Role of autoinducer-2 on the adhesion ability of Lactobacillus acidophilus

Authors

  • B.L. Buck,

    1.  Department of Food, Bioprocessing, and Nutrition Sciences; and Southeast Dairy Foods Research Center, North Carolina State University, Raleigh, NC, USA
    2.  Department of Microbiology, North Carolina State University, Raleigh, NC, USA
    Search for more papers by this author
  • M.A. Azcarate-Peril,

    1.  Department of Food, Bioprocessing, and Nutrition Sciences; and Southeast Dairy Foods Research Center, North Carolina State University, Raleigh, NC, USA
    Search for more papers by this author
  • T.R. Klaenhammer

    1.  Department of Food, Bioprocessing, and Nutrition Sciences; and Southeast Dairy Foods Research Center, North Carolina State University, Raleigh, NC, USA
    2.  Department of Microbiology, North Carolina State University, Raleigh, NC, USA
    Search for more papers by this author

Todd R. Klaenhammer, Department of Food, Bioprocessing, and Nutrition Sciences, North Carolina State University, 7624, Raleigh, NC 27695, USA. E-mail: klaenhammer@ncsu.edu

Abstract

Aims:  Lactobacilli adhere to the intestinal epithelium and this intimate association likely promotes retention in the gastrointestinal tract and communication with the immune system. The aim of this study was to investigate whether or not the quorum-sensing signalling molecule, autoinducer (AI)-2, was produced by Lactobacillus acidophilus and affected adherence to intestinal epithelial cells.

Methods:  Microarray analysis of concentrated cells of L. acidophilus NCFM revealed several genes involved in a classic stress response and potentially adhesion. Putative genes linked to the synthesis of the interspecies signalling molecule, AI-2, were overexpressed. Examination of the NCFM genome revealed the complete pathway for AI-2 synthesis. AI-2 activity from NCFM was detected using the Vibrio harveyi BB170 assay system. Using site-specific integration, an isogenic mutation was created in luxS and the resulting mutant did not produce AI-2. In addition to some minor metabolic effects, the luxS mutation resulted in 58% decrease in adherence to Caco-2 cells.

Conclusion: L. acidophilus NCFM encodes the genes for synthesis of the quorum-sensing signal, AI-2, and produces this molecule during planktonic growth.

Significance and Impact of the Study:  The ability to produce AI-2 affects the ability of L. acidophilus to attach to intestinal epithelial cells.

Introduction

Lactic acid bacteria (LAB) are naturally found in a variety of environmental niches where they exist as members of complex microbial communities. Survival in each niche depends on the ability of the organism to sense and respond to varying conditions such as temperature, pH, nutrient availability and cell population density. One major niche that LAB frequently occupy is that of the mammalian gastrointestinal tract (GIT). Multiple signal transduction pathways likely control the expression of adhesion factors, stress response genes and other genetic determinants designed to help LAB survive and exist in the various compartments of the GIT. Some intestinal lactobacilli, along with Gram-positive and Gram-negative pathogenic bacteria, have the ability to associate with the intestinal epithelium and mucosal layers therein. This association is thought to be important for the realization of certain probiotic properties including competitive exclusion, immunomodulation and the delivery of biotherapeutics, among others. The difficulties and cost of human trials, along with the dynamic and variable environment of the intestine have led to the development of in vitro model systems for the study of bacterial adherence to the intestinal epithelium, such as Caco-2 cells. Traditionally used for the study of pathogens (Darfeuille-Michaud et al. 1990; van Asten et al. 2004), probiotics (Greene and Klaenhammer 1994; Azuma and Sato 2001) and the inhibition of pathogens by probiotics (Coconnier et al. 2000), Caco-2 monolayers are also employed as models of the intestinal epithelium for the selection of adhesive probiotic strains. Although the molecular mechanisms involved with this association are not understood, it is clear that the process is complex, involving host-specific, bacterial-specific and environmental factors. Lactobacilli must maintain a balance between meeting their own growth requirements and surviving the hostile conditions, including gastric-acid shock, presented by the host defenses and competing microflora (Tannock 2005). However, little is known about how these stressors influence the ability of lactobacilli to associate with the varied substrates in the intestinal environment.

The microbiome of the healthy human intestine contains an estimated 500–1000 different species that combine to represent a uniquely vast and diverse genetic landscape. It would, therefore, not be surprising if the members of the intestinal community communicate among themselves to coordinate processes such as maintenance of the commensal microflora or disease resistance (Kaper et al. 2005). Bacterial cells frequently communicate via quorum-sensing mechanisms, which involve the density-dependant recognition of an autoinducer (AI) followed by changes in gene expression. One important quorum-sensing system that can be used to communicate among and between species is based on a furanosyl borate diester called AI-2, produced in four enzymatic steps from methionine. The gene encoding AI-2 synthase, LuxS, has been identified in many different Gram-positive and Gram-negative species. AI-2 regulates the expression of various phenotypes including virulence factors, DNA processing, cell morphology, motility, biofilm formation, toxin production, light production and cell division in numerous species (Xavier and Bassler 2003; Lebeer et al. 2007; Sztajer et al. 2008).

This study reports the identification of genes differentially expressed during concentration and acidification of mid-log phase cells of L. acidophilus. Among the induced genes identified by microarray analysis were the open reading frames (ORF) (LBA1080 and LBA1081) encoding proteins involved in the synthesis pathway for AI-2. This study investigated the role of LuxS, and AI-2, on the adhesive capability and growth of L. acidophilus.

Materials and methods

Bacterial strains and growth conditions

The bacterial strains used in this study are listed in Table 1. Lactobacillus strains were cultivated anaerobically at 37°C or 42°C in de Mann Rogosa Sharpe (MRS) broth (Difco), semi-defined medium (SDM; Kimmel and Roberts 1998) or, when appropriate, in MRS supplemented with 1·5% (w/v) agar. Escherichia coli was propagated aerobically in Luria-Bertani (LB; Difco) medium or on LB medium supplemented with 1·5% (w/v) agar at 37°C. Brain heart infusion (BHI; Difco) medium supplemented with 1·5% agar (w/v) and 150 μg ml−1 erythromycin (Em) was used for selection of E. coli transformants. Autoinducer bioassay (AB) media (Bassler et al. 1993) was used for the propagation of all Vibrio harveyi strains. When appropriate, chloramphenicol (Cm; 5·0 μg ml−1) and Em (5·0 or 150 μg ml−1) were used for selection. CFU per millilitre were determined with appropriate dilutions using a Whitley Automatic Spiral Plater (Don Whitley Scientific, Shipley, W. Yorkshire, UK).

Table 1.   Bacterial strains and plasmids used in this study
StrainsOrigin or relevant characteristicsSource or reference
Lactobacillus acidophilus
 NCFMHuman intestinal isolateSanders and Klaenhammer (2001)
 NCK 1398NCFM::pTRK685 (lacL integrant)Russell and Klaenhammer (2001)
 NCK 1765NCFM::pTRK854 (luxS integrant)This study
Vibrio harveyi
 BB170luxN::Tn5 AI-1 sensor AI-2 sensor+DeKeersmaecker and Vanderleyden (2003)
Escherichia coli
 EC1000RepA+ MC1000, Kmr, carrying a single copy of the pWV01 repA; host for pOR128-based plasmidsLaw et al. (1995)
Plasmids
 pORI28Emr, ori (pWV01), replicates only with repA provided in transLaw et al. (1995)
 pTRK669ori (pWV01), Cmr, RepA+Russell and Klaenhammer (2001)
 pTRK854471 bp internal region of luxS (LBA1081) cloned into BglII-XbaI sites of pORI28This study

DNA manipulation techniques

Total Lactobacillus genomic DNA was isolated according to the method of Walker and Klaenhammer (1994). Standard protocols were used for endonuclease restriction, ligation, DNA modification and transformation (Sambrook et al. 1989). Plasmid preparations for the purpose of screening E. coli transformants followed the method of Zhou et al. (1990). Large-scale plasmid preparations were performed with the QIAprep Spin kit according to the manufacturer’s instructions (Qiagen). Polymerase chain reactions (PCR) were performed according to the manufacturer’s recommendations using a Taq DNA polymerase PCR system (Roche Applied Science). PCR primers were synthesized by integrated DNA technologies (IDT) and, when appropriate, restriction sites were designed at the 5′ end of the primers to facilitate future cloning steps. DNA fragments were extracted from 1·0% agarose gels using the Zymoclean Gel DNA Recovery kit (Zymo Research). Electrocompetent Lactobacillus cells were prepared as described by Walker et al. (1996). Southern hybridization of genomic DNA was performed using standard protocols.

Tissue culture

The Caco-2 (ATCC HTB-37) cells were only used between the 40th and 50th passages. All reagents used in the maintenance of Caco-2 cells were obtained from Gibco (Invitrogen). The cells were routinely grown in 95% air-5% CO2 atmosphere in minimum essential medium (MEM) supplemented with 20% (v/v) inactivated (56°C, 30 min) fetal bovine serum (FBS), 0·10 mmol l−1 nonessential amino acids and 1·0 mmol l−1 sodium pyruvate. Monolayers were trypsinized for 10 min, counted using a haemocytometer, and seeded at 1·3 × 105 cells per well in 2·0 ml of the cell culture medium. Cells were grown on 15 mm Thermanox plastic coverslips (Thermofisher Scientific) in Costar 12-well tissue culture-treated plates (Corning). The medium was replaced every two days and all adherence assays were performed with cells differentiated for 14 days.

Adherence assay

Adhesion of Lactobacillus strains to Caco-2 cells was examined according to the method described previously (Buck et al. 2005). Briefly, mid-log phase bacterial cells (OD600 0·6) were prepared in 10 ml of MRS with 3·0 μg ml−1 Em to maintain selective pressure on integrants. The cells were removed by centrifugation for 10 min at 4000 g, and washed twice with phosphate-buffered saline (PBS). The bacterial pellets were resuspended in 5 ml of fresh MRS prior to adherence. Fifteen-day Caco-2 monolayers were washed twice with PBS and treated with a bacterial suspension at a concentration of 4 × 108 CFU ml−1. Bacteria were incubated on the monolayer for 1·5 h at 37°C in a mixture of MRS and cell-line culture medium. Following incubation, the monolayers were washed five times with PBS, fixed in methanol and Gram stained. Adherent bacterial cells were then enumerated microscopically. Duplicate coverslips were counted for each experiment. The final data presented collectively represents at least three independent experiments in duplicate. Total counts for each coverslip were used and adhesion was expressed as per cent (%) of the control strain NCK1398, which carries an insert in the lacL (β-galactosidase) gene (Russell and Klaenhammer 2001). Using this control, all mutant cultures and the parental control could be prepared with Em to maintain selective pressure on the LuxS integrant.

RNA isolation

A single 20-ml culture of L. acidophilus was grown in MRS to OD600 0·6, divided into two 10-ml aliquots and harvested by centrifugation for 8 min at 4000 rev min−1. One aliquot was resuspended in 10 ml of fresh MRS and the other was resuspended in 1 ml of fresh MRS. Both cultures were then incubated for 1 h in a 37°C water bath. Following incubation, the cells were harvested by centrifugation and frozen immediately in a dry ice-ethanol bath. RNA isolation was conducted using TRIzol (Invitrogen) according to the protocol described previously (Azcarate-Peril et al. 2005). RNA purity and concentration were determined by electrophoresis on agarose gels and standard spectrophotometer measurements.

Microarray hybridization and data analysis

Gene expression analysis was performed using a whole-genome DNA microarray based on PCR products of predicted ORF from the L. acidophilus NCFM genome (Azcarate-Peril et al. 2005). Briefly, identical amounts (25 μg) of total RNA were aminoallyl-labelled by reverse transcription with random hexamers in the presence of amino-allyl dUTP (Sigma Chemical), using Superscript II reverse transcriptase (Life Technologies) at 42°C overnight, followed by fluorescence-labelling of aminoallylated cDNA with N-hydroxysuccinimide-activated Cy3 or Cy5 esters (GE Healthcare). Labelled cDNA probes were purified using the PCR Purification Kit (Qiagen). Coupling of the Cy3 and Cy5 dyes to the AA-dUTP-labelled cDNA and hybridization of samples to microarrays were performed as described. Fluorescence intensities were acquired using a General Scanning ScanArray 4000 Microarray Scanner (Perkin Elmer) and processed as TIFF images. Signal intensities were quantified using the QuantArray 3.0 software package (Perkin Elmer). Two slides, each containing the complete genome of L. acidophilus NCFM spotted in triplicate, were hybridized reciprocally to Cy3- and Cy5-labelled probes per experiment (dye swap) for each of the two biological replicates. The spots were analysed by adaptive quantification. The local background was subsequently subtracted from the recorded spot intensities. Data was median normalized. The average of the six ratios per gene was recorded. The ratio between the average absolute pixel values for the replicated spots of each gene with and without treatment represented the fold change in gene expression. Confidence intervals and P-values on the fold change were also calculated with the use of a two-sample t-test. P-values of 0·05 or less were considered significant. Microarray data has been submitted to the Gene Expression Omnibus (GEO) at NCBI and can be found under accession numbers GPL1401 (platform) and GSM132762, GSM132764, GSM132766 and GSM132767 (samples).

Site-specific integration into Lactobacillus acidophilus NCFM

Using L. acidophilus NCFM chromosomal DNA as a template, a 471 bp internal fragment of LBA1081 was amplified by PCR with primers 1081-IF (5′-GATCA GATCT AAGTT AAGGC ACCTT ACG-3′) and 1081-IR (5′-GATCT CTAGA TTTCG AATGG GTCAT CAC-3′). The internal fragment was cloned onto the integrative vector pORI28 (Law et al. 1995) and subsequently transformed by electroporation into L. acidophilus NCFM containing the temperature-sensitive helper plasmid pTRK669 (Russell and Klaenhammer 2001). The steps were then performed according to Russell and Klaenhammer (2001) for the selection of integrants. Successful integration of the plasmid was confirmed by PCR and southern hybridization of the junction fragments using standard protocols.

Characterization of the LuxS mutant strain

For growth and adherence experiments involving the LuxS mutant, a derivative of L. acidophilus NCFM (NCK1398) harbouring a plasmid insertion into β-galactosidase (lacL; Russell and Klaenhammer 2001) was used as the control so that the identical antibiotic pressure could be maintained on both the LuxS mutant and the control during preparation of the bacterial cells.

The sugar utilization profile of the strains was assessed using the API50 CH test (BioMerieux) according to the manufacturer’s instructions. Growth of Lactobacillus strains was evaluated in MRS, SDM, MRS or SDM supplemented with Oxgall (0·15, 0·5 or 1%; Difco) or NaCl (5 or 15% that corresponds to a concentration of 0·85 or 2·6 mol l−1, respectively), and MRS or SDM acidified with lactic acid (pH 5·5 and 4·5) and automatically monitored by determining the changes in absorbance (A600) as a function of time using a FLUOStar OPTIMA microtitre plate reader (BMG Labtech). The maximum specific growth rate (μmax, h−1) was calculated from the slope of a linear regression line during exponential growth with a correlation coefficient (r2) of 0·99. Each point represents the mean of three independent cultures.

AI-2 detection

The detection of AI-2 from the supernatant of selected bacterial strains was performed as described previously (DeKeersmaecker and Vanderleyden 2003) with modifications as follows. The detection of AI-2 produced by LAB can be problematic owing to the decreased pH of the spent culture supernatants, and catabolite repression by glucose of the lux operon in the reporter strain V. harveyi BB170 (DeKeersmaecker and Vanderleyden 2003). Accordingly, all Lactobacillus populations used for detection of AI-2 were grown in modified MRS containing 1% galactose. At specified time points, 4-ml aliquots were collected, OD600 measured and cell-free supernatants were isolated by centrifugation at 4000 g for 10 min. The pH of the supernatant was neutralized to pH 6·5 with 2N NaOH and filtered through a 0·2-μm membrane. The reporter strain V. harveyi BB170 was grown overnight in AB medium (Greenberg et al. 1979), washed with and resuspended in fresh AB to OD600 0·5. Ten microlitre of the sterile supernatant was mixed with 90 μl of a 1/1000 dilution of the reporter strain BB170 in each well of a 96-well plate. Luminescence was measured at 30°C every 10 min for 6 h in a fluorescent microtitre plater reader (FLOUStar Optima, BMG Technologies, Durham, NC, USA). Fold induction was calculated by averaging at least six standard time points and dividing the average value obtained from the wild type (wt) by the average value obtained from the LuxS mutant. Each sample was measured in at least three independent wells and each data point represents three independent samples.

Results

Microarray analysis

In a previous study, we established that the AI-2 production gene, luxS, as well as the gene encoding the methionine synthase metE were significantly overexpressed in L. acidophilus NCFM containing an insertionally inactivated histidine protein kinase (HPK; LBA1524) at both pH 5·5 and 4·5 (Azcarate-Peril et al. 2005). Given the involvement of LuxS with events related to high cell density (Bassler 1999), we employed the existing microarray platform to compare the overall gene expression of a mid-log phase population of L. acidophilus cells either concentrated and incubated for 1 h at 37°C (test) or not concentrated but still incubated for 1 h at 37°C (control). A comparison of the results from the present study and a previous study on gene expression differences between cells at pH 4·5 and 5·5 (Azcarate-Peril et al. 2005) allowed us to identify pH-independent gene expression between the concentrated and nonconcentrated Lactobacillus cell samples. A functional classification of the genes differentially expressed under this condition is shown in Fig. 1 and the ORFs showing significant differences in expression, as determined by the Student’s t-test (P < 0·05), are listed in Supporting Information Table S1. Two-hundred and twenty three genes (c. 12% of the total number of ORF in the genome) were differentially expressed (log2 ratios concentrated/nonconcentrated >−3·2 and <3·1; −log10 (P-values) <−1·22) after cell concentration and incubation. Notably, the most represented category within overexpressed genes was [G] ‘Carbohydrate transport and metabolism’. In contrast, the most represented categories within underexpressed genes were [J] ‘Translation, ribosomal structure and biogenesis’ and [P] ‘Inorganic ion transport and metabolism’. The majority of ORF exhibited a similar expression pattern to previous microarray analysis of L. acidophilus exposed to MRS acidified with lactic acid to pH 5·5 and 4·5 (Azcarate-Peril et al. 2005). These include many of the putative stress response genes including radA, groES, groEL, bshA and dnaK, along with a putative aggregation promoting protein, myosin cross-reactive antigen and multiple genes encoding membrane-associated proteins. Sixty-five genes were differentially expressed and met the statistical criteria in response to concentration and incubation of cells but not to acidic pH. Fifteen genes were induced (>1·5-fold; –log10 (P-values) <−1·22) by the studied condition. Interestingly, clpP (LBA0694) and clpE (LBA0638 and LBA1910) genes were induced upon concentration of cells. Also induced were a permease (LBA1959), probably involved in nucleotide transport and metabolism, and cation transport ATPases (LBA0523 and LBA0999). In contrast, 50 genes were underexpressed in the concentration and incubation treatment but not in the pH treatment, most of them involved in translation, ribosomal structure and biogenesis. The two tandem ORF (LBA1080 and LBA1081) encoding proteins involved in the synthesis pathway for AI-2 were overexpressed in response to both treatments, concentration and incubation of cells and at an acidic pH of 5·5 (Azcarate-Peril et al. 2005).

Figure 1.

 Distribution of open reading frames within COG (clusters of orthologous groups; Tatusov et al. 1997). Genes were classified according to the COG domain present in the encoded protein sequence. The number of differentially expressed genes in each category is indicated.

Genomic organization of the region encoding luxS in LAB

In silico analysis of the genome of L. acidophilus NCFM (Altermann et al. 2005) revealed four ORFs (LBA1622, LBA0931, LBA0820 and LBA1081) showing homology to each gene in the biosynthetic pathway for the production of AI-2 from methionine (Fig. 2). MetK (LBA1622) converts methionine to S-adenosylmethionine (SAM). A methyl group is removed from SAM by an SAM-dependent methyltransferase (LBA931) forming S-adenosylhomocysteine (SAH), which is subsequently detoxified by an MTA/SAH nucleosidase, Pfs (LBA820), forming S-ribosylhomocysteine (SRH) and adenine. LuxS (LBA1081) converts SRH to homocysteine and 4,5-dihydroxy-2,3-pentanedione which circularizes, incorporating boron, to form AI-2 (Xavier and Bassler 2003). Homocysteine is finally methylated back to methionine by MetE (LBA1080) located directly upstream of the luxS homologue (LBA1081). A terminator between the two ORF with a free energy of −16·2 kcal suggests that LBA1080 and LBA1081 are expressed separately. Interestingly, based on predicted protein sequences, ORFs from Lactobacillus genomes (Kleerebezem et al. 2003; Pridmore et al. 2004; Altermann et al. 2005) annotated as metE (methionine synthase or cobalamin-independent homocysteine methyltransferase) grouped away from other available MetE proteins by the neighbour-joining bootstrap analysis (Fig. 3).

Figure 2.

 Pathway for the production of autoinducer (AI)-2 from methionine in Lactobacillus acidophilus NCFM. Open reading frames encoding each enzyme are listed. The S-adenosylmethionine (SAM)-dependent methyltransferase is abbreviated as SAM-MT. The final step involves the nonenzymatic circularization of 4,5-dihydroxy-2,3-pentanedione into AI-2.

Figure 3.

 Phylogenetic tree of MetK and MetE protein sequences from selected bacterial species. Protein sequence alignment and tree construction was performed using ClustalX (Thompson et al. 1997) and Mega2 (Kumar et al. 1994). MetK and MetE sequences were obtained from GenBank and represent a cross-section of available sequences with focus on the lactic acid bacteria. Open reading frames from Lactobacillus acidophilus NCFM showing similarities to the selected sequences are shown in bold.

Highly conserved homologues of the AI-2 production gene luxS are present in over 55 species of Gram-negative and Gram-positive bacteria (Kaper et al. 2005). Figure 4 shows a comparison of the region of six LAB genomes that show the highest levels of similarity in luxS to L. acidophilus NCFM. It is remarkable that, although the degree of similarity of luxS is in the range of 90% and above, the genomic organization in each strain is distinct. Specifically, the relatively high number of integrases and transposases in Lactobacillus reuteri suggests the acquisition of the gene via horizontal gene transfer. Also notable is the presence of a number of permeases and transporters with different specificities.

Figure 4.

 Gene organization of regions in selected lactic acid bacterial (LAB) strains showing the highest degree of similarity (indicated between brackets) to luxS in Lactobacillus acidophilus. LAB strains: NCFM, L. acidophilus; ATCC 11842, Lactobacillus delbrueckii subsp. bulgaricus; NCC 533, Lactobacillus johnsonii; ATCC 33323, Lactobacillus gasseri; F275, Lactobacillus reuteri; 100-23, L. reuteri; ATCC 25745, Pediococcus pentosaceus. The diagram is drawn at approximate scale. (inline image) Transporters/Permeases; (inline image) Peptidases; (inline image) Surface proteins; (inline image) Integrases; (inline image) Transposases; (inline image) Hypothetical/Conserved Hypothetical and (inline image) Pseudogenes.

AI-2 production

In order to determine when AI-2 was produced during the growth phase of L. acidophilus NCFM, supernatants were harvested at 0, 3, 6, 9, 12, 16, 20, 24, 36 and 48 h from triplicate cultures. Prepared supernatants were stored at 4°C until the end of the time course of the experiments and all samples from each culture were assayed together in separate 96-well microtitre plates. AI-2 production during the growth of L. acidophilus is shown in Fig. 5. Coinciding with a rapid drop in pH, a noticeable increase in AI-2 production occurs during the logarithmic phase of growth, at which time the abundance of AI-2 in the media maintains a constant level throughout the stationary growth phase.

Figure 5.

 Autoinducer (AI)-2 production during the growth phase of Lactobacillus acidophilus NCFM was detected from a sterile neutralized supernatant (striped bars) and expressed as mean luminescence of the parent strain/mean luminescence of the LuxS mutant strain. Growth of L. acidophilus NCFM is represented by OD600 (inline image) and the drop in pH (inline image) coincides with rapid AI-2 production during the logarithmic phase of growth. Each value represents the mean of triplicate experiments and error bars are 1 SD from the mean.

Inactivation of luxS and phenotypic characterization of the LuxS mutant strain

To investigate the potential effect of AI-2 on the adherence phenotype of L. acidophilus, an isogenic chromosomal mutation was introduced into the luxS gene, LBA1081. For insertional inactivation, a 471 bp internal fragment of LBA1081 was PCR-amplified and cloned into pORI28. The resulting plasmid, pTRK854, was then transformed by electroporation into L. acidophilus NCFM harbouring the temperature-sensitive helper plasmid pTRK669. Successful integrants of pTRK854 into LBA1081 were selected and confirmed by PCR and southern hybridization (data not shown). No AI-2 activity was detected from the supernatant of the LuxS mutant (NCK1765) at any time point tested over 48 h of growth. Additionally, no differences in the profile of carbohydrate utilization were observed between the mutant and wt strains, as determined by API50 CH tests. Furthermore, the ability of the mutant strain to grow in the presence of bile, NaCl and at acidic pH (adjusted with lactic acid) was compared with the control used in the adhesion experiments (NCFM::lacL, NCK1398). No significant differences were observed between the LuxS mutant strain and the control NCK1398 strain in the complex MRS medium (data not shown). However, the LuxS mutant showed a reduced level of final biomass in a SDM compared with the control (Fig. S1). Marginal differences in μmax values were observed between the LuxS mutant compared with the control NCK1398 in SDM at pH 5·5, and in SDM supplemented with 5% NaCl (Fig. S1).

The LuxS mutant strain, NCK 1765, was tested for the ability to adhere to Caco-2 cells using bacterial cells from the mid-log phase of growth. A significant decrease in adhesion (Student’s t-test, P < 0·001) was observed for the LuxS mutant strain (Fig. 6) compared with the control (NCK1398). No difference in adherence was observed between NCK1398 and the wt L. acidophilus NCFM (data not shown). Also, wt NCFM cells taken after holding for~1 h at 37°C under concentrated (final pH 4·5) vs non-concentrated (final pH 4·5) conditions did not exhibit a difference in adherence levels to Caco-2 cells, indicating that neither pH exposure nor cell concentration was a factor (data not shown).

Figure 6.

 Adherence ability of the LuxS mutant strain of Lactobacillus acidophilus NCFM to Caco-2 cells compared with the NCFM::lacL control strain.

Discussion

LuxS converts (SRH) into homocysteine, which is later converted into methionine by MetE. The compound 4,5-dihydroxy-2,3-pentanedione (DPD) is a product of the first reaction and then undergoes a spontaneous chemical reaction to yield AI-2. Inactivation of the luxS gene in L. acidophilus NCFM resulted in multiple effects as a consequence of its dual function, as reported in Lactobacillus rhamnosus (Lebeer et al. 2007) and Streptococcusmutants (Sztajer et al. 2008). This and a previous study (Azcarate-Peril et al. 2005) demonstrated increased expression of metE (LBA1080) and luxS (LBA1081) in acidified media, as compared with the control (NCFM::lacL; NCK1398). Moreover, both studies provided evidence for AI-2 participation in the stress response, and the current study, adhesion, in vitro. The LuxS-deficient strain of L. acidophilus did not produce AI-2 and showed a significant decrease in adhesion to Caco-2 epithelial cells when the population was harvested during the logarithmic phase of growth. No growth defects were observed in the LuxS mutant compared with the control strain in the MRS broth, a nutritionally complex medium, under any of the studied conditions. However, a metabolic effect of the mutation, probably resulting from the inability of the LuxS mutant to convert SRH into homocysteine, was detected in an SDM. The mutant showed lower final cell density in SDM and a reduced μmax when inoculated in SDM broth at pH 5·5. The results from this and previous studies suggest a potential connection between the two-component regulatory system (2CRS) and LuxS given that both mutations resulted in growth defects under varied conditions and reduced adhesion capabilities (Azcarate-Peril et al. 2005, unpublished observations). A cellular metabolic ‘shift’ with acid accumulation, coinciding with increasing cell concentration, is likely to result in cascading signalling events affecting both metabolism and environmental interactions.

The harvested bacterial cells were originally suspended in fresh MRS (pH 6·4), but the pH at the end of the 1 h incubation dropped to 4·5 in the concentrated sample and only 5·5 in the nonconcentrated sample. Wildtype NCFM cells taken from both these conditions did not exhibit a difference in adherence levels, indicating that pH exposure of the cells during growth was not a factor. A previous study reported an increase in the adhesion of Clostridium difficile when pre-exposed to pH 4 (Waligora et al. 1999). Both low-pH stress and bile-salt stress have been shown to increase enteropathogenic E. coli adhesion to Hep-2 cells (de Jesus et al. 2005). An increase in adhesion was also observed for various lactobacilli when suspended in buffer at pH 3 and 4 during the adhesion assay (Greene and Klaenhammer 1994). In contrast, adhesion of Lactobacillus fermentum increased in pH 2 buffer when compared with the same buffer at pH 7 (Henriksson et al. 1991). However, increased adhesion at a lower pH may result from other factors as high levels of adhesion of L. acidophilus BG2FO4 occurred in spent culture supernatant at pH 4·2, but not when fresh MRS was acidified to pH 4·2 (Greene and Klaenhammer 1994). Other studies have implicated the involvement of an adhesive factor secreted in the stationary phase, although this factor has not been identified (Coconnier et al. 1992; Bernet et al. 1993; Granato et al. 1999).

Among the stress genes differentially expressed in response to the concentration and incubation of cell cultures but not to pH variations were three genes encoding Clp proteins. Clp proteins play a very important role in the removal of stress-damaged proteins in the cell in low GC Gram-positive bacteria. In fact, inactivation of clp genes in Lactobacillales results in a pleiotropic range of phenotypes including general stress sensitivity, aberrant cell morphology and failure to initiate developmental programmes (reviewed by Frees et al. 2007). Additionally, Clp proteins have been implicated in biofilm formation and virulence of Staphylococcus epidermidis (Wang et al. 2007). Considering the growing number of studies that present evidence of the relationship between quorum sensing and biofilm formation (recently reviewed by Irie and Parsek 2008), a relationship between LuxS, Clp proteases and adhesion might be expected under some environmental conditions. Microarray studies to compare the overall gene expression of the LuxS mutant vs the wt strain are currently underway.

This study contributes to the functional characterization of a widely distributed signalling system. As stated by Hardie and Heurlier (2008), both products of LuxS have important functions, homocysteine in central metabolism and 4,5-dihydroxy-2,3-pentanedione in quorum sensing. Hence the disruption of LuxS can have different consequences in different strains and environments. One of the effects of the disruption of luxS in L. acidophilus resulted in a significant loss of adhesive ability to intestinal epithelial cells in vitro suggesting that the AI-2 signalling molecule is one component that likely affects the properties that mediate intimate interactions with the intestinal epithelium.

Acknowledgements

This research was supported by Cooperative State Research, Education and Extension Service (CSREES) NRI-competitive grant no. 2005-35503-16167. Additional programme support was provided by Danisco USA Inc., the North Carolina Dairy Foundation and the NIH Biotechnology Training Program. The authors wish to thank Sarah O’Flaherty for conducting adhesion experiments, Dr Eric Altermann (AgResearch, Palmerston, NZ) and Dr Jose M. Bruno-Barcena for their technical assistance and advice and Dr Bonnie Bassler for kindly providing V. harveyi BB170 and V. harveyi BB152 used in the detection of AI-2.

Ancillary