A Mutation in the luxS Gene Influences Listeria monocytogenes Biofilm Formation

Abstract

Using a Vibrio harveyi reporter strain, we demonstrated that Listeria monocytogenes secretes a functional autoinducer 2 (AI-2)-like signal. A luxS-deficient mutant produced a denser biofilm and attached to a glass surface 19-fold better than the parent strain. Exogenous AI-2 failed to restore the wild-type phenotype to the mutant. It seems that an intact luxS gene is associated with repression of components required for attachment and biofilm formation.

Listeria monocytogenes is a food-borne pathogenic bacterium which is ubiquitous in the outdoor environment. The bacterium can persist in food-processing environments over many years, which is an important cause of food contamination (4, 13, 16, 26). L. monocytogenes attaches to and forms biofilms on numerous surfaces (3, 12, 14), and therefore there is a growing interest in understanding the molecular basis of these processes. In some pathogenic bacteria, the luxS gene was found to be involved in biofilm formation (6, 7, 18, 20, 25, 28, 29). The luxS gene encodes S-ribosylhomocysteinase, an enzyme which catalyzes the hydrolysis of S-ribosylhomocysteine to homocysteine and 4,5-dihydroxy-2,3-pentadione (DPD), which serves as a precursor of autoinducer 2 (AI-2) (27). AI-2 was found to be one of the autoinducers that regulate bioluminescence in the marine bacterium Vibrio harveyi (2), and AI-2-like activity has been discovered in the culture supernatants of numerous eubacteria (21). Orthologs of luxS have been identified in various bacteria, including gram-negative and gram-positive species (19, 24, 27). These findings led to the suggestion that AI-2 is a universal language for interspecies communication (2, 9). Presently, nothing is known regarding the occurrence of AI-2-like activity in Listeria and its possible role in biofilm formation.

Preparation of cell-free culture fluids and detection of AI-2.

L. monocytogenes EGD serovar 1/2a was cultured in brain heart infusion (BHI; Difco Laboratories, Detroit, MI) broth, and growth was monitored periodically by checking the optical density at 600 nm (OD₆₀₀). At each time point, a 1-ml sample was taken and centrifuged (14,000 × g, 5 min, 4°C), and the supernatant was further clarified by filtration (0.2-μm filter). The clarified spent medium is referred to as conditioned medium (CM). The presence of AI-2-like activity in the CM was tested using the Vibrio harveyi BB170 reporter strain (AI-1 sensor negative and AI-2 sensor positive), as described previously (23). CM derived from V. harveyi BB152 (expresses AI-2 but not AI-1) served as a positive control. Vibrio strains (kindly provided by B. Bassler, Princeton University) were grown in autoinducer bioassay medium (11) at 30°C with shaking (150 rpm).

Inactivation of luxS_Lm by site-directed mutagenesis.

An in-frame deletion within the L. monocytogenes luxS ortholog (luxS_Lm) was constructed by allelic replacement, as depicted in Fig. 1 and Table 1. Plasmid pKSV7 is a temperature-sensitive, gram-positive Escherichia coli shuttle vector (22) (kindly provided by E. Gouin and P. Cossart, Institut Pasteur, France). The recombinant plasmid (pKSV7::luxS_Lm) was introducedl into E. coli DH5α by electroporation, and the presence of the specific in-frame deletion was verified by sequence analysis of the recombinant plasmid using the M13-F and M13-R primers. The truncated luxS gene contained a deletion of 258 bp and included codons for two new amino acids (leucine and glutamine) that were introduced by the addition of a new PstI site (see primers SV10 and SV11 in Table 1). The recombinant plasmid was then introduced into L. monocytogenes by electroporation, and allelic replacement was achieved essentially as described before (17). Finally, the presence of the specific deletion in the Listeria genome was confirmed by PCR amplification of chromosomal DNA, using primers derived from flanking chromosomal DNA regions (SV53/SV54) (Fig. 1), and sequence analysis.

FIG. 1.

Schematic presentation of the strategy used to construct the luxS deletion mutant by allelic replacement. The DNA region flanking the putative luxS sequence (lmo1288) is shown together with the primers used for PCR amplifications. The details of the primers are listed in Table 1. The two PCR-amplified fragments were initially cloned into pUC18, and the DNA fragments were released from the recombinant plasmids by digestion with PstI/EcoRI and PstI/BamHI (not shown). The two DNA fragments were ligated to each other and to pKSV7 precut with EcoRI and BamHI. The truncated luxS gene was introduced into the L. monocytogenes genome via allelic replacement. E, EcoRI; B, BamHI; P, PstI; X, XbaI.

TABLE 1.

Primers used in this study

Primer	Sequencea	Locationb
SV9	CAGGTGAAGAAATCCGTTCTGCTATT	3′ region of the parC gene (1312083-1312108)
SV10	CTGCAGTTTAAAGCGAACATCATATT	5′ region of luxS (1312989-1312969)
SV11	CTGCAGAGTTTAGAAGGTGCAAAAGC	3′ region of luxS (1313247-1313266)
SV12	ATCATCAATATTTGTTCGACTCAAATCTAA	Downstream of luxS (1314112-1314083)
SV53	ATTCACCACATCTTGGCTTTCT	Between luxS and lmo1289, downstream of SV12 (1313788-1313767)
SV54	AAGGCGAACCAACTTCTATTTG	3′ region of parC, upstream of SV9 (1312704-1312725)
M13-F	GTAAAACGACGGCCAGTG	M13 forward primer
M13-R	GGAAACAGCTATGACCATG	M13 reverse primer

^aNucleotides in bold represent a PstI site that was added to the 5′ region of the primer.

^bThe numbers represent nucleotide positions in the L. monocytogenes genome (GenBank/EMBL accession number NC_003210).

Expression of luxS_Lm in E. coli.

A chromosomal DNA fragment was PCR amplified using L. monocytogenes strain EGD DNA as the template and primers SV53 and SV54 (Table 1). The fragment was cut with EcoRI at two authentic sites within the L. monocytogenes genome (Fig. 1) to yield a 584-bp fragment spanning the entire luxS_Lm open reading frame (ORF). The DNA fragment was then ligated into pUC18 precut with EcoRI and transformed into E. coli DH5α. A recombinant pUC18::luxS_Lm clone was isolated and tested for the expression of listerial AI-2 following incubation in Luria-Bertani medium (Hy-Lab, Rehovot, Israel) and induction with IPTG (isopropyl-β-d-thiogalactopyranoside; 1 mM). The presence of AI-2 activity in the culture supernatant was tested as described above. Bioluminescence was determined in 0.5-ml tubes with Biocounter M1500 (Lumac, The Netherlands).

Biofilm formation assays.

L. monocytogenes was grown in BHI broth at 37°C for 24 h. The culture was adjusted to an OD₆₀₀ of 1.0, and the bacteria were diluted 1:100 in fresh BHI broth. Bacteria (0.5 ml) were transferred to a 24-well tissue culture-treated polystyrene plate (Greiner, Germany) and incubated in triplicate at 37°C for 48 h under static conditions. The wells were washed three times with 500 μl of double-distilled water (DDW), and quantitative assessment of the biofilm was performed by staining with 0.1% crystal violet (CV), essentially as described previously for 96-well plates (8). However, the volumes were adjusted to fit the larger scale, and the washed plates were dried at 65°C for 1 h. Experiments to functionally complement the luxS mutation with CM derived from the wild type (wt) were carried out in 96-well plates. CM was added at 10% (vol/vol). The absorbance of the CV solution at 595 nm was determined using a plate reader (ELx 800UV; BioTec Instruments, Inc.). The amount of biofilm is presented as the average absorbance (OD₅₉₅) of the extracted CV solutions from two independent experiments with three replicate wells each.

Attachment assay.

An overnight culture of L. monocytogenes was adjusted with BHI broth to an OD₆₀₀ of 1.0, and 2 ml was added to glass coverslips (22 by 22 cm) submerged (horizontally) in a six-well plate (Greiner). The plate was incubated for 0.5 h at 37°C, washed thrice with DDW, and stained with CV. Stained bacteria were counted under a microscope. The number of attached bacteria is presented as the average number of bacteria per field (magnification, ×40) for 14 fields.

Confocal microscopy.

For microscopic assessment, biofilms were grown at 37°C in BHI broth on glass coverslips (22 by 22 cm) submerged (horizontally) in a six-well plate (Greiner). At the indicated time points, the coverslips were washed three times, fixed (4% glutaraldehyde), and stained with acridine orange (0.1%). Samples were examined under a confocal laser scanning microscope (LSM FluoView 500; Olympus, Japan). Sections were taken every 0.5 μm, and the TIFF images of all sections were processed using AnalySIS image analysis software (Olympus) to produce a three-dimensional (3-D) structure.

Statistical analysis.

The data were analyzed by the SAS general linear model (version 8.02; SAS Institute Inc., Cary, NC). Differences were considered significant for P values of <0.05.

Identification and localization of L. monocytogenes luxS ortholog.

Recent annotation of the L. monocytogenes genome (December 2005) identified lmo1288 as S-ribosylhomocysteinase (LuxS). However, the luxS ortholog was first identified inadvertently, during a study of L. monocytogenes topoisomerase IV (15). We have shown further evidence that the lmo1288 ORF encodes the highly conserved LuxS protein. A peptide of 144 amino acids was identified within the 172 amino acids of V. harveyi LuxS which displays 41% identity and 59% similarity with the putative translated L. monocytogenes LuxS protein (Fig. 2). The 144 amino acids are encoded within an ORF spanning 467 nucleotides (nt; positions 1312857 to 1313324), which is likely luxS. The luxS gene resides downstream of the parE and parC genes and is followed by two unknown ORFs showing similarity to the internalin gene (Fig. 3). A terminator-like structure is located between parC and luxS, suggesting that luxS is not part of the parEC operon. A perfect −10 box (TATAAT) is present 26 nt upstream of the luxS ATG start codon (1312857), and a possible −35 box (ATGCAA) is found 16 nt upstream of the TATA box (data not shown). A terminator-like structure was identified within the intergenic region downstream of the luxS ORF at nt 1313333 to 1313370.

FIG. 2.

Sequence alignment of LuxS_{V. harveyi} and putative LuxS_Lm. Amino acid sequence alignment between V. harveyi LuxS (accession no. AAD17292) and the putative L. monocytogenes LuxS protein (accession no. NP_464813) was performed using the Blastp program (1) at http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi.

FIG. 3.

Linear map of the genomic DNA region flanking L. monocytogenes luxS. Known and predicted ORFs and their orientations are shown. Annotations of genes were derived from the L. monocytogenes EGD genome (accession no. NC_003210). ORFs lmo1284 and lmo1285 encode conserved hypothetical proteins similar to the Bacillus subtilis YneS and YneT proteins, respectively. ORF lmo1291 encodes a hypothetical protein which displays similarity to a B. subtilis acyltransferase (YrhL protein).

L. monocytogenes luxS gene encodes a functional AI-2 molecule.

To test whether luxS functions in L. monocytogenes, we constructed an in-frame deletion within the putative luxS sequence, as depicted in Fig. 1. The wt and the luxS deletion mutant had similar growth rates (P = 0.94), as reflected by the identical slopes of the exponential growth curves (Fig. 4). However, the OD₆₀₀ of the mutant at early stationary phase was 19% lower than that of the wt strain (P < 0.0001). The wt expressed the maximal level of AI-2-like activity at the mid-exponential growth phase, and the activity decreased when the bacterium entered the stationary phase of growth. In contrast, the luxS mutant strain had reduced AI-2-like activity compared to the wt strain, and at 5 h, the mutant activity was only 2.7% that of the wt strain (Fig. 4).

FIG. 4.

Disruption of luxS abolishes AI-2 production in L. monocytogenes. The expression of AI-2 activity, as determined through the ability to activate luminescence production of a V. harveyi BB170 reporter strain (left axis), is depicted in relation to the growth of the wt and luxS mutant strains (right axis). Sterile BHI medium and cell-free supernatants from overnight cultures of V. harveyi BB170 served as negative and positive controls, respectively. Light production was monitored with a Wallac VICTOR² 1420 multilabel counter (Perkin-Elmer). The data presented were obtained 6 h following inoculation of V. harveyi. For negative and positive controls, we used sterile BHI medium, autoinducer bioassay medium, and CM derived from V. harveyi BB152 (expresses AI-2 but not AI-1). Bioluminescence values are presented in RLU. The data presented are the means derived from triplicate samples in a representative experiment. Optical densities at 600 nm are presented as averages of four replicates. Error bars were omitted for clarity.

To further demonstrate that luxS is responsible for the synthesis of AI-2, the luxS ORF was expressed in E. coli DH5α, which is AI-2 deficient due to a frameshift mutation in luxS (24). Following induction with IPTG, the CM derived from the recombinant strain induced substantial bioluminescence (14,658 ± 192 relative light units [RLU]) upon the V. harveyi BB170 strain. In contrast, CM derived from E. coli DH5α containing the vector (pUC18) alone or no plasmid induced very low bioluminescence activity (87 ± 1 RLU). These results suggest that L. monocytogenes luxS encodes a functional AI-2 or AI-2-like molecule.

Mutation in luxS enhances attachment and biofilm formation.

The capacity of the wt and mutant strains to develop biofilms in 24-well polystyrene plates was determined using the CV assay. The luxS mutant formed considerably more biomass (OD₅₉₅, 1.84 ± 0.36) than the wt strain (OD₅₉₅, 0.51 ± 0.10). Since we used an in-frame deletion mutation, it is unlikely that the phenotype of the mutant resulted from a polar mutation. Moreover, an independently isolated luxS mutant also expressed a similar phenotype (data not shown). Thus, we concluded that luxS mutation derepresses biofilm formation. 3-D analysis of a 3-day-old biofilm confirmed that the mutant strain formed a much thicker biofilm than did the wt strain. Reconstitution of the images taken every 0.5 μm demonstrated an irregular structure of the mutant's biofilm, with regions showing a dense and relatively thick biofilm. The maximal recorded depth of the mutant's biofilm was 9 μm. In contrast, during the same period, only single wt cells were observed attached to the surface, with a small number of aggregates (Fig. 5). Microscopic images taken at 1, 2, 3, and 6 days showed that ΔluxS bacteria accumulated and formed a dense biofilm for up to 3 days. On day 6, only a few bacteria were seen attached to the surface, suggesting that detachment had occurred (Fig. 6). In contrast, the number of attached wt bacteria was much smaller and remained almost constant from 2 to 6 days. These results imply that the wt strain is repressed in its transition from the attachment and microcolony stage to mature biofilm. The decline in biofilm mass of the ΔluxS strain might be explained by the utilization of available nutrients and accumulation of toxic compounds, which might have induced bacterial detachment.

FIG. 5.

Structures of 3-day-old biofilms of luxS mutant and wild-type strains grown at 37°C on a glass surface in BHI medium. Following washing of the glass coverslips with DDW, the bacteria were stained with acridine orange and viewed under a confocal microscope. (A) Top view of a representative field (×40) composed of stacks of images showing a 3-D model of the biofilm formed by the ΔluxS mutant strain, created by AnalySIS software (Olympus). The maximum depth of the biofilm was 9 μm. Highly fluorescent (white) regions represent sites with thicker and denser biofilms, while less fluorescent areas represent sites with fewer bacteria. (B) Top view of a representative microscopic field (×40) showing very few wt bacterial cells and a small number of aggregates attached to the surface at 3 days.

FIG. 6.

Confocal microscopy of biofilm development on a glass surface by L. monocytogenes strains grown at 37°C in BHI medium. Bacteria were stained with acridine orange, and stack of images were taken every 0.5 μm at 1, 2, 3, and 6 days. The white lines in the panels represent the orientation in the x-y perspective, where z sections are shown. x-z and y-z sagittal images are shown at the bottom and right sides of images, respectively. Fluorescence is shown in gray. z-dimension scale, 8.5, 8.0, 10.0, and 6.5 μm at 1, 2, 3, and 6 days, respectively. Bar, 50 μm.

Attachment assays demonstrated that the ΔluxS cells attached significantly better (249 ± 65 cells per field) to a glass surface than did wt cells (13 ± 14 cells per field; P < 0.0001). Thus, repression of biofilm development in the wt strain might be related, at least partially, to poor attachment capacity, possibly due to repression of necessary adhesins.

There is inconsistency in the literature regarding the influence of luxS and AI-2 on the development of biofilms (7, 18, 20, 28). It was recently shown that the addition of purified AI-2 increased both the mass and thickness of E. coli biofilms (10). In contrast, a number of other studies have clearly demonstrated that mutation in the luxS gene significantly enhances biofilm formation (6, 25, 29). Experiments to functionally complement the luxS_Lm mutation with exogenous AI-2 (CM derived from wt L. monocytogenes) failed to show any significant effect on biofilm mass (data no shown). Thus, it can be concluded that the luxS gene is involved in the repression of attachment and development of a biofilm with a 3-D complex structure via a mechanism unrelated to AI-2. A recent study published during the revision process of the manuscript also reported that mutation in the L. monocytogenes luxS gene resulted in increased biofilm mass at 25°C. Interestingly, in vitro-synthesized S-ribosyl homocysteine but not AI-2 was found to moderately increase biofilm mass (5). More study is required to elucidate the precise mechanism(s) by which luxS controls L. monocytogenes attachment and biofilm formation.