The pel genes of the Pseudomonas aeruginosa PAK strain are involved at early and late stages of biofilm formation

Perrine Vasseur1, Isabelle Vallet-Gely1,{dagger}, Chantal Soscia1, Stéphane Genin2 and Alain Filloux1

1 Laboratoire d'Ingénierie des Systèmes Macromoléculaires, CNRS-IBSM-UPR9027, 31 Chemin Joseph Aiguier, 13402 Marseille cedex 20, France
2 Laboratoire de Biologie Moléculaire des Relations Plantes-Micro-organismes, INRA-CNRS, Castanet-Tolosan, France

Correspondence
Alain Filloux
filloux{at}ibsm.cnrs-mrs.fr


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Pseudomonas aeruginosa is a Gram-negative bacterium associated with nosocomial infections and cystic fibrosis. Chronic bacterial infections are increasingly associated with the biofilm lifestyle in which microcolonies are embedded in an extracellular matrix. Screening procedures for identifying biofilm-deficient strains have allowed the characterization of several key determinants involved in this process. Biofilm-deficient P. aeruginosa PAK strains affected in a seven-gene cluster called pel were characterized. The pel genes encode proteins with similarity to components involved in polysaccharide biogenesis, of which PelF is a putative glycosyltransferase. PelG was also identified as a putative component of the polysaccharide transporter (PST) family. The pel genes were previously identified in the P. aeruginosa PA14 strain as required for the production of a glucose-rich matrix material involved in the formation of a thick pellicle and resistant biofilm. However, in PA14, the pel mutants have no clear phenotype in the initiation phase of attachment. It was shown that pel mutations in the PAK strain had little influence on biofilm initiation but, as in PA14, appeared to generate the least robust and mature biofilms. Strikingly, by constructing pel mutants in a non-piliated P. aeruginosa PAK strain, an unexpected effect of the pel mutation in the early phase of biofilm formation was discovered, since it was observed that these mutants were severely defective in the attachment process on solid surfaces. The pel gene cluster is conserved in other Gram-negative bacteria, and mutation in a Ralstonia solanacearum pelG homologue, ragG, led to an adherence defect.


Abbreviations: KDO, 2-keto-3-deoxyoctonate; PST, polysaccharide transporters

{dagger}Present address: Children's Hospital and Harvard Medical School, Boston, MA 02115, USA.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacteria predominantly exist in sessile communities, rather than as free-living cells, and develop as biofilms on any surfaces. The establishment of the biofilm architecture follows a sequence of events, going from initial attachment of a single cell to formation of a mushroom-shaped mature biofilm (Stoodley et al., 2002). Such structural development of a biofilm could be dependent on environmental conditions. For example, depending on the carbon source, Pseudomonas aeruginosa forms a flat and very dynamic biofilm (citrate), or a heterogeneous biofilm containing mushroom-shaped multicellular structures separated by water-filled channels (glucose) (Klausen et al., 2003). A mature biofilm is engulfed in a matrix containing extracellular polymeric substances. The main components of this matrix are exopolysaccharides; in addition, nucleic acids and proteins are also found. The role of exopolysaccharides in biofilm formation has not been fully elucidated. On the one hand, as with Vibrio cholerae (Watnick & Kolter, 1999), mutations that abolish exopolysaccharide production are linked to a severe defect in the initial stages of attachment. On the other hand, as with Escherichia coli (Danese et al., 2000), it has been shown that colanic acid is required for biofilm architecture rather than for initial binding to surfaces. Consequently, these observations indicate that components involved in exopolysaccharide biogenesis may play a role at various stages of biofilm formation. In addition to exopolysaccharides, bacterial surface lipopolysaccharides (LPS) (Makin & Beveridge, 1996) or lipooligosaccharides (LOS) (Swords et al., 2004) are also involved in biofilm formation. Moreover, several determinants of adhesion and biofilm formation, such as pili, flagella and surface adhesins, appear to be glycosylated in several micro-organisms (Power & Jennings, 2003). The biosynthetic pathways for glycosylation of these proteins are beginning to be understood. On the one hand, the target protein has a dedicated biosynthetic pathway for protein-linked glycans. That is the case, for example, of the PAK flagellin, which requires the function of a 14-gene genomic island adjacent to the flagellin structural gene (fliC) (Arora et al., 2001). On the other hand, the glycosylation process may have a common biosynthetic origin. For example, the Pseudomonas pilin expressed by strain 1244 has an O-linked glycan that is identical to its LPS O-antigen. This observation suggests that the whole LPS glycan is attached to the target protein and that pilin glycosylation may occur as a branch of the pathway of O-antigen production (Castric, 1995). Although there are more reports on glycosylation of proteins in Gram-negative bacteria (Power & Jennings, 2003), the details of the glycosylation biosynthetic process remain to be elucidated.

Lately, P. aeruginosa has been one of the most studied micro-organisms with respect to bacterial biofilm formation. It is a ubiquitous organism found in various environments and an opportunistic pathogen that can cause disease in a wide variety of organisms. In humans, it is mainly associated with nosocomial and cystic fibrosis infections (Costerton, 2001). Moreover, there are increasing numbers of reports suggesting that the biofilm lifestyle is predominant within sites of infection, such as in the lungs of cystic fibrosis patients (Singh et al., 2000). Among the P. aeruginosa determinants that have been shown to be involved at various stages of biofilm formation, flagella and type IV pili (O'Toole & Kolter, 1998), Cup fimbriae (Vallet et al., 2001) or alginate (Davies & Geesey, 1995) are the most frequently cited. In this study, we identified a P. aeruginosa gene cluster (pel) by random screening of non-adherent mutants on abiotic surfaces. The pel gene cluster was previously identified by Friedman & Kolter (2004a). They showed that pel mutations in the P. aeruginosa PA14 strain prevents the formation of a pellicle at the air–liquid interface of bacterial cultures grown in static conditions. On the basis of this initial observation, they could show further that the pel mutants are not able to form robust and shear force-resistant biofilms on a solid surface of plastic or glass. They also suggested that the pel genes are involved in the production of a putative exopolysaccharide, which holds bacterial cells together within the mature biofilms and makes them resistant to thorough washing. Finally, they demonstrated that the PA14 matrix contains a glucose-rich structural component that is not found in the pel mutants. This component allows staining of the bacterial colony on Congo-red-containing medium. Congo red binds exopolysaccharides, including glucose-rich polymers such as cellulose. Friedman & Kolter suggested that five out of the seven pel gene products have sequence similarities with proteins involved in carbohydrate processing, which include PelA, PelC, PelD, PelE and PelF. Of these five proteins, only PelF could be recognized as a glycosyltransferase. In the present study, we further characterized the role of the pel gene cluster in the non-piliated PAK{Delta}pilA strain. Our systematic analysis of the seven pel mutants, which were engineered in this strain, revealed that they were defective for attachment on plastic and glass surfaces at very early stages of the biofilm development process. The pel genes and their genetic organization appeared well conserved in other Gram-negative bacteria, such as the plant pathogen Ralstonia solanacearum, and we were able to show that in R. solanacearum also, it serves a conserved and similar function.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains, plasmids and growth conditions.
Bacterial strains and plasmids are described in Table 1. P. aeruginosa was grown at 37 °C in Luria broth (LB). R. solanacearum was grown at 30 °C in BG medium (Bacto-Peptone 10 g l–1, yeast extract 1 g l–1, Casamino acids 1 g l–1). The plasmids were introduced in P. aeruginosa by electroporation or by conjugation, using S17-1 or an E. coli strain containing pRK2013. The recombinant bacteria were selected on Pseudomonas isolation agar (PIA). P. aeruginosa mutants obtained by using the pKNG101 suicide vector were selected on plates containing 5 % sucrose. Antibiotic concentrations (µg ml–1) were: i) E. coli: ampicillin, kanamycin, streptomycin, 50; tetracycline, 40; ii) P. aeruginosa: carbenicillin, 500; gentamicin, 50; tetracycline, 200; streptomycin, 2000; iii) R. solanacearum: streptomycin, 50; spectinomycin, 40.


View this table:
[in this window]
[in a new window]
 
Table 1. Strains and plasmids used in this study

 
Bioinformatic analysis.
The genome sequence of P. aeruginosa and its annotation is available at www.pseudomonas.com. The amino acid sequences of the pel gene products were subjected to different software techniques such as TMHMM, SignalP or PSORT, ScanProsite, BLASTP or MAXHOM, PSI-BLAST, Ssearch and T-COFFEE. Most of these programs are available at http://www.embl-heidelberg.de/predictprotein/predictprotein.html. The genome annotation of R. solanacearum is available at http://sequence.toulouse.inra.fr/R.solanacearum.html.

Construction of P. aeruginosa and R. solanacearum mutant strains.
For P. aeruginosa, 500 bp DNA fragments contiguous to the 5' and 3' ends of the target gene were PCR amplified with the High Fidelity DNA polymerase (Roche) by using PAK chromosomal DNA as template and two pairs of oligonucleotides (ma01/ma02 and mb01/mb02, Table 2). In the first round of PCR, a product (ma) corresponding to the region upstream from the 5' end of the gene was generated using ma01/ma02 and a product (mb) corresponding to the region downstream from the 3' end of the gene was obtained using mb01/mb02. The oligonucleotide sequence of ma02 included the initiation codon, whereas mb01 contained the termination codon. The oligonucleotides ma02 and mb01 were designed in such a way that the resulting ma and mb fragments were overlapping PCR products. In a second round of PCR, the ma and mb fragments were used as template and joined by their overlapping ends when the external primers ma01 and mb02 were used for amplification. The resulting ma/mb DNA fragment contained the 5' and 3' regions of the target gene linked together. Within ma/mb, only the inititation and termination codons of the gene of interest, separated by a few codons, were left. The ma/mb DNA fragment was cloned into the pCR2.1 vector (TA-cloning kit, Invitrogen), and subcloned into pKNG101, which is non-replicative in P. aeruginosa. The recombinant plasmid, pKNGmamb, was maintained in the E. coli CC118{lambda}pir strain. The plasmid was mobilized into the appropriate P. aeruginosa strain. The mutants in which double recombination events occurred, resulting in the deletion of the gene of interest, were selected on sucrose plates as previously described (Kaniga et al., 1991). Finally, the deletion event was checked by PCR, with appropriate primers. Following this strategy, the pelB, pelE and pelF genes were deleted in the PAK strain, whereas, the pelB, pelC, pelD, pelE, pelF and pelG genes were deleted in the PAK{Delta}pilA strain. For R. solanacearum, slight modifications in the strategy for the production of a joined DNA fragment containing the deleted gene of interest were introduced. Briefly, the 5' and 3' adjacent regions (named PCR1 and PCR2, approximately 1 kb in length) of the ragG gene were PCR amplified using the chromosomal DNA of strain GMI1000 as template and the Expand Long Template DNA polymerase. The oligonucleotides used included an EcoRI restriction site. The PCR products were digested with EcoRI and used in a ligation mix containing the pGEM-T (Promega) TA-cloning vector. A streptomycin/spectinomycin resistance cassette, extracted from the pHP45{Omega} plasmid, was cloned in the recombinant plasmid at the EcoRI site, localized at the junction between PCR1 and PCR2. The linear plasmid was transformed in R. solanacearum, as described previously (Boucher et al., 1985), in order to delete the ragG gene after a double recombination event occurred. The ragG mutant was selected on BG plates containing antibiotic.


View this table:
[in this window]
[in a new window]
 
Table 2. Oligonucleotides used in this study

The names of the genes for which the deletions were designed are shown. Start and stop codons are indicated in bold type. Overlapping regions of primers are in italic type. Restriction sites are underlined and restriction enzymes are given below the underlined sequence.

 
Adherence assay on inert surfaces.
For P. aeruginosa, the adherence assay was performed in 24-well polystyrene microtitre dishes, as previously described (Vallet et al., 2001). Attached bacteria were stained with 100 µl 1 % crystal violet for a period of 15 min and washed twice with water. The ring of stained bacteria was quantified after having been collected in 400 µl 95 % ethanol. Subsequently, 600 µl water was added and OD600 was measured. All quantification assays were made in triplicate.

For P. aeruginosa and for R. solanacearum, cells were grown in 5 ml M63 minimal medium supplemented with 0·2 % glucose, 0·5 % Casamino acids and 1 mM MgCl2, in a 50 ml Corning tube containing a semi-immersed glass cover slide. After a period of incubation of 6, 24 or 48 h, at 30 °C without shaking, the slide was removed and rinsed. The attached bacterial cells were visualized by phase-contrast microscopy using an Axiovert 200M microscope (magnification x100). Images were captured with a Hamamatshu type Orca ER camera.

Swimming, swarming and twitching motility assays.
For swimming and swarming assays, bacterial colonies were transferred with a sterile toothpick onto the surface of 0·3 % or 0·7 % agar plates, respectively. The plates were incubated at 37 °C for 24 h and the spreading of the colony on the agar could readily be observed. For twitching motility, plates containing 1·5 % agar were inoculated with a toothpick by stabbing the LB agar plates. After 24 h growth at 37 °C, the agar was removed and the plates were stained with 1 % crystal violet for 10 min and washed with water. Spreading of bacteria from the inoculation point could be observed by the size of the stained surface on the plastic.

Cell surface hydrophobicity.
Bacterial cells were grown overnight at 37 °C in LB medium. The cultures were centrifuged and the bacterial pellet was resuspended in PBS to adjust the OD600 to 0·8. Three millilitres of the suspension were collected in Corex tubes and mixed with 400 µl solvent (hexadecane, chloroform or ethyl acetate). After 15 min at 37 °C, the tubes were vortexed (twice for 30 s with a 5 s pause in between). The aqueous phase was collected and the OD546 determined (Rosenberg et al., 1980).

Congo red assay.
Tryptone (10 g l–1) agar (1 %) plates without salt were supplemented with Congo red (40 µg ml–1) and Coomassie brilliant blue (20 µg ml–1). Bacteria were inoculated on the surface of the plates with a toothpick and grown at 30 °C. The colony morphology and colour were analysed (Romling et al., 1998).

LPS isolation and analysis.
Bacterial cultures for analysis of LPS were grown overnight in LB at 37 °C. A sample of 10 ml was collected and centrifuged. After washing with PBS, the cell pellet was resuspended in 1·5 ml Tris/EDTA buffer (50 mM Tris/HCl, 2 mM EDTA, pH 8·5). The bacteria were lysed by sonication. The cell envelopes were further collected by centrifugation and resuspended in 70 µl Tris buffer (2 mM Tris/HCl, pH 7·8) and 100 µl loading buffer (0·125 M Tris, 4 % SDS, 10 % {beta}-mercaptoethanol, 20 % glycerol, 0·01 % bromphenol blue). The samples were incubated at 95 °C for 10 min. A 30 µl aliquot of the sample was mixed with 30 µl Proteinase K (100 µg ml–1) and incubated for 1 h at 60 °C. Ten microlitres of each sample was submitted to electrophoresis on a 15 % polyacrylamide gel containing SDS. LPS was visualized after silver staining. By using the cell envelope fraction, a colorimetric analysis of 2-keto-3-deoxyoctonate (KDO) content was performed as previously described (Weissbach & Hurwitz, 1959).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Identification of non-adherent P. aeruginosa mutants
A screen for the identification of non-adherent P. aeruginosa strains was performed as previously described (Vallet et al., 2001). The screen makes use of a collection of Tn5 mutants generated in a non-piliated PAK strain (PAK{Delta}pilA). Mutant strains defective for attachment in the wells of a polystyrene microtitre dish are called lad for lost adherence. Here we showed that out of 17 selected strains, five lad mutants had a Tn5 insertion that was located within the same gene cluster (Fig. 1a). The gene cluster encompasses seven genes, from PA3064 to PA3058 (http://www.pseudomonas.com). Inverse PCR analysis (method described by Vallet et al., 2001) indicated that one insertion was located in PA3064, two insertions were mapped in PA3063 and finally, two other insertions interrupted PA3060. This gene cluster has been recently characterized in the P. aeruginosa strain PA14 and called pelA–G, since it has been shown to be involved in the formation of a pellicle at the air–liquid interface of a static culture (Friedman & Kolter, 2004a).



View larger version (26K):
[in this window]
[in a new window]
 
Fig. 1. (a) Genetic organization of the pel gene cluster. Tn5 positions are indicated by vertical arrows and the mutant name. The numbers in the horizontal arrows correspond to the PA number of the pel gene. PA3062 and PA3060 are indicated by 62 and 3060, respectively, to fit within the arrows. (b) Hydropathy plot of PelG and the PST member RfbX2 from S. dysenteriae (Paulsen et al., 1997). The plots were generated using TOPPRED2 (http://bioweb.pasteur.fr/seqanal/interfaces/toppred.html).

 
PelG is a putative member of the PST family
Most of the pel genes have homologues involved in carbohydrate processing (Friedman & Kolter, 2004a). Among them, PelF could be identified as a cytoplasmic glycosyltransferase belonging to the type IV glycosyltransferase (GT4) family (Coutinho et al., 2003). The proteins encoded by the other pel genes had weaker similarities to proteins with functions related to polysaccharide biogenesis, and seemed to be cell envelope-located proteins (Friedman & Kolter, 2004a). Interestingly, we further analysed the amino acid sequence of PelG that was not assigned a function by Friedman & Kolter (2004a). We found that PelG has 12 predicted transmembrane domains, a total length of 456 amino acids, and a high percentage of leucine and isoleucine amino acid content. Remarkably, these features are typical of the PST family (Hvorup et al., 2003). PST proteins generally translocate glycolipid precursors, synthesized on the cytosolic face of the membrane, to the opposite side where they serve as substrates for additional processing, such as O-antigen polymerization in LPS biosynthesis. Despite a canonical topology (12–14 transmembrane domains), a similar length (346–582 amino acids) and a high percentage of leucine and isoleucine amino acid content, these proteins have little sequence identity (http://www.infobiogen.fr/services/analyseq/cgi-bin/alignp_in.pl). Among the best hits for PelG are the O-unit flippases Wzx (21 % identity) from Yersinia pseudotuberculosis (Skurnik et al., 2000) and RfbX2 (22 %) from Shigella dysenteriae (Paulsen et al., 1997). Programs, such as T-COFFEE (http://www.ch.embnet.org/software/TCoffee.html), have been developed to further characterize similarities between proteins that show a degree of identity below 30 % (Notredame et al., 2000). By comparing PelG and RfbX2, we observed similar hydropathy plots (Fig. 1b) and, using T-COFFEE, we obtained a score of 75/100, which indicates that both proteins have a significant degree of similarity (data not shown). On this basis, we could suggest PelG as being a member of the PST family.

Involvement of the pelA–G genes in biofilm formation
In order to assign a role for each of the pel genes in the biofilm formation process, the pelB, pelC, pelD, pelE, pelF and pelG genes were systematically deleted in the type IV pili-deficient PAK{Delta}pilA strain. Upon introduction of the pel mutations in PAK{Delta}pilA, all strains presented a clear defect in their biofilm phenotype after 24 h incubation (Fig. 2a). The level of attachment was quantified for each mutant, by measuring the crystal violet stain left in the microtitre dishes after washing and by comparing with the parental PAK{Delta}pilA strain (Fig. 2a). The pel mutant strains attached with an efficiency of about 5–10 % relative to the PAK{Delta}pilA parental strain. This defect could be restored by the introduction of the corresponding pel genes in trans, as shown with pelF (pUCP3059) (Fig. 2a). We concluded that all pel genes are individually involved in the biofilm phenotype analysed.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 2. Attachment of PAK{Delta}pilA and isogenic pel mutant strains on polystyrene. (a) Crystal violet staining was quantified after 24 h incubation in microtitre dishes. The bars represent OD600 from three distinct experiments. OD600 quantifies the amount of crystal violet (CV) associated with the biofilm after staining. The PAK{Delta}pilApelB–G deletion mutants showed a clear defect for attachment (left). The transposon insertion pelA mutant (Q77) from our initial screen had a similar effect (data not shown). Trans-complementation experiments using PAK{Delta}pilA{Delta}pelF containing pUCP18 or pUCP3059 (pelF) are shown on the right. (b) Biofilm formation kinetics measured by using crystal violet staining in microtitre dishes. Crystal violet was quantified at incubation times ranging from 3 to 48 h. The experiments were repeated three times and error bars are shown. Kinetics of PAK{Delta}pilA, PAK{Delta}pilA{Delta}pelE, PAK{Delta}pilA{Delta}pelB and PAK-NF (fliC) are compared.

 
pel mutants of a non-piliated PAK strain are impaired in initiation of biofilm formation
We further examined more precisely at which stages of biofilm formation the PAK{Delta}pilA{Delta}pel mutants were affected. Kinetic experiments were performed in which biofilm formation was evaluated after incubation periods ranging from 3 to 48 h (Fig. 2b). We used the pelB and pelE mutants engineered in the PAK{Delta}pilA genetic background (PAK{Delta}pilA{Delta}pelB and PAK{Delta}pilA{Delta}pelE) and compared biofilm formation levels to those of the parental PAK{Delta}pilA strain and a non-adherent mutant PAK-NF (fliC), which is devoid of flagella (Fig. 2b). Interestingly, significant differences occurred at an early stage, and could already be observed after 8 h incubation. After 12 h, the pel mutants and the fliC strain had a severely reduced biofilm formation efficiency, which corresponded to about 30 % for PAK{Delta}pilA{Delta}pelB and 5 % for PAK{Delta}pilA{Delta}pelE, compared to the parental strain. We thus concluded that the pel mutations affected early stages of biofilm formation.

pel mutants of a piliated PAK strain are impaired in formation of a mature biofilm
In the study by Friedman & Kolter (2004a), a clear difference could be observed between a pel mutant and the parental PA14 strain only at a later stage: after 18–24 h incubation. We further investigated whether the function of the pel genes, as observed in the early phase of biofilm formation, is linked to the strain (PA14 and PAK) or to the lack of type IV pili. We thus introduced pel mutations into the piliated PAK strain. PAK mutants deleted for the pelF, pelE and pelB genes were engineered and the biofilm phenotype tested and quantified with the crystal violet biofilm assay (Fig. 3a). The mutant strains were incubated for 24 h in microtitre dishes, and the difference after staining was barely detectable. However, quantification experiments revealed that the biofilm formation efficiency of the PAK{Delta}pelB, PAK{Delta}pelE and PAK{Delta}pelF mutants was only about 60–70 % relative to the parental PAK strain (Fig. 3a). Furthermore, kinetic experiments were performed in which biofilm formation was evaluated after incubation periods ranging from 3 to 48 h (Fig. 3b). The level of biofilm formed did not change significantly between the pel mutants and the PAK strain up to 12 h incubation. However, at 24 h the influence of the pel mutation could be observed and was strongly significant after 48 h with a 60 % reduction in biofilm formation efficiency. From these results we concluded that a pel mutation influences biofilm formation both in a piliated and a non-piliated PAK P. aeruginosa strain; however, the influence of the pel mutation at early stages of biofilm formation could only be observed in a non-piliated strain (PAK{Delta}pilA).



View larger version (18K):
[in this window]
[in a new window]
 
Fig. 3. Attachment of PAK and isogenic pelB, pelE and pelF mutant strains on polystyrene. (a) Crystal violet staining was quantified after 24 h incubation in microtitre dishes. The bars represent OD600 from three distinct experiments. OD600 quantifies the amount of crystal violet (CV) associated with the biofilm after staining. The quantification revealed that the level of staining in the pel mutants is about 60–70 % relative to the PAK strain. (b) Biofilm formation kinetics measured by using crystal violet staining in microtitre dishes. Crystal violet was quantified at incubation times ranging from 3 to 48 h. The experiments were repeated three times and error bars are shown. Kinetics of PAK, PAK{Delta}pelE, PAK{Delta}pelB and PAK-NF (fliC) are compared.

 
Microscopic analysis of biofilm formation in P. aeruginosa pel mutants
More detailed analysis of the defects conferred by a mutation in the pel genes was obtained through microscopic analysis of bacteria attached to glass cover slides (see Methods). Briefly, cells were grown in 5 ml M63 minimal medium supplemented with glucose, MgSO4 and Casamino acids, in a 50 ml Corning tube containing a glass cover slide. After 6, 24 or 48 h growth at 30 °C without shaking, the glass slide was removed and rinsed. The remaining cells were visualized by phase-contrast microscopy. As shown in Fig. 4(a), at 24 h a dense film of bacteria was visible when looking at the P. aeruginosa strain PAK. At 48 h incubation, several layers of bacteria were observed at the surface of the glass slide. It should be mentioned that the biofilm extended largely on both sides of the liquid–air interface at the position where the glass slide emerged from the medium. In the case of the PAK{Delta}pelE mutant (Fig. 4b), a film of bacteria was observed, but the surface covered was restricted to a thin line at the air–liquid interface position, and at 48 h incubation the position and thickness of the film did not seem to have changed. The biofilm formed by the PAK{Delta}pilA strain (Fig. 4c) did not appear to be drastically different compared to that of PAK after 48 h incubation. Strikingly, a PAK{Delta}pilA strain in which the pelE gene had been deleted (PAK{Delta}pilA{Delta}pelE) was totally unable to form a biofilm, and only a few cells were attached to the glass cover slide (Fig. 4d). These observations are consistent with the results we obtained using microtitre dishes and crystal violet staining.



View larger version (157K):
[in this window]
[in a new window]
 
Fig. 4. Microscopic analysis of biofilm formation. Bacteria grown in M63 medium supplemented with glucose, MgSO4 and Casamino acids were allowed to attach to glass cover slides over a period of 6, 24 or 48 h. The development of biofilm on a solid surface was visualized by phase-contrast microscopy. (a) PAK; (b) PAK{Delta}pelE; (c) PAK{Delta}pilA; (d) PAK{Delta}pilA{Delta}pelE.

 
Characteristics of the P. aeruginosa pel mutants and properties of the pel genes
The function of the pel genes is poorly understood and has been proposed by Friedman & Kolter (2004a) to be involved in the production of a glucose-rich matrix required for biofilm formation. Here, we confirmed that, similar to the PA14 strain, pel gene mutations in the PAK strain do not affect twitching, swarming or swimming motilities, indicating that type IV pili and flagellar functions have not been impaired, as seen with the PAK{Delta}pelB, PAK{Delta}pelE and PAK{Delta}pelF mutants (data not shown). We also checked whether the pel mutations could have any effect on the bacterial cell surface hydrophobicity. We used three different solvents, hexadecane, ethyl acetate and chloroform, and after mixing with a bacterial suspension, we evaluated the number of bacteria that were recovered from the aqueous phase (see Methods). We compared PAK{Delta}pelF and PAK{Delta}pilA{Delta}pelF to their isogenic parental strains, but no clear differences were observed in terms of surface hydrophobicity properties (data not shown). Moreover, we analysed the LPS profiles of a pelF mutant both in a PAK and a PAK{Delta}pilA genetic background. LPS was obtained from proteinase K-treated cell envelopes, separated on a 15 % polyacrylamide gel containing SDS and visualized by silver staining (data not shown). In parallel, the KDO content in the LPS samples was measured according to the method described by Weissbach & Hurwitz (1959). In both cases, no significant quantitative or qualitative differences could be observed (data not shown), suggesting that the pel genes are unlikely to be involved in LPS biogenesis. Furthermore, we confirmed, as shown by Friedman & Kolter (2004a) with the PA14 strain, that pel gene expression in the PAK strain might be associated with the production of an exopolysaccharide. We used plates containing Congo red, which is known to bind extracellular matrix components, including polysaccharides such as cellulose, to give a dark-red stained aspect to the colony (Romling et al., 1998). No obvious staining could be detected with the PAK or PAK{Delta}pilA strains, compared to the PAK{Delta}pilA{Delta}pelB mutant, even though this mutant had a smoother colony morphology (Fig. 5a). However, upon introduction of the pel genes cloned in a multicopy vector (pMO013629) to any one of these three strains, red staining became readily visible, and the colony morphology became rugose, both of which correlate with the production of exopolysaccharide (Fig. 5a). We further checked whether the overexpression of the pel genes carried on pMO013629 influenced biofilm formation. We evaluated biofilm formation with the crystal violet staining assay on microtitre plates, and could show that the plasmid complemented the attachment defect of a PAK{Delta}pilA{Delta}pelB mutant (Fig. 5b). Remarkably, when this plasmid was introduced in PAK or PAK{Delta}pilA, the level of biofilm formed was significantly increased (Fig. 5b). We concluded that the pel-dependent Congo-red-stained component that we observed is associated with the ability of the P. aeruginosa PAK strain to form biofilms.



View larger version (36K):
[in this window]
[in a new window]
 
Fig. 5. (a) P. aeruginosa strains grown on Congo red plates. (b) Influence of pel gene overexpression on biofilm formation in P. aeruginosa. The crystal violet staining after 24 h incubation in microtitre dishes was quantified and the bars represent OD600 from three distinct experiments. OD600 quantifies the amount of crystal violet (CV) associated with the biofilm after staining. The cosmid pMO013629 carries the pel gene cluster cloned into pLA2917.

 
Conservation of the pel gene cluster in other organisms
Analysis of complete and incomplete microbial genomes allowed us to identify gene clusters whose organization was similar to the P. aeruginosa pel gene cluster. Most notably, we identified a gene cluster in R. solanacearum encompassing the ORFs annotated RSc2276–RSc2270 (Salanoubat et al., 2002). The percentage of similarities between the various homologous gene products ranged from 38 to 63 % (Fig. 6a). We named the Ralstonia gene cluster rag for Ralstonia adherence glycosyltransferase, the glycosyltransferase being encoded by ragF, which is homologous to pelF.



View larger version (67K):
[in this window]
[in a new window]
 
Fig. 6. Conservation of the pel genes in R. solanacearum. (a) The percentage similarity between the P. aeruginosa Pel proteins and their homologues is indicated beneath the arrows. R. solanacearum genes are shown directly below their P. aeruginosa homologues. For R. solanacearum, the number in the arrow corresponds to the annotation at http://sequence.toulouse.inra.fr/R.solanacearum.html. (b) Glass-slide attachment of R. solanacearum after 24 h incubation. The images shown are representative of observations in multiple fields.

 
A R. solanacearum ragG mutant is deficient in attachment
It was previously shown that R. solanacearum was able to form biofilms (Kang et al., 2002). In order to determine whether the rag gene cluster was functionally related to the pel gene cluster, we engineered a R. solanacearum mutant within the ragG gene. The ability of the mutant to attach on an abiotic surface was evaluated by immersing a glass slide in a static liquid culture of the ragG mutant or the parental Ralstonia GMI1000 strain. After 24 h incubation, the glass slide was collected and microscopic analysis revealed a surface poorly colonized by the ragG mutant, with scattered clumps of cells, in contrast to the widespread coverage observed with the parental strain (Fig. 6b). We concluded that the R. solanaceareum rag genes seem to be involved in the bacterial attachment process on abiotic surfaces.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
We have identified 17 P. aeruginosa mutants in a non-piliated PAK strain that are affected in attachment in a microtitre dish assay (Vallet et al., 2001). In the present study, we characterized five of these mutants, which had a Tn5 insertion into a single chromosomal locus. The transposons were mapped in a seven-gene cluster, called pel. The fact that about one-third of the mutants we isolated carried a transposon insertion in the pel gene cluster suggested that these genes are important in the process of biofilm formation in the conditions of our assay. Previous work by Friedman & Kolter (2004a) identified the pel gene cluster in the P. aeruginosa PA14 strain and revealed its involvement in the formation of a mature and robust biofilm.

One major novel aspect that we presented in our study on the pel genes is that they influence the initiation of the biofilm formation process and also, probably, irreversible attachment on abiotic surfaces. In contrast to other studies that searched for biofilm-deficient P. aeruginosa strains, we used a non-piliated strain, PAK{Delta}pilA. In the conditions of our assay, this strain is able to form biofilms with an efficiency of 88 % relative to the PAK strain (Vallet et al., 2001). In the present study, we performed a systematic mutagenesis of the seven pel genes in the non-piliated genetic background (transposon insertion for pelA and gene deletion for pelB–G) and found that all these strains were totally deficient for initiation of attachment on plastic or glass. This was observed by quantifying the amount of crystal violet associated with the biofilm after staining (Fig. 2a) or by direct observation in phase-contrast microscopy, which revealed that hardly any bacteria are attached to the abiotic surface (Fig. 4). By using a similar screening procedure and a collection of mutants in the piliated P. aeruginosa PA14 strain, Friedman & Kolter were not able to identify pel mutants. However, they were able to identify pel mutants thanks to the observation of a pellicle at the air–liquid interface of a statically grown bacterial culture. Finally, they showed that the phenotype associated with the pel mutation could be seen in the microtitre dish crystal violet-stained biofilm assay only after an incubation time of over 18 h, which might explain why they missed it in previous screens. More precisely, after 24 h, the biofilm formed by the parental PA14 strain becomes robust and resistant to washing procedures, whereas the biofilm formed by the pel mutants is not able to resist shear forces. To check whether the differences observed between the two studies is related to the strain used (PA14 and PAK) or the piliation status, we engineered pel mutations in a piliated PAK strain. We found that PAK{Delta}pel mutants do not have an obvious defect in biofilm formation using the crystal violet staining procedure (Fig. 3a). Furthermore, we performed biofilm kinetics experiments and we revealed that the PAK{Delta}pel mutants formed less-robust biofilms after incubation periods of 24 h (Fig. 3b), similar to the results observed by Friedman & Kolter. We thus concluded that the type IV pili, which are assembled and functional in the PAK{Delta}pel mutants, may mask the influence of the pel genes in the early stages of biofilm formation.

Whereas the involvement of the pel gene cluster in attachment and biofilm formation seems to be established, the exact role of the pel gene products remains to be elucidated. Friedman & Kolter have suggested that the pel genes are involved in the production of a novel glucose-rich polysaccharide component that contributes to the formation of a matrix material that encases the cells and holds them together in a resistant biofilm. We also confirmed that in the PAK strains, expression of the pel genes is associated with the production of a matrix component that allows binding of Congo red. Since the pel mutations in PAK{Delta}pilA strongly alter the association of the bacteria with the surface, we suggest that the component synthesized by the pel system is responsible for initial cell-surface contact and is not simply required for cell–cell contact at later stages in the biofilm development process. However, the cell-surface interaction process involving the pel genes appears to be dispensable when type IV pili are present. In this case, the function of the pili is thus sufficient to allow an irreversible attachment of the bacteria to the surface. The role of expolysaccharide at early and late stages of the biofilm formation process has already been studied. It has been described for vps mutants in V. cholerae that, using the standard biofilm assay in microtitre dishes, no crystal violet staining can be observed even after 80 h incubation, whereas the parental strain reaches a maximum at 29 h (Watnick et al., 1999). This suggests that vps-encoded exopolysaccharide might have a role at quite early stages of biofilm development. In contrast, with an E. coli mutant defective in the production of colanic acid, even though the crystal violet staining observed with the mutant is significantly less intense than that observed for the wild-type strain, it increases over time, and ultimately closely approximates that of the parent. In this case, colanic acid has been suggested to be critical for the formation of the complex three-dimensional structure and depth of E. coli biofilms (Danese et al., 2000).

It was recently shown that the exopolysaccharide alginate is not a major component of the the biofilm matrix of the non-mucoid P. aeruginosa strains PA14 or PAO1, suggesting that some other exopolysaccharides might be the predominant biofilm scaffolding in these strains (Wozniak et al., 2003). Whereas the pel gene cluster may play a crucial role in the biogenesis of this yet-uncharacterized polysaccharide, an additional cluster, named psl, has also been proposed to participate in the production of exopolysaccharides that may be minor constituents of the biofilm matrix (Friedman & Kolter, 2004b; Jackson et al., 2004; Matsukawa & Greenberg, 2004). With respect to the pel cluster, Friedman & Kolter (2004a) suggested that five out of the seven pel genes encode proteins with similarities to components involved in carbohydrate processing. We agree that PelF is probably a glycosyltransferase and suggest that it belongs to the GT4 family (Coutinho et al., 2003), which might be involved in a process of glycosylation required for polysaccharide biogenesis and/or glycoprotein synthesis. Among the pel gene products, which were not assigned a homologue, we showed here that PelG shows weak sequence but strong structural similarities with members of the PST family. PSTs are integral membrane proteins that mediate export of polysaccharides across the cytoplasmic membrane. Transmembrane transporters have recently been classified, and PST members, such as PelG, will fall in the TC #2.A.66 exporter superfamily (Hvorup et al., 2003). PSTs may function together with auxiliary proteins, located within the bacterial cell envelope, that allow passage of complex carbohydrate across both membranes of the Gram-negative bacterial envelope (Paulsen et al., 1997). The putative flipping role of PST members is commonly required for LPS O-antigen repeat unit export or for extracellular (EPS) and capsular polysaccharide (CPS) export. Interestingly, three Pel proteins have transmembrane domains (PelB, PelD and PelE) and are predicted to be located within the cytoplasmic membrane, with PelB and PelE, which carry tetratricopeptide repeats (TPRs) involved in protein–protein interactions (Smith et al., 1995). More interestingly, we have identified a putative lipoprotein signal peptide in the deduced amino acid sequence of the pelC gene. The amino acid residue that follows the cysteine residue, which will be the first residue of the mature lipoprotein, is not an aspartate residue. The +2 position in the amino acid sequence of mature lipoproteins has been shown to be crucial for subcellular localization (Yamaguchi et al., 1988). In the case where the +2 residue is an aspartate, the lipoprotein is retained within the inner membrane, whereas any other residues, mostly serine, as found in PelC, direct the protein to the outer membrane. This observation strongly suggests that the putative PelC lipoprotein is targeted to the outer membrane. It is attractive to suggest that Pel proteins may constitute a novel type of multi-component polysaccharide transport system. PelG could be the PST, PelCDE the auxiliary transporter proteins, with PelC in the outer membrane and PelDE in the inner membrane. Enzymes more directly involved in carbohydrate processing could be PelF, which is the putative glycosyltransferase of the system, whereas PelA and PelB contain regions with weak similarities to domains found in glycosylhydrolase and cellulase synthase, respectively. It should be noticed that the lack of any one of the Pel components alters the function of the whole system, resulting in the associated biofilm-deficient phenotype.

Our future studies will aim to identify more precisely the nature of the components produced in a pel-dependent manner and more particularly the nature of the glycosylation event that could be supported by PelF. Moreover, the role of the P. aeruginosa pel cluster appears to be functionally conserved in other bacteria, such as the plant pathogen R. solanacearum. The characterization of the molecular targets for the pel glycosylation gene cluster will certainly reveal novel features associated with P. aeruginosa biofilms, but more significantly features associated with the virulence of bacterial animal and plant pathogens.


   ACKNOWLEDGEMENTS
 
We thank B. Henrissat and A. Imberty for advice in pel gene products analyses. We are grateful to S. de Bentzmann, C. Boucher and A. Lazdunski for stimulating discussions. A. F.'s work is supported by the European community (QLK-CT-2001-01339) and the French cystic fibrosis foundation (VLM). P. V. is supported by a grant from the French research and technology ministry (MRT).


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Arora, S. K., Bangera, M., Lory, S. & Ramphal, R. (2001). A genomic island in Pseudomonas aeruginosa carries the determinants of flagellin glycosylation. Proc Natl Acad Sci U S A 98, 9342–9347.[Abstract/Free Full Text]

Boucher, C. A., Barberis, P. A., Trigalet, A. P. & Demery, D. A. (1985). Transposon mutagenesis of Pseudomonas solanacearum: isolation of Tn5-induced avirulent mutants. J Gen Microbiol 131, 2449–2457.

Castric, P. (1995). pilO, a gene required for glycosylation of Pseudomonas aeruginosa 1244 pilin. Microbiology 141, 1247–1254.[Medline]

Costerton, J. W. (2001). Cystic fibrosis pathogenesis and the role of biofilms in persistent infection. Trends Microbiol 9, 50–52.[CrossRef][Medline]

Coutinho, P. M., Deleury, E., Davies, G. J. & Henrissat, B. (2003). An evolving hierarchical family classification for glycosyltransferases. J Mol Biol 328, 307–317.[CrossRef][Medline]

Danese, P. N., Pratt, L. A. & Kolter, R. (2000). Exopolysaccharide production is required for development of Escherichia coli K-12 biofilm architecture. J Bacteriol 182, 3593–3596.[Abstract/Free Full Text]

Davies, D. G. & Geesey, G. G. (1995). Regulation of the alginate biosynthesis gene algC in Pseudomonas aeruginosa during biofilm development in continuous culture. Appl Environ Microbiol 61, 860–867.[Abstract]

Friedman, L. & Kolter, R. (2004a). Genes involved in matrix formation in Pseudomonas aeruginosa PA14 biofilms. Mol Microbiol 51, 675–690.[Medline]

Friedman, L. & Kolter, R. (2004b). Two genetic loci produce distinct carbohydrate-rich structural components of the Pseudomonas aeruginosa biofilm matrix. J Bacteriol 186, 4457–4465.[Abstract/Free Full Text]

Hvorup, R. N., Winnen, B., Chang, A. B., Jiang, Y., Zhou, X. F. & Saier, M. H., Jr (2003). The multidrug/oligosaccharidyl-lipid/polysaccharide (MOP) exporter superfamily. Eur J Biochem 270, 799–813.[Abstract/Free Full Text]

Jackson, K. D., Starkey, M., Kremer, S., Parsek, M. R. & Wozniak, D. J. (2004). Identification of psl, a locus encoding a potential exopolysaccharide that is essential for Pseudomonas aeruginosa PAO1 biofilm formation. J Bacteriol 186, 4466–4475.[Abstract/Free Full Text]

Kang, Y., Liu, H., Genin, S., Schell, M. A. & Denny, T. P. (2002). Ralstonia solanacearum requires type 4 pili to adhere to multiple surfaces and for natural transformation and virulence. Mol Microbiol 46, 427–437.[CrossRef][Medline]

Kaniga, K., Delor, I. & Cornelis, G. R. (1991). A wide-host-range suicide vector for improving reverse genetics in gram-negative bacteria: inactivation of the blaA gene of Yersinia enterocolitica. Gene 109, 137–141.[CrossRef][Medline]

Klausen, M., Heydorn, A., Ragas, P., Lambertsen, L., Aaes-Jorgensen, A., Molin, S. & Tolker-Nielsen, T. (2003). Biofilm formation by Pseudomonas aeruginosa wild type, flagella and type IV pili mutants. Mol Microbiol 48, 1511–1524.[CrossRef][Medline]

Makin, S. A. & Beveridge, T. J. (1996). The influence of A-band and B-band lipopolysaccharide on the surface characteristics and adhesion of Pseudomonas aeruginosa to surfaces. Microbiology 142, 299–307.[Medline]

Matsukawa, M. & Greenberg, E. P. (2004). Putative exopolysaccharide synthesis genes influence Pseudomonas aeruginosa biofilm development. J Bacteriol 186, 4449–4456.[Abstract/Free Full Text]

Notredame, C., Higgins, D. G. & Heringa, J. (2000). T-Coffee: a novel method for fast and accurate multiple sequence alignment. J Mol Biol 302, 205–217.[CrossRef][Medline]

O'Toole, G. A. & Kolter, R. (1998). Flagellar and twitching motility are necessary for Pseudomonas aeruginosa biofilm development. Mol Microbiol 30, 295–304.[CrossRef][Medline]

Paulsen, I. T., Beness, A. M. & Saier, M. H., Jr (1997). Computer-based analyses of the protein constituents of transport systems catalysing export of complex carbohydrates in bacteria. Microbiology 143, 2685–2699.[Medline]

Power, P. M. & Jennings, M. P. (2003). The genetics of glycosylation in Gram-negative bacteria. FEMS Microbiol Lett 218, 211–222.[CrossRef][Medline]

Romling, U., Sierralta, W. D., Eriksson, K. & Normark, S. (1998). Multicellular and aggregative behaviour of Salmonella typhimurium strains is controlled by mutations in the agfD promoter. Mol Microbiol 28, 249–264.[CrossRef][Medline]

Rosenberg, M., Gutnick, D. & Rosenberg, E. (1980). Adherence of bacteria to hydrocarbons: a simple method for measuring cell-surface hydrophobicity. FEBS Microbiol Lett 9, 29–33.[CrossRef]

Salanoubat, M., Genin, S., Artiguenave, F. & 25 other authors (2002). Genome sequence of the plant pathogen Ralstonia solanacearum. Nature 415, 497–502.[CrossRef][Medline]

Singh, P. K., Schaefer, A. L., Parsek, M. R., Moninger, T. O., Welsh, M. J. & Greenberg, E. P. (2000). Quorum-sensing signals indicate that cystic fibrosis lungs are infected with bacterial biofilms. Nature 407, 762–764.[CrossRef][Medline]

Skurnik, M., Peippo, A. & Ervela, E. (2000). Characterization of the O-antigen gene clusters of Yersinia pseudotuberculosis and the cryptic O-antigen gene cluster of Yersinia pestis shows that the plague bacillus is most closely related to and has evolved from Y. pseudotuberculosis serotype O : 1b. Mol Microbiol 37, 316–330.[CrossRef][Medline]

Smith, R. L., Redd, M. J. & Johnson, A. D. (1995). The tetratricopeptide repeats of Ssn6 interact with the homeo domain of alpha 2. Genes Dev 9, 2903–2910.[Abstract]

Stoodley, P., Sauer, K., Davies, D. G. & Costerton, J. W. (2002). Biofilms as complex differentiated communities. Annu Rev Microbiol 56, 187–209.[CrossRef][Medline]

Swords, W. E., Moore, M. L., Godzicki, L., Bukofzer, G., Mitten, M. J. & VonCannon, J. (2004). Sialylation of lipooligosaccharides promotes biofilm formation by nontypeable Haemophilus influenzae. Infect Immun 72, 106–113.[Abstract/Free Full Text]

Vallet, I., Olson, J. W., Lory, S., Lazdunski, A. & Filloux, A. (2001). The chaperone/usher pathways of Pseudomonas aeruginosa: identification of fimbrial gene clusters (cup) and their involvement in biofilm formation. Proc Natl Acad Sci U S A 98, 6911–6916.[Abstract/Free Full Text]

Watnick, P. I. & Kolter, R. (1999). Steps in the development of a Vibrio cholerae El Tor biofilm. Mol Microbiol 34, 586–595.[CrossRef][Medline]

Watnick, P. I., Fullner, K. J. & Kolter, R. (1999). A role for the mannose-sensitive hemagglutinin in biofilm formation by Vibrio cholerae El Tor. J Bacteriol 181, 3606–3609.[Abstract/Free Full Text]

Weissbach, A. & Hurwitz, J. (1959). The formation of 2-keto-3-deoxyheptonic acid in extracts of Escherichia coli B. I. Identification. J Biol Chem 234, 705–709.[Free Full Text]

Wozniak, D. J., Wyckoff, T. J., Starkey, M., Keyser, R., Azadi, P., O'Toole, G. A. & Parsek, M. R. (2003). Alginate is not a significant component of the extracellular polysaccharide matrix of PA14 and PAO1 Pseudomonas aeruginosa biofilms. Proc Natl Acad Sci U S A 100, 7907–7912.[Abstract/Free Full Text]

Yamaguchi, K., Yu, F. & Inouye, M. (1988). A single amino acid determinant of the membrane localization of lipoproteins in E. coli. Cell 53, 423–432.[Medline]

Received 16 June 2004; revised 9 November 2004; accepted 9 November 2004.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Vasseur, P.
Articles by Filloux, A.
Articles citing this Article
PubMed
PubMed Citation
Articles by Vasseur, P.
Articles by Filloux, A.
Agricola
Articles by Vasseur, P.
Articles by Filloux, A.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2005 Society for General Microbiology.