| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Institute of Infections and Immunity, Queen's Medical Centre, C-Floor, West Block, Nottingham NG7 2UH, UK
2 School of Pharmaceutical Sciences, Nottingham University, Nottingham NG7 2RD, UK
Correspondence
Kim R. Hardie
kim.hardie{at}nottingham.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Different strategies have evolved to achieve effective quorum sensing. All comprise accumulation of signalling molecules in the extracellular milieu which modulates gene expression via cognate receptors and regulators (reviewed by Bassler, 1999
; Swift et al., 2001; Withers et al., 2001). These processes are used to control a wide variety of phenotypes in many bacterial species including production of extracellular virulence factors by Pseudomonas aeruginosa (Winzer & Williams, 2001) and Staphylococcus aureus (McDowell et al., 2001; Novick & Muir, 1999), and bioluminescence by Vibrio fischeri (Nealson et al., 1970) and Vibrio harveyi (Bassler et al., 1994).
Many Gram-negative bacteria utilize N-acylhomoserine lactone (AHL) molecules as signals, whilst Gram-positive bacteria actively export peptide signalling molecules (Winzer & Williams, 2001; Withers et al., 2001). Recently the gene luxS, possessed by both Gram-positive and Gram-negative bacteria, was identified as a component of a quorum-sensing mechanism (Surette et al., 1999). LuxS is required for the production of a signalling molecule termed autoinducer-2 (AI-2). V. harveyi responds to the presence of AI-2 by producing bioluminescence via a phosphorylation cascade involving the periplasmic sensor LuxP, the inner-membrane protein LuxQ, the cytoplasmic signal integrator protein LuxU, and the response regulator LuxO (Bassler, 1999). This arrangement is mirrored in Vibrio cholerae (Miller et al., 2002). In the only other bacterium where the mechanism of response to AI-2 has been studied, Salmonella typhimurium, AI-2 uptake appears to require an ATP-binding cassette (ABC) transporter, Lsr (Taga et al., 2001), which shows similarity to ribose uptake systems.
LuxS has been shown to convert S-ribosylhomocysteine to homocysteine and AI-2 in vitro (Schauder et al., 2001; Winzer et al., 2002a). Recent crystallographic analysis of four LuxS homologues (derived from Bacillus subtilis, Haemophilus influenzae, Deinococcus radiodurans and Helicobacter pylori) has shown that it is able to form a dimer with a Zn-binding active site compatible with this catalytic function (Hilgers & Ludwig, 2001; Lewis et al., 2001). The proposed reaction was previously shown to yield 4,5-dihydroxy-2,3-pentanedione as part of the cell's central metabolism (Duerre & Miller, 1966; Duerre et al., 1971; Miller & Duerre, 1968), but this molecule is thought to cyclize spontaneously to form a furanone. Comparison of different furanones for AI-2 activity suggests that 4-hydroxy-5-methylfuranone (MHF) is the most active (Schauder et al., 2001; Winzer et al., 2002a), and this molecule has been identified by mass spectroscopy following LuxS-dependent in vitro synthesis of AI-2 and methanol extraction (Winzer et al., 2002a). The active chemical structure of AI-2 is however suggested to be a furanosyl borate diester, since this structure maps the electron-density images generated from X-ray crystallographic analysis of the periplasmic AI-2 sensor protein of V. harveyi, LuxP, following co-crystallization with AI-2 (Chen et al., 2002). Despite uncertainty concerning LuxS function and AI-2 structure, detection of extracellular AI-2 activity has revealed that a number of bacterial species produce functionally equivalent substances capable of inducing the production of bioluminescence by the V. harveyi biosensor, which are thus likely to share a conserved chemical composition. The presence of LuxS and/or AI-2 appears to influence bioluminescence in V. harveyi (Bassler et al., 1994), levels of an ABC transporter in Salmonella typhimurium (Taga et al., 2001), type III secretion in EHEC (Sperandio et al., 1999), the virulence factor VirB in Shigella flexneri (Day & Maurelli, 2001), protease production by Porphyromonas gingivalis (Burgess et al., 2002; Chung et al., 2001) and Streptococcus pyogenes (Lyon et al., 2001), in vivo fitness of Neisseria meningitidis (Winzer et al., 2002c), iron acquisition by Actinobacillus actinomycetemcomitans (Fong et al., 2001), and multiple, but moderate, effects upon the transcription of a number of genes in Escherichia coli (DeLisa et al., 2001). Clear phenotypes dependent upon the presence of LuxS have however remained elusive for some bacteria, e.g. H. pylori (Joyce et al., 2000; Forsyth & Cover, 2000) and Proteus mirabilis (Schneider, 2002).
Currently there is much discussion in the literature concerning the primary role of LuxS in different bacteria given that it has been demonstrated to act as a component of a quorum-sensing mechanism in V. harveyi to control bioluminescence, whilst having a distinct role within central metabolism as part of the activated methyl cycle to recycle S-adenosylhomocysteine which would otherwise have toxic effects on the cell (Bassler 2002; Winzer et al., 2002a,b). To date, few of the published investigations pertaining to the role for LuxS have taken account of the latter, and consequently the basis of the influence of LuxS on cell physiology is unclear.
We set out to investigate the link between the presence of catalytically active LuxS and measurable extracellular AI-2 activity. Since there is apparent conservation of both LuxS structure and AI-2 chemistry, we took advantage of the LuxS overproduction system we have developed. This system relies on the production in E. coli of high levels of soluble LuxS derived from the curved, Gram-negative bacterium H. pylori, which colonizes the gastric epithelium of humans, causing the development of peptic ulcer disease and gastric adenocarcinoma (Blaser et al., 1995; Cover & Blaser, 1996; Mobley, 1996). In agreement with recent crystallographic investigations we demonstrate that LuxS is dimeric in vivo, and show that in the E. coli background it directs the production of AI-2 influenced by carbohydrate availability. The carbohydrate dependence however merely affects extracellular AI-2 activity, since no detectable alteration in the ability of LuxS to synthesize AI-2 in vitro was observed. We also conclude that E. coli must possess a mechanism for degradation or uptake of AI-2. Taken together this work adds weight to the arguments of Winzer et al. (2002b), which propose that the primary role of LuxS is in central metabolism rather than quorum sensing (Bassler, 2002).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and were grown in Luria–Bertani (LB) medium or on LB agar plates at 37 °C. Where required, carbenicillin (50 µg ml-1) or choramphenicol (50 µg ml-1) were added. Construction of plasmid pProEX-luxSEc is described by Winzer et al. (2002a). V. harveyi biosensor strains were grown in AB medium (Greenberg et al., 1979) at 30 °C. Where indicated glucose (0·4 %, w/v), lactose (0·6 %, w/v), maltose (2 %, w/v), potassium acetate (0·2 %, w/v), sodium citrate (0·2 %, w/v), glycerol (1 %, w/v), galactose (0·36 %, w/v), arabinose (0·4 %, w/v), pyruvic acid (0·2 %, w/v) or ribose (0·4 %, w/v) were added.
|
). Restriction enzymes (Promega UK) were used according to the manufacturer's instructions. For isolation of plasmid DNA from E. coli the Qiagen Mini and Midi kits were used. Genomic DNA was extracted from H. pylori according to Atherton (1997). Standard methods were used for preparation of competent cells, and for the electroporation of plasmids into E. coli (Sambrook et al., 1989).
Following PCR amplification (5 min 95 °C; 30 cycles of 30 s 95 °C/30 s 55 °C/1 min 72 °C; 5 min 72 °C) of genomic DNA from H. pylori strain 26695 with primers luxSF (cgcatgAAGCTTaaaccaatcaaacccc) and luxSR (gctaatGAATTCgcatccctaaaacgc), the ORF HP0105 plus 150 bp 5' flanking DNA was cloned into pBluescript (creating pKH4) and the lower copy number pHG327 (creating pKH5). Clones were confirmed by restriction digestion and sequencing. Automated non-radioactive sequencing reactions were carried out using the BigDye terminator cycle sequencing kit in conjunction with a 373A automated sequencer (Perkin Elmer Applied Biosystems). Plasmid pcrp encoding high levels of Crp was a gift from Steve Busby (Bell et al., 1990).
Antibody generation.
E. coli strain DH5
(pKH4) was grown in LB+glucose overnight and cells were harvested into SDS-PAGE sample buffer. Following separation through 15 % SDS-PAGE, LuxS was excised from the gel and electroeluted into 50 mM ammonium bicarbonate/0·1 % (w/v) SDS. Rabbit polyclonal antibodies were raised according to a protocol based on Harlow & Lane (1988). New Zealand White rabbits were immunized subcutaneously biweekly with between 50 and 400 µg mixed 50 : 50 with Freund's complete adjuvant on the first immunization and subsequently with Freund's incomplete adjuvant. A test bleed was carried out following the third immunization. A final fourth immunization was carried out before finally obtaining complete bleeds from the rabbits. Absorption of non-specific antibodies was carried out using a lysate of E. coli DH5(pBluescript) at 37 °C for 60 min.
SDS-PAGE and Western blotting.
These were performed as described by Hardie et al. (1996), except phosphate-buffered saline with 0·5 % (v/v) Tween 20 (PBST) replaced TBST. The primary antibody, Anti-LuxSHp, was used at dilutions of 1 : 2000. Western blots were developed using the enhanced chemiluminescence kit (ECL, Amersham) according to the manufacturer's instructions. SDS-polyacrylamide gels were either stained with 0·1 % (w/v) Coomassie blue/45 % (v/v) methanol/9 % (v/v) acetic acid and destained in 20 % (v/v) methanol/7 % (v/v) acetic acid, or stained with silver stain (Wray et al., 1981).
Analysis of AI-2 production and degradation.
AI-2 production was essentially analysed as described by Bassler et al. (1997) using 20 µl AI-2 extract and 180 µl 1 : 5000 diluted overnight cultured V. harveyi biosensor BB170 in AB medium. Changes in bioluminescence upon addition of AI-2 were determined at 30 °C every 60 min using an automated luminometer (VICTOR2, 1420 multilabel counter, Wallac). For a single experiment, the V. harveyi bioassay was performed at least in duplicate for each sample. Experiments were repeated at least three times.
Protein cross-linking.
Protein cross-linking with formaldehyde and dithiobis(succinimidylpropionate) (DSP) was performed as described by Hardie et al. (1996).
In vitro generation of AI-2.
Cells were harvested by centrifugation at 10 286 g, 15 min, 4 °C and resuspended in 1/5 vol. 10 mM sodium phosphate pH 7 and frozen at -70 °C. Following two passes through a French press (SIM AMINCO Spectronic Instruments) at 2000 p.s.i. (13·8 MPa), cell debris was removed by centrifugation at 2571 g for 10 min at 4 °C, leaving the cell-free extract. To generate AI-2 in vitro, 10 µl 200 mg S-adenosylmethionine ml-1/10 mM sodium phosphate pH 7 was added to 1 ml cell-free extract and incubated at room temperature for 60 min. Control incubations were performed with the addition of 10 µl 10 mM sodium phosphate pH 7. All reactions were filtered through a 0·2 µm filter before addition to the V. harveyi biosensor BB170 bioassay to remove any contaminating bacteria, which was verified by viable counts following serial dilution of extracts. Control extracts to which no substrate (S-adenosylmethionine) was added were assayed in parallel to confirm that the AI-2 detected originated from catalysis by LuxS during the incubation.
Purification of LuxS.
Cells were harvested from overnight cultures of E. coli DH5(pKH4) by centrifugation at 10 286 g for 10 min at 4 °C, and resuspended in 20 mM Tris/HCl pH 8·0. Following two passes through the French press (SIM AMINCO Spectronic Instruments) at 2000 p.s.i. (13·8 MPa), cell debris was removed by centrifugation at 2571 g for 10 min at 4 °C, leaving the cell-free extract, which was filtered through a 0·2 µm filter before loading onto a monoQ ion-exchange column in 20 mM Tris/HCl pH 8·0. For elution, a gradient of 1–10 % 1 M NaCl/20 mM Tris/HCl pH 8·0 was first applied and held until no further peaks of absorbance at 280 nm were detected. The first peak of absorbance at 280 nm eluting after reinitiating a gradient with the same buffer contained the majority of LuxS. Eluted LuxS was pooled and applied to a Superdex 200 column in 20 mM Tris/HCl pH 8·0. LuxS eluted at a purity of approximately 90 %, as judged by eye from Coomassie- and silver-stained SDS-PAGE, at a predicted molecular mass of 40 kDa. Western blots confirmed the identity of the purified protein as LuxS. All column chromatography was achieved using the Pharmacia Biotech ÄKTA explorer.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Filtered spent supernatants from early stationary-phase cultures of DH5(pKH4) and DH5(pKH5) were assayed for the presence of LuxS-dependent autoinducers (AI-2) with the biosensor strain V. harveyi BB170 (Fig. 1a, b). HP0105 clearly directed the synthesis of autoinducers capable of inducing light production in the biosensor strain V. harveyi BB170.
|
(pKH4) with those of E. coli DH5(pBluescript) revealed one major overproduced protein (Fig. 1c
). The identity of the overproduced protein was confirmed as H. pylori LuxS by N-terminal sequencing, and it was purified by electroelution and used to inoculate a rabbit as described in Methods to generate a specific polyclonal antiserum. High levels of LuxS were produced even in the absence of the inducer of the plasmid-borne promoter, IPTG.
Carbon source affects extracellular AI-2 levels
Surette & Bassler (1998) reported that the presence of AI-2 activity in spent culture supernatants of E. coli and S. typhimurium depended on addition of glucose to the growth medium. Interestingly, addition of glucose to the growth medium affected the measurable AI-2 activity in supernatants harvested from stationary-phase cultures of E. coli DH5(pKH4). Spent culture supernatants were harvested from E. coli strains DH5(pBluescript), DH5(pKH4) and MG1655(pBluescript) grown in either LB or LB+0·4 % glucose at the indicated times (Fig. 2). Similar levels of AI-2 were detectable by the V. harveyi BB170 bioassay in DH5(pKH4) and MG1655(pBluescript) supernatants after 6·5 h growth in both the presence and the absence of glucose, but following longer periods of growth (24 h) detectable AI-2 persisted in the supernatants harvested from these strains grown with supplemented glucose, but not without. From this point onwards, the ability of glucose to sustain AI-2 activity in stationary-phase culture supernatants will be referred to as GRAIL (Glucose Retention of AI-2 Levels). As expected, supernatants harvested from E. coli DH5(pBluescript) remained devoid of detectable AI-2 in all conditions (Fig. 2). Other carbon sources (see Methods) which had a similar effect to glucose included gycerol, maltose, galactose, ribose and L-arabinose. In contrast, lactose, D-arabinose, acetate, citrate and pyruvate did not sustain high levels of AI-2 activity in stationary-phase cultures despite maintaining high-level LuxS production (Table 2). In each case results obtained with E. coli DH5(pKH4) were mirrored by those of E. coli MG1655(pBluescript), indicating that the origin of LuxS (whether from E. coli or H. pylori) was of little consequence. As expected, E. coli MG1655 bearing either pBluescript or pKH4 exhibited blue colonies on agar plates containing 40 µg X-Gal ml-1 and 0·1 mM IPTG, indicating that the inability of lactose to mimic GRAIL was not due to an inactive 
-galactosidase, which cleaves lactose into glucose and galactose.
|
|
(pKH4) cells grown in the presence or absence of 0·4 % glucose for 24 h. Between 5·3 h and 6·5 h of growth, LuxS levels were higher in the presence of 0·4 % glucose (Fig. 2c). However, after this point in the absence of glucose LuxS levels continued to rise slightly as AI-2 levels declined. In contrast, in the presence of 0·4 % glucose the levels of LuxS declined slightly despite the retention of high AI-2 activity. Similar levels of LuxS were also detected by Western blots for E. coli DH5(pKH4) grown for 24 h with all the other carbon sources described above (data not shown).
The influence of exogenous glucose upon extracellular AI-2 activity is not affected by lesions in metabolic or regulatory systems known to involve glucose
To determine whether GRAIL resulted from metabolism of the carbon source, the activity of extracellular AI-2 following growth of a range of E. coli mutants in LB or LB+0·4 % glucose was compared. E. coli mutants were assessed whilst containing pKH4 or pBluescript, and compared to parent strains. Where data are presented (Table 3, Fig. 5) the legend specifies whether pKH4 was present for each particular strain. The complete glycolytic pathway is not required for GRAIL as E. coli mutants defective in the glycolytic enzymes (phosphoglycerate kinase, phosphoglycerate mutase and enolase, Fig. 4) display wild-type GRAIL (Table 3). As galactose is able to substitute for glucose, the effect of galactose epimerase (GalU) was analysed and it was found that GRAIL could be mediated in its absence by exogenous glucose or galactose, but not by lactose.
|
|
|
). Each of these mechanisms was investigated for its role in GRAIL using mutants in E. coli defective for their function.
Since GRAIL was observed in E. coli strains bearing mutations in ptsI, ptsII (crr) and hrp and containing plasmid pKH4 or pBluescript, a functional PTS (Postma et al., 1996) was not required. A mechanism dependent upon cAMP and CRP was also ruled out since GRAIL was observed in crp and cya (adenylate cyclase) mutants of E. coli containing plasmid pKH4 or pBluescript.
The regulatory mechanisms believed to participate in cAMP-independent catabolite repression are numerous, and include (i) the global regulatory protein FruR (fructose repressor, which shows similarity to several periplasmic sugar receptors including those specific for ribose, arabinose and galactose), (ii) the stationary-phase-specific sigma factor 
s (RpoS), (iii) the sensor kinase–response regulator CreBC which is presumed to be involved in regulating gene expression in response to environmental catabolites, and (iv) (p)ppGpp as increased levels correlate with the onset of starvation (otherwise known as the stringent response, and reliant upon relA and spoT: Cashel et al., 1996). Another global regulator, the carbon storage regulator (CsrA), impacts strongly upon the metabolic carbon flow by repressing gluconeogenesis, glycogen biosynthesis and glycogen catabolism whilst activating glycolysis (Lin et al., 1997). Since GRAIL was observed in fruR, rpoS, creC, relA, spoT and csrA mutants of E. coli containing plasmid pKH4 or pBluescript, no role for these regulators in the control of AI-2 synthesized from luxS was indicated (Table 3; see Saier et al., 1996, for description of each mechanism of regulation).
Having found no role for most of the well-known metabolic pathways, and since the phosphotransacetylase–ATP : acetate phosphotransferase (Pta–AckA) pathway plays a critical catabolic role during aerobic growth on excess glucose or other glycolytic intermediates (Chang et al., 1999), we investigated the effect of mutants in this pathway upon GRAIL. Under aerobic conditions, when the carbon flux into cells exceeds the amphibolic capacity of the central metabolic pathways, e.g. the TCA cycle, E. coli cells adjust by moving acetyl-CoA through the Pta–AckA pathway, excreting acetate and generating ATP. Later, as they begin the transition to stationary phase, cells undergo the metabolic switch to acetyl-CoA synthetase (Acs) and resorb acetate (see Fig. 4, Pruss et al., 1994). Under conditions that result in mixed acid fermentation, cells also convey acetyl-CoA through the Pta–AckA pathway to gain ATP via substrate-level phosphorylation. E. coli mutants in this pathway containing either pKH4 or pBluescript were however still able to demonstrate GRAIL (Table 3).
LuxS is dimeric in vivo, and dimer formation is not dependent upon exogenous glucose
Cross-linking of cellular proteins in vivo with formaldehyde and DSP revealed the presence of multimeric forms of H. pylori LuxS overproduced in E. coli (Fig. 3). DH5(pBluescript) and DH5(pKH4) were grown in LB with and without supplementation with 0·4 % glucose to stationary phase, protein cross-linking was performed, and complexes containing LuxS revealed by Western blotting. Patterns of cross-linked products were not affected by the presence or absence of glucose in the growth medium. Formaldehyde cross-linking yielded a number of high-molecular-mass complexes containing LuxS including two distinct complexes, one which migrated at the predicted dimeric size, and one at a size intermediate between monomeric and dimeric forms of LuxS (marked with an asterisk in Fig. 3). Following partial disruption of formaldehyde cross-links by boiling, monomeric and dimeric forms of LuxS remained. Similarly with DSP, multiple forms of multimeric LuxS were observed, the most prominent being the dimeric form, which was stable following disruption of cross-links in reducing conditions. The cross-linked product asterisked in Fig. 3(a) that was seen following cross-linking with formaldehyde was not observed when complexes were cross-linked with DSP. Interestingly, separation of whole-cell extracts by non-reducing SDS-PAGE followed by Western blotting revealed a similar pattern of multimeric proteins containing LuxS as was observed following cross-linking with DSP, suggesting that these multimers may be stabilized by disulphide bridges. The observation of more abundant high-molecular-mass complexes in non-reducing SDS-PAGE following addition of DSP confirms that additional cross-links were formed under these conditions, as does the presence of the dimer following DSP treatment and resolution using reducing SDS-PAGE.
|
(pBluescript), DH5(pKH4), DH5(pProEX-luxSEc), MG1655(pBluescript) and MG1655(pKH4) were prepared after 7 h and 24 h growth in LB alone, or LB+0·4 % glucose. None of these cell extracts contained detectable levels of AI-2 activity. Incubation of these cell extracts in the presence of S-adenosylmethionine (SAM), S-adenosylhomocysteine (SAH) or S-ribosylhomocysteine (SRH) resulted in the production of AI-2 as measured by bioluminescence of the V. harveyi BB170 biosensor from all 7 h extracts except that prepared from E. coli DH5(pBluescript) (data not shown). As seen in previous experiments, following 24 h growth in the absence of glucose in the culture medium, reduced levels of AI-2 were detected in the culture supernatants of E. coli strains DH5(pKH4), DH5(pProEX-luxSEc), MG1655(pBluescript), and MG1655(pKH4) compared with growth in the presence of glucose (Fig. 2, Table 4 and Table 5). Despite this, the ability of LuxS contained within cell-free extracts prepared from 24 h cultures of E. coli strains DH5(pKH4), DH5(pProEX-luxSEc), MG1655(pBluescript) and MG1655(pKH4) to synthesize AI-2 in vitro was unaffected by the addition of 0·4 % glucose to growth medium (Tables 4 and 5, and data not shown). Control incubations containing the cell extracts, but lacking substrate, did not result in detectable levels of AI-2 activity. Western blots revealed production of LuxS (derived from either E. coli or H. pylori, as appropriate) to comparable levels in all strains whether glucose was present in the medium or not (data not shown). The H. pylori LuxS contained within crude cell-free extracts of E. coli directed the production of AI-2 in vitro using SAM, SAH or SRH as substrates; however following purification (see Methods, Fig. 1b) the addition of the Pfs enzyme was required if SAH was used as a substrate (data not shown). These observations are similar to those reported for other LuxS homologues (Schauder et al., 2001; Winzer et al., 2002a), and thus support our prediction that LuxS homologues derived from different sources possess similar properties.
|
|
reported that S. typhimurium use an ABC transporter to import AI-2 which resembles the ribose uptake transporter. E. coli possesses two machineries which function to take up ribose. The ribose transporter encoded by the genes rbsA, rbsB, rbsC, rbsD, rbsK and rbsR has the highest affinity for ribose, and has been extensively studied. Mutants interrupted in rbsR (encoding a regulatory protein that functions to repress transcription of the rbs operon in the absence of ribose: Mauzy & Hermodson, 1992), rbsB (encoding a periplasmic-binding protein that recognizes ribose and interacts with RbsC: Park et al., 1999), and rbsD (encodes a protein with no known function: Bell et al., 1986) were obtained from the E. coli Genetic Stock Center. All mutant strains exhibited GRAIL in stationary phase. All strains were tested whilst bearing either pBluescript or pKH4, and where the strain contained an active, chromosomally encoded LuxS, the same pattern of AI-2 activity was recorded, although only the data in the presence of pKH4 are shown (Fig. 5). Furthermore, the presence or absence of an active chromosomally encoded LuxS was mirrored in parent strains.
The ribose- and galactose-binding proteins of E. coli (RbsB and MglB respectively) exhibit more similarity to each other (26 % similarity/44 % identity) than any other pair of sugar-binding proteins in E. coli, and both interact with the inner membrane translocator, Trg (Falke et al., 1997). Although mutations in trg did not alter the manifestation of GRAIL (Fig. 5b), a complete deletion of the mgl operon resulted in lower, shorter-lived levels of AI-2 activity in the absence of exogenous glucose. For completeness, E. coli mutants of genes encoding nucleoside uptake components and regulators of such machineries (cytR, deoR, tsx, nupC) were assessed, and none exhibited altered GRAIL (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
), emphasizing the functional relatedness of this family of proteins. Our data also support recent crystallographic studies (Lewis et al., 2001) by using protein cross-linking approaches to reveal that LuxS associates into dimers in vivo. Analysis of AI-2 production throughout growth of E. coli strains MG1655 and DH5(pKH4) has extended previous observations for S. typhimurium LT2 and E. coli AB1157 (Surette & Bassler, 1998), which reported that the sustained presence of active AI-2 in stationary-phase culture supernatants requires the presence of glucose in the growth medium (GRAIL). Our observation that LuxS presence, functionality and dimerization is independent of exogenous glucose suggests (i) that AI-2 is actively removed or degraded in the absence of glucose during stationary phase, and (ii) that if LuxS becomes incapable of producing AI-2 in vivo in stationary-phase E. coli cells grown in the absence of glucose, it is not a result of degradation or inactivation of LuxS. Rather, it is likely to be due to a lack of substrate availability. The absence of increased levels of AI-2 upon overproduction of LuxS in E. coli is in agreement with limited substrate availability within cells as suggested for S. typhimurium by Beeston & Surette (2002).
In accordance with the carbon source dependence of AI-2 detection in supernatants of S. typhimurium and E. coli strain AB1157 (Surette et al., 1999; Surette & Bassler, 1999), detectable AI-2 is found in 24 h cultures of E. coli MG1655 and E. coli DH5(pKH4) in the presence of both PTS (glucose, maltose) and non-PTS (galactose and arabinose) sugars. However, in contrast to the data published for S. typhimurium (Surette & Bassler, 1999), addition of glycerol (which feeds into glycolysis at the level of glyceraldehyde 3-phosphate) mimicked GRAIL. Supplementary carbon sources that feed into the TCA cycle directly (pyruvate, citrate, acetate) did not result in extracellular AI-2 activity in stationary-phase cultures. This is particularly surprising since the addition of acetate to nutrient broth has been reported to induce the production of LuxS (Kirkpatrick et al., 2001). Mutations in the glycolytic enzymes which link glyceraldehyde 3-phosphate to pyruvate (encoded by pgk, pgm and eno) did not prevent GRAIL. Together, this evidence suggests that substrate availability, but not breakdown via the glycolytic pathway, determines AI-2 activity in culture supernatants.
Surprisingly, lactose (a disaccharide containing glucose and galactose, both of which can induce GRAIL) was unable to mimic glucose in cells which are -galactosidase-positive and therefore capable of breaking it down into its constituent monosaccharides. This finding argues that glucose implements its effect via a specific uptake pathway. Mutations in the phosphotransferase proteins Hpr, PtsI and PtsII which constitute the specific route of glucose entry into cells did not however prevent its effect upon AI-2 activity, ruling this out as the basis for GRAIL. Supporting this is the ability of strains deficient in different mechanisms of achieving catabolite repression and non-PTS transported carbon sources to display GRAIL.
The potential loss in production of AI-2 during stationary phase in the absence of certain carbon sources is unlikely to be due to a conformational change in LuxS since cross-linking patterns were unaltered, which is suggestive of maintenance of quaternary structure. This is consistent with retention of LuxS activity in vitro, and limiting substrate availability. The multimeric forms of LuxS detected in these experiments are likely to be the dimers also discovered during crystallographic studies. The presence of dimers noted here in non-reducing conditions can also be explained following analysis of the crystal structures since they revealed that cysteine residues are located at the interface of adjacent molecules.
The loss of detectable AI-2 in stationary-phase culture supernatants grown in the absence of glucose may result from uptake of AI-2 by the cells, or inactivation of AI-2. An absence of extracellular degrading agents produced by E. coli DH5(pKH4) is suggested by the inability of filtered stationary-phase culture supernatants grown without glucose to diminish AI-2 activity following incubation for up to 2 h at a range of temperatures (20–42 °C) with AI-2-containing culture supernatant (data not shown). It is however possible that degradation requires the presence of cells. Surette et al. (1999) stated that E. coli DH5 does not degrade AI-2 based on evidence presented in Surette & Bassler (1998); however the latter article does not contain results to this effect. Our results described here clearly show that E. coli DH5 is capable of removing AI-2 from the supernatant of stationary-phase cultures in the absence of glucose, indicating that E. coli DH5 cells either degrade or take up AI-2. Currently we can not distinguish between these two possibilities. Uptake of AI-2 may be a mechanism of monitoring the population density (quorum sensing), but is also consistent with a role in central metabolism (Winzer et al., 2002a, b). In the latter context, given its proposed structure as a furanosyl borate diester, AI-2 may be acting as a source of carbon or boron (Coulthurst et al., 2002; Winans, 2002; Winzer et al., 2002b). The production of the precursor of AI-2 (4,5-dihydroxy-2,3-pentanedione, DPD) is linked to the flux through the activated methyl cycle; however the production of AI-2 does not correlate with the activity of pfs in the absence of glucose (Beeston & Surette, 2002), suggesting that the production of DPD is a separate event from export of AI-2 activity. AI-2 activity in the culture supernatant is higher in carbohydrate-rich media; thus, like acetate, AI-2 can be produced and excreted in the presence of preferred carbon sources, and utilized when preferred nutrients are exhausted, which in turn suggests that uptake of AI-2 is another independent event, unlinked to synthesis and export.
The ABC transporter Lsr in S. typhimurium is proposed to import AI-2 (Taga et al., 2001). In E. coli, two transporters show similarity to the S. typhimurium Lsr transporter, the products of the rbs genes (24 % similarity/42 % identity for the periplasmic sugar-binding protein), and the products of the unstudied operon b1513 (77 % similarity/82 % identity for the periplasmic sugar-binding protein): Blattner et al. (1997); Iida et al. (1984). We show here that mutations in different rbs genes do not prevent uptake of AI-2 produced by LuxS. This suggests that the products of operon b1513, and not rbs, may perform a function analogous to Lsr in E. coli. Promotion of GRAIL by galactose in a strain mutated in the gene encoding galactose epimerase may indicate that it exerts its effect at the level of binding to the transport protein, MglB, in the periplasm, rather than following metabolism. In support of a link between sugar uptake and detectable AI-2 activity, mutation of the mgl operon encoding the galactose uptake system (which displays the highest level of similarity to ribose uptake systems: periplasmic sugar binding proteins RbsB and MglB are 26 % identical/44 % similar) resulted in lower, shorter-lived levels of AI-2 activity. The common denominator is however unlikely to be the inner membrane translocator, Trg, to which both RbsB and MglB bind (Falke et al., 1997) since mutation of trg did not alter GRAIL.
Regulation of AI-2 uptake is unlikely to result from conventional catabolite repression as mutations in crp (cyclic AMP receptor protein), rpoS (starvation-induced sigma factor), cya (adenylate cyclase), csrA (carbon storage regulator), fruR (fructose regulator: Saier et al., 1996), effectors of the stringent response (relA, spoT) and creC (the catabolite-repression-linked sensor kinase–response regulator) did not alter the ability of exogenous glucose to induce clearance of extracellular AI-2 activity (Kolb et al., 1993; Saier et al., 1996). Further studies are under way to identify potential AI-2 uptake mechanisms in E. coli, and determine whether they are regulated by the presence of glucose.
Further analysis is required to determine the basis of carbon source dependence of AI-2 activity. This paper has highlighted that there is likely to be a degradation or uptake mechanism for AI-2 in E. coli, which may facilitate its recycling into central metabolism (discussed by Winzer et al., 2002a, b, Winans, 2002: suggested by Lewis et al., 2001; Fong et al., 2001). It is perhaps interesting to speculate that a role in central metabolism is more important to some bacteria than a role in quorum sensing. V. harveyi luminesence is clearly a quorum-sensing phenomenon; however no obvious phenotypes could be attributed to loss of luxS in other organisms, suggesting that it may play a more subtle role in their life cycle. Pertinently, many of these studies were performed in nutrient-rich medium, where recycling of homocysteine is less important. This duel functionality dictates that studies to show the effects of luxS upon bacterial physiology must incorporate complementation with both the gene (to discount effects caused by second-site mutation) and purified AI-2 (to determine whether effects are likely to be mediated via quorum sensing).
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bassler, B. L. (1999). How bacteria talk to each other: regulation of gene expression by quorum sensing. Cur Opin Microbiol 2, 582–587.[CrossRef][Medline]
Bassler, B. L. (2002). Small talk: cell-cell communication in bacteria. Cell 109, 421–424.[CrossRef][Medline]
Bassler, B. L., Wright, M. & Silverman, M. R. (1994). Multiple signalling systems controlling expression of luminescence in Vibrio harveyi: sequence and function of genes encoding a second sensory pathway. Mol Microbiol 13, 273–286.[Medline]
Bassler, B. L., Greenberg, E. P. & Stevens, A. M. (1997). Cross-species induction of luminescence in the quorum sensing bacterium Vibrio harveyi. J Bacteriol 179, 4043–4045.
Beeston, A. L. & Surette, M. G. (2002). pfs-dependent regulation of autoinducer 2 production in Salmonella enterica serovar typhimurium. J. Bacteriol 184, 3450–3456.
Bell, A. W., Buckel, S. D., Groarke, J. M., Hope, J. N., Kingsley, D. H. & Hermodson, X. (1986). The nucleotide sequences of rbsD, rbsA, and rbsC genes of Escherichia coli K 12. J Biol Chem 261, 7652–7658.
Bell, A., Gaston, K., Williams, R., Chapman, K., Kolb, A., Buc, H., Minchin, S., Williams, J. & Busby, S. (1990). Mutations that alter the ability of Escherichia coli cyclic AMP receptor protein to activate transcription. Nucleic Acids Res 18, 7243–7250.
Berman, M. & Lin, E. C. C. (1971). Glycerol-specific revertants of a phophoenolpyruvate phosphotransferase mutant: suppression by the desensitization of glycerol kinase to feedback inhibition. J Bacteriol 105, 113–120.
Bitner, R. M. & Kuempel, P. L. (1981). P1 transduction map spanning the replication terminus of Escherichia coli K12. Mol Gen Genet 184, 208–212.[Medline]
Blaser, M. J., Perez-perez, G. I., Kleanthous, H., Cover, T. L., Peek, R. M., Chyou, P. H., Stemmermann, G. N. & Nomura, A. (1995). Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res 55, 2111–2115.
Blattner, F. R., Plunkett, G., 3rd, Bloch, C. A. & 14 other authors (1997). The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1474.
Burgess, N. A., Kirke, D., Williams, P., Winzer, K., Hardie, K. R., Meyer, N. L., Aduse-Opoku, J., Curtis, M. A. & Cámara, M. (2002). LuxS-dependent quorum sensing in Porphyromonas gingivalis modulates protease and haemagglutinin activities but is not essential for virulence. Microbiology 148, 763–772.
Busby, S., Kotlaz, D. & Buc, H. (1983). Deletion mutagenesis of the Escherichia coli galactose operon promoter region. J Mol Biol 167, 259–274.[Medline]
Casadaban, M. J. (1976). Transposition and fusion of the lac genes to selected promoters in Escherichia coli using bacteriophage lambda and Mu. J Mol Biol 104, 541–555.[CrossRef][Medline]
Cashel, M., Gentry, D. R., Hernandez, V. J. & Vinella, D. (1996). The stringent response. In Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn, vol. 1, pp. 1458–1463. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.
Chang, D.-E., Shin, S., Rhee, J.-S. & Pan, J.-G. (1999). Acetate metabolism in a pta mutant of Escherichia coli W3110: importance of maintaining acetyl coenzyme A flux for growth and survival. J Bacteriol 181, 6656–6663.
Chen, X., Schauder, S., Potier, N., Van Dorsselaer, A., Pelczer, I., Bassler, B. L. & Hughson, F. M. (2002). Structural identification of a bacterial quorum-sensing signal containing boron. Nature 415, 545–549.[CrossRef][Medline]
Chung, W. O., Park, Y., Lamont, R. J., McNab, R., Barbieri, B. & Demuth, D. R. (2001). Signaling system in Porphyromonas gingivalis based on a LuxS protein. J Bacteriol 183, 3903–3909.
Coulthurst, S. J., Whitehead, N. A., Welch, M. & Salmond, G. P. C. (2002). Can boron get bacteria talking? Trends Biochem Sci 27, 217–219.[CrossRef][Medline]
Cover, T. L. & Blaser, M. J. (1996). Helicobacter pylori infection, a paradigm for chronic mucosal inflammation: pathogenesis and implications for eradication and prevention. Adv Intern Med 41, 85–117.[Medline]
Crasnier-Medansky, M., Park, M. C., Studley, W. K. & Saier, M. H., Jr (1997). Cra-mediated regulation of Escherichia coli adenylate cyclase. Microbiology 143, 785–792.[Abstract]
Day, W. A. & Maurelli, A. T. (2001). Shigella flexneri LuxS quorum sensing system modulates virB expression but is not essential for virulence. Infect Immun 69, 15–23.
DeLisa, M. P., Valdes, J. J. & Bentley, W. E. (2001). Mapping stress-induced changes in autoinducer AI-2 production in chemostat-cultivated Escherichia coli K-12. J Bacteriol 183, 2918–2928.
Duerre, J. A. & Miller, C. H. (1966). Cleavage of S-ribosyl-L-homocysteine by extracts from Escherichia coli. J Bacteriol 91, 1210–1217.
Duerre, J. A., Baker, D. J. & Salisbury, L. (1971). Structure elucidation of a carbohydrate derived from S-ribosylhomocysteine by enzymatic cleavage. Fed Proc 30, 88.
Falke, J. J., Bass, R. B., Butler, S. L., Chervitz, S. A. & Danielson, M. A. (1997). The two-component signaling pathway of bacterial chemotaxis: a molecular view of signal transduction by receptors, kinases, and adaptation enzymes. Annu Rev Cell Dev Biol 13, 457–512.[CrossRef][Medline]
Fong, K. P., Chung, W. O., Lamont, R. J. & Demuth, D. R. (2001). Intra-and interspecies regulation of gene expression by Actinobacillus actinomycetemcomitans LuxS. Infect Immun 69, 7625–7634.
Forsyth, M. H. & Cover, T. L. (2000). Intercellular communication in Helicobacter pylori: luxS is essential for the production of an extracellular signaling molecule. Infect Immun 68, 3193–3199.
Fox, C. F. & Wilson, G. (1968). The role of a phophoenolpyruvate-dependent kinase system in -glucoside catabolism in Escherichia coli. Proc Natl Acad Sci U S A 59, 988–995.
Fraenkel, D. G. (1996). Glycolysis. In Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn, vol. 1, pp. 189–195. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.
Greenberg, P., Hastings, J. W. & Ulitzur, S. (1979). Induction of luciferase synthesis in Beneckea harveyi by other marine bacteria. Arch Microbiol 120, 87–91.[CrossRef]
Guyer, M. S., Reed, R. R., Steitz, J. A. & Low, K. B. (1981). Identification of a sex-factor-affinity site in E. coli as gamma delta. Cold Spring Harbor Symp Quant Biol 45, 135–140.
Hammer-Jespersen, K. & Nygaard, P. (1976). Multiple regulation of nucleoside catabolizing enzymes in Escherichia coli: effects of 3 : 5' cyclic AMP and CRP protein. Mol Gen Genet 148, 49–55.[CrossRef][Medline]
Hammer-Jespersen, K. & Munch-Petersen, A. (1975). Multiple regulation of nucleoside catabolizing enzymes: regulation of the deo operon by the cytR and deoR gene products. Mol Gen Genet 137, 327–335.[CrossRef][Medline]
Hardie, K. R., Lory, S. & Pugsley, A. P. (1996). Insertion of an outer membrane protein in Escherichia coli requires a chaperone-like protein. EMBO J 15, 978–988.[Medline]
Harlow, E. & Lane, D. (1988). Antibodies: a Laboratory Manual, pp. 139–243. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Hilgers, M. T. & Ludwig, M. L. (2001). Crystal structure of the quorum-sensing protein LuxS reveals a catalytic metal site. Proc Natl Acad Sci U S A 98, 11169–11174.
Hoffmeyer, J. & Neuhard, J. (1971). Metabolism of exogenous purine bases and nucleosides by Salmonella typhimurium. J Bacteriol 106, 14–24.
