Pseudomonas aeruginosa produces extracellular DNA which functions as a cell-to-cell interconnecting matrix component in biofilms. Comparison of extracellular DNA and chromosomal DNA by the use of polymerase chain reaction and Southern analysis suggested that the extracellular DNA is similar to whole-genome DNA. Evidence that the extracellular DNA in P. aeruginosa biofilms and cultures is generated via lysis of a subpopulation of the bacteria was obtained through experiments where extracellular β-galactosidase released from lacZ-containing P. aeruginosa strains was assessed. Experiments with the wild type and lasIrhlI, pqsA, pqsL and fliMpilA mutants indicated that the extracellular DNA is generated via a mechanism which is dependent on acyl homoserine lactone and Pseudomonas quinolone signalling, as well as on flagella and type IV pili. Microscopic investigation of flow chamber-grown wild-type P. aeruginosa biofilms stained with different DNA stains suggested that the extracellular DNA is located primarily in the stalks of mushroom-shaped multicellular structures, with a high concentration especially in the outer part of the stalks forming a border between the stalk-forming bacteria and the cap-forming bacteria. Biofilms formed by lasIrhlI, pqsA and fliMpilA mutants contained less extracellular DNA than biofilms formed by the wild type, and the mutant biofilms were more susceptible to treatment with sodium dodecyl sulphate than the wild-type biofilm.
Bacteria in natural, industrial and clinical settings predominantly live in surface-associated communities called biofilms (Costerton et al., 1995). Bacteria in biofilms are resistant to biocides, antibiotics and host immune responses, and biofilm formation therefore leads to a variety of problems in industry and to chronic or lethal infections in human and animals (Costerton et al., 1999). During the last 10 years there has been a rapidly increasing recognition of microbial biofilms as a highly significant topic in microbiology with relevance for many important areas in modern society. It has become apparent that detailed knowledge about the developmental process from single bacteria scattered on a surface to the formation of thick structured biofilms is essential in order to create strategies to control biofilm development.
The opportunistic pathogen Pseudomonas aeruginosa has become a model organism in biofilm research. Of considerable interest is the nature of the components that hold the bacteria together in biofilms, and the cues which induce their synthesis. It appears that extracellular DNA, exopolysaccharide and proteinaceous compounds can all function as matrix components in P. aeruginosa biofilms, and that their relative importance as structural component may depend on the environmental conditions, on the age of the biofilm and on the particular P. aeruginosa strain forming the biofilm. Evidence for a role of extracellular DNA as cell-to-cell interconnecting compound in P. aeruginosa biofilms has been presented both for the P. aeruginosa PAO1 reference strain and for clinical P. aeruginosa isolates (Whitchurch et al., 2002; Nemoto et al., 2003). DNase treatment dissolved young P. aeruginosa PAO1 biofilms whereas established PAO1 biofilms were only marginally affected by the DNase treatment, suggesting that the cells in young PAO1 biofilms are held together by extracellular DNA whereas the cells in older PAO1 biofilms are held together primarily by other compounds (Whitchurch et al., 2002). In accord, Matsukawa and Greenberg (2004) recently reported that, although DNA makes up a substantial fraction of the matrix material in mature PAO1 biofilms, exopolysaccharides encoded by the psl genes appears to be the critical matrix component. A role for the psl-encoded exopolysaccharide as biofilm matrix component was also reported for the P. aeruginosa ZK2870 strain (Friedman and Kolter, 2004a). In addition, it was reported that the P. aeruginosa PA14 strain produces a glucose-containing exopolysaccharide encoded by the pel genes as the main structural component of the biofilm matrix (Friedman and Kolter, 2004b). In contrast to the findings that extracellular DNA is not the primary cell-to-cell interconnecting compound in mature biofilms formed by the P. aeruginosa reference strains, Nemoto et al. (2003) found that mature biofilms formed by four independent clinical P. aeruginosa isolates could be dissolved by DNase treatment, suggesting that extracellular DNA is the primary cell-to-cell interconnecting compound in mature biofilms formed by these P. aeruginosa strains. Long before biofilms became a central research area a Japanese group reported the characterization of extracellular ‘slime’ produced by P. aeruginosa (Murakawa, 1973a,b). They investigated the chemical composition of slimes from 20 clinical P. aeruginosa isolates, and found that slimes from 18 strains consisted primarily of DNA, while two strains with a mucoid phenotype produced slimes composed primarily of polyuronic acid. Besides extracellular DNA and exopolysaccharide, proteinaceous polymers may be important matrix components. Evidence was recently provided that cell appendages, termed CupA fimbria, are an important constituent of the extracellular matrix for the P. aeruginosa strains PAK, PAO1 and PA14 (Vallet et al., 2001; D’Argenio et al., 2002; Friedman and Kolter, 2004b). In addition, type IV pili have various conditional roles in P. aeruginosa biofilm formation (O’Toole and Kolter, 1998; Klausen et al., 2003a,b).
Quorum sensing, a mechanism by which bacteria can monitor their cell population density through the extracellular accumulation of signalling molecules, has been shown to play a role in structural development and stabilization of P. aeruginosa biofilms (Davies et al., 1998). P. aeruginosa employs the quorum-sensing systems lasRI and rhlRI to control the expression of a wide range of virulence factors (Passador et al., 1993; Latifi et al., 1995; 1996). The rhlRI system is to some extent controlled by the lasRI system (Latifi et al., 1996). The signal molecules involved are primarily 3-oxo-C12-homoserine lactone (3-oxo-C12-HSL) in case of the lasRI system, and primarily C4-HSL in case of the rhlRI system (Latifi et al., 1995; 1996). Quorum-sensing systems involving such acylated homoserine lactone (AHL) signalling molecules can be inhibited by halogenated furanone compounds isolated from the macroalgae Delisea pulchra (Givskov et al., 1996; Hentzer et al., 2002; 2003). A third signalling system based on 2-heptyl-3-hydroxy-4-quinolone, designated the Pseudomonas quinolone signal (PQS), has been shown to be part of the quorum-sensing regulatory network in P. aeruginosa (Pesci et al., 1999). The production of PQS depends on lasR (Pesci et al., 1999), and exogenous PQS strongly induces expression of elastase B and rhlI in a lasR mutant background (McKnight et al., 2000). The PQS signalling system is therefore believed to operate between the las and the rhl systems in the quorum-sensing regulatory network (McKnight et al., 2000).
In the present article we report a characterization of the release of extracellular DNA in P. aeruginosa cultures and biofilms. Although the study was initiated in order to gain insight into the role of extracellular DNA in P. aeruginosa biofilm development, we found it worthwhile to begin with a characterization of DNA release in planktonic P. aeruginosa cultures. The experiments with planktonic cultures are followed by studies of DNA release in microtitre trays where the bacteria are present both in the planktonic and in the biofilm mode of growth. Finally, DNA release and the spatial distribution of extracellular DNA are studied in flow chamber-grown biofilms.
Extracellular DNA in P. aeruginosa cultures
The accumulation of extracellular DNA in the medium in growing P. aeruginosa PAO1 cultures was assessed by spectrophotometric measurements of light absorbance by DNA, and by fluorometric measurements of PicoGreen bound to DNA. Both methods showed that the P. aeruginosa cultures contained a low basal level of extracellular DNA in the initial and mid-log phase of growth, and that a large amount of DNA was released in the late-log phase of growth (Fig. 1). The sudden increase of extracellular DNA in the late-log phase was followed by a steep decrease suggesting that the cultures contained extracellular DNase activity. The finding that a large amount of extracellular DNA was released specifically in the late-log phase of growth suggested that quorum sensing might have a role in DNA release. Accordingly, measurements of the accumulation of extracellular DNA in the medium of growing P. aeruginosa lasIrhlI cultures showed that the quorum-sensing mutant did not release a large amount of DNA in the late-log phase (Fig. 1). To further substantiate a role of quorum sensing in DNA release, we followed the accumulation of extracellular DNA in a P. aeruginosa lasIrhlI culture supplemented with 3-oxo-C12-HSL and C4-HSL signal molecules, and in a P. aeruginosa wild-type culture supplemented with the synthetic compound furanone C-30, which specifically inhibits quorum sensing (Hentzer et al., 2002; 2003). In support of the suggestion that generation of extracellular DNA in P. aeruginosa late-log phase cultures is a quorum sensing-regulated process, we found that accumulation of extracellular DNA in the late-log phase was restored in the P. aeruginosa lasIrhlI culture supplemented with signal molecules, and that accumulation of extracellular DNA was inhibited in the wild-type culture supplemented with furanone (Fig. 1). Although the signal molecules used for chemical complementation of quorum sensing in the P. aeruginosa lasIrhlI culture were present also in the log phase, generation of large amounts of extracellular DNA occurred only in the late-log phase, suggesting that the quorum sensing-regulated DNA-release process is subject to regulation by other systems than las and rhl, a phenomenon that is also true for other quorum sensing-regulated processes such as rhamnolipid synthesis (Medina et al., 2003). The graphs presented in Fig. 1 show the result of a representative experiment; however, for reasons which are addressed below, the amount of extracellular DNA detected in the cultures varied somewhat between experiments and is probably underestimated.
Because extracellular DNA can function as a cell-to-cell interconnecting compound in P. aeruginosa biofilms (Whitchurch et al., 2002; Nemoto et al., 2003), we investigated whether release of extracellular DNA in P. aeruginosa cultures resulted in clumping. Staining and microscopic investigation of samples from a late-log phase culture of wild-type P. aeruginosa showed that they contained large clumps of cells with extracellular DNA intertwined between the cells (Fig. 2). In late-log phase cultures of the lasIrhlI mutant, and in wild-type cultures supplemented with DNase I, such large clumps were not found. In addition, late-log phase cultures of the lasIrhlI mutant, and late-log phase cultures of the wild-type supplemented with DNase I, reached higher optical densities than late-log phase cultures of the wild-type without DNase I (data not shown), presumably because homogeneous cultures scatter light different from cultures where a fraction of the cells are located in clumps. Taken together, these experiments indicate that extracellular DNA may function as cell-to-cell interconnecting compound also in planktonic cultures. Because of the association of the extracellular DNA with the bacteria, the DNA measurements presented in Fig. 1 probably underestimate the amount of extracellular DNA, as the supernatant samples were simply separated from the bacteria by centrifugation before the measurements of DNA content.
Excretion of DNA from the surface of intact P. aeruginosa cells has been reported in a previous study based on electron microscopy (Hara and Ueda, 1981). In addition, in strains of Neisseria gonorrhoeae, which contain a specific genetic island, DNA excretion was shown to occur via a type IV secretion pathway (Dillard and Seifert, 2001). Moreover, the P. aeruginosa PAO1 genome contains prophage genes (Stover et al., 2000). It could therefore not be excluded that the extracellular DNA generated by P. aeruginosa could consist of only part of the chromosome, or could be phage DNA. In order to investigate the nature of the extracellular DNA generated by P. aeruginosa, we purified extracellular DNA from late-log phase culture medium, and chromosomal DNA from P. aeruginosa cells, and compared the two types of DNA by the use of standard molecular methods. The fact that the extracellular DNA could be degraded by DNase I but not S1 nuclease suggested that it is double stranded (data not shown). The genes fliC, gacA, rhlA, lasB and pilA, which are distributed in different regions of the P. aeruginosa chromosome, could be amplified by polymerase chain reaction (PCR) from both chromosomal and extracellular DNA (Fig. 3). Random amplified polymorphic DNA (RAPD) PCR, which generates a fingerprint of DNA by amplification of several DNA fragments guided by short primers, gave largely the same patterns when applied to chromosomal and to extracellular DNA (Fig. 4). As PCR is non-quantitative, and contamination of the extracellular DNA with small amounts of chromosomal DNA might affect the results obtained through PCR, we also used the more quantitative Southern blot technique. Chromosomal DNA and extracellular DNA were cut with the restriction enzymes SphI or BglII, gel electrophorized, and probed with labelled chromosomal DNA or labelled extracellular DNA. As shown in Fig. 5, the Southern analysis gave the same patterns in all cases, indicating that the extracellular DNA is similar to whole-genome DNA. These experiments, however, do not exclude minor differences such as the presence of more phage DNA in the extracellular DNA than in the chromosomal DNA.
Measurement of release of β-galactosidase in cultures of lacZ-containing strains has been used to infer that release of extracellular DNA occurs via cell lysis in the case of Acinetobacter calcoaceticus and Streptococcus pneumoniae (Palmen and Hellingwerf, 1995; Steinmoen et al., 2002). To begin to understand the mechanism by which DNA is released in P. aeruginosa cultures and biofilms, we cloned the Escherichia coli lacZ gene in a miniTn7 construct and inserted miniTn7::lacZ on the chromosome of the PAO1 wild type and the lasIrhlI mutant. Because P. aeruginosa has no lacI gene we anticipate that the inserted lacZ gene is constitutively expressed. Release of cytoplasmic enzymes such as β-galactosidase concomitant with the DNA release would indicate that the release occurs via cell lysis, or alternatively via lysis of vesicles released from the bacteria. We measured the β-galactosidase activity in the medium (i.e. without cells) as a fraction of the β-galactosidase activity in the culture (i.e. medium and cells) in growing cultures of P. aeruginosa::lacZ and P. aeruginosa::lacZ lasIrhlI. Because the absolute β-galactosidase activity values were low initially, and the fraction of extracellular β-galactosidase activity therefore could not be determined with certainty, we started assessment of the fraction of extracellular β-galactosidase when the cultures reached mid log phase. As shown in Fig. 6, the fraction of extracellular β-galactosidase activity was relatively high in both the P. aeruginosa::lacZ and the P. aeruginosa::lacZ lasIrhlI cultures, but was highest specifically in P. aeruginosa::lacZ late-log phase cultures. The experiment therefore suggested that the extracellular DNA is released via lysis of a small subpopulation of the cells, or alternatively via lysis of DNA-containing vesicles released from the bacteria.
Identification of quorum sensing-controlled factors involved in DNA release
In order to identify quorum sensing-controlled factors which are involved in the generation of extracellular DNA by P. aeruginosa, we developed a high-throughput assay. In this assay P. aeruginosa was grown in the wells of microtitre plates in medium supplemented with propidium iodide, which fluoresces when it is bound to DNA and does not penetrate live bacteria, and the extracellular DNA in the small cultures was measured by the use of a microtitre plate reader. The extracellular DNA measured in the assay was therefore released both from planktonic cells and from cells which formed a biofilm on the surface of the wells. A significant difference between the amount of extracellular DNA in wells with wild type and lasIrhlI mutants was found after 24 h of growth under these conditions (Fig. 7). When the lasIrhlI mutants grew in medium supplemented with 3-oxo-C12-HSL and C4-HSL, extracellular DNA was released to the same level as in wells with wild-type bacteria (Fig. 7). Using this high-throughput assay to screen mutants with defects in quorum sensing-regulated factors we found that a pqsA mutant generated low amounts of extracellular DNA (Fig. 7). The pqsA mutant is deficient in the production of PQS, and a pqsL mutant, which overproduces PQS (D’Argenio et al., 2002), was found to release large amounts of extracellular DNA (Fig. 7). PQS has previously been shown to be involved in lysis of P. aeruginosa cells (D’Argenio et al., 2002), and it was proposed that this might occur via induction of a prophage. Induction of prophage in P. aeruginosa has previously been shown to be quorum sensing-regulated (Hentzer et al., 2004). Because prophage-mediated cell lysis in old (9 days or more) P. aeruginosa biofilms has been shown to be dependant on flagella and type IV pili (Webb et al., 2003) we tested a mutant, fliMpilA, devoid of these cell appendages, and found that it released low amounts of extracellular DNA (Fig. 7). This suggested that DNA release might occur via induction of a prophage in a few cells followed by flagella/pili-dependent phage propagation and lysis of a subpopulation of the cells in the culture. In accordance with the suggestion that the pqsA mutant is deficient in generation of extracellular DNA because it lacks PQS-mediated prophage induction, and that the fliMpilA mutant is deficient in generation of extracellular DNA because it lacks flagella/pili-dependent phage propagation, we found that the pqsA and fliMpilA mutants complemented each other in a 1:1 mixture and released large amounts of extracellular DNA (Fig. 7). In order to obtain more direct evidence for a role of the prophage in DNA release, we have attempted to knock out the chromosomal prophage by the use of allelic displacement. So far, however, we have not succeeded in this, presumably because the prophage also exists in a replicative (plasmid) form.
Extracellular DNA in P. aeruginosa biofilms
We have previously shown that P. aeruginosa PAO1 biofilms grown in flow chambers on glucose minimal medium develop mushroom-shaped multicellular structures via a sequential process which involves a stalk-forming sessile subpopulation and a cap-forming migrating subpopulation (Klausen et al., 2003a), and were interested in mapping the location of the extracellular DNA in relation to these subpopulations. The extracellular DNA in P. aeruginosa PAO1 biofilms was recently visualized by staining with the specific fluorescent double-stranded DNA stain PicoGreen (Matsukawa and Greenberg, 2004). Because PicoGreen penetrates and stains live bacteria it is not suitable for detailed mapping of the location of extracellular DNA in biofilms. In order to map the extracellular DNA in the heterogeneous P. aeruginosa biofilm, we grew Gfp-tagged PAO1 biofilm in flow chambers irrigated with glucose minimal medium, stained the biofilms with DNA stains that do not penetrate the membrane of live bacteria and performed microscopic investigations by the use of confocal laser scanning microscope (CLSM). Figure 8A–I shows CLSM images acquired in 4-day-old biofilms stained with propidium iodide (Figure 8A–C), ethidium bromide (Figure 8D–F) and DDAO [7-hydroxy-9H-(1,3-dichloro-9,9-dimethylacridin-2-one)] (Figure 8G–I). As far as we know, DDAO has not previously been used as a DNA stain, but we tested this compound due to its excellent fluorescent properties and its structural resemblance with the DNA stain acridine orange. The top-down views (Fig. 8A, D and G) show extracellular DNA as a red grid-like structure on the substratum, and mushroom caps which appear green because they are devoid of visible extracellular DNA. The horizontal section shown in Fig. 8B is located just above the substratum so that it visualizes bacteria in the stalk portion of the mushroom-shaped structure, and it is evident that the propidium iodide-stained extracellular DNA is located in the outer part of the stalk. The horizontal section shown in Fig. 8E is located further away from the substratum and visualizes a thin layer of cap bacteria in the periphery surrounding the stalk bacteria which have the majority of the ethidium bromide-stained extracellular DNA located in the outer part of the stalk forming a border between the stalk subpopulation and the cap subpopulation. The horizontal section shown in Fig. 8H is located even further away from the substratum and visualizes more of the cap bacteria in the periphery surrounding the stalk bacteria with DDAO-stained extracellular DNA of which the majority is located in the outer parts of the stalk. The horizontal sections in Fig. 8C, F and I visualize only the extracellular DNA, and provide further evidence that the extracellular DNA is located in the stalk portion of the mushroom-shaped structures with an increased concentration in the outer parts of the stalks.
The use of propidium iodide and ethidium bromide to visualize extracellular DNA in biofilms required that the sensitivity of the confocal laser scanning microscope was increased to a very high level, and images of good quality were difficult to obtain. DDAO, on the other hand, was easier to detect, and we therefore used this stain for further studies of the spatial distribution of extracellular DNA in our P. aeruginosa biofilms. Figure 8J–L shows microcolonies in a 2-day-old biofilm upon which caps have not yet formed. The microcolonies contain extracellular DNA, and a large amount of extracellular DNA is located in their outer part as well as upon them and between them forming a grid-like structure on the substratum. Figure 8M and N shows mushroom-shaped structures in a 6-day-old biofilm. From the top-down view (Fig. 8M) it is evident that the extracellular DNA at this stage extends up through the cap, and from the horizontal sections (Fig. 8N and O) it is apparent that the extracellular DNA at this stage is located with high concentrations in discrete layers.
In some cases a few bacterial cells were stained with propidium iodide, ethidium bromide or DDAO and appeared bright fluorescent. These bright fluorescent cells were most likely permeable to the DNA stains because they were dead. Besides these few stained cells, however, the stained DNA did not colocalize with the bacteria, as shown clearly in Fig. 9. In a series of experiments, which will be reported elsewhere, we used propidium iodide to monitor killing by antimicrobial compounds in P. aeruginosa biofilms. In that study we used FACS sorting and plating to ascertain that green (gfp-tagged) bacteria were alive and that red (propidium iodide-stained) bacteria were dead. Importantly in the present context, the study showed that all of the bacteria (except for a small subpopulation) in 4-day-old P. aeruginosa biofilms were alive before treatment with the antimicrobials.
Because our experiments with P. aeruginosa in planktonic culture indicated that the lasIrhlI, pqsA and fliMpilA mutants produce less extracellular DNA than the wild type, we investigated DNA release in flow chamber-grown biofilms formed by these mutants. We grew the PAO1 wild type and the isogenic mutants in flow chambers and stained the biofilms with propidium iodide or with DDAO. Under our conditions the difference in structural biofilm development between the wild type and the quorum-sensing mutant was less pronounced than reported by Davies et al. (1998). The biofilms formed by the quorum-sensing mutant, however, contained less extracellular DNA than the biofilms formed by the wild type (Fig. 10A–D). In accordance with Diggle et al. (2003) the pqsA mutant could not form structured biofilms; in our flow chamber set-up it formed only a flat and thin biofilm (Fig. 10E). Staining of the thin pqsA mutant biofilm with DDAO showed that this mutant biofilm also contained only little extracellular DNA (Fig. 10E). The fliMpilA mutant formed biofilms with irregularly shaped microcolonies, and the biofilm formed by this mutant also contained only little extracellular DNA (Fig. 10F).
In order to investigate whether β-galactosidase was released in biofilms formed by the P. aeruginosa::lacZ and the P. aeruginosa::lacZ lasIrhlI strains, we used the compound DDAOG [9H-(1,3-dichloro-9,9-dimethylacridin-2-one-7-yl)β-d-galactopyranoside], which is a conjugate of DDAO and β-galactoside. DDAOG is cleaved by β-galactosidase to β-galactoside and DDAO, and only DDAO (not DDAOG) has the far-red fluorescence properties detectable by CLSM. In biofilms formed by both the P. aeruginosa::lacZ and the P. aeruginosa::lacZ lasIrhlI strains, addition of DDAOG led to staining of the extracellular DNA with DDAO (data not shown), suggesting that extracellular β-galactosidase activity was present in both the P. aeruginosa::lacZ and the P. aeruginosa::lacZ lasIrhlI biofilms. When DDAOG was added to biofilms formed by P. aeruginosa strains without the lacZ gene, the extracellular DNA was not stained (data not shown). These experiments therefore suggested that cell lysis occurs in the biofilm formed by the wild type, as well as in the biofilm formed by the lasIrhlI mutant.
Because our experiments suggested that biofilms formed by the lasIrhlI, pqsA and fliMpilA mutants all contained less extracellular DNA than biofilms formed by the P. aeruginosa wild type, we found it of interest to determine the resistance of these mutant biofilms towards treatment with the detergent SDS. As shown in Fig. 11, the lasIrhlI, pqsA and fliMpilA mutant biofilms were all more sensitive to SDS treatment than the wild-type biofilm, which might indicate that the extracellular DNA stabilizes the wild-type biofilm. In order to obtain more direct evidence for a role of extracellular DNA in stabilization of P. aeruginosa biofilms, we pre-treated a 4-day-old wild-type biofilm for a short time with DNase before the SDS treatment, and found that the DNase-treated wild-type biofilm was more sensitive to SDS treatment than the wild-type biofilm which had not been treated with DNase (Fig. 11).
Many strains of P. aeruginosa, including PAO1, have previously been shown to produce large amounts of extracellular DNA (e.g. Goto et al., 1971; Murakawa, 1973a,b; Hara and Ueda, 1981; Muto and Goto, 1986), which may function as a matrix component in biofilms (Whitchurch et al., 2002; Nemoto et al., 2003). In the present report we provide evidence that this extracellular DNA is organized in distinct patterns in P. aeruginosa biofilms. In 2-day-old P. aeruginosa biofilms, which contain small microcolonies, the extracellular DNA is present in the microcolonies and as a grid-like structure on the substratum, but the highest concentration of extracellular DNA appears to be located on the surface of the microcolonies. In 4-day-old P. aeruginosa biofilms, which contain mushroom-shaped structures, the extracellular DNA is located on the substratum and in the stalk portion of the mushroom-shaped structures with the highest concentration in the outer parts of the stalks forming a border between the stalk subpopulation and the cap subpopulation. In 6-day-old biofilms the extracellular DNA is located throughout the mushroom-shaped structures with high concentrations in discrete layers. The formation of the mushroom-shaped structures in glucose-grown P. aeruginosa biofilms was previously shown to occur in a sequential process involving a non-motile bacterial subpopulation that forms the initial microcolonies (which later become mushroom stalks) by growth in certain foci of the biofilm, and a migrating bacterial subpopulation which subsequently forms the mushroom caps via a process which requires type IV pili (Klausen et al., 2003a). It is currently not understood how the migration of the motile cells is co-ordinated so that they form mushroom caps. However, because type IV pili mediate bacterial migration by an extension-grip-retraction mechanism (Skerker and Berg, 2001), and because type IV pili bind to DNA (Aas et al., 2002; Van Schaik et al., 2005), it is tempting to speculate that the high concentration of extracellular DNA on the mushroom stalks might cause accumulation of the migrating bacteria resulting in the formation of mushroom caps. It is further tempting to speculate that quorum sensing-induced DNA release could occur in a periodic pattern in the accumulating cap subpopulation. Experiments to test these hypotheses are underway in our laboratory.
The PCR and Southern analysis presented here suggested that the extracellular DNA present in P. aeruginosa late-log phase cultures is similar to whole-genome DNA. In agreement, it has previously been shown that different markers, including his+, leu+ and trp+, could be transferred by transformation of CaCl2-treated P. aeruginosa cells with extracellular DNA, at the same frequencies as when transformation was performed with an equivalent amount of purified intracellular DNA (Hara et al., 1981; Muto and Goto, 1986). In accord with the suggestion that the extracellular DNA is similar to whole-genome DNA, our experiments with lacZ-containing P. aeruginosa strains suggested that DNA release occurs through cell lysis, or through lysis of released DNA-containing membrane vesicles. The fact that 4-day-old P. aeruginosa biofilms contained relatively few cells which were stained with the DNA stains, and therefore judged dead, suggested that, if cell lysis is involved in DNA release, it occurred shortly after the cells became permeable to the DNA stain.
Our results suggested that the extracellular DNA observed in P. aeruginosa cultures and biofilms may be generated via at least two different pathways. A basal level of extracellular DNA present in P. aeruginosa cultures and biofilms appears to be generated via a pathway which is not linked to quorum sensing. The DNA which is released specifically in P. aeruginosa late-log phase cultures, and the increased amount of extracellular DNA found in P. aeruginosa wild-type biofilms in comparison with lasIrhlI biofilms, are evidently linked to quorum sensing. Our findings that biofilms formed by the wild type and quorum-sensing mutant differ in particular with respect to the amount of extracellular DNA in the portion of the biofilm located close to the substratum is in agreement with previous studies which showed that the expression of lasI and rhlI in P. aeruginosa biofilms were highest in the portion of the biofilm closest to the substratum (DeKievit et al., 2001). Similar to our findings of the generation of extracellular DNA primarily in the stalks of the mushroom-shaped structures in P. aeruginosa biofilms, it was recently shown that synthesis of rhamnolipid, an established quorum sensing-regulated process, occurs primarily in the stalks of P. aeruginosa biofilm mushroom-shaped structures (Lequette and Greenberg, 2005).
Our screen of DNA release from mutants which carry lesions in quorum sensing-regulated genes identified pqsA as a gene involved in DNA release. The pqsA mutant is deficient in production of PQS, and in support of a direct role of PQS in DNA release, a pqsL mutant, which overproduces PQS, showed elevated release of extracellular DNA. Quinolone compounds have previously been shown to induce prophages in bacteria (Phillips et al., 1987; Froshauer et al., 1996), and recent studies by Webb et al. (2003) and Hentzer et al. (2004) suggested that quorum sensing-regulated DNA release might be linked to phage induction in biofilms causing cell lysis. In support of a role of phage lysis in DNA release we found that a fliMpilA mutant, which has been shown previously not to undergo phage-mediated cell lysis (Webb et al., 2003), was deficient in DNA release. In support of the suggestion that PQS-mediated prophage induction and flagella/pili-dependent phage propagation are involved in DNA release, we found that the pqsA and fliMpilA mutants complemented each other in a 1:1 mixture and released the same amount of extracellular DNA as the wild type. We propose that PQS produced by the fliMpilA mutant induced the prophage, and that flagella/pili-dependent phage propagation in the pqsA mutant resulted in lysis of a subpopulation of the cells and DNA release. However, two very recent findings suggest that membrane vesicles might have a role in DNA release as an alternative to prophages. P. aeruginosa releases membrane vesicles which have bacteriolytic effects and contain DNA (Kadurugamuwa and Beveridge, 1996; Renelli et al., 2004), and extracellular DNA might be released either from vesicles that eventually lyse, or through the bacteriolytic activity of the vesicles which might lyse a small subpopulation of the P. aeruginosa cells. Very recently it was shown that PQS is necessary for vesicle formation in P. aeruginosa (Mashburn and Whiteley, 2005), and evidence was presented that type IV pili and flagella are necessary for quorum sensing in P. aeruginosa (Hassett, 2005). The involvement of PQS and type IV pili and flagella in DNA release therefore should not be regarded as strong evidence for the involvement of prophage.
Contrary to our suggestion, a DNA release mechanism which does not involve lysis was suggested in a study which employed electron microscopy to visualize what was interpreted as excretion of double-stranded DNA from the surface of intact P. aeruginosa cells (Hara and Ueda, 1981). Studies of other bacterial species indicate that although they all produce extracellular DNA (for a review, see Lorenz and Wackernagel, 1994), the DNA-release mechanisms vary between species. While DNA release caused by lysis of a fraction of the bacteria was reported to occur in case of S. pneumoniae (Steinmoen et al., 2002) and A. calcoaceticus (Palmen and Hellingwerf, 1995), DNA release without cell lysis was reported to occur in case of Bacillus subtilis (Lorenz et al., 1991) and N. gonorrhoeae (Dillard and Seifert, 2001). In strains of N. gonorrhoeae, which contain a specific genetic island, DNA excretion was shown to occur via a type IV secretion pathway (Dillard and Seifert, 2001).
On top of a basal level of DNA release it appears that many bacteria, especially those that are able to develop natural competence, possess a specific DNA-release programme. For example, a correlation between DNA release and competence development has been established in S. pneumoniae (Steinmoen et al., 2002), B. subtilis (Lorenz et al. (1991), A. calcoaceticus (Palmen and Hellingwerf (1995), N. gonorrhoeae (Dillard and Seifert, 2001) and Pseudomonas stutzeri (Stewart et al., 1983). In all these cases DNA release and competence development occur in the late-log phase in liquid cultures, and in some of the cases competence development has been shown to be regulated through a quorum-sensing mechanism (Magnuson et al., 1994; Pestova et al., 1996). P. aeruginosa is not known to be naturally transformable, but as the other members of the P. aeruginosa DNA homology subgroup (i.e. P. stutzeri, Pseudomonas mendocina, Pseudomonas alcaligenes and Pseudomonas pseudoalcaligenes) were all found to be naturally transformable (Carlson et al., 1983), it is possible that some P. aeruginosa strains has this capability. It is possible therefore that DNA release in some cases both allows exchange of genetic material to take place and induces biofilm structure formation and stabilization. The relatively long-lasting physical proximity of bacteria in biofilms enable the constituent cells to establish long-term relationships with each other, and biofilms appear to be optimal environments for gene transfer to occur via transformation. Biofilm-grown Streptococcus mutans cells were shown to be transformed at rates 10- to 600-fold higher than planktonic S. mutans cells (Li et al., 2001). The transfer of chromosomal genes in P. stutzeri cell mats were 1000-fold more efficient when the DNA was released from donor cells in comparison with when the same amount of DNA was provided from a DNA preparation (Stewart et al., 1983). Transfer of a non-conjugative plasmid from Treponema denticola to Streptococcus gordonii growing in a mixed species biofilm was demonstrated to occur with high efficiency (Wang et al., 2002), and finally Wuertz and co-workers showed that transformation occurs with high efficiency in biofilms of an Acinetobacter sp. (Hendrickx et al., 2003).
Quorum sensing has previously been shown to play a role in structural development of P. aeruginosa biofilms (Davies et al., 1998), but with the exception of rhamnolipid, which has been shown to play a role in maintaining the channels between the mushroom-shaped structures (Davey et al., 2003), the actual quorum sensing-controlled factors that play a role in P. aeruginosa biofilm development have not been identified. Several recent studies have demonstrated that biofilms which develop in the absence of functional quorum-sensing systems (caused by either inhibition or mutation) are vulnerable to SDS treatment, shear forces, kanamycin treatment, tobramycin treatment, H2O2 treatment and polymorphonuclear neutrophils phagocytosis (Davies et al., 1998; Hassett et al., 1999; Hentzer et al., 2002; 2003; Shih and Huang, 2002; Bjarnsholt et al., 2005; Rasmussen et al., 2005). In the present study we found that release of large amounts of extracellular DNA in P. aeruginosa biofilms is dependent on quorum sensing, and that P. aeruginosa biofilms without large amounts of extracellular DNA, formed by the lasIrhlI, fliMpilA and pqsA mutants or obtained through DNase treatment of wild-type biofilm, are vulnerable to treatment with SDS. Although several hundred genes in P. aeruginosa have been shown to be quorum sensing-regulated (Wagner et al., 2003), and it is possible that mutations in the quorum-sensing control systems might affect many gene activities which have roles in biofilm development, the present study suggests that one of the quorum sensing-controlled factors that might have a role in biofilm development by P. aeruginosa is programmed DNA release resulting in extracellular DNA which functions as a biofilm matrix component.
Bacterial strains and growth conditions
Pseudomonas aeruginosa PAO1 (Holloway and Morgan, 1986) was used as the model organism in this study. All experiments involving wild-type PAO1 were performed with three different sublines, one obtained from the Pseudomonas Genetic Stock Center (strain PAO0001), one obtained from the laboratory of Barbara Iglewski and one obtained from the laboratory of John Mattick. The three PAO1 wild-type sublines did not differ with respect to the investigated phenotypes. The lasIrhlI derivative was constructed by allelic displacement in the PAO0001 subline as described (Hentzer et al., 2003). The fliMpilA derivative was constructed in the PAO0001 subline and in the PAO1 subline from John Mattick's laboratory as described (Klausen et al., 2003b), and both fliMpilA derivatives had the same phenotypes. The pqsA and pqsL mutants were constructed by transposon insertion in the PAO1 subline from Barbara Iglewski's laboratory as described (D’Argenio et al., 2002). The strains were fluorescently tagged at an intergenic neutral chromosomal locus with gfp in a miniTn7 construct (Klausen et al., 2003b). AB medium (Clark and Maaløe, 1967) was used supplemented with 30 mM glucose for batch cultures, and with 0.3 mM glucose for biofilm cultivation. Biofilms and batch cultures were grown at 30°C. DNase I (Sigma) was used at a concentration of 90 Kunitz units per ml in medium supplemented with MgCl2 (5 mM). Propidium iodide, ethidium bromide and DDAO were used at a concentration of 1 µM to stain extracellular DNA in biofilms. DDAOG was used at a concentration of 1 µM to assess β-galactosidase activity in biofilms. 3-oxo-C12-HSL and C4-HSL were used at a concentration of 1 µM. Furanone C-30 was used at a concentration of 10 µM.
Construction of P. aeruginosa strains with lacZ inserted on the chromosome
A miniTn7-PUV5–lacZ delivery plasmid was constructed by cloning a 3.3 kb NotI fragment containing a PUV5–lacZ cassette into a NotI-digested miniTn7-strepR vector based on pUC19. The miniTn7-strepR-PUV5–lacZ transposon was inserted into the chromosome of the PAO1 wild type and lasIrhlI mutant by four parental matings using E. coli HB101 with the delivery plasmid, and the helper strains E. coli HB101/RK600 and E. coli HB101/pUX-BF13 as described previously (Klausen et al., 2003b).
Measurements of extracellular DNA in planktonic cultures
Culture samples were centrifuged (3 min, 10.000 r.p.m.) and the supernatant was transferred to a new eppendorf tube. NaCl was added to the supernatant to a concentration of 0.25 M, and the extracellular DNA was precipitated by adding 2:1 volume of ethanol. The precipitated extracellular DNA was dissolved in TE buffer, and the DNA concentration was determined by spectrophotometry (OD260/OD280), or by using the PicoGreen® dsDNA Quantitation Kit (Molecular Probes).
Measurements of extracellular DNA in microtitre tray cultures
Overnight cultures grown in AB (0.5% glucose) medium were diluted to OD600 = 0.001 in AB medium supplemented with 0.5% glucose and 0.05 mM propidium iodide. The diluted cultures were transferred to wells of polystyrene microtitre plates (150 µl of cultures per well) and incubated for 24 h at 37°C, whereupon propidium iodide absorbance was measured at OD480 and cell density was measured at OD600 by the use of a Wallac microplate reader. In the experiment with homoserine lactone complementation, 3-oxo-C12-HSL and C4-HSL were added to the medium to a final concentration of 10 µM.
Measurements of β-galactosidase activity in planktonic cultures
The fraction of extracellular β-galactosidase in planktonic cultures was assessed by measuring β-galactosidase activity in culture samples and in culture supernatants essentially as described by Steinmoen et al. (2002).
Cultivation of biofilms
Biofilms were grown in flow chambers with individual channel dimensions of 1 × 4 × 40 mm. The flow system was assembled and prepared as described previously (Møller et al., 1998). The flow chambers were inoculated by injecting 350 µl of overnight culture diluted to an OD600 of 0.001 into each flow channel with a small syringe. After inoculation flow channels were left without flow for 1 h, after which medium flow (0.2 mm s−1) was started using a Watson Marlow 205S peristaltic pump.
Microscopy and image acquisition
All microscopic observations and image acquisitions were performed with a Zeiss LSM 510 CLSM (Carl Zeiss, Jena, Germany) equipped with detectors and filter sets for monitoring of Gfp, propidium iodide, ethidium bromide and DDAO fluorescence. Images were obtained using a 63×/1.4 objective or a 40×/1.3 objective. Simulated three-dimensional images and sections were generated using the IMARIS software package (Bitplane AG, Zürich, Switzerland).
SDS treatment of biofilms
In order to assess the sensitivity of wild type and mutant P. aeruginosa biofilms to SDS treatment, 2-day-old biofilms were irrigated for 2 h with medium containing 0.01% SDS, and CLSM images were acquired before and after the SDS treatment. In order to investigate the effect of pre-treatment with DNase I on the sensitivity of the wild-type biofilm to DNase treatment, a 4-day-old wild-type biofilm was treated with DNase I (100 µg ml−1) for 45 min before the SDS treatment.
PCR and RAPD PCR
All PCR reactions were performed using Taq DNA Polymerase (Sigma) and 10× reaction buffer with MgCl2 (Sigma) + 0.25 mM each of ATP, CTP, GTP and TTP in the presence of 5% DMSO (Merck). Primers for amplification of fliC, gacA, rhlA, lasB and pilA sequences were used at a concentration of 5 µM. The PCR reaction was performed on a T3 Thermocycler (Biometra); 45 s at 95°C, 1 min at 52°C, 1 min at 72°C and cycled 30 times. The RAPD PCR reactions were performed with 10-mer primers, the annealing temperature was 31°C and the reactions were run for 45 cycles. Sequences of the used primers are available on request.
Southern blot analysis
Southern analysis was performed using standard protocols. Chromosomal and extracellular DNA probes were prepared with the DIG DNA Labeling Kit (Roche), which inserts digoxigenin-dUTP by random primed DNA labelling using Klenow enzyme. Before use the probe was denatured by boiling for 10 min. The probes were mixed with hybridization buffer and allowed to hybridize with the membrane overnight at 55°C. The membrane was then washed with decreasing concentrations of SSC buffer at 55°C, and the hybridized DIG-labelled probes were bound by anti-DIG alkaline phosphates. Finally the bands were visualized with NBT/BCIP dissolved in Tris buffer at pH = 9.5 according to the manufacturer (Roche).
We thank Janus Haagensen for help with confocal microscopy and Anne Nielsen for performing Southern blots. This work was supported by a grant from the Danish Technical Research Council to T.T.N. (26-03-0234).