|Home | About | Journals | Submit | Contact Us | Français|
Listeria monocytogenes is a food-borne pathogen that is capable of living in harsh environments. It is believed to do this by forming biofilms, which are surface-associated multicellular structures encased in a self-produced matrix. In this paper we show that in L. monocytogenes extracellular DNA (eDNA) may be the only central component of the biofilm matrix and that it is necessary for both initial attachment and early biofilm formation for 41 L. monocytogenes strains that were tested. DNase I treatment resulted in dispersal of biofilms, not only in microtiter tray assays but also in flow cell biofilm assays. However, it was also demonstrated that in a culture without eDNA, neither Listeria genomic DNA nor salmon sperm DNA by itself could restore the capacity to adhere. A search for additional necessary components revealed that peptidoglycan (PG), specifically N-acetylglucosamine (NAG), interacted with the DNA in a manner which restored adhesion. If a short DNA fragment (less than approximately 500 bp long) was added to an eDNA-free culture prior to addition of genomic or salmon sperm DNA, adhesion was prevented, indicating that high-molecular-weight DNA is required for adhesion and that the number of attachment sites on the cell surface can be saturated.
The food-borne pathogen Listeria monocytogenes is known to persist in food processing plants (28, 48), and it has been reported that some strains of this species are capable of forming biofilms (2, 16). The mechanisms of biofilm formation have not been elucidated, but this process seems to depend on factors such as temperature and inducing compounds (14). One inducing compound is NaCl (22), but ethanol, isopropanol (14), quorum sensing (36), and an increasing temperature (8, 14, 38) also seem to enhance attachment and biofilm formation, whereas an acidic pH reduces adhesion (17, 38, 43). Furthermore, at 30°C flagellum-based motility seems to be a specific determinant for the initial adhesion (23, 42) and biofilm formation (23); however, it has recently been reported that in time nonflagellated mutants can produce hyperbiofilms (42).
Since bacteria adhering to surfaces, both in biofilms and as single cells, exhibit increased resistance to sanitizers and antimicrobial agents (10, 41), examining the essential steps in adhesion and biofilm formation is important in order to develop new and improved sanitation processes.
Extracellular DNA (eDNA) is a ubiquitous component of the organic matter pool in soil, marine, and freshwater habitats (26), but it is also found in environments as diverse as tissue cultures and the blood of mammals (11, 25). The presence of eDNA in the matrix of multicellular structures has recently been reported to influence the initial attachment and/or biofilm structure of Pseudomonas (1, 47), Streptococcus (29), and Staphylococcus (21, 33, 34) species.
The prevalence of eDNA in nature appears to be associated with both lysis of cells and active secretion. The concentrations of eDNA released can be up to 2 μg g−1 soil (30) and up to 0.5 g (m2)−1 in the top few centimeters of deep-sea sediment (where more than 90% of the DNA is extracellular) (5). In the deep sea eDNA plays a key role in the ecosystem, functioning as a nitrogen and phosphorus reservoir (5). At present, there are different theories concerning both the function and the release of eDNA in multicellular structures. The presence of eDNA could be a result of either cell lysis (33, 34) or vesicle release (47), whereas active transport is a more speculative explanation. The role of eDNA in biofilm structure has not been revealed yet, but various functions, including a role as a structural component, an energy and nutrition source, or a gene pool for horizontal gene transfer (HGT) in naturally competent bacteria, can be envisaged.
Until now there have been no studies of L. monocytogenes eDNA as a possible matrix component in relation to adhesion and biofilm development. In this study, we determined for the first time the presence of L. monocytogenes eDNA, its origin, and its role as a matrix component for both single-cell adhesion and biofilm formation using static assays, as well as flow cell systems. Furthermore, we showed that an additional component is necessary for eDNA-mediated adhesion.
Previously, an analysis of L. monocytogenes cell aggregation (14) was conducted with the first fully annotated strain, L. monocytogenes strain EGDe (12), as the model organism. Hence, the EGDe strain was chosen as a reference strain in this study, but other isolates having diverse origins were included to determine possible variations in the species (Table (Table1).1). Isolate 412 was also selected for further study since it produced larger-than-average amounts of biofilm at 37°C (14).
To view biofilm formation in real time, the EGDe and 412 strains were both transformed with plasmid pNF8 (9), an optimized green fluorescent protein (GFP) expression vector. Listeria strains were made competent by using a protocol described by Monk et al. (27).
The complex medium used was brain heart infusion (BHI) broth (Oxoid Ltd.), and the defined medium used was Hsiang-Ning Tsai medium (HTM) (44) with 3% (wt/vol) glucose (23). For the flow cell biofilm system we used AB minimal medium (32) supplemented with BHI at a final concentration of 2.5% (vol/vol). Erythromycin (Sigma-Aldrich) was used at a concentration of 10 μg ml−1 for selection for pNF8 (9); however, because of the continuous flow, a concentration of only 5 μg ml−1 was used in the flow cell biofilm system. DNase I (Sigma-Aldrich) was used at a concentration of 100 μg ml−1 in the static adhesion assay and at a concentration of 10 μg ml−1 for the flow cell system.
Overnight (ON) cultures were incubated at 37°C with shaking (200 rpm) in BHI medium unless stated otherwise. The ON cultures were grown for approximately 11 to 14 h.
To quantify the effect of DNase I on the initial attachment, an ON culture was diluted to obtain an optical density at 600 nm (OD600) of 0.5 in phosphate-buffered saline (PBS) using a UVmini-1240 spectrophotometer (Shimadzu). One milliliter was then transferred to a coverglass cell culture chamber (Nunc) and incubated statically for 30 min at 37°C, using enzymes for the treated cultures and PBS for the control. After incubation the chambers were washed three times with PBS to remove planktonic cells. The sessile cells were visualized with the SYTO9 stain (Molecular Probes), and images were acquired using a Zeiss LSM510 microscope (Carl Zeiss Microscopy) and a 63×/0.95 objective. Proteinase K (Roche), RNase A (Sigma), and heat-inactivated DNase I (2 h at 65°C) were employed as controls for the enzymatic effect of DNase I on attachment. All of the enzymes were used at a final concentration of 100 μg ml−1.
eDNA from ON cultures of 41 L. monocytogenes isolates were precipitated by standard ethanol precipitation methods (37). The presence of DNA was verified using an agarose gel. The purified product was treated with DNase I (Sigma-Aldrich) and RNase (Amersham Bioscience) to verify that the bands corresponded to the nucleic acids. RNase was used at a final concentration of 100 μg ml−1.
The origin of eDNA in biofilms formed by the EGDe and 412 strains was investigated by performing PCRs for six genes distributed evenly over the genome of L. monocytogenes EGDe. The genes examined were lmo0003, lmo0461, lmo0971, lmo1467, lmo1952, and lmo2430.
A 150-μl portion of HTM with 3% (wt/vol) glucose was placed in each well of U-bottom polystyrene 96-well microtiter plates (Sterilin), and eight replicate wells were used for each analysis. An ON culture (prepared as described above) was added to obtain a final OD600 of approximately 0.05, and the plates were incubated at 37°C. At 0, 9, 24, and 48 h, DNase I was added, and the plates were incubated again. After 51 h the adhered cells were assayed with a Biomek 2000 laboratory automation workstation (Beckman Coulter). To assess growth, the OD595 was determined. A pipetting robot removed the culture, washed the cells twice with 200 μl of a 0.9% (wt/vol) NaCl solution, and added 170 μl of a 0.1% (wt/vol) crystal violet solution in 0.9% (wt/vol) NaCl. After 15 min of staining the crystal violet was removed, two wash cycles were performed with 200 μl of a 0.9% (wt/vol) NaCl solution, and finally 170 μl 96% (vol/vol) ethanol was added. After 1 h the optical density at 595 nm of the eluted crystal violet was determined using a DTX800 multimode detector (Beckman Coulter). Both the DNase I-treated cultures and the controls were incubated for a total of 51 h prior to quantification of the crystal violet; hence, the duration of the DNase I treatment ranged from 3 to 51 h. However, a microtiter biofilm that was not treated with DNase I was also assessed after 24 h to review the progress of biofilm formation.
Three independent experiments were conducted for all of the strains examined. Outliers in each experiment were removed by a Grubbs outlier test.
To ensure that the effect was not temperature dependent, the EGDe and 412 strains were also assayed at 25°C; however, the incubation time was adjusted due to the reduced growth rate. In this experiment DNase I was added at 0, 13.5, 36, and 72 h, and staining was performed 75 h after inoculation for all of the conditions evaluated.
A modified version of the biofilm setup described by Sternberg and Tolker-Nielsen (40) was employed. To ensure better biofilm formation, the incubation time with the glass surface pointed down was increased to 48 h, and there was no flow during the first hour. After 48 h the flow chambers were turned so the glass surface was pointing up. The treated biofilm and the control biofilm were visualized after 3 and 4 days. DNase I treatment was performed for 18 h prior to visualization on day 4 (78 h after inoculation). Images were acquired with a Zeiss LSM510 microscope (Carl Zeiss Microscopy) and a 40×/1.3 objective. The images were then treated using the image program IMARIS. To evaluate the biomass under each condition, nine images from each of two channels were quantified using COMSTAT (18).
We examined the possibility that attachment was stimulated by addition of DNA to an exponentially growing culture. This growth phase was selected to avoid cell death and hence reduce the amount of eDNA in the culture. An ON culture (prepared as described above) was reinoculated into BHI medium to obtain an OD600 of 0.05 and grown to an OD600 of 0.5 (UVmini-1240 spectrophotometer; Shimadzu) at 37°C with shaking (200 rpm). This step was repeated to ensure that exponential growth occurred. DNase I (100 μg ml−1) was added 30 min before the cells in a fraction of the culture were harvested to remove any eDNA in the suspension (a control without DNase I was also included). Cells were harvested at an OD600 of 0.5 and washed twice with PBS buffer to remove traces of the medium and enzyme using gentle centrifugation at 3,000 × g to avoid cell destruction. Two milliliters was then transferred to a coverglass cell culture chamber (Nunc) and incubated statically for 30 min at 37°C. After incubation the chambers were washed three times with PBS to remove planktonic cells. The sessile cells and DNA were visualized with a Zeiss LSM510 microscope (Carl Zeiss Microscopy) and a 63×/0.95 objective. The DNA was stained with SYTOX orange (Molecular Probes).
The DNA added was precipitated as previously described and purified from a 1% (wt/vol) agarose gel using a GFX PCR DNA and gel band purification kit (GE Healthcare). Purified DNA from 2 ml of an ON culture was added to each attachment chamber to obtain a DNA concentration that was in the same range as the concentrations in the other experiments.
The supernatant added was obtained from a 48-h-old culture in BHI medium incubated at 37°C with shaking at 200 rpm. The enzymatically treated supernatant was treated with either DNase I (Sigma) or proteinase K (Roche) or with both enzymes at 37°C for 3 h without shaking. When the supernatant was treated with both enzymes, the DNase I was added 1 h before the proteinase K was added. The enzymes were added at a final concentration of 100 μg ml−1. The amount of supernatant added to the attachment chambers was 2 ml to mimic the concentration used in the other experiments. All of the attachment experiments were repeated three times.
In order to reduce the number of steps for each strain and each condition and to ensure more similar shear stresses during washing, we employed a manual flow cell system in a subsequent attachment assay. In contrast to the flow cell system described above, in the system used for this experiment a 50-ml syringe was used instead of a pump, and the bubble traps were removed. The cells were grown and washed as described above for the other attachment assay and were subsequently resuspended in PBS buffer with and without DNA and with and without other chemicals, all at a final concentration of 3 μg ml−1. They were inoculated into a flow cell chamber as described by Sternberg and Tolker-Nielsen (40) with the substratum facing down. Like the wells in the microtiter adhesion assay, the flow cells contained HTM with 3% (wt/vol) glucose. After 15 min of adhesion, medium was pushed through the flow cell with a syringe to remove any nonadhering cells from the substratum. The adhered cells were visualized with a Zeiss LSM510 microscope (Carl Zeiss Microscopy) and a 40×/1.3 objective. Outliers in each experiment were removed with a Grubbs outlier test, and a paired Student t test was performed.
We examined different types of DNA in combination with peptidoglycan (PG) from Bacillus subtilis (Fluka) or with the building blocks N-acetylglucosamine (NAG) and N-acetyl muramic acid (NAM) (Sigma-Aldrich) or their precursors (glucose and lactic acid, respectively). Genomic DNA was isolated from a cell lysate that was prepared as described by Grimberg et al. (15). DNA from this cell lysate was extracted by standard ethanol precipitation methods (37). The genomic DNA was digested ON with two restriction enzymes, MspI and RsaI. A 200-bp fragment of DNA was also generated, employing the lmo2430 PCR primer previously used to describe the origin of eDNA. The attachment sites were saturated by pretreating the cells with short DNA fragments for 5 min prior to addition of long DNA fragments.
The impact of eDNA on initial cellular attachment in static chambers was investigated by treating a stationary-phase culture with DNase I and heat-inactivated DNase I. It was observed that eDNA had a pronounced effect on the initial attachment of cells in the static chamber assay (Fig. (Fig.1),1), since removing the eDNA enzymatically severely reduced the attachment of cells to the glass surface. Heat-inactivated DNase I did not affect the attachment, which supported the hypothesis that the active enzyme has a role in obstructing attachment (data not shown). Neither treatment with RNase A nor treatment with proteinase K had a significant impact on attachment. Furthermore, centrifuging and washing the cells with PBS buffer in order to remove unbound eDNA had no effect on attachment (data not shown). The latter control results indicate that the eDNA was associated with the cells prior to adhesion in the chamber, which was also observed when the cells were stained with SYTOX orange (data not shown).
All attachment assays with both the EGDe and 412 strains (data for the latter strain are not shown) were performed with and without the GFP plasmid to ensure that there were no adverse effects on adhesion when the pNF8 plasmid was introduced into the two strains. The capacity to attach was not affected by the strain or by plasmid insertion (data not shown).
For characterization of the eDNA released from L. monocytogenes, the nucleic acids in supernatants from stationary-phase cultures of 41 L. monocytogenes isolates were precipitated with ethanol. After the nucleic acids were precipitated and analyzed, it was evident that all of them produced three bands on an agarose gel (Fig. (Fig.2).2). Subsequent enzymatic treatment of nucleic acids from the gel revealed that the top band corresponded to eDNA, whereas the two lower bands contained RNA (data not shown).
To verify that the purified eDNA had a chromosomal origin, PCRs for six genes evenly distributed over the genome of EGDe were performed. These PCRs confirmed that all six genes, regardless of position, were amplified from the purified DNA, and we therefore assumed that the purified eDNA had a chromosomal origin.
The impact of eDNA at different stages of biofilm formation in microtiter biofilm assays was monitored by adding DNase I at various time points. These assays showed that especially early addition of DNase I had a pronounced effect on both attachment and subsequent biofilm stability (Fig. (Fig.3).3). Addition of DNase I up to 24 h after inoculation completely prevented biofilm formation and removed the existing multicellular structures (Fig. (Fig.33).
Between 24 and 48 h the capacity of DNase I to remove biofilm material was reduced. However, DNase I treatment after 48 h could still remove more than 55% of the EGDe biofilm material (Fig. (Fig.3).3). The reduction in the removal effect indicates that other factors besides eDNA may also stabilize the biofilm. However, extensive removal of cell material at late stages of biofilm formation has not been described previously for other bacteria. The strain least affected by DNase I treatment was the 412 strain, but even for this strain there was a >20% reduction in biofilm material when it was treated with DNase I after 48 h.
Regardless of the origin of the isolates examined, their attachment and biofilm formation were sensitive to DNase I treatment (Fig. (Fig.3),3), although there were individual differences.
Based on the crystal violet assay results, we continued our study with strains EGDe and 412. The former organism is a reference laboratory strain, and the latter was the strain least affected by DNase I treatment at the later stages of adhesion and biofilm formation.
When the three-dimensional biofilm structures of the two strains were visualized after 3 and 4 days (Fig. (Fig.4),4), the images showed that for EGDe strain there were large void areas on the substratum, and the microcolonies of this strain were better defined than the flatter and rougher structures produced by the 412 strain. Quantification of the biomass with COMSTAT showed that at 37°C the 412 strain is a better biofilm producer than the EGDe strain, producing more than twice as much biomass after 3 days of growth (Fig. (Fig.4).4). After 4 days the difference was smaller, although the 412 strain still produced considerably more biofilm. Despite the different structures and biomasses, it is clear that treatment with DNase I could remove large amounts of the biofilms, although not all cell material was removed (see Video S1 in the supplemental material). When the biomass was treated with DNase I, the amount was significantly reduced (two-tailed Student t test) to an amount that was less than 80% of the amount of biomass on days 3 and 4.
The requirements for eDNA during adhesion of exponentially growing cells and eDNA-free cells supplied with exogenous DNA in static culture chambers were compared. During all growth phases eDNA could be precipitated. The amount of eDNA present increased over time, reaching a maximum in stationary phase. By use of exponentially growing cells the amount of eDNA transferred from lysed cells was assumed to be minimal. The eDNA-free cell fraction was treated with DNase I and subsequently washed with PBS to remove remnants of DNase I. While exponentially growing cells adhered readily, the cells without eDNA could not adhere to glass (Fig. (Fig.5A).5A). Time series images (Fig. (Fig.5B)5B) showed that the cells dispersed as an immediate response to the DNase I treatment.
Interestingly, this assay also showed that supplying eDNA-free cells with exogenous DNA is not enough to restore the capacity to attach (Fig. (Fig.5A).5A). Attachment could be restored only if the eDNA-free cells were supplied with both exogenous DNA and supernatant. The supernatant was enzymatically treated prior to addition to ensure that the effect was not a result of further exogenous DNA supplementation. Proteins in the supernatant were also degraded by proteinase K, and, as shown in Fig. Fig.1,1, there was no effect on adhesion when proteins from the supernatant were hydrolyzed enzymatically.
The attachment assays showed that if DNA was not present in the inoculum, bacteria were almost incapable of adhering (P 0.001) (Fig. (Fig.6)6) and that addition of salmon sperm DNA alone to a DNase I-treated culture did not restore the capacity to adhere (P < 0.05). Not only did the presence of DNA affect adhesion, but the properties of the DNA also had an impact. When the effects of normally prepared genomic Listeria DNA were compared with those of shorter DNA fragments (either partially digested or produced as 200-bp PCR DNA fragments), it was clear that DNA fragments had a certain length to confer the capacity to adhere (P 0.001 and P 0.001, respectively). Furthermore, pretreatment of an eDNA-free inoculum with short DNA fragments prior to addition of genomic DNA inhibited the adhesion of L. monocytogenes cells (P 0.001).
These results suggest that compounds other than exogenous DNA are necessary to induce adhesion. The results of the static attachment assay further indicated that such a compound is present in the supernatant of an old culture and that it is neither DNA or a protein. We therefore assayed additional adhesion factors which could induce adhesion in combination with eDNA.
Addition of peptidoglycan (PG) to an eDNA-free culture almost fully restored adhesion compared to the early-exponential-phase control. This could suggest that PG alone confers the ability to adhere. However, microscope images showed that the eDNA-free culture was not completely devoid of eDNA (data not shown), which could explain the level of adhesion. When PG was added along with salmon sperm DNA to an eDNA-free culture, attachment was induced (P < 0.05). However, replacing the salmon sperm DNA with normally prepared genomic DNA did not induce adhesion at a level that was significantly different from that obtained with PG alone. Analysis of the two DNA products by gel electrophoresis suggested that the genomic DNA contained fractions of low-molecular-weight DNA and that the molecular weight of the genomic DNA was lower than that of salmon sperm DNA.
We also examined whether either of two PG building blocks, N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM), could restore adhesion when it was added to an eDNA-free culture, but neither NAG nor NAM had the capacity to restore adhesion when it was added as the sole component (P 0.001 and P 0.001, respectively). However, we found that when genomic DNA was added in combination with either NAG or NAM, adhesion could be restored to the level of an early-exponential-phase culture only when NAG was present.
Addition of the precursors of NAG and NAM (glucose and lactic acid, respectively) with genomic DNA to an eDNA-free culture could not restore adhesion (P 0.001).
In the combination adhesion assays combining short DNA fragments with either PG or one of its precursors (NAG or NAM) did not induce adhesion. All of the experiments showed that the level of adhesion was similar to that observed for the eDNA-free culture.
Biofilms have long been defined as surface-associated assemblies of microorganisms embedded in a hydrated self-produced polymeric matrix (3), and extracellular polysaccharides are the most-studied component of this matrix. However, when a BLAST analysis was performed with genes encoding extracellular polysaccharides, such as the genes encoding alginate in Pseudomonas aeruginosa or poly-N-acetylglucosamine (PNAG) in Staphylococcus and E. coli, the results obtained for L. monocytogenes (12) suggested that genes of this organism may not encode such polysaccharides. Based on the results described here, we propose that high-molecular-weight DNA together with N-acetylglucosamine (NAG) can form a polymer with properties resembling those of PNAG of Staphylococcus and E. coli. Such a polymer could support adhesion and biofilm formation.
Supernatants of L. monocytogenes cultures were found to contain genomic DNA, and we suggest that this eDNA has an essential role in efficient attachment of cells to a surface. This property of eDNA has been documented for other bacteria (46) based mainly on the observation that addition of DNase I significantly reduced cellular attachment, resulting in reduced biofilm formation. By using heat inactivation we showed that attachment was hindered due to the enzymatic effect of DNase I rather than the protein itself. Using the chamber attachment assay it was established that the eDNA could not be removed simply by washing the cells with PBS buffer, indicating that the eDNA is not loosely associated with the cells.
Other studies have shown that the impact of eDNA on adhesion decreases as the adhered cells develop into a multicellular structure (33). For Listeria, it was observed that even after 24 h of growth the attached multicellular biomass was completely dispersed by addition of DNase I, and even after 48 h much of the attached biomass could still be removed. This was true for all of the isolates examined, and experiments performed at 25°C confirmed that it was also true at temperatures at which flagella are expressed (data not shown), although several studies have shown that flagellar motility is critical for L. monocytogenes biofilm formation (23, 42). Since we did not examine motility at 25°C, we cannot exclude the possibility that flagella were present only on the cell surface and that the cells were nonmotile. However, we believe that based on our knowledge of flagella and expression these results indicate that DNase I can disperse attached biofilms independent of motility or the presence of flagella.
The finding that DNase I may dissolve even a multicellular biofilm structure has not been described previously for any other bacteria. In the case of Staphylococcus epidermidis biofilms the effect of DNase I decreased with time, and addition more than 12 h after inoculation had little effect (33). The stronger reliance of L. monocytogenes on eDNA for adhesion and further biofilm development could be a consequence of this organism lacking the potential to produce structural components, such as extracellular polysaccharides, like those produced by S. epidermidis. The observed decrease in the efficiency of DNase I for dispersal over time could indicate that the bacteria produce an additional structural component at later times, which hinders complete dissolution of the multicellular structure. Alternatively, in older biofilms eDNA could be less susceptible to DNase I or the enzyme could be trapped in the biofilm. The efficiency of DNase I is even more pronounced in a flow cell than in a microtiter tray. Even after 72 h of biofilm development addition of DNase I dissolved more than 80% of the sessile community. The difference could be due to the continuous flow of DNase I compared to the static nature of a batch culture in a microtiter tray. Depending on the model system employed, the structure of the biofilm varies (35), which could also affect the efficacy of the DNase I treatment.
The importance of eDNA was further established by the finding that enzymatic removal of both RNA and proteins did not alter the capacity to attach, in accordance with observations for S. epidermidis (33). However, in another study proteins were found to have a role in the initial attachment of L. monocytogenes (39). In that study Smoot and Pierson showed that when 0.01% trypsin was added to the attachment medium, the adhesion to stainless steel and Buna-N rubber was reduced by 99.9%. The difference between the two results could have been due to strain, surface, and/or protease variation.
The adhesion assays also showed that the length of eDNA is very important for attachment of L. monocytogenes. By adding DNA having different molecular weights to eDNA-free cells and comparing the results, it was found that whenever low-molecular-weight (LMW) DNA was employed, adhesion was eliminated (Fig. (Fig.6).6). Furthermore, purified high-molecular-weight (HMW) DNA added alone was observed not to support adhesion of L. monocytogenes, independent of the origin (Fig. (Fig.55 and and6).6). The latter finding was supported by the observation that in combination with HMW DNA, supernatant that was enzymatically treated with DNase I and proteinase K could complement adhesion (Fig. (Fig.5),5), suggesting that another component of the supernatant fraction combined with HMW DNA complements adhesion. Here it was found that PG and, more specifically, the NAG part of PG, both of which were present in the supernatant as debris from cell lysis, could act as such a component (Fig. (Fig.6).6). Based on the observations described here, we propose that HMW DNA together with NAG can form a polymer that perhaps resembles PNAG of Staphylococcus or E. coli. Such a polymer could support adhesion and biofilm formation whenever other intercellular matrix components are scarce or absent. The hypothesis that polymerized NAG is required was further supported by the finding that PG with only very small amounts of eDNA, resulting from mechanical treatment prior to adhesion, could restore the adhesive properties to the same level as the control, whereas the nonpolymeric NAG could not (Fig. (Fig.6).6). PNAG purified from S. epidermidis has been shown to contain at least 130 residues (24), indicating that the adhesion is due in part to the size of the molecule, which supports the hypothesis that HMW DNA is required to fully polymerize NAG. The bonds between the unbranched artificial polysaccharide, adjacent polysaccharide chains, the cell wall, and/or lectins should increase the van der Waals interactions and the number of hydrogen bonds (24), thereby increasing the adhesive forces. Another molecule known to induce the adhesive properties of bacteria is the 240-kDa large biofilm-associated protein (BAP) of Staphylococcus aureus (4). Based on all the data described above it seems that macromolecules in general are necessary for the adhesion of bacteria to surfaces and also for intercellular adhesion. These data all support the hypothesis that the size of the intercellular molecules matters and that reducing the size of these molecules to the single components in turn eliminates adhesion.
This work was supported by grant 274-05-0073 from the Danish Research Council for Technology and Production Sciences.
We thank B. Knudsen for her work on searching for genes encoding extracellular polysaccharides in the genome of L. monocytogenes EGDe. We are grateful to P. R. Jensen for his idea concerning saturating the adhesion sites with short DNA fragments.
Published ahead of print on 5 February 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.