Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Dec 28.
Published in final edited form as: Mol Microbiol. 2010 Jun 28;77(4):815–829. doi: 10.1111/j.1365-2958.2010.07267.x

A bacterial extracellular DNA inhibits settling of motile progeny cells within a biofilm

Cécile Berne 1, David T Kysela 1, Yves V Brun 1,*
PMCID: PMC2962764  NIHMSID: NIHMS219767  PMID: 20598083

Abstract

In natural systems, bacteria form complex, surface-attached communities known as biofilms. This lifestyle presents numerous advantages compared to unattached or planktonic life, such as exchange of nutrients, protection from environmental stresses and increased tolerance to biocides. Despite such benefits, dispersal also plays an important role in escaping deteriorating environments and in successfully colonizing favorable, unoccupied habitat patches. The α-proteobacterium Caulobacter crescentus produces a motile swarmer cell and a sessile stalked cell at each cell division. We show here that C. crescentus extracellular DNA (eDNA) inhibits the ability of its motile cell type to settle in a biofilm. eDNA binds to the polar holdfast, an adhesive structure required for permanent surface attachment and biofilm formation, thereby inhibiting cell attachment. Since stalked cells associate tightly with the biofilm through their holdfast, we hypothesize that this novel mechanism acts on swarmer cells born in a biofilm, where eDNA can accumulate to a sufficient concentration to inhibit their ability to settle. By targeting a specific cell type in a biofilm, this mechanism modulates biofilm development and promotes dispersal without causing a potentially undesirable dissolution of the existing biofilm.

Keywords: Caulobacter, biofilm, extracellular DNA, inhibition, attachment, holdfast

Introduction

Biofilms are multicellular, surface-associated complexes formed by different microorganisms. Bacteria are predominantly organized in such structured communities in natural environments (Hall-Stoodley et al., 2004). The chemical composition within biofilms is highly dynamic, promoting solute gradient formation and nutrient exchange (Stewart & Franklin, 2008, Spormann, 2008, Moons et al., 2009). Furthermore, biofilms represent a protected mode of growth, as bacteria within a biofilm typically exhibit greater resistance to deleterious agents, such as antibiotics, biocides or predators, than their planktonic counterparts (Davey & O'Toole G, 2000, Harrison et al., 2007).

An extracellular matrix provides the structural basis and many of the protective properties of mature biofilms (Tart & Wozniak, 2008). The distribution of matrix compounds varies from one microorganism to another (Kolter & Greenberg, 2006), but it typically comprises primarily exopolysaccharides, though proteins, lipids and nucleic acids also serve important roles (Goller & Romeo, 2008, Tart & Wozniak, 2008, Karatan & Watnick, 2009). Work on Pseudomonas aeruginosa first identified an essential role of extracellular DNA (eDNA) in stabilizing the biofilm matrix; addition of DNase to P. aeruginosa biofilms inhibits biofilm formation and dissolves existing biofilms (Whitchurch et al., 2002). This stabilizing eDNA derives from cell lysis within the biofilm (Allesen-Holm et al., 2006). Numerous other studies have since reported similar properties of eDNA in biofilms formed by a wide array of gram-positive and gram-negative bacteria (Spoering & Gilmore, 2006, Karatan & Watnick, 2009, Lappann et al., 2010), suggesting that bacterial biofilm stabilization by eDNA is widespread.

Though biofilm-associated bacteria may enjoy certain benefits not available to planktonic counterparts, cells must nonetheless disperse at least occasionally in order to colonize new habitat patches. Dispersal becomes particularly important upon deterioration of local habitat quality within the biofilm, e.g. due to over-crowding, phage infection, or a change in the physical environment (Spormann, 2008). Ideally, bacteria should adjust their behavior to suit the current environmental context, contingent on the ability to detect and respond to environmental cues in a timely manner (Meyers & Bull, 2002). Such behaviors are indeed observed, whereby accumulation of nitric acid or depletion of oxygen or nutrients triggers bacterial dispersal from a biofilm (Karatan & Watnick, 2009). Specific biofilm inhibitory molecules often mediate the dispersal response, including quorum-sensing signaling analogs, excreted proteins and polysaccharides (Karatan & Watnick, 2009). Though details of the dispersal process remain unresolved for many bacterial systems, described mechanisms typically involve regulatory changes in response to an environmental cue, whereby signal transduction pathways effect inhibition of exopolysaccharide synthesis and/or induction of motility genes (Spormann, 2008, Goller & Romeo, 2008). Thus, biofilm-associated cells achieve a balance between attachment and planktonic dispersal (Moons et al., 2009), and this balance may be specifically tuned in response to the current conditions encountered in the biofilm.

Caulobacter crescentus is an oligotrophic α-proteobacterium commonly found in aquatic environments. This stalked bacterium has a dimorphic life cycle where each cell division produces a motile swarmer cell and a sessile stalked cell (Fig. 1). A flagellum and several pili occupy a single pole of the nascent swarmer cell, which cannot initiate DNA replication. Irrespective of whether cells reside in suspension or attached to a surface, differentiation into a division-competent stalked cell begins after a developmentally programmed delay. The swarmer cell sheds its flagellum, retracts its pili and elaborates a membranous stalk at the former flagellar pole. Like many α-proteobacteria, C. crescentus adheres to surfaces by means of a polar polysaccharide called the holdfast (Brown et al., 2009). The holdfast is synthesized in the late stage of the swarmer phase and resides at the tip of the stalk following differentiation (Levi & Jenal, 2006). The holdfast exhibits extremely high adhesive forces (Tsang et al., 2006) and is required for permanent adhesion to surfaces and for biofilm formation (Bodenmiller et al., 2004, Entcheva-Dimitrov & Spormann, 2004, Levi & Jenal, 2006).

Figure 1. C. crescentus life cycle.

Figure 1

Open in a new tab

Each cell division produces a motile swarmer cell and a sessile stalked cell. After an obligatory delay during which it is unable to start the next cell cycle, swarmers cell differentiate into a division-competent stalked cells. The stalked cell has a polar adhesin called the holdfast that is required for permanent adhesion to surfaces and for biofilm formation. The holdfast is synthesized at the flagellated pole in the late stage of the swarmer phase and is found at the tip of the stalk once the stalk is synthesized at the same pole during cell differentiation.

In this study, we describe a mechanism that modulates the balance between biofilm development and dispersal by directly inhibiting the ability of Caulobacter swarmer cells to adhere. Lysis in cell cultures or biofilms releases eDNA that binds specifically to the adhesive holdfast, preventing newborn swarmer cells from attaching to surfaces or settling into a biofilm, thereby promoting their dispersal. The inhibition is specific to Caulobacter and does not occur with non-Caulobacter DNA. By acting via direct interaction between eDNA and holdfast, this dispersal mechanism differs substantially from those described previously in that it requires no intermediate intracellular response, and our findings identify a novel role of eDNA as a specific inhibitor of bacterial biofilm formation. Finally, the specific targeting of the swarmer cell type allows the modulation of biofilm development and the promotion of dispersal without causing a potentially undesirable deleterious dissolution of the biofilm.

Results

eDNA inhibits biofilm formation in C. crescentus

To determine whether C. crescentus produces factors that enhance dispersal from a biofilm, we tested the effect of filtered spent medium from saturated planktonic cultures and found that it inhibited biofilm formation. In an attempt to identify and characterize the molecules involved in biofilm inhibition, we tested different conditions that could remove or destroy extracellular compounds. The inhibitory activity was not dialysable using two different molecular weight cut-offs (3,500 and 10,000 Da) (Fig. S1A). The pH of the spent medium was 6.4 as compared to a pH of 7.0 for fresh medium. Adjusting the pH of the spent medium to 7.0 or that of the fresh medium to pH 6.4 did not affect the biofilm inhibition activity (Fig. S1B). The inhibitory effect persisted after both heat treatment of the spent medium (100°C for 15 minutes) and both proteinase K and pronase treatment (Fig. 2A). However, treatment with DNase I completely abolished the inhibition, whereas treatment of the spent medium with RNase had no impact on its inhibitory activity, suggesting that the inhibitory molecule was DNA (Fig. 2A). Treatment of the spent medium with nuclease S1 partially abolished the biofilm inhibition activity, indicating that single-stranded DNA has some inhibitory activity (Fig. 2A).

Figure 2. The biofilm formation inhibitor is sensitive to DNase I.

Figure 2

Open in a new tab

(A) Effect of different enzymatic treatments of the spent M2G medium on the biofilm inhibitory activity. (B) Agarose gel showing eDNA purified from 1 ml of spent medium (lane 2). Lanes 1 and 3 show HyperLadder II (Bioline) and 1 kb ladder (Invitrogen), respectively. (C) Low molecular weight DNA fragments from C. crescentus inhibit biofilm formation in C. crescentus. Effect of C. crescentus CB15 gDNA addition on biofilm formation. gDNA was partially digested with HpaII (30 min at 37°C, using different amounts of HpaII) followed by HpaII heat inactivation (70°C for 20 min). gDNA was added to static biofilm assays to obtain a final concentration 15 µg/ml of DNA. (D) 0.8% agarose gel of the gDNA used in the static biofilm assays. Lane 1: 1 kb ladder (Invitrogen); 2: no digestion; 3: 0.25 U HpaII digestion; 4: 0.5 U HpaII digestion; 5: 0.75 U HpaII digestion; 6: 1 U HpaII digestion. The static biofilm assay was performed for 24 h at 30°C and the biofilm attached to the coverslips was quantified by crystal violet staining. The results are expressed as a percentage of biofilm formation in the absence of gDNA. The error bars represent the S.E.M. of 3 independent experiments run in duplicate.

Purified eDNA from the spent medium had a relatively low molecular weight, mostly below 500 bp (Fig. 2B). Addition of intact purified genomic DNA (gDNA) from C. crescentus cells did not inhibit biofilm formation (Fig. 2C). Progressive digestion of gDNA increased its inhibitory activity in a manner inversely proportional to its average molecular weight (Fig. 2C–D). We conclude that low molecular weight DNA from C. crescentus is an inhibitor of biofilm formation in C. crescentus.

We examined the growth rate of C. crescentus in the presence of DNA in order to determine whether reduced biofilm formation resulted from a decline in reproductive rate. The growth rate of C. crescentus was not affected by the presence of 66% or less spent medium (containing ~ 20 µg/ml eDNA) or 20 µg/ml or less purified, sheared gDNA (Fig. S2), despite strong biofilm inhibition at similar concentrations of DNA or spent medium (see below, Fig. 3B). This result indicates that the biofilm inhibition did not result from bacteriostatic or bacteriocidal activity. eDNA inhibited biofilm formation in a variety of media with different carbon sources at different concentrations, demonstrating that inhibition of biofilm formation by eDNA operates both under good and poor growth conditions of this facultative oligotroph (Fig. S3).

Figure 3. Correlation between cell death, ecDNA release and biofilm inhibition.

Figure 3

Open in a new tab

(A) Correlation between cell death, eDNA release and biofilm inhibition over time. Wild-type C. crescentus CB15 was grown for 48 h in M2G at 30°C under constant shaking. Dead cells shown as a percentage of total cells (open triangles) and concentration of DNA released in the spent medium (solid diamonds) were monitored over time. Static biofilm assays were performed using spent medium harvested at the different time points (kept at −20°C before use). Overnight biofilm formation (open circles) is expressed as a percentage of biofilm formation compared to biofilm formation in the absence of spent medium. (B) Effect of different concentrations of spent medium and DNA on C. crescentus CB15 biofilm formation. Various volumes of spent M2G medium (solid diamonds), corresponding amounts of eDNA purified from C. crescentus CB15 culture spent medium (open circles), or sheared C. crescentus CB15 gDNA (open triangles) were added to M2G for a static biofilm assay. The results are expressed as a percentage of biofilm formation in the absence of spent medium, eDNA, or sheared gDNA. The results are expressed as a percentage of biofilm formation in the absence of gDNA. The error bars represent the S.E.M. of 3 independent experiments run in duplicate.

eDNA is released during cell death

As shown in Fig. 3A, measurable cell death occurs during growth of C. crescentus and is correlated with the accumulation of eDNA and biofilm inhibitory activity. Filter-sterilized spent medium, eDNA purified from the spent medium, or sheared genomic DNA (gDNA) purified from C. crescentus cells gave indistinguishable dose-response curves (Fig. 3B). This result indicates that components other than eDNA are not required for biofilm inhibition, and supports the conclusion that eDNA is a product of cell death rather than secretion of specific DNA fragments.

Small amounts of free DNA are detectable in the environment, up to 20 ng/ml for aquatic environments (Lorenz & Wackernagel, 1994), and C. crescentus eDNA concentrations likely remain low outside of biofilms. In biofilms, however, high cell density and local diffusion may result in sufficient C. crescentus eDNA accumulation to inhibit new attachment. Live/dead staining of a static biofilm grown on a plastic coverslip shows that substantial cell death occurs during biofilm formation even under favorable growth conditions (Fig. 4A). Quantification of eDNA in the biofilm reveals that ~ 6 µg/ml is present within the biofilm matrix after 24 hours (Fig. 4B); this concentration is sufficient to inhibit biofilm formation (Fig. 3B). The amount of eDNA measured under these conditions is an underestimate of the total eDNA present in the biofilm since the quantification method requires the complete removal of the supernatant, thereby removing some of the diffusible eDNA, and since it only measures double-stranded DNA using PicoGreen, whereas we have shown above using nuclease S1 treatment that single-stranded DNA also inhibits biofilm development (Fig. 2A). Indeed, extensive prewashing of the biofilms prior to quantification of eDNA removes most of the eDNA from the biofilm matrix (Fig. 4B), showing that eDNA is not tightly bound to the biofilm extracellular matrix. These results also demonstrate that, unlike species where eDNA plays an important role in biofilm structural integrity, firmly bound eDNA is not an abundant component of the C. crescentus biofilm matrix, as discussed below.

Figure 4. Cell death and eDNA release during biofilm formation.

Figure 4

Open in a new tab

(A) Biofilm attached to PVC coverslips stained with the BacLight Live/Dead stain at different times of a static biofilm assay. Images represent overlays of green (live cells) and red (dead cells) signals collected by epifluorescence microscopy. (B) eDNA concentration in the biofilm or attached to the extracellular matrix over time. Total eDNA within the biofilm (representing diffusible eDNA and eDNA attached to the extracellular matrix) is represented in red. To quantify the eDNA bound to the extracellular matrix, biofilm were washed prior to matrix isolation to remove diffusible eDNA; the amount of eDNA bound over time to the matrix is shown in black. Data are the average of 3 replicates of 2 independent experiments and the error bars represent the S.E.M. (C) Biofilm formation in the presence of DNase I. Biofilm formation in the presence of DNase I added at t = 0. The bacteria were incubated at 30°C with or without 30 units of DNase I, and the biofilm attached to the wells was quantified by crystal violet staining. The error bars represent the S.E.M of 3 independent experiments run in triplicate. (D) Effect of DNase I on established biofilms. Static biofilm assays were performed using 2 ml microtubes sealed with AeraSeal tape (1 ml final). The bacteria were incubated at 30°C for various times and then treated with 20 µg/ml DNase I for 1 hour. The biofilm attached inside the tubes was then quantified by crystal violet staining. The crystal violet staining of the non-treated samples and DNase I treated sample are shown in red and black respectively. The error bars represent the S.E.M of 2 independent experiments run in triplicate.

eDNA is not required for the stability of C. crescentus biofilm

Numerous studies indicate that eDNA is a major component of the biofilm extracellular matrix in many gram-positive and gram-negative bacteria and is required for biofilm stability (Karatan & Watnick, 2009). In all the reported cases, DNase I treatment causes inhibition and/or dispersion of the biofilm. However, C. crescentus biofilm formation is not inhibited by the presence of 20 µg/ml DNase I in a static biofilm assay (Fig. 4C). In fact the biofilm matures faster in the presence of DNase I than in its absence (Fig. 4C). We suggest that in the presence of DNase, eDNA is degraded as soon as it is released by cell lysis and cannot prevent biofilm formation, whereas without DNase, endogenously produced eDNA inhibits biofilm formation. Following maturation at ~ 30 h, biofilm dispersion appears more effective in the sample without DNase I treatment (Fig. 4C). Furthermore, DNase I treatment is not able to disperse established biofilms, as shown when biofilms at different stages of maturation were treated with DNase I (Fig. 4D). Lastly, quantification of the eDNA firmly bound to the extracellular matrix reveals only traces of DNA (Fig. 4B). We therefore conclude that diffusible eDNA has a net inhibitory effect on biofilm formation and persistence rather than serving as a stabilizing component of the biofilm matrix.

eDNA inhibits swarmer cell attachment by inhibiting holdfast adhesiveness

The initial and reversible stage of C. crescentus attachment to surfaces is mediated by flagellar motility and pili, whereas permanent attachment requires the holdfast (Bodenmiller et al., 2004). Previous research has identified a role for pilin-eDNA interactions in biofilm integrity (Allesen-Holm et al., 2006, Jurcisek & Bakaletz, 2007), albeit a net stabilizing effect of eDNA rather than the inhibitory effect reported here. eDNA still inhibited biofilm formation of a pilus- and flagellum-deficient mutants at a magnitude comparable to wild-type, suggesting that eDNA does not inhibit the reversible attachment stage mediated by these structures (Fig. 5).

Figure 5. The biofilm inhibition is not due to interaction between eDNA and pili or eDNA and flagellum.

Figure 5

Open in a new tab

Static biofilm assays were performed using 10 µg/ml sheared gDNA extracted from CB15 (black solid bars), CB15 ΔpilA (stripped bars) or CB15 ΔflgE (dotted bars) in 3 ml M2G (12-well plates). Control biofilm grown in the absence of eDNA are represented with white bars. Tested bacteria were incubated for 16 hrs at 30°C and the biofilm attached to the coverslips was quantified with crystal violet. Results are given as the biofilm score (amount of biofilm formed on the coverslip corrected by the culture cell density). Despite the fact that the ΔpilA and ΔflgE mutants are already deficient in biofilm formation, the addition of eDNA reduces their biofilm formation by ~70 %, as is the case for wild-type. The error bars represent the S.E.M of 2 independent experiments run in duplicate.

In order to determine how the initial attachment of cells to surfaces was affected by eDNA, we monitored single cell attachment to glass by microscopy. eDNA significantly decreased the binding efficiency of synchronized single swarmer cells (Fig. 6A), even in the presence of kanamycin, indicating that new protein synthesis is not required for the inhibition (Fig. S4). In contrast, eDNA had no effect on the reversible attachment of a holdfast synthesis mutant (Fig. 6B), indicating that attachment inhibition by eDNA requires the holdfast.

Figure 6. eDNA prevents the reversible attachment of cells and binds specifically to the C. crescentus holdfast.

Figure 6

Open in a new tab

(A) Attachment of wild-type C. crescentus CB15 synchronized swarmer cells to a glass surface in the presence (open circles) or absence (solid diamonds) of 10 µg/ml eDNA. The swimming cells are given as a percentage of the overall cell population at various times. Results are averages from duplicate samples in 3 independent experiments. (B) Swimming cell quantification of the holdfast-deficient C. crescentus CB15 ΔhfsDAB mutant in the presence (open circles) or in the absence (solid diamonds) of 10 µg/ml eDNA. Results are given as a percentage of swimming cells at t = 0 for 2 independent replicates. (C) Biofilm and planktonic cells of C. crescentus wild-type strain CB15 and a holdfast-minus ΔhfsDAB mutant in the presence of AF 488-labeled eDNA purified from C. crescentus CB15 (green), and AF 594-labeled WGA (red), to visualize holdfast. (D) Coverslip binding of cells of C. crescentus CB15 and holdfasts left on the surface following the removal of the holdfast shedding mutant CB15 ΔhfaB and fluorescence localization of eDNA (green) and holdfast (red).

In order to determine whether eDNA inhibits C. crescentus adhesion by interacting with the holdfast, we labeled sonicated C. crescentus genomic DNA (gDNA) or purified eDNA using AlexaFluor (AF) 488 and determined their localization by fluorescence microscopy. Similar results were obtained with both types of DNA (not shown). When added to a preformed biofilm, the AF 488-labeled DNA was associated with cells in the biofilm and exhibited a punctate staining pattern (Fig. 6C). Co-staining of the holdfast using AF 594-labeled wheat-germ agglutinin (WGA) indicated that DNA co-localized with the holdfast in the biofilm (Fig. 6C). Observation of planktonic cells confirmed that eDNA bound at the location of the holdfast in stalked cells and in swarmer cells harboring a holdfast, but eDNA failed to bind to a holdfast synthesis mutant (Fig. 6C), showing that the holdfast is necessary for eDNA binding to the tip of the stalk or pole of Caulobacter cells.

We used the C. crescentus ΔhfaB holdfast-shedding mutant strain to cover a surface with holdfasts without attached cells (Hardy et al., 2010). We found that eDNA is co-localized with WGA on the isolated holdfasts (Fig. 6D), revealing that eDNA strictly interacts with holdfasts. Finally, we observed surface attachment of purified holdfasts in suspension to determine the strict holdfast/eDNA interaction and subsequent attachment inhibition. This approach specifically examines the direct interaction between holdfast and eDNA while eliminating confounding factors in biofilm formation, such as reversible attachment and motility. C. crescentus eDNA once again strongly inhibited holdfast attachment, with almost no holdfast binding for eDNA concentrations higher than 25 µg/ml (Fig. 7A).

Figure 7. Different eDNA can inhibit purified holdfasts binding, but a strong inhibition only occurs in response to DNA from Caulobacter and close relatives.

Figure 7

Open in a new tab

(A) Surface attachment of purified holdfasts in the presence of various concentrations of eDNA. Holdfasts purified from the holdfast shedding mutant C. crescentus CB15 ΔhfaB were incubated in suspension in the presence of eDNA from C. crescentus CB15 (solid circles), B. diminuta (solid triangles), A. biprosthecum (open squares), and R. palustris (open triangles). Holdfasts were then allowed to bind to a glass coverlip for 4 h. AF 488-labeled WGA was used to visualize holdfasts by fluorescence microscopy. The number of holdfast attached per field of view was quantified using the ImageJ analysis software. The results are expressed as a percentage of the number of holdfasts attached in the absence of eDNA. The error bars represent the S.E.M of 10 samples from at least 3 independent experiments. (B) Phylogenetic distribution of bacterial species whose DNA was tested in purified holdfast attachment and biofilm inhibition assays. The tree represents a maximum likelihood phylogeny based on 1370 aligned positions from 16S ribosomal RNA gene sequences obtained from GenBank. PAUP* v4.0b10 (Swofford 2003) obtained the maximum likelihood reconstruction via a heuristic search using the HKY+I+G substitution model, with relevant parameters estimated by maximum likelihood on an initial neighbor-joining tree. C. crescentus CB13 was tested but not added to this tree, as its 16S ribosomal RNA sequence is not available.

We conclude that the holdfast is necessary and sufficient for the binding of the inhibitory eDNA to cells and that eDNA limits swarmer cell attachment by inhibiting holdfast adhesiveness. In addition, these results and the results of flow cell experiments presented later indicate that an adhesive contact already made with a surface cannot be readily disrupted by eDNA or WGA-lectin when they bind to exposed areas of the holdfast.

Biofilm inhibition by eDNA is specific to Caulobacter species eDNA

To determine whether the observed biofilm inhibition is specific to C. crescentus DNA, we tested four additional Caulobacter strains and fourteen other bacterial species (Fig. 7B). Only spent media from Caulobacter strains, eDNA purified from these spent media, or sonicated gDNA from these strains inhibited C. crescentus CB15 biofilm formation (Table 1). Thus, C. crescentus biofilm inhibition is a specific property of Caulobacter DNA and not a general property of DNA or negatively charged molecules.

Table 1.

Species-specificity of the biofilm inhibition.
Biofilm formation (%)a
Tested bacterium CB15 + tested
bacterium spent
mediumb 		CB15 + tested
bacterium eDNAc 		CB15 + tested
bacterium gDNAd
C. crescentus CB15 44.7 (± 1.3) 44.8 (± 5.3) 40.1 (± 1.1)
C. crescentus NA1000 40.1 (± 5.2) 43.9 (± 6.1) 41.6 (± 3.9)
C. crescentus ATCC
19089		 42.4 (± 5.7) 39.8 (± 3.5) 47.8 (± 4.4)
Caulobacter sp. K31 41.0 (± 5.6) 36.7 (± 9.2) 50.6 (± 0.7)
C. crescentus CB2A 42.0 (± 9.6) 43.1 (± 3.2) 51.6 (± 2.6)
C. crescentus CB13 56.8 (± 6.3) 45.8 (± 5.8) 55.0 (± 2.4)
A. biprosthecum C19 90.8 (± 5.3) 93.5 (± 3.8) 91.1 (± 5.7)
A. excentricus C48 101.2 (± 11) 99.1 (± 9.3) 93.0 (± 1.7)
B. diminuta 107.7 (± 11) 102.3 (± 9.5) 90.4 (± 1.7)
B. subvibrioides 93.7 (± 17.1) 98.4 (± 6.4) 94.5 (± 2.0)
B. alba 103.4 (± 4.0) 94.9 (± 10.3) 98.1 (± 2.5)
B. bacterioides 98.2 (± 5.7) 103.1 (± 8.7) 94.5 (± 4.9)
A. tumefaciens C58 96.7 (± 6.9) 99.1 (± 7.4) 101.1 (± 3.4)
R. capsulatus SB1003 95.1 (± 1.0) 89.7 (± 13.2) 90.0 (± 3.6)
R. sphaeroides 2.4.1 n.te n.t 94.4 (± 2.1)
R. palustris CGA009 n.t n.t 84.7 (± 6.6)
B. subtilus ATCC
35854		 n.t n.t 101.8 (± 0.4)
E. coli K12 n.t n.t 93.1 (± 5.4)
P. putida n.t n.t 98.4 (± 1.7)
S. meliloti n.t n.t 94.5 (± 5.1)
Open in a new tab
a

Biofilm formation was calculated as a percentage of biofilm formation in the absence of spent medium or sheared gDNA. Percentages were calculated from at least three independent experiments done in duplicate. S.E.M are indicated between parentheses.

b

The concentration of eDNA was normalized to provide 15 µg/ml of eDNA for all the biofilm assays.

c

eDNA was extracted from spent media and added to a final concentration of 15 µg/ml for all biofilm assays.

d

The final concentration of purified and sonicated gDNA was 15 µg/ml for all the biofilm assays.

e

Only purified sheared gDNA addition was tested for these strains.

C. crescentus CB15 DNA has a high GC content (67 %); however the inhibitory activity is not specific to high GC DNA since DNA from Rhodobacter capsulatus or Brevundimonas diminuta, whose genomes also have a GC content of 67%, do not inhibit C. crescentus biofilm formation (Table 1). Biofilm assays using hemi-methylated DNA extracted from a C. crescentus strain depleted of the DNA methylase CcrM (Stephens et al., 1996) indicated that the inhibition specificity did not result from DNA modification (Fig. S5).

Although DNA from non-Caulobacter species failed to inhibit biofilm formation, we sought to determine whether any inhibition might be detected by the more sensitive assay using isolated holdfasts in suspension. eDNA from the closely related species B. diminuta inhibits holdfast binding to some extent in the purified holdfast adhesion assay, with around 60% inhibition at 25 µg/ml (Fig. 7A). A weaker attachment inhibition can also be detected with the other tested eDNA, with a positive correlation between phylogenetic distance and the degree of inhibition (Fig. 7B). The results indicate that the magnitude of Caulobacter holdfast binding inhibition correlates directly with the relatedness of the gDNA donor bacterium, resulting in strong binding inhibition only in response to DNA from Caulobacter and close relatives.

We examined bacterial genomic sequence data for enrichment of particular motifs that may impart the observed specificity of biofilm inhibition by Caulobacter DNA. We sought to identify any correlations between the genomic abundance of particular sequences and the magnitude of holdfast binding inhibition (Fig. 7A) across four bacterial species: C. crescentus CB15, R. palustris CGA009, B. diminuta and A. biprosthecum. In order to identify candidate inhibitory sequences, we used MEME 4.3.0 (Bailey & Elkan, 1994) to search for repeated sequence elements in the C. crescentus genome. We then compared the frequency of each motif in the various query genomes as determined by MAST. None of the motif abundance patterns reflects the large differences in holdfast binding inhibition observed among the corresponding gDNA treatments (Table S1). Even in cases where the relative ordering of genomes based on motif abundance agrees with their ordering based on holdfast inhibition (e.g. Motif Max16-7), the differences in motif frequencies appear inconsistent with the corresponding magnitude of inhibition (Fig. 7A). Quite possibly, an enriched inhibitory sequence exists in Caulobacter, but the search space rapidly becomes intractable with increasing sequence length and complexity (e.g. mixtures of specific and non-specific sites in a signature sequence and/or variable spacing between specific residues).

eDNA prevents the attachment of newborn cells but does not disperse existing biofilm

We have shown that eDNA acts as a competitive inhibitor of holdfast adhesion by coating its adhesive surface, thereby reducing its ability to bind to new surfaces, but that eDNA does not dislodge previously bound holdfast from a surface. This suggests that eDNA might specifically prevent adhesion of newborn cells in a biofilm without disrupting stalked cells already bound by their holdsfast. Indeed, static biofilm experiments showed that eDNA does not cause dispersal of pre-established biofilms (Fig. 8A), but instead blocks the increase of biofilm biomass over time (Fig. 8B). Furthermore, as described earlier, biofilm dispersion is enhanced in the presence of eDNA compared to a DNase I treated biofilm (Fig. 4C), showing that newborn cells are prevented from settling within the biofilm when the eDNA concentration is sufficient.

Figure 8. Effect of spent medium on different stages of biofilm formation.

Figure 8

Open in a new tab

(A) Addition of spent medium to mature biofilms. Biofilms of C. crescentus were grown on coverslips for 24 hours in fresh M2G medium (before addition) and then placed in the presence of spent medium or fresh M2G medium for an additional 24 hours. (B) Time course study of the effect of C. crescentus CB15 spent medium on C. crescentus CB15 biofilm formation under static conditions. 10 µg/ml of eDNA was added (open circles) or not (solid diamonds) to M2G for static biofilm assays. The bacteria were incubated for various times at 30°C and the biofilm attached to the coverslips was quantified with crystal violet. The error bars represent the S.E.M of at least 3 independent experiments run in duplicate.

Under dynamic-flow conditions, no biofilm development could be detected in the presence of eDNA; only a few cells were found on the surface of flow-cells irrigated with a spent:fresh medium mixture even after 4 days (Fig. 9A, compare flow-cells a and e), indicating that eDNA inhibits new adhesion to the surface as shown by surface coverage quantification (Fig. 9B). eDNA stopped the increase of biomass regardless of the maturation stage of the biofilm, but did not disrupt established biofilms (Fig. 9C). Therefore, eDNA acts in a cell-specific manner to prevent growth of a biofilm without causing its dissolution.

Figure 9. Effect of spent medium on biofilms at different stages of biofilm formation under dynamic flow conditions.

Figure 9

Open in a new tab

Four flow-cells were inoculated with GFP-labeled C. crescentus CB15, irrigated with M2G, and incubated at 30°C. After 24 (b), 48 (c) or 72 (d) h, the irrigating medium was switched to a 33% spent:fresh M2G mixture and the incubation continued for 96 h. Two control flow-cells were irrigated with a 33% spent:fresh M2G mixture (a) and with fresh M2G (e) for the entire 96 h to serve as references. (A) Microscopy images of the cells at various stages. AutoCOMSTAT was used to determine the surface coverage (B) and the total biomass (µm3 of fluorescence per µm2 of surface area) (C) in the five flow-cells over time. Values were average for 10 image stacks (5 image stacks from 2 different flow-cell channels for each sample). The arrows indicate the switch of medium irrigating the flow-cells from M2G to a 33% spent:fresh M2G mixture.

Discussion

Association with a sessile biofilm and dispersal into new environments provide different benefits to bacteria that depend on the current ecological context. Mechanisms that maintain a balance between these two strategies should enhance bacterial fitness, particularly when such mechanisms tune bacterial behavior to suit the current environmental state. Specifically, favorable local conditions should tend to favor biofilm maturation and persistence, whereas deterioration of the local environment due to crowding, predation, or other deleterious factors should promote dispersal into alternate habitat patches. Here we report that eDNA can modulate this balance by specifically stimulating the dispersal of newborn swarmer cells in a C. crescentus biofilm. We show that eDNA inhibits single cell attachment and biofilm growth irrespective of the stage of biofilm maturation but that it does not dissolve existing biofilm. eDNA binds specifically to the adhesive holdfast and the holdfast is both necessary and sufficient for eDNA binding. Interestingly, our observation that eDNA can inhibit biofilm formation stands in stark contrast with previous studies showing that eDNA often plays an important role in biofilm formation and stabilization as a major compound of the extracellular biofilm matrix in many bacterial species (Allesen-Holm et al., 2006, Spoering & Gilmore, 2006, Karatan & Watnick, 2009).

Since cell death occurs at a sufficient frequency in a C. crescentus biofilm even under favorable growth conditions, we hypothesize that the inhibitory effect of eDNA increases dispersal of swarmer cells born in a biofilm. The specific targeting of the swarmer cell type in the developing biofilm allows the stimulation of dispersal without the simultaneous destruction of the biofilm. Holdfast-bearing stalked cells are non-motile and associate tightly with the biofilm through their holdfast, and we have shown that pre-bound holdfasts cannot be dislodged from a surface by eDNA. Therefore stalked cells are not able to disperse, even when the eDNA concentration increases within the biofilm, explaining why mature biofilms are not disrupted. Our results indicate that eDNA functions by coating the adhesive surface of the holdfast of swarmer cells, thereby reducing their ability to bind to surfaces. Even if a dispersing swarmer cell with bound eDNA will not be able to attach efficiently to a surface, it will eventually differentiate into a stalked cell and produce a progeny swarmer cell free of eDNA and capable of settling in a new environment.

This method of biofilm inhibition permits rapid tuning of the degree of dispersal in response to environmental cues. Previous studies have identified various dispersion mechanisms, including active degradation of the extracellular matrix, induction of motility, production of surfactants or cell lysis, in each case mediated by signal transduction and/or regulatory changes within the cell (Karatan & Watnick, 2009, Goller & Romeo, 2008). In contrast, our study shows that eDNA acts via direct interference with cell attachment and requires no intermediate intracellular response. Since the magnitude of inhibition is roughly proportional to the eDNA concentration, this mechanism may provide a rheostat control of biofilm inhibition: cells should be less affected by small quantities of DNA released through sporadic cell death, but they should be strongly inhibited under conditions when cell death is more frequent. This mechanism can also promote swarmer cell dispersal without causing a potentially undesirable dissolution of the biofilm. Particularly in the case of over-crowding, cells currently residing in the biofilm may continue to occupy a productively colonized habitat patch, albeit at a saturating density; whereas inhibition of new attachment by eDNA promotes dispersal of newborn swarmer cells that would otherwise stretch the biofilm beyond its carrying capacity.

While our data support the role of eDNA as a cue that prevents settling of swarmer cells, it is unclear if this mechanism constitutes cooperation (Nadell et al., 2009), or is simply a consequence of random cell death occurring within the biofilm. Cooperative behaviors involve a fitness cost to the actor while benefiting other individuals (Nadell et al., 2009). In biofilm inhibition by eDNA, such behaviors could include active lysis via programmed cell death (Claverys & Havarstein, 2007, Bayles, 2007, Rice & Bayles, 2008, Thomas et al., 2009, Lopez et al., 2009) or production of endonucleases that specifically digest the released DNA. On the other hand, if eDNA only arises from non-regulated cell death and is not cleaved by an endonuclease produced specifically for that process, there is no direct cost to the dying cells and therefore no cooperation.

The species-specific nature of biofilm inhibition by eDNA provides the ability to respond to close relatives, an advantage since DNA is an abundant molecule in all environments that support life (Lorenz & Wackernagel, 1994). Therefore, the species-specificity of the inhibition should allow C. crescentus to colonize permissive environments that contain eDNA from other organisms. Furthermore, this specificity may increase the likelihood that Caulobacter cells avoid settling in environments already densely inhabited by closely related bacteria, thereby reducing intraspecific resource competition.

The mechanism underlying species specificity of the eDNA-holdfast interaction remains undetermined. The holdfast is composed in part of β-1,4-N-acetylglucosamine polysaccharides (Merker & Smit, 1988), but its detailed chemical composition is not known. Interestingly, β-1,3-glucan polysaccharides can specifically interact with certain polynucleotides by forming triple-stranded and helical macromolecular complexes (Sakurai et al., 2005); similar DNA-polysaccharide interactions may mediate the sequence-specificity of the eDNA-holdfast interaction. The next challenge is to determine how species-specific inhibition is achieved.

Experimental procedures

Bacterial strains, plasmids, and growth conditions

Bacterial strains used in this study are listed in Table 2. C. crescentus was grown at 30°C in complex peptone-yeast extract (PYE) medium (Poindexter, 1964) or in minimal M2 medium supplemented with 0.2% glucose (M2G) (Johnson & Ely, 1977). For some experiments, different concentrations of glucose (0.2%, 0.1%, 0.05%, 0.02% and 0.01% w/v) were used in M2G as well as xylose and alanine as carbon sources in M2 medium. PYE diluted 50% v/v (PYE 1/2) was used in some experiments.When using GFP-expressing C. crescentus CB15 (YB4789), kanamycin was added to the culture medium at a final concentration of 5 µg/ml. YB4789 was constructed by φCr30 mediated transduction (West et al., 2002) of the miniTn7-gfp from AS110 (Entcheva-Dimitrov & Spormann, 2004) into C. crescentus CB15.

Table 2.

Strains and plasmids used in this study

Strain or plasmid Description Source or reference
Strains
C. crescentus
  CB15 (YB135) Wild-type (Poindexter, 1964)
  CB15 ΔhfaB (YB4251) Clean deletion of hfaB (Hardy et al., 2010)
  CB15 ΔhfsDAB (YB2857) Clean deletion of hfsDAB C.S. Smith, unpublished
  AS110 (YB4784) CB15∷miniTn7gfp (Entcheva-Dimitrov & Spormann, 2004)
  YB4789 CB15∷miniTn7gfp This study
  YB127 NA1000, holdfast deficient (Evinger & Agabian, 1977)
  ATCC 19089 (YB1360) Wild-type ATTCa
  CB2A (YB1336) Wild-type (Smit & Agabian, 1984)
  CB13 (YB234) Wild-type (Lagenaur & Agabian, 1977)
  LS2144 (YB5795) NA1000 ΔccrM/pCS226 (Stephens et al., 1996)
Caulobacter sp. K31 (YB4571) Wild-type (Mannisto et al., 1999)
Other bacteria
  Asticcacaulis biprosthecum C19 (YB642) Wild-type (Larson & Pate, 1975)
  Asticcacaulis excentricus C48 (YB258) Wild-type (Pate & Ordal, 1965)
  Agrobacterium tumefaciens C58 (YB2951) Wild-type (Watson et al., 1975)
  Brevundimonas alba ATCC 15265
(YB3925)		 Wild-type ATCCa
  Brevundimonas bacteroides ATCC 19090
(YB3922)		 Wild-type ATCCa
  Brevundimonas dimunuta ATCC 11568
(YB5193)		 Wild-type ATCCa
  Brevundimonas subvibrioides ATCC
15264 (YB5185)		 Wild-type ATCCa
  Rhodobacter capsulatus SB1003 (YB4650) Wild-type C.E. Bauer
  Rhodobacter sphaeroides 2.4.1 (YB4653) Wild-type S. Kaplan
  Rhodopseudomonas. palustris CGA009
(YB4651)		 Wild-type (Larimer et al., 2004)
  Bacillus subtilus ATCC 35854 (YB3881) Wild-type ATCCa
  Escherichia coli K12 (YB279) Wild-type (Ackerley et al., 2006)
  Pseudomonsa putida ATCC 17422
(YB1309)		 Wild-type ATCCa
  Sinorhizobium meliloti 1021 (YB113) Wild-type (Meade et al., 1982)
Open in a new tab
a

American Type Culture Collection

Preparation of spent medium

Bacteria were grown for 24 hours with shaking at 30°C in M2G or in PYE medium. At that point, cultures were in stationary phase (OD600 of 1.2 – 1.8). Bacteria were removed by centrifugation and the spent medium was filter-sterilized using a 0.2 µm filter and kept at −20°C until needed.

DNA extraction and sonication

Genomic DNA was extracted from 1 ml overnight culture in PYE using the Bactozol kit for Bacterial DNA Isolation (Molecular Research Center) and was resuspended in 500 µl dH2O. DNA was sonicated using a Misonix 3000 sonicator apparatus and a cup horn attachment filled with iced water for 2 cycles (10 seconds on, 2 seconds off) at power 2.

Fluorescent labeling

Extracellular DNA (eDNA) and genomic DNA (gDNA) were labeled with AF 488 using the ULYSIS Nucleic Acid labeling system (Molecular Probes) according to the manufacturer’s instructions. C. crescentus holdfasts were labeled using AF 594 conjugated WGA (Molecular Probes). WGA binds specifically to the N-acetyl glucosamine residues of the holdfast (Merker & Smit, 1988). Fluorescent and phase contrast microscopy were performed using a Nikon Eclipse E800 microscope.

Static biofilm experiments

Bacteria were grown to mid-log phase in PYE or M2G medium and diluted to A600 = 0.05 in the same medium. Biofilm assays were conducted in two different experimental set-ups: 1. in 12-well polystyrene plates containing polyvinylchloride (PVC) coverslips placed vertically in the wells where the biofilm formed on the coverslip is measured or 2. in 24-well polystyrene plates without coverslips where the biofilm formed inside the well is measured. For biofilms grown in 12-well plates with coverslips, the final volume of diluted cultures was 3 ml. The final volume in the 24-well plates without coverslips was 500 µl. Both methods gave comparable results. When specified, 33% of filtered spent medium, purified eDNA, or sheared DNA resuspended in dH2O were added directly to the medium in the wells. Plates were incubated at 30°C overnight. After incubation, coverslips or wells were rinsed with dH2O to remove planktonic cells, stained with a 0.1% crystal violet solution for 5 minutes, and rinsed again with dH2O to remove excess crystal violet. The crystal violet was dissolved using 10% acetic acid and quantified by measuring the absorbance at 600 nm (A600).

Hydrodynamic biofilm experiments

Five sterile flow cells (three 40 mm × 5 mm × 1 mm channels) were equilibrated for 24 h at 30°C prior to inoculation (4 flow-cells were equilibrated with M2G and 1 with a mixture 33% spent medium:M2G). Mid-log phase cultures of C. crescentus CB15∷miniTn7gfp (YB4089) were adjusted to A600 = 0.05 and 200 µl were inoculated per chamber. Initial attachment was performed in the absence of flow for 1 h, followed by a constant flow of 3 ml/h. Surface colonization of the glass surface covering the flow cells was monitored for 5 days at 30°C. The medium was switched from M2G to the spent medium mixture at different times after inoculation (0, 24, 48 and 72 h). Images of the biofilms grown in the flow-cells were recorded by confocal scanning laser microscopy and quantitative analyses were performed using autoCOMSTAT, a modified version of the COMSTAT software (Merritt et al., 2007). Each experimental condition was performed twice in parallel triplicate chambers.

Coverslip binding assay

Coverslip binding assays were performed as described (Cole et al., 2003) with some modifications. Sterilized coverslips (22 × 22 mm) were placed at the bottom of a 6-well polystyrene plate containing 1.5 ml M2G–grown bacteria diluted to A600 = 0.2. Plates were then incubated at 30°C for 4 h under constant agitation (50 rpm). After incubation, coverslips were rinsed with dH2O to remove all unattached cells. AF 488-labeled sheared C. crescentus CB15 gDNA (10 µg/ml) and/or AF 594-labeled WGA (5 µg/ml) were added to the rinsed coverslips and incubated in the dark for 30 min at room temperature. Coverslips were then rinsed with dH2O, mounted on a microscopy glass slide with a drop of M2G under a large coverslip (24 × 50 mm) and sealed with nail polish.

To determine if the DNA binds directly to the holdfast, we used a C. crescentus ΔhfaB holdfast shedding mutant, YB4251, that can synthesize and export holdfast polysaccharide but cannot keep it attached to the tip of the stalk (Kurtz & Smith, 1992, Cole et al., 2003, Hardy et al., 2010), to coat microscope coverslips with holdfasts without cells. We then incubated the coverslips with labeled C. crescentus DNA and stained the holdfast with WGA as below.

Cell synchronization and initial attachment of C. crescentus single cells to surfaces

Small scale synchronization of C. crescentus CB15 was performed as described previously (Degnen & Newton, 1972) with some modifications. An overnight culture of C. crescentus CB15 in M2G was diluted 10 times in fresh M2G and incubated shaking at 50 rpm at room temperature in a 150 mm diameter Petri dish. After overnight incubation, the medium was removed and the monolayer biofilm formed in the dish was thoroughly washed. 60 ml of fresh M2G medium was added to the dish and the incubation was carried out for 4 h under the same conditions. The dish was then rinsed 10 times using 10 ml of fresh M2G medium to remove unattached cells from the biofilm. 1 ml of M2G medium was added to the dish and the incubation was carried out 5 min under the same conditions to release the newly born swarmer cells. The homogeneity of the synchronized swarmer population was checked by microscopy.

Initial attachment of C. crescentus cells to surfaces was recorded by dark-field microscopy. A 1 µl aliquot of synchronized swarmer cells was observed by dark-field microscopy (10X objective, stage warmed at 30°C). A stack of 100 frames (100 ms exposure time) was recorded every min during 60 min. To be able to track attached cells over time, images were processed as follows: the average of the 100-frame stack at t = 0 was calculated. This average file was subtracted from the entire stack to eliminate background noise. The maximum intensity of the 10 first frames of each stack was calculated in order to visualize motile cells with a swimming trajectory. All the calculations were performed using the built-in functions in the Metamorph version 7 software package (Molecular Devices).

The attachment of C. crescentus CB15 ΔhfsDAB cells (YB2857) was studied with non synchronized cells, as this holdfast-minus mutant cannot be synchronized under the conditions mentioned above. An aliquot of cells grown to exponential phase was used instead of the synchronized swarmer cells.

Purified holdfast attachment assay

This assay was performed using 12 well glass multitest slides (MP Biomedicals). To maximize holdfast binding, the multitest slides were soaked overnight in 100% Micro-90 detergent, carefully rinsed with dH2O and cleaned with ethanol prior to use.

The holdfast-shedding mutant C. crescentus CB15 ΔhfaB strain (YB4251) was grown to late exponential phase in M2G and cells were pelleted by centrifugation (30 min at 4,000 g). The supernatant contains free holdfasts shed by the ΔhfaB mutant. Different concentrations of purified sonicated gDNA were added to 100 µl of holdfast-containing supernatant. 50 µl samples were spotted on a precleaned 12 well glass multitest slide and incubated for 2 to 5 h at room temperature in a humid chamber. After incubation, the slides were rinsed with dH2O to remove unbound material. AF 488-labeled WGA (50 µl at 5 µg/ml) was added to the rinsed wells and incubated in the dark for 20 min at room temperature. Slides were then rinsed with dH2O, toped with a large glass coverslip (24 × 50 mm) and sealed with nail polish. Holdfast attachment was visualized by epifluorescence microscopy and quantified using Image J analysis software.

Biofilm extracellular matrix and DNA extraction

C. crescentus CB15 wt cells were grown in 2 ml microtubes (1 ml final culture in M2G, starting OD600 = 0.05) and sealed with a piece of sterile breathable sealing film (AeraSeal, Excel Scientific). After incubation at 30°C for different amounts of time, the planktonic phase was carefully removed and biofilm extracellular matrix was extracted as described previously (Schooling & Beveridge, 2006) with some modifications. 1 ml M2 was added and the treated tube was incubated with DNase I for 1 h at 37°C to degrade eDNA. The other tube received the same treatment but without DNase I. Both tubes were heated at 70°C for 30 min to inactivate DNase I in the treated tube. Controls indicated that heating did not change the eDNA quantification results. The biofilms formed inside the microtubes were resuspended in sterile 0.9% NaCl by vortexing 3 min. Crystal violet staining was used to confirm that the biofilms were completely resuspended. The tubes were then centrifuged at 12,000 g at 4°C for 20 min. The supernatants were filtered through 0.2 µm filter (Millipore) to remove any remaining cells. eDNA was quantitated using a PicoGreen assay (Molecular Probes) following the manufacturer’s instructions. The total eDNA concentration was determined as the difference between the non-treated and the DNase I treated samples at each time point and includes both diffusible eDNA and eDNA attached to the matrix. The amount of eDNA attached to matrix was determined using the same procedure but after the biofilms were washed with sterile water three times prior to the procedure.

The effect of DNase I treatment on biofilm formation and stability was determined in similar experiments in which the biofilm was quantified after DNase I treatment using crystal violet as described above.

Live/dead staining

Biofilms were grown on plastic coverslips as described above and incubated for different times. After incubation, the coverslips were carefully rinsed with dH2O. 0.5 µl of Live/Dead stain mixture (Live/Dead BacLight Bacterial Viability kit L7007, Molecular Probes) was added to 500 µl dH2O, placed on top of the coverslip and incubated for 15 min at room temperature in the dark. The stained coverslips were then rinsed with dH2O and observed under epifluorescent microscopy.

Bioinformatic analysis

Bioinformatic searches utilized genome sequences of C. crescentus CB15 (NC_002696), R. palustris CGA009 (NC_005296), B. diminuta (GDSub11081) and A. biprosthecum (GDSub11080).

To select candidate biofilm inhibitory sequences from Caulobacter, we sampled a 24 kb region of the C. crescentus CB15 genome and searched for repeated sequence elements using MEME 4.3.0 (Bailey & Elkan, 1994) to identify the 20 best-scoring elements with 10 to 500 occurrences in the sampled region. Since the MEME scoring strategy favors longer matches, we iterated the analysis with maximum motif lengths of 6, 9, 12, 16, 24 and 50 bases in order to consider a range of target sequence lengths, also constraining minimum match lengths to avoid overlap with searches that used shorter maxima. MAST (Bailey & Gribskov, 1998) then determined the frequency of each motif in the query genomes.

Supplementary Material

Supp Fig legend
Supp Fig s1
Supp Fig s2
Supp Fig s3
Supp Fig s4
Supp Fig s5

Acknowledgements

We thank A.M. Spormann for providing strain AS110. Guanglai Li kindly shared an unpublished method for detecting motile Caulobacter cells using darkfield microscopy. We are grateful to the members of our laboratory, Dan Kearns, Armin Moczek, Greg Velicer, Curt Lively, Tom Platt, Jean-Marc Ghigo, and Clay Fuqua for their critical reading of this manuscript and helpful discussions, and to Sarah Schooling and Sima Setayeshgar for helpful discussions.

References

  1. Ackerley DF, Barak Y, Lynch SV, Curtin J, Matin A. Effect of chromate stress on Escherichia coli K-12. J Bacteriol. 2006;188:3371–3381. doi: 10.1128/JB.188.9.3371-3381.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allesen-Holm M, Barken KB, Yang L, Klausen M, Webb JS, Kjelleberg S, Molin S, Givskov M, Tolker-Nielsen T. A characterization of DNA release in Pseudomonas aeruginosa cultures and biofilms. Mol Microbiol. 2006;59:1114–1128. doi: 10.1111/j.1365-2958.2005.05008.x. [DOI] [PubMed] [Google Scholar]
  3. Bailey TL, Elkan C. Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proceedings of the Second International Conference on Intelligent Systems for Molecular Biology. 1994:28–36. [PubMed] [Google Scholar]
  4. Bailey TL, Gribskov M. Combining evidence using p-values: application to sequence homology searches. Bioinformatics. 1998;14:48–54. doi: 10.1093/bioinformatics/14.1.48. [DOI] [PubMed] [Google Scholar]
  5. Bayles KW. The biological role of death and lysis in biofilm development. Nat Rev Microbiol. 2007;5:721–726. doi: 10.1038/nrmicro1743. [DOI] [PubMed] [Google Scholar]
  6. Bodenmiller D, Toh E, Brun YV. Development of surface adhesion in Caulobacter crescentus. J Bacteriol. 2004;186:1438–1447. doi: 10.1128/JB.186.5.1438-1447.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Brown PJ, Hardy GG, Trimble MJ, Brun YV. Complex regulatory pathways coordinate cell-cycle progression and development in Caulobacter crescentus. Adv Microb Physiol. 2009;54:1–101. doi: 10.1016/S0065-2911(08)00001-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Claverys JP, Havarstein LS. Cannibalism and fratricide: mechanisms and raisons d'etre. Nat Rev Microbiol. 2007;5:219–229. doi: 10.1038/nrmicro1613. [DOI] [PubMed] [Google Scholar]
  9. Cole JL, Hardy GG, Bodenmiller D, Toh E, Hinz A, Brun YV. The HfaB and HfaD adhesion proteins of Caulobacter crescentus are localized in the stalk. Mol Microbiol. 2003;49:1671–1683. doi: 10.1046/j.1365-2958.2003.03664.x. [DOI] [PubMed] [Google Scholar]
  10. Davey ME, O'Toole G A. Microbial biofilms: from ecology to molecular genetics. Microbiol Mol Biol Rev. 2000;64:847–867. doi: 10.1128/mmbr.64.4.847-867.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Degnen ST, Newton A. Chromosome replication during development in Caulobacter crescentus. J Mol Biol. 1972;64:671–680. doi: 10.1016/0022-2836(72)90090-3. [DOI] [PubMed] [Google Scholar]
  12. Entcheva-Dimitrov P, Spormann AM. Dynamics and control of biofilms of the oligotrophic bacterium Caulobacter crescentus. J Bacteriol. 2004;186:8254–8266. doi: 10.1128/JB.186.24.8254-8266.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Evinger M, Agabian N. Envelope-associated nucleoid from Caulobacter crescentus stalked and swarmer cells. J Bacteriol. 1977;132:294–301. doi: 10.1128/jb.132.1.294-301.1977. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Goller CC, Romeo T. Environmental influences on biofilm development. Curr Top Microbiol Immunol. 2008;322:37–66. doi: 10.1007/978-3-540-75418-3_3. [DOI] [PubMed] [Google Scholar]
  15. Hall-Stoodley L, Costerton JW, Stoodley P. Bacterial biofilms: from the natural environment to infectious diseases. Nat Rev Microbiol. 2004;2:95–108. doi: 10.1038/nrmicro821. [DOI] [PubMed] [Google Scholar]
  16. Hardy GG, Allen RC, Toh E, Long M, Cole JL, Brun YV. A localized multimeric anchor attaches the Caulobacter holdfast to the cell pole. Mol Microbiol. 2010;76:409–427. doi: 10.1111/j.1365-2958.2010.07106.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Harrison JJ, Ceri H, Turner RJ. Multimetal resistance and tolerance in microbial biofilms. Nat Rev Microbiol. 2007;5:928–938. doi: 10.1038/nrmicro1774. [DOI] [PubMed] [Google Scholar]
  18. Johnson RC, Ely B. Isolation of spontaneously derived mutants of Caulobacter crescentus. Genetics. 1977;86:25–32. doi: 10.1093/genetics/86.1.25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jurcisek JA, Bakaletz LO. Biofilms formed by nontypeable Haemophilus influenzae in vivo contain both double-stranded DNA and type IV pilin protein. J Bacteriol. 2007;189:3868–3875. doi: 10.1128/JB.01935-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Karatan E, Watnick P. Signals, regulatory networks, and materials that build and break bacterial biofilms. Microbiol Mol Biol Rev. 2009;73:310–347. doi: 10.1128/MMBR.00041-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kolter R, Greenberg EP. Microbial sciences: the superficial life of microbes. Nature. 2006;441:300–302. doi: 10.1038/441300a. [DOI] [PubMed] [Google Scholar]
  22. Kurtz HD, Jr, Smith J. Analysis of a Caulobacter crescentus gene cluster involved in attachment of the holdfast to the cell. J Bacteriol. 1992;174:687–694. doi: 10.1128/jb.174.3.687-694.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lagenaur C, Agabian N. Caulobacter flagellins. J Bacteriol. 1977;132:731–733. doi: 10.1128/jb.132.2.731-733.1977. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lappann M, Claus H, van Alen T, Harmsen M, Elias J, Molin S, Vogel U. A dual role of extracellular DNA during biofilm formation of Neisseria meningitidis. Mol Microbiol. 2010;75:1355–1371. doi: 10.1111/j.1365-2958.2010.07054.x. [DOI] [PubMed] [Google Scholar]
  25. Larimer FW, Chain P, Hauser L, Lamerdin J, Malfatti S, Do L, Land ML, Pelletier DA, Beatty JT, Lang AS, Tabita FR, Gibson JL, Hanson TE, Bobst C, Torres JL, Peres C, Harrison FH, Gibson J, Harwood CS. Complete genome sequence of the metabolically versatile photosynthetic bacterium Rhodopseudomonas palustris. Nat Biotechnol. 2004;22:55–61. doi: 10.1038/nbt923. [DOI] [PubMed] [Google Scholar]
  26. Larson RJ, Pate JL. Growth and morphology of Asticcacaulis biprosthecum in defined media. Arch Microbiol. 1975;106:147–157. doi: 10.1007/BF00446517. [DOI] [PubMed] [Google Scholar]
  27. Levi A, Jenal U. Holdfast formation in motile swarmer cells optimizes surface attachment during Caulobacter crescentus development. J Bacteriol. 2006;188:5315–5318. doi: 10.1128/JB.01725-05. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lopez D, Vlamakis H, Losick R, Kolter R. Cannibalism enhances biofilm development in Bacillus subtilis. Mol Microbiol. 2009;74:609–618. doi: 10.1111/j.1365-2958.2009.06882.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lorenz MG, Wackernagel W. Bacterial gene transfer by natural genetic transformation in the environment. Microbiol Rev. 1994;58:563–602. doi: 10.1128/mr.58.3.563-602.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mannisto MK, Tiirola MA, Salkinoja-Salonen MS, Kulomaa MS, Puhakka JA. Diversity of chlorophenol-degrading bacteria isolated from contaminated boreal groundwater. Arch Microbiol. 1999;171:189–197. doi: 10.1007/s002030050698. [DOI] [PubMed] [Google Scholar]
  31. Meade HM, Long SR, Ruvkun GB, Brown SE, Ausubel FM. Physical and genetic characterization of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J Bacteriol. 1982;149:114–122. doi: 10.1128/jb.149.1.114-122.1982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Merker RI, Smit J. Characterization of the Adhesive Holdfast of Marine and Freshwater Caulobacters. Appl Environ Microbiol. 1988;54:2078–2085. doi: 10.1128/aem.54.8.2078-2085.1988. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Merritt PM, Danhorn T, Fuqua C. Motility and chemotaxis in Agrobacterium tumefaciens surface attachment and biofilm formation. J Bacteriol. 2007;189:8005–8014. doi: 10.1128/JB.00566-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Meyers LA, Bull JJ. Fighting change with change: adaptive variation in an uncertain world. Trends in Ecology & Evolution. 2002;17:551–557. [Google Scholar]
  35. Moons P, Michiels CW, Aertsen A. Bacterial interactions in biofilms. Crit Rev Microbiol. 2009;35:157–168. doi: 10.1080/10408410902809431. [DOI] [PubMed] [Google Scholar]
  36. Nadell CD, Xavier JB, Foster KR. The sociobiology of biofilms. FEMS Microbiol Rev. 2009;33:206–224. doi: 10.1111/j.1574-6976.2008.00150.x. [DOI] [PubMed] [Google Scholar]
  37. Pate JL, Ordal EJ. The fine structure of two unusual stalked bacteria. J Cell Biol. 1965;27:133–150. doi: 10.1083/jcb.27.1.133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Poindexter JS. Biological Properties and Classification of the Caulobacter Group. Bacteriol Rev. 1964;28:231–295. doi: 10.1128/br.28.3.231-295.1964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rice KC, Bayles KW. Molecular control of bacterial death and lysis. Microbiol Mol Biol Rev. 2008;72:85–109. doi: 10.1128/MMBR.00030-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sakurai K, Uezu K, Numata M, Hasegawa T, Li C, Kaneko K, Shinkai S. Beta-1,3-glucan polysaccharides as novel one-dimensional hosts for DNA/RNA, conjugated polymers and nanoparticles. Chem Commun (Camb) 2005:4383–4398. doi: 10.1039/b506673p. [DOI] [PubMed] [Google Scholar]
  41. Schooling SR, Beveridge TJ. Membrane vesicles: an overlooked component of the matrices of biofilms. J Bacteriol. 2006;188:5945–5957. doi: 10.1128/JB.00257-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Smit J, Agabian N. Cloning of the major protein of the Caulobacter crescentus periodic surface layer: detection and characterization of the cloned peptide by protein expression assays. J Bacteriol. 1984;160:1137–1145. doi: 10.1128/jb.160.3.1137-1145.1984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Spoering AL, Gilmore MS. Quorum sensing and DNA release in bacterial biofilms. Curr Opin Microbiol. 2006;9:133–137. doi: 10.1016/j.mib.2006.02.004. [DOI] [PubMed] [Google Scholar]
  44. Spormann AM. Physiology of microbes in biofilms. Curr Top Microbiol Immunol. 2008;322:17–36. doi: 10.1007/978-3-540-75418-3_2. [DOI] [PubMed] [Google Scholar]
  45. Stephens C, Reisenauer A, Wright R, Shapiro L. A cell cycle-regulated bacterial DNA methyltransferase is essential for viability. Proc Natl Acad Sci U S A. 1996;93:1210–1214. doi: 10.1073/pnas.93.3.1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Stewart PS, Franklin MJ. Physiological heterogeneity in biofilms. Nat Rev Microbiol. 2008;6:199–210. doi: 10.1038/nrmicro1838. [DOI] [PubMed] [Google Scholar]
  47. Tart AH, Wozniak DJ. Shifting paradigms in Pseudomonas aeruginosa biofilm research. Curr Top Microbiol Immunol. 2008;322:193–206. doi: 10.1007/978-3-540-75418-3_9. [DOI] [PubMed] [Google Scholar]
  48. Thomas VC, Hiromasa Y, Harms N, Thurlow L, Tomich J, Hancock LE. A fratricidal mechanism is responsible for eDNA release and contributes to biofilm development of Enterococcus faecalis. Mol Microbiol. 2009;72:1022–1036. doi: 10.1111/j.1365-2958.2009.06703.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tsang PH, Li G, Brun YV, Freund LB, Tang JX. Adhesion of single bacterial cells in the micronewton range. Proc Natl Acad Sci U S A. 2006;103:5764–5768. doi: 10.1073/pnas.0601705103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Watson B, Currier TC, Gordon MP, Chilton MD, Nester EW. Plasmid required for virulence of Agrobacterium tumefaciens. J Bacteriol. 1975;123:255–264. doi: 10.1128/jb.123.1.255-264.1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. West L, Yang D, Stephens C. Use of the Caulobacter crescentus genome sequence to develop a method for systematic genetic mapping. J Bacteriol. 2002;184:2155–2166. doi: 10.1128/JB.184.8.2155-2166.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Whitchurch CB, Tolker-Nielsen T, Ragas PC, Mattick JS. Extracellular DNA required for bacterial biofilm formation. Science. 2002;295:1487. doi: 10.1126/science.295.5559.1487. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Fig legend
NIHMS219767-supplement-Supp_Fig_legend.doc (70KB, doc)
Supp Fig s1
NIHMS219767-supplement-Supp_Fig_s1.eps (79.7KB, eps)
Supp Fig s2
NIHMS219767-supplement-Supp_Fig_s2.eps (218.1KB, eps)
Supp Fig s3
NIHMS219767-supplement-Supp_Fig_s3.eps (253.3KB, eps)
Supp Fig s4
NIHMS219767-supplement-Supp_Fig_s4.eps (107.9KB, eps)
Supp Fig s5
NIHMS219767-supplement-Supp_Fig_s5.eps (62.6KB, eps)

RESOURCES