Correspondence: George L. Mendz, School of Biotechnology and Biomolecular Sciences, The University of New South Wales, Sydney, NSW 2052, Australia. Tel.: +61 2 9385 2042; fax: +61 2 9385 1483; e-mail: firstname.lastname@example.org
Campylobacter jejuni is a pathogen that colonizes the intestinal tract of humans and some animals. The in vitro responses of the bacterium to ox-bile were studied using proteomics to understand the molecular mechanisms employed by C. jejuni to survive bile stress. Its in vitro tolerance to bile was determined by growing the bacterium for 18 h in liquid cultures containing different bile concentrations. Significant growth inhibition was observed in the presence of 2.5% bile, and a decrease of 1.12 log units was measured at a bile concentration of 5%. Protein expression profiles of bacteria grown with and without bile were compared using two-dimensional polyacrylamide gel electrophoresis. Proteins with differential intensities greater than two-fold were identified using tandem mass spectrometry. Nuclear magnetic resonance spectroscopy and spectrophotometry were employed to measure enzyme activities in cell extracts from bacteria grown with and without bile. Together with proteins known to be involved in C. jejuni bile tolerance, the presence of bile modulated the expression of proteins such as elongation factors, ferritin, chaperones, ATP synthase and others, previously unknown to be implicated in the response of the bacterium to bile.
Campylobacter jejuni is a Gram-negative enteric human pathogen, and the leading bacterial cause of human gastroenteritis in the developed world (Allos, 2001). This bacterium colonizes the intestinal tract, where it can invade the epithelium. Chickens serve as a natural reservoir for the bacterium, and undercooked poultry represents the primary means for infection (Harris et al., 1986). The bacterium is also present in other animal species, including cattle, without any signs of disease. Notwithstanding this asymptomatic presence in many cases, there have been reports of C. jejuni-mediated enteritis in calves (Prescott & Munroe, 1982) and of bovine abortions (Welsh, 1984; Larson et al., 1992); whether or not it is pathogenic to cattle remains uncertain. Infection with the bacterium has been linked to the debilitating autoimmune system disorders Miller–Fischer syndrome and Guillain–Barré syndrome; the latter can leave patients with severe neurological damage (Ropper, 1986). Despite significant advances in the understanding of Campylobacter pathogenesis factors, the knowledge about C. jejuni infection is still limited (Wassenaar & Blaser, 1999).
Bile is produced in the liver, and stored in the gallbladder. From there it is secreted into the small intestine via the bile duct. Bile acids have antimicrobial activities owing to their amphipathic nature and, as such, represent an environmental stress to which bacteria of the gut flora must adapt. Campylobacter jejuni is exposed to bile as it colonizes and proliferates in the bovine gut, and therefore needs to tolerate the bactericidal effects of bile acids. The response regulator CbrR has been shown to modulate bile resistance and, as such, chicken colonization ability (Raphael et al., 2005). The bacterium uses the multidrug efflux pump CmeABC as a mechanism of bile resistance (Lin et al., 2003), the expression of which is modulated by a transcriptional repressor factor encoded by cmeR (Lin et al., 2005). It has been determined that C. jejuni shows complex interactions with bile and its constituents, for example, chemoattractive behaviour and upregulation of important virulence factors, such as the flagellin FlaA and the Campylobacter invasion antigens Cia (Rivera-Amill et al., 2001; Allen & Griffiths, 2001). Nonetheless, the current knowledge about the molecular responses of C. jejuni to bile remains very limited. Many genes and proteins are known to participate in the interactions of other bacterial species with bile; thus, a more global study of C. jejuni protein expression when exposed to bile will serve to increase our knowledge about the adaptation of this bacterium to the intestine.
This work investigates the global response of C. jejuni to ox-bile, the type commonly used to elucidate bacterial adaptations to bile. The objective of the study was to obtain a more complete view of the role of bile in the infection of cattle by C. jejuni, and thus to contribute to a better understanding of its pathogenicity in these and other hosts. Bacteria were grown in medium supplemented with different bile concentrations to determine the sublethal concentrations at which growth was only partially inhibited. Proteomic analyses served to identify proteins whose expression was modulated by the presence of bile. Proteins differentially expressed in the presence of bile revealed metabolic pathways involved in the adaptation of C. jejuni to bile. The results of this work agree with previous findings and provide new insights into the bile tolerance mechanisms of this bacterium.
Materials and methods
Bacterial strain and growth conditions
Campylobacter jejuni NCTC 11168, isolated from humans, was grown on Blood Agar Base 2 solid medium (Oxoid, Heidelberg West, Vic., Australia) supplemented with 6% defibrinated horse blood (Oxoid) and containing 5 μg mL−1 trimethoprim, 0.32 μg mL−1 polymyxin B, 10 μg mL−1 vancomycin and 2 μg mL−1 Fungizone® (Bristol–Myers Squibb, Noble Park, Vic., Australia). Bacteria were incubated in a Sanyo O2/CO2 Tri-Gas Incubator (Quantum Scientific, Lane Cove West, NSW, Australia) in an atmosphere of 5% O2, 5% CO2 and 90% N2 at 37°C. Liquid cultures were grown in vented flasks using 50-mL cultures of Brain–Heart Infusion Broth (Oxoid) supplemented with ox-bile (Sigma-Aldrich, Castle Hill, NSW, Australia) at concentrations between 0% and 5%. Cultures were tested for purity using phase contrast microscopy.
Preparation of cell-free protein extracts
For proteomic studies, cells were grown in the absence of bile (control culture) or in 2.5% or 5% (w/v) ox-bile. Chloramphenicol was added to bacterial cultures after 18 h of incubation to a final concentration of 128 μg mL−1. Cultures were centrifuged at 2879 g at 4°C for 25 min, and the cell pellets were washed twice with 0.2 M ice-cold sucrose. The pellets were then disrupted by twice freezing and thawing, and resuspended in 1 mL of TSU buffer [50 mM Tris, pH 8.0; 0.1% sodium dodecyl sulphate (SDS); 2.5 M urea]. Cell debris was removed by centrifugation at 14 000 g at 4°C for 20 min. Protein concentrations were determined using the bicinchoninic acid method employing a microtitre protocol (Pierce, Rockford, IL). Absorbances at 595 nm were measured using a Beckman Du 7500 spectrophotometer.
Two-dimensional polyacrylamide gel electrophoresis and image analysis
Typically, a 110-μg protein suspension was adjusted to 490 μL using a rehydration buffer containing 8 M urea, 100 mM dithiothreitol (Astral Scientific, Gymea, NSW, Australia), 65 mM 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulphonate (CHAPS), 40 mM Tris-HCl, pH 8.0, and 10 μL pH 4–7 IPG (immobilized pH gradient) buffer (Amersham Biosciences, Melbourne, Vic., Australia). Nuclease buffer (10 μL) was added, and the mixture was incubated at 4°C for 20 min. The sample was then centrifuged at 14 000 g at 4°C for 20 min, and the supernatant was loaded on to an 18-cm Immobiline DryStrip pH 4–7 (Amersham Biosciences), which was incubated sealed at room temperature for 20 h. pH 7–11 NL Immobiline DryStrips were loaded with 340 μL of Destreak Rehydration Solution (Amersham Biosciences) containing 0.5% pH 7–11 IPG buffer, and incubated at room temperature for 20 h. Protein samples were cup-loaded during isoelectric focusing, which was performed using a flat-bed Multiphor II Unit (Amersham Biosciences). For pH 4–7 strips, the unit was programmed for 2 h at 100 V, followed by 0.5-h steps at 500, 1500 and 2500 V, and a final step of 3500 V for 18 h. For pH 7–11 NL strips, the programme used was 1 h at 100 V, followed by 0.5-h steps at 500, 1500 and 2500 V, and a final step of 3500 V for 12.5 h. Focused Immobiline DryStrips were equilibrated in two sequential buffers containing 6 M urea, 20% (w/w) glycerol, 2% (w/v) SDS and 375 mM Tris-HCl, the first with 130 mM dithiothreitol and the second with 135 mM iodoacetamide (Sigma–Aldrich). SDS-polyacrylamide gel electrophoresis was performed on 11.5% acrylamide gels using the Protean II System (Bio-Rad, Regents Park, NSW, Australia) at 50 V for 1 h, followed by 64 mA for 5 h. Gels were fixed individually in 0.2 L of fixing solution containing 50% (v/v) methanol and 10% (v/v) acetic acid for a minimum of 1 h, and were subsequently stained using a sensitive ammoniacal silver method. Gels were imaged using a UMAX Powerlook 1000 flat-bed scanner (Fujifilm, Brookvale, NSW, Australia). For comparative image analysis, statistical data were acquired and analysed using z3 software (Compugen, Sunnyvale, CA).
Identification of proteins by MS
Excised gel slices with protein spots were washed twice with 0.2 mL of 100 mM NH4HCO3 for 10 min, reduced with 50 μL of 10 mM dithiothreitol at 37°C for 1 h, alkylated in 50 μL of 10 mM iodoacetamide at 37°C for 1 h, washed three times for 10 min with 0.2 mL of Milli-Q water, washed with 0.2 mL of 100 mM NH4HCO3 for 10 min, dehydrated in acetonitrile and rehydrated in a buffer containing 10.5 ng μL−1 trypsin (Promega, Annandale, NSW, Australia). After digestion for 14 h at 37°C, peptides were extracted by washing the gel slice with 25 μL of 1% formic acid for 15 min, followed by dehydration with acetonitrile. Digests were dried in vacuo, resuspended in 10 μL of 1% formic acid, and separated by nano-liquid chromatography (nano-LC) using an Ultimate/Famos/Switchos System (LC Packings, Dionex, Lane Cove, NSW, Australia). Samples (5 μL) were loaded on to a C18 precolumn (Micron; 500 μm × 2 mm) with H2O–CH3CN (98 : 2), 0.1% formic acid (buffer A) at 25 μL min−1. After a 4-min wash, the flow was switched into line with a C18 RP analytical column (PEPMAP, 75 μm × 15 cm) and eluted using buffer A to H2O–CH3CN (40 : 60), 0.1% formic acid at 200 nL min−1 over 30 min. The nano-electrospray needle was positioned c. 1 cm from the orifice of an API QStar Pulsar I tandem mass spectrometer (ABI, Foster City, CA). The QStar was operated in information-dependent acquisition mode. A time-of-flight mass spectrometry (TOF MS) survey scan was acquired (m/z= 350–1700, 0.5 s), and the two largest precursors (counts>10) were selected sequentially by Q1 for tandem mass spectrometry (MS/MS) analysis (m/z=50–2000, 2.5 s). A processing script generated data suitable for submission to database search programs. Collision-induced dissociation (CID) spectra were analysed using the Mascot MS/MS ion search tool (Matrix Science, Boston, MA) with the following parameters: trypsin digestion allowing up to one missed cleavage, oxidation of methionine, peptide tolerance of 0.25 Da and MS/MS tolerance of 0.2 Da. Searches were performed on the National Center for Biotechnology Information (NCBI) nr database, or using the Sanger C. jejuni 11168 sequence FTP database.
Nuclear magnetic resonance (NMR) spectroscopy
Suspensions of bacterial lysates or cell-free extracts were placed in 5- or 10-mm tubes (Wilmad, Buena, NJ), the appropriate substrates were added and enzyme activity measurements were carried out at 37°C. Proton NMR (1H-NMR) spectroscopy free induction decays were collected using a Bruker DMX-600 spectrometer, operating in the pulsed Fourier transform mode with quadrature detection. Phosphorus-31 NMR (31P-NMR) spectroscopy free induction decays were collected using a Bruker DMX-500 spectrometer, operating as above. The instrumental parameters have been described previously (Kaakoush & Mendz, 2004).
The time evolution of substrates and products was followed by acquiring sequential spectra of the reactions. Progress curves were obtained by measuring the integrals of the compounds at different points in time. Maximal rates were calculated from good fits to straight lines (correlation coefficients ≥0.95) of the data for 30–60 min. Calibration of the peaks arising from the substrates was performed by extrapolating the resonance intensity data to zero time and assigning to each intensity the appropriate concentration value.
2-Phosphoglycerate, 3-phosphoglycerate and phosphoenol pyruvate were obtained from Sigma-Aldrich. The reversible reactions of the glycolysis pathway, catalysed by phosphoglycerate mutase and enolase, were initiated by adding 2-phosphoglycerate to C. jejuni lysate suspensions. Fumarate, malate, oxaloacetate and NADH were obtained from Sigma-Aldrich. The tricarboxylic acid (TCA) cycle reactions, catalysed by fumarate reductase, fumarase or malate dehydrogenase, were initiated by adding fumarate, malate or oxaloacetate and NADH, respectively, to the bacterial lysate suspensions.
Spectrophotometric assays and lysate preparation
Succinyl-coenzyme A (succinyl-CoA) activities in C. jejuni cell-free extracts were measured as described previously (Gibson et al., 1967) using a Cary-100 UV–visible spectrophotometer and quartz cuvettes (path length, 1 cm). Succinyl-CoA and ADP were purchased from Sigma-Aldrich. Succinyl-CoA synthetase was measured by adding succinyl-CoA, ADP and inorganic phosphate to C. jejuni cell-free extracts.
Effect of ox-bile on C. jejuni growth and viability
Growth of the bacterium in media containing ox-bile at concentrations in the range 1.25–5% (w/v) was compared with that of the bacterium in bile-free medium. Figure 1 shows the change in the number of viable cells. Bacterial growth rates decreased with increasing bile concentration in the medium, but C. jejuni still proliferated in cultures containing 5% (w/v) bile and, at this concentration, a decrease of only 1.12 log units was measured relative to that of cultures without bile.
Two-dimensional gel analyses
The predicted isoelectric points of most C. jejuni proteins lie between 4 and 11. Therefore, two pH gradients were selected for isoelectric focusing: pH 4–7 and pH 7–11. The second dimension electrophoresis was performed using 11.5% reducing SDS acrylamide gels. In this study, the expression of a protein was considered to have changed significantly if its spot in gels from cultures with and without bile showed a two-fold or greater difference in intensity. Figure 2 shows the two-dimensional gels of C. jejuni protein extracts from cells grown in the absence of bile or with 2.5% (w/v) bile. Forty-eight spots were found to be significantly different, 34 in the pH 4–7 range and 14 in the pH 7–11 range.
CID spectra generated by micro-LC/electrospray ionization-MS/MS were obtained for 15 protein spots. All identifications were made using the C. jejuni genome sequence database with mascot software. The number of peptide matches for the identified proteins ranged from 1–16. Three of the protein spots selected for MS/MS analysis contained more than one protein, i.e. spots 2, 7 and 15. The theoretical pI values for most of the proteins identified closely matched those observed on the gels, with the exception of spot 14, which showed a difference of 0.88 pH units (observed pI, 6.22; predicted pI, 5.34). This discrepancy may be a result of posttranslational modifications, such as phosphorylation.
Functional classification of identified proteins
Proteins identified as differentially expressed at 2.5% (w/v) bile were grouped according to function (Table 1). Flagellin A was found in three spots, suggesting that it probably undergoes various posttranslational modification events. Flagellin A and flagellin B were upregulated in the presence of 2.5% (w/v) bile. This result was the same at 5% (w/v) bile (data not shown).
Table 1. Campylobacter jejuni proteins differentially expressed in the presence of 2.5% ox-bile, identified by mass spectrometric analyses
Transcriptional factors were regulated by the presence or absence of bile. At 2.5% (w/v) bile, the expression of elongation factor G (EF-G) was upregulated. In the case of EF-P, bile also appeared to influence posttranslational differences when compared with cells grown in the absence of bile. Similarly, at 5% (w/v) bile, EF-TU appeared to undergo significant modifications, and showed shifts in both molecular weight and pI range depending on the cell growth conditions (data not shown).
Two putative periplasmic proteins were upregulated in the presence of 2.5% (w/v) bile, one of which contained an ATP/GTP-binding site motif. An outer membrane protein found to be upregulated at 2.5% (w/v) bile contained a bacterial surface antigen domain. At 5% (w/v) bile, the major outer membrane protein showed increased expression.
Two of the proteins identified as differentially expressed were involved in iron storage. Ferritin was upregulated in the presence of 2.5% (w/v) bile, whereas bacterioferritin was downregulated under the same conditions.
An increase in expression of the F1 sector ɛ-subunit of ATP synthase and downregulation of UTP-glucose-1-phosphate uridyltransferase (GalU) were observed at 2.5% (w/v) bile. The glycolytic enzyme enolase and the TCA cycle enzymes succinyl-CoA synthetase (β-chain) and malate dehydrogenase were downregulated at 5% (w/v) bile.
The heat shock proteins GroEL were downregulated by 2.5% (w/v) bile, together with CheV, a chemotaxis protein.
To verify the downregulation of enolase expression observed in bacteria grown in the presence of 5% bile, the activities of this and another glycolytic enzyme were measured in lysates from cells grown with or without bile. 31P-NMR spectroscopy was employed to determine the activities of enolase and phosphoglycerate mutase in lysates of cells grown in 2.5% and 5% (w/v) bile. Figure 3 shows the time evolution of the conversion of 2-phosphoglycerate to 3-phosphoglycerate and phosphoenol pyruvate by phosphoglycerate mutase and enolase, respectively. The data indicated a decrease in the enzyme rates in bacterial lysates from cells grown in the presence of bile relative to the activities in cell lysates of control cultures grown without bile.
The activities of the TCA cycle enzymes fumarate reductase, fumarase and malate dehydrogenase were examined using 1H-NMR spectroscopy. The results showed decreases in the activities of all three enzymes in cells grown in the presence of bile relative to the controls. Spectrophotometric measurements showed that the activity of succinyl-CoA synthetase was lower in cell-free extracts of cultures grown in the presence of bile.
Microorganisms that inhabit the gastrointestinal tract are exposed to various toxic factors, one of which is bile salts. These potent antimicrobials can disrupt cell membranes, and different bacterial species have developed varying levels of resistance against bile salts. In this study, C. jejuni was shown to possess a level of tolerance for bovine bile above 0.3% (w/v), which is considered to be the threshold of resistance to bile salts (Andrews, 1998). At ox-bile concentrations of up to 5% (w/v), the bacterium was still capable of growing at 0.74 log units in 18 h, although this represented a 20-fold decrease relative to the growth of the control culture with no bile.
Novel proteins involved in the response of C. jejuni strain 11168 to bile were identified, together with proteins previously implicated in bile tolerance (Allen & Griffiths, 2001). Flagellin proteins have been studied in detail in Campylobacter, and have been shown to be an important virulence factor of the bacterium (Golden & Acheson, 2002; Konkel et al., 2004; Song et al., 2004). Both flagellin A and flagellin B were upregulated by 2.5% (w/v) bile, with similar results found for flagellin A in the presence of 5% (w/v) bile.
Iron is an essential nutrient for all living organisms, and a lack of iron delays bacterial growth, which could prove detrimental to the survival of C. jejuni in vivo. Bacteria have two classes of iron storage protein: ferritin and bacterioferritin; the latter differs from the former in that it has a protohaem in addition to the nonhaem iron core. Ferritin is coded by the cft gene in C. jejuni, and Cft serves as an intracellular iron deposit able to sequester several thousand iron atoms in its central core. Iron acquisition represents a major determinant in the development of a pathogen within its host (Ratledge & Dover, 2000). In its gastrointestinal niche, the bacterium must compete for nutrients, such as iron, with a vast array of other microorganisms. A higher expression of ferritin on exposure to bile may give the bacterium an advantage for sequestering iron from the gastrointestinal tract, and ensure that C. jejuni can outcompete other microorganisms for successful colonization. Recent studies have found the expression of ferritin to be regulated by iron (Palyada et al., 2004; Holmes et al., 2005), again indicating that the bacterium may upregulate ferritin production to maximize its utilization of any available iron. Thus, bile may act as an environmental signal for the bacterium to increase its iron storage capacity.
By contrast, unused iron can be toxic for microorganisms owing to its participation in the formation of hydroxyl radicals. Campylobacter jejuni cft-negative mutants are more sensitive than the wild type to the superoxide stress inducer paraquat and the peroxide stress inducer H2O2 (Wai et al., 1996). Accordingly, bacteria require a fine regulation of iron homeostasis, and excess iron needs to be stored in an inactive form. The genome of C. jejuni strain 11168 contains the gene cj1534c which encodes a possible bacterioferritin. Downregulation of the expression of this protein on exposure to bile seems to indicate a shift in the iron storage mechanism of C. jejuni from bacterioferritin to ferritin.
The enzyme GalU is responsible for the production of UDP-glucose, which is involved in the synthesis of different surface structures. The galU gene is a major virulence factor in many bacterial pathogens, and galU mutants have disrupted bacterial surfaces (Chang et al., 1996; Rioux et al., 1999; Nesper et al., 2001; Bonofiglio et al., 2005). Such disruptions can involve the capsule, lipopolysaccharides, the core oligosaccharide and/or outer membrane proteins. Vibrio cholerae galU mutants show a much higher sensitivity than wild-type strains to bile, and also demonstrate significantly reduced ability to colonize the mouse small intestine (Nesper et al., 2001). GalU of C. jejuni is downregulated in the presence of 2.5% bile. If C. jejuni mutants demonstrate a similar increase in sensitivity to bile as is the case with V. cholerae, the component of ox-bile responsible for the downregulation of this protein expression may be exploited to hinder C. jejuni pathogenicity.
Increased expression of the heat shock proteins GroEL in the presence of bile has been shown in other organisms (Rince et al., 2003; Sanchez et al., 2005). In this study, GroEL proteins were identified in spots 5 and 15; the intensity of the former was decreased in the presence of 2.5% (w/v) bile, and the intensity of the latter was increased (Table 1). Flagellin A was identified in spot 15, together with GroEL, and from the differential expression of flagellin A determined in other spots, it was found to be upregulated by bile. Thus, it is possible that the increased protein expression measured in spot 15 may have stemmed from flagellin A only and not from GroEL. Alternatively, 2.5% (w/v) bile may have caused a modification of the heat shock proteins, inducing a shift in their position in the gels relative to their position in gels for cells grown in the absence of bile.
In addition to their known roles in translation, transcriptional elongation factors (EF) have chaperone roles, including protein folding and protection from stress (Caldas et al., 2000, 1998). In this study, the expression levels of various EF were affected by the presence of bile, with 5% (w/v) bile causing apparent modifications of EF-TU structure, and 2.5% (w/v) bile causing changes in EF-P. Bile was shown to regulate the expression of transcriptional factors in Bifidobacterium longum when the bacterium was exposed to bile (Sanchez et al., 2005).
The outer membrane protein Cj0129c and two putative periplasmic proteins Cj0092 and Cj0998c were upregulated in cells exposed to bile. The functions of these proteins are currently unknown; however, their upregulation suggests that they may play a role in enhancing the efflux of bile from C. jejuni.
The expression of the chemotaxis protein CheV was downregulated by 2.5% (w/v) bile. The exact role of this protein is not currently known (Marchant et al., 2002); however, the observed change in its expression could be related to the chemoattraction of C. jejuni to bile, even though bile acids by themselves have a repellent effect on the bacterium (Hugdahl et al., 1988). The upregulation of ATP synthase may contribute to the energy needs of enhanced flagellar motion required for chemotaxis. Similarly, an increase in activity of the multidrug efflux pump CmeABC in order to tolerate bile salt stress may require that extra ATP be generated.
The activities of two glycolysis and four TCA cycle enzymes decreased in bacteria grown in the presence of bile. These enzymes are involved in the production of energy and biosynthesis precursors in the cells, and the decrease in their activity was reflected in the inhibition of growth measured in the presence of bile.
This work presents the first insight into changes in the global protein expression of C. jejuni in response to bile exposure. The expression and/or activities of 19 proteins were modulated by the presence of bile. It provides the first proteomic profile of the bacterium in relation to an intestinal stress, and identifies proteins belonging to a variety of functional pathways previously unknown to be involved in the bile adaptation of C. jejuni. It also suggests new avenues for physiological studies on the interaction of C. jejuni with bile.
This work was made possible by the support of the Australian Research Council.