptide transport systems were also affected by bile. A gene cluster encoding an ABC oligopeptide transport system was clearly downregulated in the presence of bile (Table 2). In addition, the OppD subunit of a second ABC oligopeptide transport system was detected as being significantly less abundant in the proteomic analysis of the membrane fraction (Table 2). The transcriptomic analysis showed lower ratios in all Opp genes, although the differences were significant only for the cluster LCABL_22420–LCABL_22460. Hamon et al. (2012) also observed strain-dependent variations in proteins involved in peptide and amino acid transport and metabolism. Downregulation of Opp-encoding genes has also been observed in L. rhamnosus GG, although differences were not statistically significant (Koskenniemi et al., 2011). The effect of bile on the expression of putative amino acid transporters was mixed (Tables 1 and 2). Unfortunately, transport of amino acids in L. casei has received little attention. Among the induced transporters, GbuA encodes the ATPase subunit of a putative glycine betaine/lysine ABC transporter. The homologous proteins Gbu and OpuA have been shown to be induced under salt stress in Listeria monocytogenes (Duché et al., 2002) and Bacillus subtilis (Petersohn et al., 2001), respectively.
Other transporters
In addition to the cases already discussed, bile induced significant changes in the transport capabilities of L. casei BL23 (Tables 1 and 2). These changes included the downregulation of four uncharacterized transport systems (Table 2), the induction of a putative phosphate transport system and of a number of uncharacterized transporters (Table 1). Among them, it is worth noting the strong induction of the expression of a putative multidrug ABC transporter encoded by genes LCABL_20970 and LCABL_20980, and of a transcriptional regulator present in the same gene cluster (LCABL_20990; Table 1). A strong induction of the homologous gene cluster in L. rhamnosus GG has also been detected (Koskenniemi et al., 2011), indicating that this transport system possibly plays a relevant role in the bile response of these micro-organisms. The ATPase subunits of another three putative multidrug ABC transporters were found to be significantly more abundant in the L. casei BL23 membrane fraction (Table 1). The corresponding homologue of L. rhamnosus GG (LGG_0979) of one of them, LCABL_11790, has also been determined to be significantly upregulated (Koskenniemi et al., 2011). In summary, as observed in other lactobacilli (Koskenniemi et al., 2011; Pfeiler & Klaenhammer, 2009; Whitehead et al., 2008), L. casei BL23 induced the expression of a number of multidrug transporters in response to bile.
Uncharacterized or unknown proteins
A total of 22 genes or proteins of unknown function were detected as differentially expressed in the presence of bile (Tables 1 and 2). Among them, it is worth noting the strong induction of the putative operon constituted by genes LCABL_10560, LCABL_10570 and LCABL_10580. The orthologous genes LGG_00914–LGG_00916 were also strongly upregulated in L. rhamnosus GG in response to bile (Koskenniemi et al., 2011). Unfortunately, there is no indication concerning the functional role of these proteins. Strong induction of LCABL_16090 (Table 1) was also observed. This gene is transcribed in the opposite direction from LCABL_16100, which was also significantly upregulated (Table 1) and is of unknown function as well. LCABL_16090 encodes a protein belonging to the haemolysin III family (PF03006). Members of this family are integral membrane proteins with seven transmembrane domains. No homologue has been characterized in lactic acid bacteria, but homologues in other bacteria possess haemolytic activity (Baida & Kuzmin, 1995), whereas in eukaryotes they have been described as receptor proteins with a wide range of ligand specificities (Tang et al., 2005). LCABL_16100 belongs to the DegV family (Pfam 02645) of unknown function although the resolution of the structure of a homologous protein of Thermotoga maritima suggests that they may have fatty acid-binding activity (Schulze-Gahmen et al., 2003).
Concluding remarks
A combination of proteomic and transcriptomic approaches has been used to characterize the response to bile of L. casei BL23. The analysis of the results shows that bile had a deep effect on L. casei physiology, including cell envelope biosynthesis, carbon metabolism and cell protection. Our results show noticeable differences from those reported for other L. casei strains (Hamon et al., 2012; Wu et al., 2010). There may be a number of reasons for these discrepancies: utilization of different methods, strain variability and different experimental designs. Earlier studies focused on bile-adapted cells, whereas this study characterized the response of L. casei BL23 to a bile challenge. This study establishes a basis for detailed studies of the mechanisms enabling L. casei BL23 to cope with bile stress and to adapt to the intestinal habitat.
Acknowledgements
This work was financed by funds from the Spanish Ministry of Science and Innovation (AGL2007-60975 and Consolider Fun-C-Food CSD2007-00063). We thank Amalia Blasco for technical assistance and Abelardo Margolles for critical reading of this manuscript.