Abstract
Incidences of bacterial foodborne illness caused by ingestion of fresh produce are rising. Instead of this being due to incidental contamination, the animal pathogen Salmonella enterica utilizes specific molecular mechanisms to attach to and colonize plants. This work characterizes two S. enterica genes of unknown function: a putative periplasmic protein, STM0278, and a putative protein with a hydrolase in the C-terminus, STM0650. STM0278 and STM0650 are important for seedling colonization but appear to have different roles during the process of colonization. Mutants of either STM0278 or STM0650 showed reduced colonization of alfalfa seedlings at 24 h, and the STM0278 mutant also showed reduced colonization at 48 h. Both genes were expressed in planta at 4 h following inoculation of 3-day-old seedlings and at 72 h after seed inoculation. This suggests that the role of STM0650 in seedling colonization is less important later in the process or is duplicated by other mechanisms. Mutants of STM0278 and STM0650 were defective in swarming. The STM0278 mutant failed to swarm in 24 h, while swarming of the STM0650 mutant was delayed. Addition of surfactant restored swarming of the STM0278 mutant, suggesting that STM0278 is involved in surfactant or osmotic agent production or deployment. Alfalfa seed exudates as the sole nutrient source were capable of perpetuating S. enterica swarming. Sequence analysis revealed sequences homologous to STM0278 and STM0650 in plant-associated bacteria, but none in Escherichia coli. Phylogenetic analysis of STM0650 showed similar sequences from diverse classes of plant-associated bacteria. Bacteria that preferentially colonize roots, including S. enterica, may use a similar hydrolase for swarming or biofilm production on plants. Multicellular behaviours by S. enterica appear central to plant colonization. S. enterica genes involved in plant colonization and survival outside of a host are most likely among the ‘function unknown’ genes of this bacterium.
Edited by: P. H. Everest
INTRODUCTION
Incidences of bacterial foodborne illness caused by ingestion of fresh produce are rising. Salmonellosis accounts for more than half of foodborne disease outbreaks, and the most common food commodity to which outbreak-related cases were attributed was produce (CDC, 2009). Instead of this being due to incidental contamination, the animal pathogens Salmonella enterica, Escherichia coli and Listeria monocytogenes utilize specific molecular mechanisms to attach to and then colonize plant seedlings and tissues (Gorski et al., 2003; Jeter & Matthysse, 2005). S. enterica, in particular, appears to employ an arsenal of surface molecules for attachment and colonization. During seedling colonization, S. enterica can reach large populations, similar to plant-associated bacteria (Barak et al., 2002).
S. enterica uses several surface molecules regulated by AgfD to attach to and colonize seedlings (Barak et al., 2005, 2007). AgfD regulates a multicellular behaviour that produces a matrix composed of thin aggregative fimbriae (Tafi) and cellulose. Mutations in the Tafi cell-bound minor subunit, cellulose biosynthesis, or a double mutant of both reduced S. enterica colonization of seedlings. However, genetic analysis of S. enterica to date has not revealed a single mechanism that abolishes plant attachment or colonization. We hypothesize that S. enterica employs many mechanisms either in unison or in parallel and they may include those produced by genes not characterized for other systems, animal pathogenicity or laboratory growth.
Previously, we created a transposon library of attachment-deficient S. enterica strains (Barak et al., 2005). Of the unique insertions, 65 % were in uncharacterized genes; this work further characterizes two insertions. A blast search of the affected genes against S. enterica serovar Typhimurium strain LT2 identified a putative periplasmic protein, STM0278, and a putative protein with a hydrolase in the C-terminus, STM0650.
METHODS
Bacterial strains, media and culture conditions.
The strains used in this study are listed in Table 1⇓. Complemented strains were constructed by amplifying wild-type copies of the affected genes (primers listed in Table 2⇓) from S. enterica Newport, cloning into pCR2.1 (Invitrogen), and electroporating the resulting plasmids into the appropriate mutant strain. A 1149 bp fragment was amplified and cloned for STM0278 (pJDB3) and a 1235 bp fragment for STM0650 (pJDB4). Each plasmid was transformed into its respective mutant and used for phenotype analysis.
Bacterial strains used in this study
PCR primers used in this study
Bacteria were grown in, or plated on, Luria–Bertani (LB) medium. When necessary, antibiotics were used at the following concentrations: ampicillin, 100 μg ml−1; kanamycin, 50 μg ml−1. Salmonella-Shigella (SS) agar, a Salmonella semi-selective indicator medium, was used to determine S. enterica populations. Alfalfa medium (AE) was made by sprouting alfalfa seeds in sterile water at a ratio of 1 g per 10 ml. Exudates were removed daily for 3 days, filter-sterilized (0.2 μm), and stored at 4 °C until used, but for no longer than 1 week. Sterile water was added to the seeds following removal of exudates.
We tested the ability of each strain to swarm or swim using swarm plates (LBG) composed of LB with 0.5 % (w/v) glucose and 0.5 % (w/v) agar, and swim plates composed of LB with 0.3 % (w/v) agar. Overnight bacterial cultures grown on LB agar were suspended in sterile water at a concentration of ∼109 c.f.u. ml−1. Twenty microlitres of the bacterial suspension was spotted in the middle of a swarm or swim plate, and the spot allowed to dry for 1 h at room temperature. All plates were incubated at 28 °C for 24 h or as noted otherwise. The swarming or swimming distance was measured from the centre of the original spot to the furthest distance that the bacteria travelled. To determine if swarming deficiencies were related to a lack of surfactant, Bacillus subtilis surfactin (2.5 mg ml−1) was added to LBG plates and the rate of travel was measured over time as described above.
Flagella were counted from swarm cells grown on LBG plates for 24 h at 37 °C. Sterile water was placed on a glass slide, swarm cells lifted from the agar surface with a plastic loop, and the loop touched to the water drop to transfer the cells to the slide. Slides were left in a laminar flow biosafety hood to dry. A drop of flagella stain (Becton Dickinson) was then added to the bacteria, which were incubated at room temperature for 3.5 min, rinsed with sterile water, and left to dry. Fifty bacterial cells were examined and flagella per cell were counted. Fifty flagella were chosen randomly and their lengths were recorded.
To test if alfalfa exudates could trigger swarming, a second type of swarm plate was made, composed of alfalfa exudates at a ratio of 2 : 1 (alfalfa exudates : water agar). Water agar was made to a final concentration of 0.5 % (w/v) agar. Hybrid LBG-AE swarm plates were prepared by a slight modification of the method of Kim & Surette (2004). Briefly, a small Petri dish was inverted in the middle of a large Petri dish and the inverted plate was sealed to the bottom of the large plate with Vaseline. Semi-solid AE was poured around the inverted dish. When the medium had solidified, the small dish was removed and the space was filled with LBG to the height of the AE. The hybrid plates were allowed to dry overnight and inoculated and incubated as described above.
Biofilm assays.
Biofilm assays were conducted using 5 ml LB made without salt (LBNS) in 13×100 mm test tubes. A bacterial suspension was added to the medium to give a starting concentration of approximately 105 c.f.u. ml−1. Tubes were incubated stationary at 28 °C. After 3 days, media and pellicles were poured from the tubes, and the tubes were rinsed three times with sterile water. A solution of 30 μl crystal violet (0.5 g per 100 ml in 20 % methanol) was added to each tube with 5 ml sterile water and the tubes were left standing for 15 min. The solution was poured out and the tubes rinsed with sterile water until no remaining crystal violet was rinsed from the biofilms. Thirty per cent acetic acid was added to each tube, and the tubes placed in a spectrophotometer, where transmission was read at 540 nm.
Alfalfa seedling attachment and colonization assays.
Alfalfa seedling colonization assays were performed as described previously (Barak et al., 2005). All experiments were performed at least three times. Briefly, alfalfa seeds were surface sanitized, sprouted in Petri plates, and irrigation water was exchanged daily. For initial attachment assays, 3-day-old uninoculated seedlings were used that ranged from 0.01 to 0.03 g (mean 0.0182±0.004 g). Twelve seedlings were transferred to 50 ml conical tubes containing bacterial suspensions (∼104 c.f.u. ml−1) and incubated for 4 h at 25 °C with constant shaking at 40 r.p.m. Individual seedlings were rinsed in 1 ml sterile water for 30 s and then homogenized. The homogenate for each seedling was dilution plated on SS agar; the plates were incubated overnight at 37 °C, and the colonies counted.
For colonization and microscopy assays, seeds were incubated in a bacterial suspension of ∼106 c.f.u. ml−1. After 1 h, the suspension was replaced with sterile water and the seeds were sprouted for 3 days at room temperature. To determine bacterial populations, seedlings were removed after the daily water change at 24 and 48 h and sampled as described above.
Microscopy.
Putative promoter regions of the STM0278 and STM0650 genes were PCR amplified (PCR primers listed in Table 2⇑) from S. enterica Newport and cloned into pCR2.1. All constructs were confirmed and promoter orientation determined by sequencing. Putative promoter regions were excised with HindIII/XbaI and cloned into pProbe NT (Miller et al., 2000). Each pProbe NT construct was electroporated into S. enterica Newport wild-type. These strains were used to determine whether STM0278 or STM0650 was expressed on alfalfa seedlings.
Statistics.
The mean S. enterica populations on seedlings (c.f.u. g−1) were calculated and the 24 and 48 h data were log-transformed prior to statistical analysis. To determine whether the swarming distance or rate differed between wild-type and mutant, 10 plates were used per strain for each experiment. The mean distance was calculated for each time point. The mean swarming rate was calculated for each strain tested and separately for each medium. Determination of the S. enterica populations on alfalfa seedlings and swarm experiments was repeated three times. Differences between means for wild-type and mutant, and when appropriate the complemented strains, were analysed by unpaired t-tests with Welch correction.
Phylogenetic analysis.
For the phylogenetic analysis of STM0278 and STM0650, sequences were downloaded from NCBI (). Amino acid sequences from taxa identified in the STM0650 blast analysis were aligned using clustal w in megalign 5.08 (Lasergen, DNASTAR) in slow-accurate mode with a pairwise alignment gap-opening penalty of 10 and a gap-extension penalty of 0.1, and multiple alignment gap-extension penalty of 0.2. Phylogenetic affiliations based on the alignment matrix were inferred by maximum-parsimony and distance using paup v 4.0b10. For parsimony analyses, informative characters were used in a heuristic search with random-stepwise addition of taxa and tree bisection–reconnection branch swapping (1000 replicates). Tree topology and statistical estimates of branch support were based on 1000 bootstrap replicates, with group retention set at >70 %.
RESULTS
Hypothetical proteins are involved in different phases of plant colonization
Individual S. enterica strains were inoculated either on 3-day-old seedlings to ascertain initial attachment capacity or on seeds to determine the ability to colonize alfalfa seedlings over time. In comparison with the wild-type, significantly fewer cells of STM0278 and STM0650 were detected on the 3-day-old seedlings (Fig. 1a⇓). The complemented strain of STM0278 had populations higher than the original strain and similar to those of the wild-type. The complemented strain of STM0650 had higher populations than the original strain but smaller populations than the wild-type. In the continued colonization assays, mutants of STM0278 and STM0650 both had populations smaller than the wild-type at 24 h. At 48 h the STM0278 mutant population was significantly smaller than wild-type (Fig. 1a⇓). The complemented strain of STM0278 had populations higher than the original strain and similar to those of the wild-type at both 24 h and 48 h. The complemented strain of STM0650 had higher populations than the original strain and similar to those of the wild-type at 24 h. At 48 h, the populations of both the STM0650 mutant and its complemented strain were similar to those of the wild-type.
Role of STM0278 and STM0650 in S. enterica colonization of alfalfa seedlings. (a) Data from a representative experiment, showing the mean c.f.u. per seedling (±sd) at 4 h post-inoculation of 3-day-old alfalfa seedlings or the mean log c.f.u. per seedling at 24 and 48 h post-inoculation of alfalfa seeds. Different letters in columns signify significant differences between S. enterica populations (P<0.05). (b) Micrographs of confocal laser scanning microscopy projected z-series of promoter-probe assays of S. enterica colonizing alfalfa seedlings. Seedlings were visualized using the nucleic acid stain SYTO59 (blue). S. enterica cells were intrinsically labelled by the plasmid pProbeNT, which carries the gene encoding the green fluorescent protein transcribed from STM0278 or STM0650. Scale bar, 50 μm.
To visualize gene expression in vivo, the promoters of STM0278 and STM0650 were cloned upstream of the green fluorescent protein gene in pProbe-NT and the plasmids were transformed into S. enterica Newport cells. Green fluorescent cells were not observed when the bacteria were grown on LB or SS plates. The reporter plasmids did not appear to inhibit bacterial colonization since similar numbers of S. enterica Newport cells colonized individual seedlings regardless of the plasmid carried (data not shown). When seedlings were examined microscopically, green fluorescent cells were observed on alfalfa roots from both promoter reporters (Fig. 1b⇑). For both constructs green fluorescent cells were observed attached to alfalfa roots at 4 h post-inoculation, and dense aggregates of green fluorescent cells colonizing roots were observed at 72 h post-inoculation.
Plant attachment and colonization genes also play a role in swarming
To characterize mutants of STM0278 and STM0650, their ability to swim and swarm was tested. Mutants of STM0278 and STM0659 swam similarly to wild-type (data not shown). Within 24 h, the wild-type swarmed to the perimeter of the plate while the STM0278 mutant was unable to swarm, and the STM0650 mutant was reduced in swarming ability. The complemented STM0650 swarmed to wild-type levels, and the STM0278 complement actually swarmed faster than the wild-type (Fig. 2a, b⇓). The mean number of flagella per cell on cells from swarm-inducing medium was similar for all strains (data not shown). Flagellar length was also examined and no difference was found among the strains (data not shown).
Mutants of STM0278 and STM0650 are defective in swarming. (a, b) Data from representative experiments showing the mean colony radius (±sd) of cultures inoculated on Luria–Bertani supplemented with glucose (LBG) semi-solid agar. (c) Mean swarming rate with and without the addition of surfactant (Sur) to LBG semi-solid agar. Different letters in columns signify significant differences between S. enterica populations (P<0.05).
To further characterize the role that STM0278 and STM0650 play in the ability of S. enterica to swarm, strains were spotted on medium with B. subtilis surfactin. Both mutants swarmed more slowly over LBG compared to wild-type. The STM0278 mutant swarmed at the same rate as the wild-type when surfactin was present in the medium, whereas the STM0650 mutant still showed reduced swarming (Fig. 2c⇑).
To determine if S. enterica would swarm on alfalfa seedlings grown in a hydroponic environment, alfalfa seedling exudates were used as the sole carbon source to induce or maintain swarming. S. enterica could not be induced to swarm on AE when inoculated at high density. However, when the bacteria were induced to swarm on LBG, swarming was perpetuated on AE to the plate edge of some alfalfa exudates (Fig. 3⇓). Alfalfa exudates from 24 h seedling failed to maintain swarming (Fig. 3b⇓) while exudates from older, 48 and 72 h, seedlings did maintain swarming (Fig. 3c, d⇓).
Swarming maintenance of cells induced on Luria–Bertani supplemented with glucose (LBG) semi-solid agar. Hybrid medium consists of LBG in the inner circle, with the outer ring containing LBG (a), 24 h alfalfa exudates (b), 48 h alfalfa exudates (c), or 72 h alfalfa exudates (d).
Mutants of STM0278 and STM0650 are altered in biofilm formation, in a temperature-sensitive manner
To further characterize STM0278 and STM0650, the ability of the mutants to form biofilms was tested. Wild-type biofilms were more developed at 30 °C than 37 °C as measured by transmission of light through crystal violet. Both STM0278 and STM0650 mutants developed less biofilm at 30 °C compared to wild-type and more biofilm at 37 °C (Table 3⇓). The mutant phenotype was more pronounced in STM0278. The complemented strain of STM0278 had a staining intensity similar to wild-type; the complemented strain of STM0650 did not.
Light transmission (540 nm; mean±sd) through crystal violet suspensions from S. enterica biofilms grown in LBNS
Significant differences are shown as in Figs 1(a) and 2(c).
Homologues of the hypothetical proteins
To compile information related to STM0278 and STM0650, several sequence analyses were performed. STM0278 is the gene following STM0277 in a two-gene operon. STM0277 is also an uncharacterized gene, whose product may be a putative cytoplasmic protein. A protein blast analysis of STM0278 identified homologous sequences in S. enterica, Burkolderia ambifaria and Enterobacter sakazakii (Table 4⇓). Sequences homologous to STM0277 were found by protein blast in S. enterica, Erwinia tasmaniensis, Vibrio sp., Acinetobacter baumannii, Pseudomonas syringae pv. syringae and Burkholderia cenocepacia. A multi-genome alignment revealed no orthologues in S. enterica serovar Typhi or serovar Choleraesuis, or in Escherichia coli. Orthologues were identified for both STM0277 and STM0278 in S. enterica serovars Typhimurium and Paratyphi, and in Ps. syringae pv. syringae (data not shown). However, the Ps. syringae pv. syringae operon has a third gene of unknown function following the STM0278 orthologue. Single-gene orthologues of STM0278 not found in operons were identified in Yersinia pestis and Ps. syringae pv. tomato.
Top five results from blastp analysis
STM0650 is the second gene of a two-gene operon, and the gene product may contain a hydrolase in the C-terminus. STM0649 is also an uncharacterized gene whose product may contain a hydrolase in the N-terminus. A protein blast of STM0650 identified homologous sequences in all S. enterica serovars available, Desulfitobacterium hafniense, Carboxydothermus hydrogenoformans, Clostridium botulinum, Clostridium difficile, Thermosinus carboxydivorans, Thermoanaerobacter pseudethanolicus and Geobacillus kaustophilus (Table 4⇑). A protein blast of STM0649 identified homologous sequences in the same bacterial species as that of STM0650, except those of Clostridium (data not shown). A multi-genome alignment identified orthologues for STM0649 and STM0650 in S. enterica serovars Typhimurium, Paratyphi A and Typhi, but not in other Enterobacteriaceae, e.g. Escherichia coli and Yersinia pestis. STM0649 and STM0650 are within a 4 kb region found in S. enterica and not in E. coli.
Numerous plant-associated taxa were identified in the protein blast analysis of STM0650, and a subset were used in a phylogenetic analysis of the S. enterica serovar Newport STM0650 homologue. The subset was chosen to include both plant pathogens and symbionts and at least one sequence to represent each species identified in the original blast analysis. The S. enterica serovar Newport amino acid sequence was identical to other S. enterica sequences and most closely related to sequences from the enteric plant pathogens Pectobacterium atrosepticum and Pectobacterium carotovorum subsp. carotovorum (Fig. 4⇓). Orthologues of STM0650 were identified in plant-associated bacterial genera, both pathogenic and symbiotic, from various taxonomic classes, including Agrobacterium tumefaciens, Ralstonia solanacearum and Bradyrhizobium japonicum.
Maximum-likelihood tree of plant-associated taxa and the S. enterica serovar Newport STM0650 homologue. The numbers at each node are bootstrap values based on 1000 replicates using heuristic research based on parsimony with substitutions weighted according to the instantaneous rate matrix. The names of taxa have been abbreviated as follows: Agrobacterium tumefaciens (At), Bradyrhizobium japonicum (Bj), Burkholderia cenocepacia (Bc), Pectobacterium carotovorum subsp. carotovorum (Ec), Pectobacterium atrosepticum (Pca), Pseudomonas syringae pv. syringae (Pss), Pseudomonas syringae pv. tomato (Pst), Pseudomonas syringae pv. phaseolicola (Psp), Ralstonia solanacearum (Rs), Salmonella enterica serovar Newport (SEN), Salmonella enterica serovar Typhi (SE typhi), and Xanthomonas campestris pv. campestris (Xcc); NCBI accession numbers follow taxa.
DISCUSSION
The uncharacterized genes of S. enterica examined in this study were both important for seedling colonization but may have different roles over time. Both the STM0278 and STM0650 mutants showed a reduction in initial attachment and colonization of alfalfa seedlings at 24 h, but the STM0278 mutant also showed reduced alfalfa seedling colonization at 48 h compared to the wild-type. To our knowledge, STM0650 is the first S. enterica gene to be identified as important for initial attachment, but not continued colonization of seedlings. The other mechanisms identified as important for S. enterica initial attachment, namely thin aggregative fimbriae, O-antigen capsule, and cellulose (Barak et al., 2005, 2007), are also important for continued colonization. The promoters for both STM0278 and STM0650 were expressed at 4 h following inoculation of 3-day-old seedlings and at 72 h after inoculation of alfalfa seeds. Since STM0650 is expressed at 72 h post-inoculation but its mutant did not show reduced colonization at 48 h, we hypothesize that STM0650 is less important or is duplicated by other mechanisms at this time point. In contrast, STM0278 is necessary for both initial attachment and continued colonization. Our work to date, including this study, supports the conclusion that several mechanisms are utilized by S. enterica to persist on seedlings. What is novel in this study is the discovery that different mechanisms are employed over time.
Examining the role of STM0278 and STM0650 on seedlings led us to examine other phenotypes in our mutant analysis. Flagella and motility play a role in fitness and colonization of alfalfa seedlings by L. monocytogenes (Gorski et al., 2009). Neither the STM0278 nor the STM0650 mutant was affected in the ability to swim, but neither could swarm comparably to wild-type. In S. enterica, swarm cells are elongated, hyper-flagellated cells that can migrate across semi-solid surfaces and more viscous media (Harshey & Matsuyama, 1994), which may be encountered in the sprouting system. Since both mutants could swim similarly to wild-type, we concluded that they did not have flagellar assembly defects. The length and number of flagella were similar between mutants and wild-type. Thus the failure to swarm was not due to an inability to hyper-flagellate or an obvious defect in the flagella.
The swarming dysfunction of each mutant was distinct. The STM0278 mutant failed to swarm, while swarming of the STM0650 mutant was delayed. Swarming bacteria produce a surfactant as a wetting agent (Toguchi et al., 2000), or an osmotic agent that extracts water from the underlying agar and provides a thin layer of fluid to aid the movement of cells across a semi-solid surface (Rauprich et al., 1996). Addition of surfactant to swarm-competent medium allowed the STM0278 mutant to swarm at the same rate as wild-type. These results suggest that STM0278 plays a role in the production or deployment of a surfactant or osmotic agent. Although the hydroponic growing conditions of alfalfa seedlings would appear to be aqueous, and thus only the capacity to swim a necessity for bacteria to colonize seedlings, the microscopic environment encountered by individual bacterial cells may in fact be less fluid than expected. S. enterica preferentially colonizes alfalfa roots (Fig. 1⇑ and Barak et al., 2007), forming large aggregates around root hairs (Charkowski et al., 2002) and in mucilage near the root tip (Mohle-Boetani et al., 2009). Mucilage (polysaccharides) and proteins compose the largest proportion of root exudates (Bais et al., 2006), suggesting that swarming may be useful to reach preferred colonization sites along roots that can be accomplished only by traversing this viscous environment.
To determine whether S. enterica swarms on alfalfa seedlings, we used alfalfa exudates to induce swarming. Alfalfa exudates added to low-percentage agar failed to induce swarm cell differentiation. S. enterica cells that were induced to swarm maintained their swarming ability and actively migrated across alfalfa exudates (Fig. 3⇑). Swarm-cell differentiation and maintenance in S. enterica require different essential nutrients (Kim & Surette, 2004). When propagated on M9 minimal medium supplemented with glucose (M9G), vegetative cells do not undergo swarmer differentiation; however, differentiated swarmers can maintain their ability to migrate on M9G, unlike LB without the addition of glucose. Our data support the conclusion that alfalfa exudates from 48 h and 72 h seedlings can maintain efficient utilization of the full tricarboxylic acid cycle and enhanced de novo biosynthetic pathways for cells to maintain the swarm state, similar to what was shown for M9G (Kim & Surette, 2004).
S. enterica swarmers also display an increased production of the universally recognized signalling molecule AI-2 and sensitivity to N-acylhomoserine lactones (AHLs), signalling molecules that Salmonella does not produce (Kim & Surette, 2006). Although AHLs have not been identified in the human gastrointestinal tract, where S. enterica causes disease, AHL mimics have been detected from rice, soybean, tomato, crown vetch and Medicago truncatula (Teplitski et al., 2000). sdiA, the Salmonella AHL receptor, is not active during transit through chicks, pigs, mice, guinea pigs, calves or rabbits (Smith et al., 2008); however, it may be active outside an animal host. S. enterica swarmers also display an adaptive antibiotic resistance to a wide variety of structurally and functionally distinct classes of antibiotics (Kim et al., 2003), which would be advantageous among the diverse plant microbiota and as protection from the plant itself. We hypothesize that swarming in the plant environment would offer Salmonella an ecological advantage: (1) motility to explore and locate preferential colonization sites; (2) heightened sensitivity for intra- and interspecies communication; and (3) protection from antimicrobials.
In addition to swarming, mutants of STM0278 and STM0650 are affected in biofilm formation. S. enterica forms biofilm on the roots and root hairs of alfalfa seedlings (Charkowski et al., 2002), the roots of mung bean seedlings (Mohle-Boetani et al., 2009), and lettuce roots (Klerks et al., 2007). We have previously identified S. enterica genes, agfA and bcsA, that play a role in seedling colonization via biofilm formation (Barak et al., 2007). We hypothesize that STM0278 and STM0650 may be involved with seedling colonization through swarming and biofilm formation, but at different times. Swarming and biofilm formation seem in contrast, motile versus sedentary. Bacteria on plant leaf surfaces can be either solitary or in aggregates – biofilms –and appear to play different roles in these modes. On leaves it is hypothesized that the aggregated population serves to protect bacteria under stressful conditions and the solitary cells may foster spread to newly colonizable sites (Boureau et al., 2004). S. enterica has been reported to require flagella production and movement to travel from the roots of Arabidopsis thaliana along the surface of the plant to its leaves and flowers (Cooley et al., 2003). Our data identify STM0278 and STM0650 as two of the S. enterica genes involved in both swarming and biofilm formation and suggest that these are the mechanisms used during plant colonization.
We identified nucleotide sequences similar to both STM0278 and STM0650 in plant-associated bacteria. We interpret this as further evidence of the importance of these genes in the life cycle of S. enterica outside a warm-blooded host. A phylogentic analysis of the amino acid sequence of STM0650 revealed that plant-associated bacteria from diverse classes, i.e. Alpha- and Betaproteobacteria, have similar sequences (Fig. 4⇑). We hypothesize that these putative hydrolase proteins in Agrobacterium tumefaciens, Ralstonia solanacearum and Bradyrhizobium japonicum may also be involved in root colonization of these plant-associated bacteria, pathogens and symbiont.
As more information is learned about the molecular mechanisms utilized by S. enterica to colonize and survive on seedlings, multicellular behaviour appears to be important. Previous studies have shown the importance of AgfD-regulated behaviour that produces a matrix of Tafi and cellulose (Barak et al., 2005). This study identified a second multicellular behaviour, swarming, that appears to be involved in seedling colonization. Plant roots may be a natural part of the life cycle of S. enterica. Biofilm formation is by its nature a multicellular behaviour. Mutations in agfD, or either of the matrix components, disrupt biofilm formation at 30 °C (Barak et al., 2007) while mutations in STM0278 and STM0650 lead to increased biofilm formation at this temperature (this study). Thus, careful coordination of several multicellular behaviours allows S. enterica to successfully attach to and colonize seedlings.
Successful colonization of plants that may be eaten by a desirable host may be an additional benefit to S. enterica. Plants have become a significant vector for S. enterica to move from host to host, as seen by the numerous salmonellosis outbreaks caused by the consumption of contaminated produce, including seedlings. The S. enterica genes involved in plant colonization and survival outside a warm blooded host are most likely among the ‘function unknown’ genes, as the mechanisms used in these environments differ greatly from the well-studied environment of animals or laboratory media.
Acknowledgments
We thank J. Palumbo for helpful discussions and critical reading of the manuscript. This research was partially funded by the US Department of Agriculture (Agricultural Research Service CRIS project number 5325-42000-044-00D).