|Home | About | Journals | Submit | Contact Us | Français|
AdnA in Pseudomonas fluorescens, an ortholog of FleQ in P. aeruginosa, regulates both motility and flagellum-mediated attachment to various surfaces. A whole-genome microarray determined the AdnA transcriptome by comparing the gene expression pattern of wild-type Pf0-1 to that of Pf0-2x (adnA deletion mutant) in broth culture. In the absence of AdnA, expression of 92 genes was decreased, while 11 genes showed increased expression. Analysis of 16 of these genes fused to lacZ confirmed the microarray results. Several genes were further evaluated for their role in motility and biofilm formation. Two genes, Pfl01_1508 and Pfl01_1517, affected motility and had different effects on biofilm formation in Pf0-1. These two genes are predicted to specify proteins similar to the glycosyl transferases FgtA1 and FgtA2, which have been shown to be involved in virulence and motility in P. syringae. Three other genes, Pfl01_1516, Pfl01_1572, and Pfl01_1573, not previously associated with motility and biofilm formation in Pseudomonas had similar effects on biofilm formation in Pf0-1. Deletion of each of these genes led to different motility defects. Our data revealed an additional level of complexity in the control of flagellum function beyond the core genes known to be required and may yield insights into processes important for environmental persistence of P. fluorescens Pf0-1.
Pseudomonas fluorescens is a common soil bacterium that is an important part of the soil ecosystem. Various strains have unique characteristics which make them suitable for use in biocontrol or bioremediation (16, 31, 51, 56). The P. fluorescens strain Pf0-1 was isolated during a study of fitness in soil (10) and has been studied extensively to identify genes involved in soil colonization (14, 15, 33, 52–54). Two transposon insertion mutants were found to be deficient in adhesion to quartz sand, soil, and seeds (14, 15). Both insertions were within one gene, which was termed adnA (8). adnA mutants were also found to be deficient in colonization, spread, and persistence in live soil (33), in biofilm formation, and in motility (8). These previous studies indicate that the lack of flagella is a limiting factor for soil colonization, motility, and biofilm formation, which is an effect of the adnA mutation in the wild-type P. fluorescens Pf0-1 background.
AdnA is a homolog of FleQ (84% amino acid identity) of P. aeruginosa PAb1, (8), which is the master transcriptional regulator of flagellar biogenesis in P. aeruginosa (3, 13). Homologs of FleQ have been identified in Legionella pneumophila (1), Vibrio cholerae (28), Vibrio parahaemolyticus (26), and Helicobacter pylori (58) as major regulators of flagellar synthesis. Transcriptional regulation of flagellar biogenesis in P. aeruginosa is separated into a four-tiered regulatory circuit (13) involving several transcription regulators and sigma factors which are required for the regulation of flagellum-associated genes: FleQ (3, 13), FleSR (13, 47), FliA (σ28) (12), σ54 (59), and σ70 (12, 65).
It is not known whether FleQ is under the control of environmental stimuli or if the transcription of fleQ is enhanced by any other regulatory proteins. However, it is known that σ70 is required for transcription of fleQ (12) and that FleR, FliA, and FleQ do not have an effect on the transcription of fleQ (13). Several proteins have been shown to negatively regulate fleQ in vitro in P. aeruginosa by either binding to the DNA upstream of the gene (AlgT through AmrZ [AlgZ] [19, 63, 64] and Vfr ) or interacting directly with FleQ (FleN ). Vfr is a homolog of cyclic AMP (cAMP) receptor protein (CRP), which has been identified in many organisms such as Salmonella enterica serovar Typhimurium (30), Vibrio cholerae (57), and P. aeruginosa (4, 67). It appears to be involved in the regulation of virulence factors, including flagella.
The Pseudomonas genus demonstrates two types of flagellum-based motility: swimming and swarming. Swimming motility is defined as random movement, which is within the agar (0.2% to 0.4% agar), while swarming motility is a coordinated movement on the agar surface (0.5% to 2% agar) (reviewed in reference 22).
A two-component regulatory system, SadC/B, has been shown to influence both swarming motility and biofilm formation in P. aeruginosa (5, 6, 36). A sadC mutation increases swarming while decreasing biofilm formation (36). In P. fluorescens F113, mutations in sadB, gacS, or a third gene, wspR, result in increased swimming and swarming motility relative to wild-type F113 (39). In double and triple mutants, there is an additive effect on both swimming and swarming motility, while differences in biofilm formation were not tested (39). In addition to the effects on motility in P. fluorescens F113, the two proteins SadB and GacS were also found to indirectly repress the expression of fleQ (39). More recently, GacA has also been found to be involved in the swarming motility in P. fluorescens Pf-5 and exerts positive regulational control on fleQ (23; J. Loper, personal communication).
Mutations affecting flagellar regulation, biogenesis, or modification in Pseudomonas have been shown to affect the ability of organisms to move through the environment (33), display chemotaxis toward attractants (48), and form biofilms (8, 61). The basic details underlying flagellar assembly and motility in a broad range of bacteria are well understood, but additional genes continue to reveal aspects of motility in different species. The genes fliT and fleP in P. fluorescens F113 and P. aeruginosa PAO1, respectively, are located in the same genetic locus, but each seems to have a different function in these species (7). Both FliT and FleP are required for motility, but the PAO1 fleP mutant showed detached flagella (13), while the F113 fliT mutant showed flagella with morphologies similar to those of the wild-type strain (7). fleP is regulated by FleQ in P. aeruginosa PAO1 (13), and the fliT ortholog in P. fluorescens Pf0-1 is under the control of AdnA. However, Redondo-Nieto et al. have shown that fliT and another gene, flaG, are both transcribed independently of FliA and FleQ in P. fluorescens F113 (46). The differences in P. fluorescens strains are not limited to their regulation of genes but occur in their genomic content as well (51).
It is clear that there are still details that remain to be determined regarding the control of flagellar motility in terms of the genes involved and how they are regulated. The example of fliT highlights the fact that different strains within the same genus and even species may have diverged from the general scheme of regulation and use specialized regulatory mechanisms.
In this study, we examined the AdnA regulon from P. fluorescens Pf0-1 by comparing the gene expression patterns of wild-type Pf0-1 and the adnA deletion mutant Pf0-2x (48) using a whole-genome microarray. Our goal was to identify novel genes involved in motility and biofilm formation by examining genes in the AdnA regulon of Pf0-1 not previously associated with motility. Genes were selected on the basis of four criteria: (i) genes that are controlled by AdnA in Pf0-1 but not FleQ in P. aeruginosa, (ii) genes that have unknown functions, (iii) genes coding for putative regulators, and (iv) genes not predicted to be involved in motility and biofilm formation in Pseudomonas spp.
The bacterial strains and plasmids used in this study are listed in Table 1. P. fluorescens and Escherichia coli were grown at 30°C and 37°C, respectively. E. coli was routinely cultured in Luria-Bertani (LB) medium (49), while P. fluorescens strains were grown in either LB medium or Pseudomonas minimal medium (PMM) (27). Antibiotics and supplements were added as required at the following concentrations: ampicillin, 100 μg/ml (Ap100; for E. coli) and 50 μg/ml (Ap50; for P. fluorescens); kanamycin, 50 μg/ml (Kan50; for E. coli) and 25 μg/ml (Kan25; for Pf0-1); gentamicin, 20 μg/ml (Gm20); tetracycline, 7.5 μg/ml (Tc7.5) for both plasmids and insertions into the Pf0-1 genome and 15 μg/ml (Tc15) for plasmids in E. coli; and X-Gal (5-bromo-4-chloro-3-indoyl-β-d-galactopyranoside), 40 μg/ml. E. coli strains were transformed by electroporation with a Bio-Rad Micropulser following the recommendations of the manufacturer, while DNA was introduced into P. fluorescens strains by conjugation from E. coli S17-1 λ pir.
Restriction and DNA-modifying enzymes were purchased from Invitrogen, Inc. (Carlsbad, CA), and New England BioLabs (Ipswich, MA). Plasmid DNA was prepared by a QIAprep spin miniprep kit (Qiagen, Valencia, CA), and total genomic DNA was prepared using a Wizard genomic DNA purification kit (Promega, Madison, WI). Plasmid DNA was recovered from agarose gel slices by a QIAquick gel extraction kit (Qiagen), and restriction digests of plasmid DNA were purified with a QIAquick PCR purification kit (Qiagen). PCR was performed in a GeneAmp PCR system 2700 (Applied Biosystems, Foster City, CA). PCR products were cloned with pGEM-T (Promega) according to the manufacturer's instructions. Screening of potential mutants was performed using purified genomic DNA from potential isolates. Oligonucleotides were synthesized by IDT (Coralville, IA), and DNA sequences were determined by the Tufts University Core Facility (Boston, MA).
P. fluorescens Pf0-1 (adnA+) and Pf0-2x (adnA deletion mutant) were grown overnight in 50 ml of PMM. Five milliliters of each culture at the same optical density (OD; 600 nm [OD600]) was treated with 10 ml of RNAprotect bacteria reagent (Qiagen). The cells were centrifuged at 16,000 × g for 3 min. RNA was extracted from the cell pellets using an RNeasy minikit (Qiagen) according to the manufacturer's instructions. The extracted RNA was treated with RQ1 RNase-free DNase (Promega). Before creating cDNA, RNA samples were quantified using a NanoDrop spectrophotometer (Isogen Life Science, The Netherlands). The quality of the RNA was verified using an Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA). First-strand cDNA was synthesized from 10 mg total RNA with random hexamer primers from Invitrogen (Carlsbad, CA) using SuperScript double-stranded cDNA synthesis kits (Invitrogen). The cDNA synthesis was carried out according to the NimbleGen protocol (Roche NimbleGen, Inc., Madison, WI) for synthesis of double-stranded cDNA. The P. fluorescens Pf0-1 microarray was developed by Roche NimbleGen. The cDNA was labeled with Cy3 dye, and hybridization was performed by Roche NimbleGen. Two independent samples of each strain were analyzed using the NimbleGen microarray as previously described (18). Genes that showed a statistically significant difference from wild-type Pf0-1 (95% confidence interval) are reported in Table S1 in the supplemental material.
Internal fragments of 500 bp of each gene were amplified via PCR, introducing flanking BglII and SpeI restriction sites, and cloned into pGEM-T (Promega) for sequencing. The confirmed fragments were then cloned into the suicide plasmid pUIC3 (45). The resulting plasmids were introduced into both P. fluorescens Pf0-1 (adnA+) and Pf0-2x (adnA deletion mutant) by conjugation (2:1 ratio of recipient to donor). Homologous recombination between the 500 bp cloned in the plasmid and the Pf0-1/2x chromosome results in a single-crossover event creating a transcriptional fusion of the gene of interest to a promoterless lacZ carried by pUIC3 (45). The transconjugants were selected on PMM containing Tc7.5 and Ap50. Pf0-1/2x transconjugants were then isolated on fresh PMM containing Tc7.5 and Ap50. Colonies were checked for the production of β-galactosidase (blue pigment) by patching them on LB containing Tc15, Ap50, and X-Gal. Blue colonies were further examined for the gene disruption by PCR.
Deletions were made by using splicing by overlap extension-PCR (SOE-PCR) (24) and allelic exchange using the plasmid pSR47s (35). In short, SOE-PCR is a two-part process: (i) two 500-bp products are amplified using PCR from the chromosome flanking the gene of interest, and (ii) a second round of PCR is performed using the two 500-bp products, which both contain 35 bp from each product that overlap, in order to produce a full-length 1-kb fragment. This overlap region was engineered to contain an EcoRI restriction site. The amplified products were cloned into pGEM-T and sequenced. The resultant plasmids were digested with EcoRI, and a gene conferring resistance to the antibiotic gentamicin was cloned into this region. This Gm-resistance cassette was cloned from pBBR1MCS4 (29). The presence and orientation of the resistance gene were verified via PCR. The full-length fragments carrying a resistance cassette were cloned into the BamHI site of pSR47s (35). The resulting plasmids were transferred into P. fluorescens Pf0-1 as described above. Transconjugants were selected on PMM containing Gm20. Further isolation was done to screen for Kan-sensitive and sucrose-resistant colonies on PMM; the final sucrose concentration was 10%.
Strains were grown overnight in LB broth at 30°C. β-Galactosidase assays were performed on both Pf0-1 and Pf0-2x mutants grown in PMM as described elsewhere (37), with one exception: the mutants Pfl01_0596 and Pfl01_4889 were grown overnight in LB at 30°C and subcultured into PMM and grown overnight into late stationary phase and then were assayed. Assays were performed in three independent trials in duplicate.
Motility was evaluated on PMM containing 0.3% agar. Plates were allowed to solidify overnight at room temperature and then placed at 30°C for 5 to 10 min to remove any excess moisture before use. Cultures of all Pf0-1 mutant strains were grown overnight in LB at 30°C and then were diluted to an OD600 of 0.500 in PMM, and 10 μl of each strain was spotted on PMM containing 0.3% agar and Ap50. The motility plates were incubated for 20 to 22 h at 30°C overnight before measuring the diameter of motility in millimeters. Further evaluation of the deletion mutants was done by monitoring their motility over a 48-h period. Assays were performed in three independent trials in duplicate.
Chemotaxis was evaluated using PMM (without sodium succinate) containing 0.2% agar. The assays were performed as previously described (48) at 26°C. The formation of a ring in the direction of the chemoattractant, 10% Casamino Acids solution, was recorded as a positive result. Samples were plated in duplicate in three independent trials.
Biofilm formation was assayed using the method of O'Toole and Kolter (41). Experiments were performed using both borosilicate glass (hydrophilic surface) and polyvinyl chloride (PVC; hydrophobic surface). Borosilicate glass has previously been used in our laboratories to examine biofilm formation of P. fluorescens Pf0-1 (8, 48). Both PVC and borosilicate glass have also been used for studies of P. fluorescens biofilm formation (41). Both of these types of material are easier to visualize then others. PVC is also an environmentally relevant plastic since it is used in household plumbing.
Biofilm formation was evaluated at 30°C for all deletion mutants alongside the controls, Pf0-1 and Pf0-2x (adnA deletion mutant). All strains were grown overnight in PMM at 30°C. Then, each strain was inoculated into fresh medium and allowed to grow to mid-log phase. For borosilicate glass assays, samples of each strain (50 μl) were added to 1 ml of fresh PMM containing 0.2% glucose to a final OD600 of 0.025. For PVC assays, each strain was diluted in PMM containing 0.2% glucose, and 150 μl of each was added to 4 wells of a 96-well microtiter dish. All samples were allowed to incubate for 24 or 48 h at 30°C without agitation. All samples were then rinsed with double-distilled H2O (ddH2O) one time and stained with 1% crystal violet. The samples were incubated at room temperature for 20 min, and the samples were washed twice with ddH2O to remove any unbound stain. Ethanol (95%) was added to the tubes and wells for 20 min to remove the crystal violet dye. The samples were removed, and the absorbance was measured at 600 nm. The values were corrected for the negative controls, which consisted of borosilicate glass tubes or PVC wells incubated with PMM culture medium for the times indicated above. Biofilms were evaluated in reference to wild-type Pf0-1. Assays were performed in three to four independent trials in duplicate for borosilicate glass and in triplicate for PVC.
Cultures were grown overnight in PMM at 30°C for RNA extraction. The RNA was stabilized and protected by using RNAprotect bacteria reagent (Qiagen). RNA was extracted from P. fluorescens Pf0-1 and the Pfl01_1572 and Pfl01_1573 deletion mutants using an RNeasy minikit and on-column DNase I digestion (Qiagen). RNA so recovered was treated with RQ1 DNase (37°C, 1 h; Promega) and subsequently purified using an RNeasy minikit column. To access transcription of various gene products, reverse transcriptase PCR was performed. First-strand synthesis for reverse transcription-PCR (RT-PCR) was carried out using Superscript III (Invitrogen) at 52°C for 1 h. Gene-specific primers for first-strand synthesis are as follows: for detection of the Pfl01_1572/Pfl01_1573 transcript, primer Pfl1573R (5′-ACTAGTAGCCTGGGTCAACCCGCAGG-3′) was used. For detection of Pfl01_1572 and Pfl01_1573 in Pfl01_1573 and Pfl01_1572, respectively, primers Pfl01_1573R (5′-GCATGCGATCTGGTCTTTGCTTCAGC-3′) and 1572CompR (5′-AAGCTTCTTGAATCAATCAGATGTTCTCCAG-3′) were used. The cDNAs were amplified using the primers 1572RTF (5′-GTTGCTGATCCTGGTCGAG-3′, for the Pfl01_1572/Pfl01_1573 cDNA), RT1572F (5′ CTAGGGTATGAACGACAAGGCTAC 3′, for the Pfl01_1572 cDNA), and RT1573F (5′-CATCTGATTGATTCTCGAGGTAGC-3′, for the Pfl01_1573 cDNA) in the respective strains. A negative control consisting of a reverse transcriptase-free reaction mixture was used for each experiment. The RT-PCR experiments were performed in at least three independent trials for all target transcripts.
Each of the deletion mutants was complemented in trans using a wild-type copy of the gene in the plasmid pMQ71B under the control of an arabinose-inducible promoter (50). The gene fragments were cloned into pGEM-T for sequencing. These fragments included the wild-type gene, which was preceded by the ribosomal binding site (RBS), CACGGAGGAC, 8 base pairs upstream of the translational start site. This RBS sequence was determined on the basis of the alignment of upstream sequences of several Pf0-1 genes and has previously been used in our laboratories (unpublished data). The correct fragments were cloned into the plasmid pMQ71B (50) and transformed into E. coli. Plasmids were isolated from E. coli and transformed into P. fluorescens via electroporation. Electroporations were performed as previously described (9), with two exceptions: 6- to 8-ml cultures and 500 ng of plasmid DNA were used for each of the strains.
Complemented mutants were evaluated for motility in the same manner as their mutant counterparts as described above on soft agar plates with and without the addition of 0.25% l-arabinose. Chemotaxis experiments were performed as mentioned above with and without the addition of 0.25% l-arabinose. Biofilm assays were also performed as mentioned above with and without the addition of 0.25% l-arabinose. All complementation experiments were in two to three independent trials in duplicate for motility assays and in triplicate for biofilms.
All strains were grown overnight on LB plates containing the appropriate antibiotics at 30°C. Bacteria were transferred to copper-coated grids as follows: isolated colonies were covered with 1 drop of 1× phosphate-buffered saline (PBS) for 2 to 3 min. Grids were floated on the suspension for 1 min. Samples were then washed two times in 1× PBS and stained with uranyl acetate (2% in distilled H2O [dH2O]) for 25 s. Samples were viewed using a Philips CM-10 transmission electron microscope (TEM; Tufts University School of Medicine, Department of Anatomy, Electron Microscopy Facility, Boston, MA).
The soil used in these experiments was a gamma-irradiated fine loam from Sherborn, MA. The texture, nutrient, and mineral concentrations of the soil were determined at the Soil and Plant Tissue Testing Laboratory of the University of Massachusetts Amherst. The soil was 4.9% (dry weight) organic carbon, had a total nitrogen content of 0.224%, and had a pH of 5.3. The particle size distribution of the soil sample was 54.5% sand, 40% silt, and 5.5% clay. Nutrients in parts per million (ppm [wt/wt]) were as follows: calcium, 706; potassium, 105; aluminum, 71; nitrate, 69; magnesium, 62; iron, 14.3; phosphorus, 8; zinc, 5.4; manganese, 1.8; copper, 0.5; boron, 0.1; and nickel, 0.1. There was less than 2 ppm of lead, cadmium, and chromium.
The soil growth experiments were performed as previously described (52–54). In short, bacterial strains were grown overnight in PMM at 30°C with appropriate antibiotics. Cultures were standardized to an optical density (600 nm) of 1.00. Cells were diluted to approximately 1 × 105 CFU/ml in sterile distilled H2O. For survival experiments, 1 ml of the diluted cell suspension was mixed with 5 g of soil, achieving a water-holding capacity of approximately 50%. For competition experiments, cultures were adjusted to an optical density (600 nm) of 1.00 prior to dilution, and then 500 μl of each diluted competing strain was combined and mixed with soil as described for the survival experiments. Inoculated soil samples were transferred to a 15-ml conical polypropylene Falcon tube. The initial recoverable population was established by removal of 0.5 g of soil after 30 min and recovery and enumeration of bacteria from that sample. Briefly, 1 ml of sterile water was added to the 0.5-g soil sample, followed by vigorous vortexing for 30 s. After the soil particulates had settled, bacteria in suspension were recovered and enumerated by CFU determination. The initial populations of wild-type and mutant strains were approximately equal. Strain populations were monitored over 9 days by extraction of bacteria from the soil as mentioned above and determination of numbers by counting colonies plated on PMM containing the appropriate antibiotics. The wild-type strain in competition experiments was Pf0-1 Kmr (52). All assays were performed in three independent trials plated in duplicate.
A two-tailed Student's t test assuming equal variance was performed using Microsoft Excel software.
Microarray data were submitted to the Gene Expression Omnibus, http://www.ncbi.nlm.nih.gov/geo/ (accession number GSE33865).
Using microarray analysis, we compared the expression patterns of genes from P. fluorescens Pf0-1 (adnA+) and P. fluorescens Pf0-2x (adnA deletion mutant) grown in PMM. This comparison showed that 103 coding sequences (CDSs) were affected by the absence of AdnA (see Table S1 in the supplemental material). These genes represent the 95% confidence interval based on two independent trials. A majority of those CDSs, 92, showed decreased expression in the absence of AdnA, while the remaining 11 CDSs had increased expression (see Table S1 in the supplemental material). Four of the CDSs detected had previously been found in our laboratories to be regulated by AdnA: Pfl01_0657, Pfl01_1504 (flgI), Pfl01_1514, and Pfl01_4250 (flgC) (48).
Of those 92 CDSs that had decreased expression in the absence of AdnA, 48 encode predicted orthologs of proteins involved in either flagellar synthesis/export or chemotaxis in several Pseudomonas species. The remaining 44 CDSs encode orthologs of proteins involved in various enzymatic activities, a putative regulator, putative lipoproteins, heat shock proteins, and hypothetical proteins (see Table S1 in the supplemental material). None of these 44 CDSs is predicted to be involved in motility or chemotaxis in Pseudomonas.
From these 103 predicted CDSs, 13 CDSs (Table 2) were selected for further study on the basis of several criteria: genes that are controlled by AdnA in Pf0-1 but not FleQ in P. aeruginosa, genes with unknown functions, genes encoding putative regulators, and genes not related to motility in Pseudomonas spp. Three additional genes, all controlled by FleQ in P. aeruginosa, were chosen as controls. Two control genes, fliC and flgJ, are both required for motility (17, 38), and a third gene, Pfl01_1573, encodes a conserved hypothetical protein whose ortholog (PA1465) has no known function and is located in a putative chemotaxis operon containing two other genes. The genes upstream of Pfl01_1573 and PA1465 both encode proteins that have domains similar to those found in the chemotaxis protein CheW. CheW is a link between the methyl-accepting carrier proteins (MCPs) and CheA (20). The homolog of the P. fluorescens gene Pfl01_1572, PA1464 (cheW), has been shown to affect chemotaxis in P. aeruginosa (25).
Two of the genes chosen, Pfl01_0596 and Pfl01_4889, are negatively regulated by AdnA. Genes coding for two putative regulators, Pfl01_0623 and Pfl01_4889, were found (Table 2). Pfl01_0623 specifies a protein with putative GGDEF and EAL domains, and Pfl01_4889 encodes a proposed transcriptional regulator in the LuxR family. Two CDSs (Pfl01_1508 [fgtA] and Pfl01_1517) were annotated as glycosyl transferases on the basis of protein sequence identity (Table 2; see Fig. S1 in the supplemental material). The remaining nine genes specify either hypothetical proteins or ones of predicted function not related to motility.
The microarray results were verified for all 16 genes using lacZ transcriptional fusions. β-Galactosidase assays were performed, and the results from all of the genes examined are similar to the results from the microarray analysis (Table 2); still, some differences were observed when examining RNA versus protein synthesis. For instance, the results from the β-galactosidase assays showed that the Pfl01_0596 fusion had a 2-fold increase in the absence of AdnA (Table 2), while the microarray data showed a 45-fold increase in the absence of AdnA (Table 2; see Table S1 in the supplemental material).
All insertion mutants were tested for motility defects. Six mutants, fliC, flgJ, Pfl01_1508, Pfl01_1516, Pfl01_1517, and Pfl01_1573, showed defects. The two fliC and flgJ strains were used as controls for the β-galactosidase assays and had the expected motility-deficient phenotype, as seen in other Gram-negative bacteria (38, 61). Deletion mutants were made to test the importance of the products of the following genes for motility: Pfl01_1508, Pfl01_1516, Pfl01_1517, and Pfl01_1573. Motility was evaluated for each of the deletion mutants in comparison to wild-type Pf0-1. After 20 h, the motility was reduced or eliminated in mutants with each of the target genes (Fig. 1). Further investigation showed some motility at 48 h (Fig. 2A) in all strains except the Pfl01_1572 deletion mutant. This motility was not apparent at 20 to 22 h (Fig. 1) but was significantly less than that of Pf0-1 (Fig. 1 and and2A).2A). (At 48 h, the P values for each deletion mutant were as follows: P < 5 × 10−5 for ΔPfl01_1508, ΔPfl01_1516, ΔPfl01_1517, and ΔPfl01_1573 and P < 7 × 10−6 for ΔPfl01_1572.) Interestingly, the movement of two mutant strains was not in the concentric circles produced by wild-type Pf0-1. Pf0-1 ΔPfl01_1573 produced a fried egg type of colony, while Pf0-1 ΔPfl01_1516 showed irregular movement in the form of a small offshoot of motility after 48 h (Fig. 2A). The offshoots from ΔPfl01_1516 were subcultured onto fresh PMM containing 0.3% agar plates. The results from these experiments showed an increased motility from these offshoots compared to the initial inocula (Fig. 2C). This finding suggests the possibility of a suppressor mutation, which is under investigation. Mutants defective in swimming motility showed defects in chemotaxis as well (Table 2).
The wild-type Pf0-1 and the 16 mutants mentioned above were examined for biofilm formation on two different surfaces: borosilicate glass (hydrophilic) and polyvinyl chloride (PVC; hydrophobic) at 30°C (Table 2). PVC and borosilicate glass have been used by others to examine biofilm formation defects in P. fluorescens (8, 41, 48). The fliC and flgJ mutants failed to form biofilms on either surface, confirming the importance of flagella for biofilm formation by Pf0-1. One other mutant, Pfl01_0623, a putative GGDEF/EAL domain protein, showed a lack of biofilm production on PVC only (Table 2).
On borosilicate glass, three of the four deletion mutants, Pfl01_1508, Pfl01_1516, and Pfl01_1573, showed decreased biofilm production, while the fourth mutant, Pfl01_1517, showed an increased biofilm production compared to Pf0-1 at 24 and 48 h (Fig. 3A; for Pfl01_1517, P < 0.002 at 24 h and P < 0.031 at 48 h). Visual observation of the tubes after 24 h showed that only Pf0-1 and the mutants Pfl01_1517 and Pfl01_1573 produced biofilms, while the other mutants, Pfl01_1508 and Pfl01_1516, showed greatly decreased biofilms (data not shown). Very little dye binding was seen for either the Pfl01_1508 or Pfl01_1516 deletion mutants (Fig. 3A). Pfl01_1573 mutants seemed to produce a biofilm similar to that of Pf0-1 at 24 h with respect to the dye binding. After 48 h of incubation, Pf0-1 and the deletion mutants Pfl01_1508, Pfl01_1516, and Pfl01_1517 showed increased biofilm formation, as detected by the increased dye binding, while the Pfl01_1573 deletion mutant did not have an increase in dye binding (Fig. 3A). Both Pfl01_1508 and Pfl01_1516 deletion mutants showed about half the dye binding of Pf0-1, while the Pfl01_1517 deletion mutant seemed to produce a denser biofilm than Pf0-1.
On PVC, visual observation of the PVC plates after 24 h showed results similar to those observed on borosilicate glass. The Pf0-1 and the Pfl01_1517 deletion mutants produced biofilms on PVC, while the other mutants showed biofilm defects (Fig. 3B). After 48 h, only Pf0-1 and the Pfl01_1517 deletion mutant showed increased dye binding. The Pfl01_1508 and Pfl01_1516 deletion mutants resembled the negative-control strain, Pf0-2x, at 24 h (Fig. 3B; P = 0.51 and 0.77, respectively). However, at 48 h, the Pfl01_1508 deletion mutant had lower dye binding than Pf0-2x (P < 0.006), while the Pfl01_1516 deletion mutant had similar dye binding as Pf0-2x (P = 0.475). The Pfl01_1573 mutant had approximately the same dye binding on both borosilicate glass and PVC (Fig. 3).
Complemented strains were evaluated for restoration of motility, chemotaxis, and biofilm formation under the same conditions used to examine the deletion mutants (Table 3). Experiments using the pMQ71B expression vector (50) showed that each of the complementing genes restored motility to 35 to 89% of that of wild-type Pf0-1 (Fig. 4A). Complementation of swimming motility was also examined using clones in the expression vector pHERD26T (44), which restored motility to 53 to 90% of wild-type Pf0-1 (Fig. 4B). Even though there was not complete restoration of the motility defects using the pMQ71B vector (Fig. 4A), the wild-type colony phenotype was restored in all mutants (Fig. 2B).
In the Pfl01_1573 deletion mutant, the Pfl01_1573 gene did not complement the motility defect (Fig. 4). However, a clone carrying either Pfl01_1572 or both Pfl01_1572 and Pfl01_1573 did show complementation (Fig. 4A and B). Due to the procedures taken to create these mutations, the Pfl01_1572 gene was sequenced in this mutant in order to identify any possible disruptions that could have occurred using the SOE-PCR method; no such mutation was identified. RT-PCR data confirmed that these two genes are cotranscribed into a single mRNA (Fig. 5A).
A Pfl01_1572 deletion mutant was constructed and examined for complementation in the same manner as the Pfl01_1573 deletion mutant (Fig. 4B). The results show that only the addition of a wild-type copy of Pfl01_1572 or copies of both Pfl01_1572 and Pfl01_1573 was able to complement both Pfl01_1572 and Pfl01_1573 deletion mutants. Chemotaxis was restored for all of the mutants to 53 to 92% of the wild-type Pf0-1 level (Table 3). Biofilm formation was also restored in all the mutants (Table 3), with the exception of the Pfl01_1517 mutant, which did not show a biofilm defect (Fig. 3).
Mutants were evaluated for flagella using TEM. Wild-type Pf0-1 and all deletion mutants examined produced one to two flagella per cell (Table 3). Nearly all of the cells in the ΔPfl01_1573 samples lacked attached flagella, strongly suggesting that there may be attachment or structural problems.
In order to further our knowledge about soil survival and colonization, the mutants created in this study were evaluated in a sterile soil environment as previously described (52–54). Several deletion mutants, Pfl01_1508, Pfl01_1516, Pfl01_1517, and Pfl01_1573, were used for these experiments. No defects in soil growth were observed for any mutant used in these experiments compared to Pf0-1 (see Fig. S2 in the supplemental material). Competition experiments between Pf0-1 (kanamycin-resistant mutant) and two deletion mutants, Pfl01_1508 and Pfl01_1516, were performed in both live and sterile soil. These experiments showed no effect on soil colonization and survival in a laboratory test tube analysis (data not shown).
AdnA/FleQ is a master regulator for both flagellum-based motility and surface attachment in Pseudomonas spp. We have identified 103 genes in the AdnA regulon of P. fluorescens Pf0-1 by using a whole-genome microarray (see Table S1 in the supplemental material). About half of these genes are predicted orthologs of genes known to be involved in some aspect of motility. Of the remaining genes, 39 do not have orthologs in P. aeruginosa PAO1. The majority of these 39 genes are predicted to specify hypothetical proteins that do not have defined functional domains (see Table S1 in the supplemental material). By comparing the regulons of AdnA and FleQ in their respective organisms, one notes that while both are involved in regulation of motility genes, they are also involved in the regulation of a vastly different set of genes.
By investigating four AdnA-controlled genes not previously associated with motility in Pf0-1 (Pfl01_1508, Pfl01_1516, Pfl01_1517, and Pfl01_1572), we have identified new features of flagellum-dependent motility. This study and others (7, 13) have added to the list of genes that are involved in some aspect of flagellar synthesis and/or motility, highlighting the complexity of flagellar motility in Gram-negative bacteria. The ability of Pseudomonas to swarm or swim can also affect its ability to form biofilms (40, 42, 43; current study). This finding is not necessarily due to the known genes related to flagellar synthesis or transport (42).
Two of the genes, Pfl01_1508 and Pfl01_1517, identified have protein sequence identity with glycosyl transferases. Orthologs of these protein products, found in P. syringae and P. aeruginosa, affect virulence, motility, and biofilm formation in P. syringae (60–62) and virulence in P. aeruginosa (2). P. syringae possesses two flagellar glycosyl transferases, FgtA1 and FgtA2; deletion of either gene affects swarming motility in this strain (61). Deletion of fgtA1 has the same effect on swimming motility in viscous media (60) and decreased adhesion to polystyrene compared to the wild-type strain (61). In P. aeruginosa, FgtA is not involved in motility (2) and its role in biofilm formation has not been reported. A phylogenetic tree of Pseudomonas glycosyl transferases identified in GenBank was constructed (see Fig. S1 in the supplemental material). We have named Pfl01_1508 and Pfl01_1517 fgtA1 and fgtA2 based on their protein sequence identity to FgtA1 (57%, Psyrps6_010100004493) and FgtA2 (50%, Psyrps6_010100004488) from P. syringae pv. syringae 642. Many other studies have been performed using different genera and have indicated that flagellar glycosylation is involved in the production and assembly of flagella, not just motility and virulence (reviewed in reference 32).
The Pfl01_1508 (fgtA1) deletion mutant is different from P. syringae carrying its ortholog, in that it has motility defects on semisolid medium (0.3% agar). fgtA1 mutants also show biofilm defects, which may result from an inability to interact with borosilicate glass (hydrophilic) or PVC (hydrophobic) because of the altered properties of flagella that are not glycosylated appropriately. In P. syringae pv. tabaci 6605, an fgtA1 mutant has defects in binding to another hydrophobic surface, polystyrene (61).
In contrast to the mutants lacking the second putative glycosyl transferase, the Pfl01_1517 (fgtA2) mutant produces wild-type biofilm on PVC and has increased biofilm formation on borosilicate glass when grown at 30°C (Fig. 3). A deletion of either fgtA1 or fgtA2 in Pf0-1 results in the loss of swimming motility, unlike deletion of their counterparts in P. syringae, where only a mutant with disruption of fgtA1 has that phenotype in viscous medium, 0.6% agar or higher (60). Motility on agar surfaces with agar concentrations of 0.6% or higher is usually associated with swarming motility (22). Further investigation is required to determine whether indeed these two proteins have glycosyl transferase functions and which residues of FliC they glycosylate.
Pfl01_1516 was annotated as a cephalosporin hydroxylase but clearly has an important role in motility and biofilm formation and was therefore named flhH. Deletion of flhH resulted in greatly reduced motility (Fig. 1 and and2),2), even though the strain still produced flagella that are visually indistinguishable from those of Pf0-1 (data not shown). Biofilm formation on both borosilicate glass and PVC is also impaired in this mutant (Fig. 3).
Colonies of the Pfl01_1573 deletion mutant have a fried egg phenotype on motility plates (Fig. 2A). TEM showed that this deletion mutant produced flagella which were detached from the P. fluorescens cells (data not shown). While the loss of some flagella may be due to sample preparation-facilitated detachment, the complete lack of attached flagella was not seen with wild-type Pf0-1 or any other mutants prepared under similar conditions. Since the Pfl01_1573 mutant was still motile (albeit motility was reduced), the function is not in attachment of the flagella per se but in the tightness of the flagellar attachment. Clearly, flagella are made and are functional (Fig. 2A and data not shown); however, the inability to maintain attachment leads to reduced motility.
By complementing each of the motility mutants with its respective gene, the wild-type Pf0-1 colony phenotype was restored (Fig. 2B). However, the motility of each of these strains exhibited on soft agar was only 35 to 90% of wild-type levels (Fig. 4). The one exception is the Pfl01_1573 deletion mutant, which required a wild-type Pfl01_1572 gene, as described below. Biofilm formation and chemotaxis were also restored in these mutants as well (Table 3). The restoration of chemotaxis in the mutants ranged from 53 to 92% of that of the wild type.
In order to complement the motility defect of the Pfl01_1573 deletion mutant, a wild-type copy of the Pfl01_1572 gene was needed. During replacement of the wild-type Pfl01_1573 gene by a gentamicin-resistance cassette, both Pfl01_1572 and Pfl01_1573 could be affected since they are part of the same transcript (Fig. 5A). RT-PCR results showed that the Pfl01_1572 transcript was present in the Pfl01_1573 deletion mutant (Fig. 5B). The presence of the transcript verifies that the Pfl01_1572 mRNA transcript is stable (Fig. 5B); however, translation of the Pfl01_1572 protein may still be the limiting factor.
To help determine if the phenotype that we see in the Pfl01_1573 deletion mutant is linked to a disruption of translation of Pfl01_1572 or a deletion of Pfl01_1573, we complemented the strain with wild-type Pfl01_1572. Motility was restored in the Pfl01_1573 deletion mutant (to 55% of that of wild type using pMQ71B) bearing Pfl01_1572 (Fig. 4A). To further investigate these findings, a Pfl01_1572 deletion mutant was created. Removal of Pfl01_1572 completely abolished motility, unlike what was seen in the Pfl01_1573 deletion mutant (Fig. 1 and and2A).2A). Biofilm experiments were also carried out using the Pfl01_1572 deletion mutant. The results from this assay showed similar trends in both the Pfl01_1572 and Pfl01_1573 deletion mutants (Fig. 3). Chemotaxis results showed a drastic reduction in chemotaxis in the Pfl01_1573 mutant and a lack of chemotaxis in the Pfl01_1572 mutant (Table 2 and data not shown). We performed RT-PCR on the Pfl01_1572 deletion mutant in order to determine if Pfl01_1573 was transcribed downstream of the gentamicin-resistance cassette (Fig. 5B). The results showed that Pfl01_1573 is transcribed in this mutant.
The complementation results indicate that Pfl01_1572 is responsible for the motility defects seen in the Pfl01_1573 mutant. Pfl01_1572 specifies a putative CheW protein, and the ortholog in P. aeruginosa PAO1 was shown to be involved in chemotaxis (25). Kato et al. (25) also noted no motility defects in the cheW mutant on soft agar, but the effects on biofilm formation were not examined. To our knowledge, this is the first report of a CheW mutant affecting biofilm formation and completely abolishing swimming motility.
Both Pfl01_1572 and Pfl01_1573 deletion mutants exhibited reduced biofilm development compared to Pf0-1 over a 48-h period on both borosilicate glass (Fig. 3A) and PVC (Fig. 3B). Flagella are important for attachment by a number of organisms (reviewed in reference 66). The TEMs show a Pfl01_1573 mutant with detached flagella, which, in conjunction with other results, stems from the lack of a functional Pfl01_1572.
Recently, a protein, FlgT, which was identified to be an anchoring protein for flagella, was found in Vibrio cholerae (34). flgT mutations result in detached flagella. FlgT was found to be localized in the periplasm, which could indicate an interaction with either the P or L rings, which are part of the flagellar basal body (34). The mutant with the deletion of Pfl01_1573 in Pf0-1 had a similar phenotype, with the detachment of flagella, but no protein sequence identity between FlgT and Pfl01_1572 or Pfl01_1573 was seen (data not shown). This phenotype may also be related to the loss of a functional Pfl01_1572 in the Pfl01_1573 mutant.
To date our laboratories have identified only four genes important for colonization/survival in soil either under natural field conditions, adnA (33), or in soil competition experiments in a laboratory setting, cosA (52) and both ppk and pst (54). So far, the only gene to increase soil colonization of wild-type Pf0-1 has been cosA (52). Under the current laboratory conditions, the adnA mutant does not exhibit the same soil colonization defects (53) seen under natural field conditions (33). These differences in results show that there is a greater level of complexity in the natural field environment that cannot be duplicated in the laboratory. Thus, our results using the current soil methods (see Fig. S2 in the supplemental material) may not completely reflect the entire story that is happening in the natural soil environment. Further experiments are needed in order to determine each of these gene's roles under natural conditions.
The identification of novel genes involved in both motility and biofilm formation and the differences seen between organisms in the same genus and even species indicate that we do not fully understand motility in the Pseudomonas genus. Further investigation in other Pseudomonas species may help to better define the functional basis for these differences.
We thank Sarah Craven for her help in technical aspects of the study and the laboratory of George O'Toole for providing the pMQ71B expression vector.
This work was supported by National Research Initiative competitive grant 2006-35604-16673 (to S.B.L.) and Agriculture and Food Research Initiative competitive grant 2010-65110-20392 (to S.B.L. and M.W.S.) from the USDA's National Institute of Food and Agriculture, Microbial Functional Genomics Program.
Published ahead of print 6 April 2012
Supplemental material for this article may be found at http://aem.asm.org/.