Background

Results

Conclusions

Motility is considered to be a beneficial trait for microbes, as it solves many of the challenges these microorganisms face in their habitats, allowing them to find hosts, explore local environments, access protected niches and needed nutrients, and disperse to new locations. The quantitative relationship between motility and virulence, particularly of plant-associated microorganisms, remains unclear. This study has dissected the quantitative contributions of these factors in the behavior of the plant pathogenic bacterium Pseudomonas syringae within plant tissues as well as on leaf surfaces, where it commonly occurs as an epiphytic resident of leaves. P. syringae is the only known plant pathogen with pili from the widespread chaperone-usher fimbrial family as well as all subfamilies of type IV pili. These families of pili are important factors in bacterial virulence, and implicated in adhesion, invasion, and biofilm formation in mammalian systems. However, the role of pili are less well characterized in plant pathogens. As such, P. syringae is a useful model system for studying these processes in plant colonization. The availability of reproducible, quantitative variation in nearly isogenic P. syringae strains, coupled with readily generated mutants altered in expression of adhesive pili, made P. syringae an obvious choice for investigating the relative contributions of adhesive factors and motility to the behavior of plant pathogens.

Nine nearly isogenic isolates of P. syringae pv. syringae strain B728a exhibited stable and reproducibly different swimming and swarming motility in culture. Swimming and swarming motility were positively correlated among the variants, suggesting that similar traits contributed to both processes. A greater rate of motility of variants was associated with their incitation of a higher incidence of lesion formation on bean leaves after topical application of bacteria. The size of lesions formed by variants after direct injection into bean pods, thus bypassing the need for invasion, was also positively correlated with both the relative rate of motility in culture and their ability to incite leaf lesions. The differences in motility among the variant strains was not associated with any difference in biosurfactant production, although mutants blocked in syringafactin production exhibited a deficiency in ability to move within the apoplast of bean pods. Re-sequencing of the variant strains using high-throughput (Illumina) sequencing followed by genomic analyses revealed five single nucleotide polymorphisms (SNPs) and two short simple repeat regions (SSRs) differing among the nine P. syringae variant strains. None of the SNPs or SSRs were associated with genes known to be directly involved in flagellar or fimbrial biogenesis and regulation. Site-directed knockout mutants in those loci altered in variant strains recapitulated the reduced swarming motility exhibited by a given strain. Phylogenetic reconstruction of the collection of strains based on variation in genome features was in agreement with historical relationships of the strains. Additional analyses of two P. syringae variant strains that were selected as having either higher or lower swarming motility than their parental strains revealed additional genetic elements involved with modulating swarming motility. A mutation leading to early truncation of ispA, encoding a protein involved in cell septation, was associated with the variant exhibiting reduced motility and thus reduced virulence to bean. While no genetic reason could be identified in the selected variant that exhibited enhanced motility, the fact that it exhibited traits specifically beneficial to motility in culture may explain why it was less virulent than expected based on its high degree of motility in vitro.

Further investigation of factors affecting motility were explored through bioinformatic analyses of putative fimbrial biogenesis pathways. Five putative fimbrial types were identified which had structural similarities to either chaperone-usher (CU) fimbriae, type IVa fimbriae, type IVb fimbriae, or the type IVb monophyletic sub-family known as Flp fimbriae. Additional analyses of strains deficient in the production of the major structural pilin of type IVa fimbriae (PilA), CU fimbriae (FimA), and Flp fimbriae (Flp) revealed that twitching motility was not a major contributor to P. syringae motility. However, flp mutants and to a lesser extent fimA mutants exhibited higher swimming and swarming motility, suggesting that adhesive features of these fimbriae suppressed motility. Consistent with such a role, over-expression of Flp pilin resulted in a significant suppression in motility of the wild-type P. syringae strain. Co-inoculation of wild-type and fimbrial mutants onto leaf surfaces revealed that Flp fimbriae are involved in surface attachment, as the flp mutant exhibited a decreased attachment to bean leaves when compared with the wild-type strain. Attachment by Flp fimbriae appears to be an early step in the colonization process, as the reduced attachment of the flp mutant was much less pronounced after a longer incubation period. fimA and pilA mutants exhibited similar abilities to attach to leaf surfaces as the wild-type strain. Topical application of cell suspensions of pilA mutants onto bean leaves resulted in the formation of fewer lesions than conferred by the wild-type strain. When simulated rain was applied to leaves after inoculation, both the pilA and fimA mutant incited as much as 2-fold fewer lesions than the wild-type strain, suggesting that they adhered less tenaciously and thus were retained less well after the rain event. fimA and flp mutants but not a pilA mutant formed larger lesions than the wild-type strain when directly inoculated into plant tissue, suggesting that their decreased adhesiveness contributed to their ability to move within the plant after inoculation. Taken together, Flp and CU fimbriae appear to have significant roles in adhesion that affect their motility in planta, and type IVa fimbriae appear to be important in inciting disease on bean, presumably by aiding local movement to sites suitable for invasion of the leaf.

Bacterial ability to multiply endophytically (inside leaves) is an intrinsic trait of bacterial plant pathogens such as Pss B728a and Pto DC3000. Type III secretion system (TTSS)-possessing pathogens multiply endophytically to large population sizes (>108-109cfu/g). Non-pathogenic strains such as the saprophytes Pantoea agglomerans and Pseudomonas fluorescens, and hrp/hrc cluster mutants (TTSS mutants) mostly fail to multiply endophytically. The growth-restricting activity of the plant defense surveillance system is thought to restrict growth of non-pathogens inside the leaves through a series of responses that induce deposition of callose, reactive oxygen species production, and production of a variety of antimicrobials. This process is collectively known as PTI (pathogen-associated molecular pattern-triggered immunity). However we found several treatments that induce high sugar availability in the apoplast to elicit a dramatic growth of various non-pathogenic strains. Saprophytes and hrp mutants growing in both host and non-host plants multiplied up to 10,000-fold either when co infiltrated into plants with utilizable (i.e. metabolized substrates in vitro) carbon sources, infiltrated with water alone while water-soaking conditions maintained or when co-infiltrated with a TTSS-proficient pathogen. Pathogens were not limited in growth by carbon sources since they carry TTSS effectors that they can deploy to provide them with needed resources. We thus propose that similarly to bacteria growing in the surface of leaves, availability and distribution of sugars are the main limiting factors for endophytic bacteria, and while TTSS-proficient pathogens can alter the plant to obtain apoplastic nutrients, non-pathogenic strains cannot.

Bacterial colonization of a plant host requires both specialized and general traits. On the leaf surface and in the leaf interior (apoplast), bacteria encounter resources needed to support growth but are also exposed to multiple abiotic and biotic stresses that they must overcome. The foliar plant pathogen Pseudomonas syringae can establish large epiphytic populations on leaf surfaces before moving through stomata or other openings such as wounds into the apoplast, where it can multiply to very high numbers before inciting disease symptoms. Strain B728a is pathogenic to common bean (Phaseolus vulgaris), lima bean (P. lunatus), pepper (Capsicum annuum), Nicotiana benthamiana, and I show that it is rather promiscuous, capable of growth within other plant species. Strains of P. syringae are commonly found as epiphytes on a variety of both host- and non-host plants in agricultural systems, native plants, and in various aquatic environments. P. syringae has thus presumably evolved or acquired a diversity of traits needed for success in these varied habitats, but we lack an understanding of the identity and role of these genes and the degree to which they are linked to fitness in a given context.

To examine the context-dependent fitness contributions of genes in P. syringae for growth on or in leaves, I created a barcoded transposon insertional mutant library in strain B728a. Randomly-barcoded TnSeq (RB-TnSeq) enables the same heterogeneous mutant population to be used to interrogate gene contribution to bacterial fitness in a variety of distinct hosts or environments. I determined the contributions of 4,296 nonessential genes to fitness on the leaf surface and in the apoplast of common bean. Genes within the functional categories of amino acid and polysaccharide biosynthesis contributed most to fitness both on the leaf surface and in the apoplast, while genes involved in type III secretion and syringomycin synthesis were important primarily in the apoplast. Interestingly, for most genes no relationship was seen between fitness in planta and either the magnitude of their expression or degree of induction in planta compared in this host plant to in vitro conditions measured in other studies. Given the robust bacterial growth of P. syringae also in lima bean and pepper, I utilized the barcoded transposon library to identify genes that differentially contributed to apoplastic colonization in those hosts. Most genes contributed to apoplastic fitness at similar magnitudes in each of the plant hosts. However, 87 genes significantly differed in their fitness contribution in at least one host, and 50 of these genes were highly important for apoplastic colonization. Genes in the functional categories of amino acid metabolism, phytotoxin biosynthesis, alginate biosynthesis, and cofactor metabolism significantly differed in contributions to fitness in the plant species tested. In addition, six genes that encoded hypothetical proteins contributed differentially to growth in these hosts.

Because P. syringae colonizes such a large number of plant species and various habitats it presumably is exposed to a wide variety of inhibitory compounds that it must tolerate. To test the hypothesis that plants would differ in the intensity and nature of chemical defenses against bacterial colonists I used a mexB mutant of P. syringae strain B728a that was blocked in expression of an important efflux pump as an indicator of the chemical defenses of various plants. While MexB was required for full virulence in common bean and pepper, it was dispensable for apoplastic growth in lima bean, Nicotiana benthamiana, sunflower (Helianthus), and tomato (Solanum lycopersicum). Within these plants, B728a grew to varying population sizes, indicating differential susceptibilities among these compatible host plant species. P. syringae strain B728a encodes 668 predicted transport proteins. However, the substrates of a majority of these transporters are unknown. Many of these transporters confer resistance to diverse toxicants present in the environment, but due to a high level of functional redundancy, it is difficult to identify those that are of most importance in conferring resistance to specific compounds. To identify transporters in B728a that are important for tolerance of external toxicants but redundant with MexB, I exposed the WT barcoded transposon mutant library as well as a library constructed in the mexB deletion strain to diverse toxic compounds in vitro to identify mutants with reduced ability to grow in the presence of those compounds. The homologous transporters MexCD and MexEF were partially redundant with MexAB, sharing a subset of substrates tested. I hypothesized that inner membrane transporters contribute to substrate specificity, particularly when functioning cooperatively with outer membrane transporters or porins. In support of this model, the inner membrane transporters Psyr_0228 and Psyr_0541 were highly substrate specific, and contributed to growth independent of mexB genotype. This study of isogenic transposon libraries revealed the substrates of efflux transporters that are masked by the dominant activity of MexAB.

These studies provide a more comprehensive insight into the chemical ecology of P. syringae. It is clear that plant species differ in their chemical defenses, as seen in the differential fitness of the hyper-susceptible mexB efflux pump mutant in various host plants. The fitness of various auxotrophs within a transposon library in planta also indicates resources that are often in low amounts in the host, such as amino acids and cofactors. These studies identified genes contributing to competitive fitness in diverse environmental conditions, including tolerance of toxicants that might be encountered by P. syringae. The resulting large datasets represent a valuable tool for the identification of genes required for compatibility with a given plant host. Improving our understanding of genes that contribute to bacterial fitness in diverse environments, including the phyllosphere, will improve our understanding of those habitats and the bacterial traits needed to survive and thrive.

Pseudomonas syringae is an important member of the phyllosphere microbial community and has been studied for many decades both as a model saprophyte, residing epiphytically, and a pathogen, living in the plant apoplast. While much is known about the traits that contribute to P. syringae's success as a phyllosphere microbe, one area that is not well understood, is the contribution of temperature dependent gene regulation, or thermo-regulation, to P. syringae's epiphytic colonization and survival strategies.

Flagellar-mediated motility is a trait conserved among plant-associated Pseudomonads. In P. syringae, motility contributes both to epiphytic colonization and survival, as well as pathogenicity by allowing cells to invade into the leaf interior. We have found that multiple forms of flagellar-mediated motility, including swarming and swimming motility, are thermo-repressed at around 30 °C. Repression of swarming results from reduced expression of both flagellar genes, including flagellin, encoded by fliC, as well genes involved in regulation and biosynthesis of syringafactin, the major surfactant produced by P. syringae B728a. Thermo-regulation of the flagellum is context dependent, being influenced by the nutrient status of the agar, which together with temperature contribute to a heterogeneous swimming phenotype at elevated temperatures. The heterogeneous swimming phenotype may represent a so-called "bet-hedging" strategy, which may be an important strategy for colonization of the leaf surface under varying and unpredictable weather conditions.

Investigation into the molecular determinants of thermo-regulation revealed several regulators, which are all conserved across the Pseudomonas genus, potentially indicating these are conserved thermo-regulators with in this important group of organisms. flgM as well as fleN were both involved in thermo-regulation of the flagellum. Additionally, fleN appeared to be important for the heterogeneous swimming phenotype, as mutations in Spontaneous Hot Swimming (SHS) mutants, which swim at elevated temperatures with a cool, diffuse swimming phenotype, commonly mapped to this gene, as determined by genome resequencing. Two separate loci, an acyl-CoA dehydrogenase and a nudix hydrolase, together contributed to thermo-regulation of genes involved in the regulation and synthesis of syringafactin. Interestingly, mutations that contributed to thermo-regulation of the flagellum did not affect thermo-regulation of syringafactin, while mutations affecting thermo-regulation of syringafactin did not affect thermo-regulation of the flagellum, indicating that thermo-regulation of these two traits are independently controlled.

As our understanding of host-associated microbial communities (microbiomes) deepens, there is a simultaneous revelation of key gaps in our understanding of these systems. Among these is knowledge of the forces underlying the assembly of, selection within, and dynamics among microbiota. These questions relate to broad principles that are shared across host species, and synthesis across these systems will identify conserved principles in the larger field of microbiome research. The work presented here seeks to identify and explore the relative importance of multiple forces simultaneously shaping the microbiome, specifically, that of the phyllosphere (above-ground surfaces of plants).

This research begins by investigating the importance of vertically transmitted (parent to offspring) microbes in seedling health. Currently, there is little understanding of the ecological importance of commensal or mutualistic bacteria that are transmitted on or within the seeds. Using a tomato study system and combination of classic microbiological techniques and next generation sequencing, we found that the vertically transmitted seed microbiome is capable of protecting seedlings against a common plant pathogen. This work provides evidence that the seed microbiome plays an important ecological role in early seedling life, and very likely shapes the development of the microbiome both directly through priority effects and indirectly through interactions with the plant host.

This work then explores the importance of host genotype and environmental selection in shaping the phyllosphere microbiome in tomato plants. A successive passaging experiment was used to address this question by selecting upon the phyllosphere microbiome. Beginning with a diverse microbial community generated from field-grown tomato plants, replicate plants across five plant genotypes were inoculated for four eight-week long passages, and the microbial community was sequenced at each passage. We observed consistent shifts in both the bacterial (16S amplicon sequencing) and fungal (ITS amplicon sequencing) communities across replicate lines over time, as well as a general loss of diversity over the course of the experiment suggesting that much of the naturally observed microbial community in the phyllosphere in an open environment in outdoor setting is likely transient or poorly adapted. We found that both host genotype and environment shape microbial composition, but the relative importance of genotype declines through time. Furthermore, using a community coalescence experiment, we found that the bacterial community from the end of the experiment was robust to invasion by the starting community. These results highlight that selecting for a stable microbiome that is well adapted to a particular host environment is indeed possible, emphasizing the great potential of this approach in agriculture and other systems.

In the final chapters of this work, the importance of bacteriophages (viruses that infect bacteria) in the phyllosphere is examined. This question was addressed by transferring microbial communities from field-grown tomato plants to juvenile plants grown under mostly sterile conditions in either the presence or absence of their associated bacteriophage (phage) community. In three separate experiments, we found that the presence of phages affects overall bacterial abundance during colonization of new host plants. Furthermore, bacterial community analysis (16S amplicon sequencing) shows that phages significantly alter the relative abundance of dominant community members and can influence both within- and among-host diversity. This is then extended through the measuring of the impact of phages on bacterial communities weeks after initial inoculation. This work describes the impact of disrupting bacteria-phage temporal dynamics on bacterial communities. Together, these results underscore the importance of both lytic and lysogenic phages in host-associated microbiomes but how they can have fluctuating impacts over relatively short timescales.

Together, these results contribute to fundamental knowledge gaps by demonstrating the ecological importance of vertically transmitted microbes, determining that microbiomes can be adapted to their host and environment, and uncovering the temporal variability of key driving forces underlying microbiome structure.

Transcriptional profiles of P. syringae pv. syringae B728a (Pss) were obtained during both epiphytic and endophytic phases of colonization of its host plant Phaseolus vulgaris as well as the conditions of water stress, oxidative stress, iron limitation, and nitrogen limitation in culture, mimicking the challenges faced in an on plants. The transcript profiles of Pss B728a support a model in which leaf surface sites specifically favor flagellar motility, swarming motility requiring HAA surfactant production, and chemosensing and chemotaxis, indicating that cells actively relocate on leaves after immigration. Phenylalanine degradation is also induced in cells on leaves, which may help them counteract phenylpropanoid-based defenses prior to leaf entry. In contrast, intercellular sites favor the high-level expression of genes for gamma-amino butyric acid (GABA) metabolism, the degradation of which would attenuate GABA-dependent host defenses. In addition, the synthesis of phytotoxins, two novel secondary metabolites, and syringolin A is induced in Pss cells in intercellular spaces, apparently contributing to its virulence. Syringolin A may also suppress host plant defenses as well as stomatal closure. The prominent catabolism of various non-sugar compounds, and the elaborations of secondary metabolites by Pss underscores the complexity of the chemical interface between it and host plants. A comparison of the transcriptome of Pss recovered from plants to that of cells exposed to osmotic stress, oxidative stress, and iron and nitrogen limitation in culture revealed that gene expression of cells in and on leaves was most similar to that of cells experiencing water stress in culture. Surprisingly, water was apparently more limiting in the apoplast than on the leaf surface. Global gene expression of mutants of Pss in which the global regulators gacS, retS, salA, rpoN, rpoE, rpoS, hrpL, ahlR and aefR were disrupted was also assessed in these same seven conditions to determine the roles of these regulators in mediating transcriptional responses to such stimuli. Complex patterns of gene expression were observed that revealed that the regulons differed greatly in size, and had variable overlap between other regulons and with the various stimulons. Genes enabling responses to water stress such as those for compatible solute synthesis and other osmoprotection mechanisms, which are strongly expressed both on the surface of leaves and within leaves, are rpoE -dependent. This reveals that rpoE is a key regulator mediating water stress responses and fitness on plants.

Because of our particular interest in quorum sensing, we focused our analyses upon the ahlR and the aefR regulons. AefR was found to be a repressor of a variety of efflux pumps. The most prominent feature of the aefR regulon is that it oppositely regulates two sets of genes; while the ahlR regulon is positively regulated by aefR , genes encoding orthologs of the MexEF-OprN Resistance Nodulation Division multidrug efflux pump are repressed by aefR . This juxtaposition is especially striking in light of finding that mexS is a positive regulator of ahlI and thus quorum sensing, as mexS mutants induce expression of mexEF-oprN via MexT in P. aeruginosa . MexT/MexS is responsive to disulphide stress, a condition distinct from that conferred by reactive oxygen intermediates. Another regulator (Psyr_1729), likely a repressor of the adjacent genes encoding efflux pumps, positively regulates the expression of ahlI . The probable role of such efflux pumps is to export toxic, electrophilic oxidized products from cells. Such a scenario leads to the model that expression of multidrug efflux pumps, under the control of aefR , modulates quorum sensing indirectly through a change in cellular redox status. aefR is unique from other Pss regulators studied, in that its regulon overlaps little with that of global regulators such as gacS, salA , or rpoN . A distinct stimulus for the aefR regulon remains elusive, while we expect such a stimulus would be present during host interactions since an aefR mutant was impaired in apoplastic growth in plants.

The ahlR (Psyr_1622) regulon was small, with only eight genes that are positively regulated, and all were in two adjacent, divergent operons. One operon is initiated by the ahlI promoter, to form a transcript sense to ahlI that extends into genes encoding a truncated pyruvated dehydrogenase E1 subunit, pdhT (Psyr_1623) as well as one encoding a full length pyruvate dehydrogenase E1 subunit, pdhQ (Psyr_1625), thus producing a transcript of over 5 kb that is antisense to both ahlR and qrpR . Pyruvate accumulated in a pdhQ mutant and pyoverdine-dependent fluorescence was altered or quenched. The other ahlR -regulated operon originates from the same intergenic region as ahlI (Psyr_1621), and extends, on the opposite strand, through five genes (Psyr_1620-Psyr_1616). This second operon was termed the pao operon (paoABCDE ), as it was necessary for the accumulation of pyruvate that occurs when pdhQ is deleted. GC-MS and HPLC-MS profiling revealed that at least five unidentified compounds, as well as pyruvate accumulated in cells in the absence of pdhQ , but only when the pao operon was expressed. High accuracy mass spectrometry reveals that at least one of these compounds with the chemical formula C ₁₀ H ₈ O ₂ is probably an oxidized phenolic. Elevated levels of oxidized glutathione, methionine sulfoxide, and methanethiol in an ahlR ahlI double mutant in stationary phase cultures are consistent with the AhlR regulon contributing to cellular redox equilibrium.

Both ahlR and qrpR have their own active promoters and their expression is not responsive to AHL. AhlR can positively regulate its target promoter(s) even in the absence of its ligand 3oxoC6-AHL, but its activity is greatly enhanced upon ligand interaction. Given the position of AhlR in the phylogeny of LuxR homologs, at a clade branch equidistant between apo-repressors and ligand-dependent activators, it is noteworthy that it has ligand-independent, yet ligand-responsive activator activity. QrpR is a MarR-family repressor of the promoter of pdhQ . The ligand needed to de-repress QrpP is not known, but it is not 3oxoC6-AHL. In the absence of AhlR and AhlI, pdhQ expression is strongly de-repressed in the presence of a paoABCDE transcript. We posit that a compound originating from these enzymes is the ligand that elicits QrpR-dependent de-repression of pdhQ . Exogenous application of pyruvate did not de-repress pdhQ . Phylogenetic assessment indicates that QrpR orthologs form a clade distinct from other well-described MarR homologs.

The second ortholog of the pyruvade dehydrogenase complex E1 subunit in Pss B728a, Psyr_0518, exhibits constitutive expression in a variety of genetic backgrounds and growth conditions. The pH of the medium is greatly decreased in a Psyr_0518 mutant but this phenotype is reversed by expression of pdhQ , in trans. Mutants blocked in expression of ahlR -dependent genes did not exhibit altered growth on the surface of host plants, or in the interior of host or non-host plants, while mutants in the housekeeping pyruvate dehydrogenase E1 ortholog (Psyr_0518) had greatly impaired apoplastic growth which could be restored by overexpression of the quorum-controlled ortholog pdhQ .

These data taken together are consistent with a model whereby a redox stress, a potential inducing stimulus of the ahlR regulon, can be eliminated by expression of efflux pumps under the control of MexEF-OprN at low cell densities, whereas at high cell densities efflux is inadequate for this process since cellular activity needed for the process is low and the products exported by the pumps will accumulate in the vicinity of the cells. Under such circumstances, the AhlR regulon mediates metabolism of these toxic compounds. The presumed toxic effects of certain breakdown products on the pyruvate dehydrogenase E1 subunit are circumvented by expression of a toxin-resistant quorum-controlled enzyme PdhQ. The regulation of this metabolism by AHL signaling, QrpR, and transcriptional read-through events is exquisitely complex, as may be required to ensure cell density- or environmental porosity-dependent expression of this metabolism in anticipation of cellular stress, while avoiding endogenously generated cytotoxicity when more effective strategies are appropriate.

We also dissected the effects of iron on ahlI expression using both promoter-probe fusions and transcript abundance, to ascertain independently, the effects of iron on AHL production and AHL perception, and to resolve the effects of iron on these processes over time. Initial iron limitation does not alter ahlI expression, but another event linked with iron starvation that occurs after the central iron homeostasis limitation response is initiated, influences ahlI expression. Site-directed mutagenesis of loci encoding central iron homeostasis regulators Fur, PrrF1, and PrrF2, reveals that they are not required for the dependence of ahlI on iron availability.

Leaf surfaces are a harsh habitat for microbes since they are subjected to high fluxes of ultraviolet radiation, fluctuations in temperature, low nutrient availability, and frequent drying. Despite this, leaves can be heavily colonized by bacteria. I explored the ability of bacteria to modify their local environment to better tolerate or avoid the water stress they would experience on leaves by addressing the role of the biosurfactant syringafactin produced by the model epiphytic bacterium Pseudomonas syringae B728a. Syringafactin is very hygroscopic, binding nearly three times its weight in water, but only in atmospheres of high relative humidity. The water stress status of bacteria differing in syringafactin production was assessed in different settings by measuring the fluorescence of cells harboring a plasmid containing the promoter of proU, encoding the synthesis of the compatible solute proline, fused to a promoterless gfp reporter gene. The wild-type syringafactin-producing strain experienced less water stress than the mutant strain on membrane filters exposed to high relative humidity, but not on filters exposed to low relative humidity. Syringafactin production was also associated with lower water stress in bacteria on leaves, irrespective of the dryness of the air in which the plants were incubated. The high relative humidity-dependent water-binding capabilities of syringafactin are consistent with a model in which the laminar boundary layer surrounding leaves is nearly saturated with water vapor, irrespective of the dryness of air away from the leaf. Thus, by producing syringafactin, P. syringae hydrates its local microenvironment by sequestering liquid water, on an otherwise dry leaf, from the humid air trapped at the leaf surface.

The production of syringafactin requires syfA, encoding a synthase necessary for syringafactin production. Therefore, as a readout for syringafactin production, I used a P. syringae strain harboring a plasmid containing the syfA promoter fused to a promoterless gfp reporter gene. GFP fluorescence exhibited by P. syringae, was much higher in cells cultured on nutrient agar surfaces compared to in cells in the corresponding broth culture. These observations suggested that expression of this trait was dependent on contact of cells with a solid surface. Immobilization of this strain on various surfaces, including various membranes, plastic surfaces, and on leaves, resulted in rapid increases in apparent expression of syfA within two hours compared to that of cells that were maintained in a planktonic state in broth cultures. These findings indicated that immobilization of cells was sufficient to induce the production of this biosurfactant. These findings also suggested that P. syringae might either experience different environmental conditions upon immobilization or might use contact with the surface as a cue to express certain traits in anticipation of their necessity for optimum fitness upon change from a liquid habitat, in which cells might exist in a planktonic state, to a dry habitat, such as the leaf surface. To explore what additional traits besides syringafactin production might be expressed in a contact-dependent fashion in P. syringae, I compared the global transcriptome of planktonic cells grown in a broth medium with those immobilized for two hours on a filter surface placed on a similar medium. A large fraction of the genes exhibited either higher (26.63%) or lower expression (33.83%) upon immobilization on the filter surface. Genes in the functional categories of translation, siderophore synthesis and transport, nucleotide metabolism and transport, flagellar synthesis and motility, lipopolysaccharide synthesis and transport, energy generation, transcription, chemosensing and chemotaxis, replication and DNA repair, iron-sulfur proteins, peptidoglycan/cell wall polymers, terpenoid backbone synthesis, iron metabolism and transport, and cell division were significantly more likely to be up-regulated upon immobilization while those in the categories of quaternary ammonium compound metabolism and transport, compatible solute synthesis, carbohydrate metabolism and transport, organic acid metabolism and transport, phytotoxin synthesis and transport, amino acid metabolism and transport, and secondary metabolism were repressed on surfaces compared to in liquid.

While surface contact is increasingly being recognized as an important cue for the differential gene expression in various microorganisms, little is known of the mechanisms by which cells perceive that they are immobilized on such a surface. The mechanisms by which P. syringae B728a mediates surface contact-dependent production of syringafactin, and other-contact dependent processes are largely unknown. I identified genes that alter expression of syfA on agar surfaces using random transposon mutagenesis of P. syringae, harboring a syfA:gfp fusion. Putative regulatory genes contributing to activation of syfA expression identified by this method included efeO, tsr, and mexB while those contributing to its repression included psrA, cyoC, and algW. Some but not all of these putative regulatory genes also exhibited differential contact- dependent expression in immobilized cells compared to in planktonic cells. The KEGG database was used to analyze the metabolic pathways in which these genes might participate. Furthermore, I formulated possible pathways that could account for the surface-dependent regulation of syfA and potentially the regulation of other surface-regulated genes. One of these pathways is a stress sensing pathway that AlgW is part of which may play a role in activating genes involved in adhesion. Another pathway includes Tsr which may regulate genes involved in swarming, a type of bacterial motility that can only occur on a surface. These potential pathways reveal that a variety of mechanisms may be involved in regulating surface-induced genes in bacteria.

The phyllosphere epiphyte Pantoea agglomerans 299R synthesizes indole-3-acetic acid (IAA), an important plant hormone. IAA production was previously shown to confer a small but significant fitness advantage to Pa299R cells inoculated onto bean (Phaseolus vulgaris) leaves, but the mechanism by which bacterial IAA exerts this effect is unknown. In this work, we investigated several hypotheses regarding how bacterial IAA enhances the growth and survival of leaf epiphytic microbes such as Pa299R.

We first tested the hypothesis that bacterial IAA enhances the availability of plant sugars to phyllosphere bacteria, thereby relieving carbon limitation of bacterial growth on the leaf surface (Chapter Two). We inoculated sucrose- and fructose-inducible biosensor bacteria onto bean leaves, and investigated the effects of auxin availability on sugar sensing. We manipulated auxin availability in two ways: first, we compared sugar sensing by wild-type Pa299R (IAA+) biosensors to that of the isogenic, IAA- mutant PaMX149; and second, we compared sugar sensing by PaMX149 bacteria inoculated with exogenous NAA to bacteria inoculated without NAA. The presence of auxin--whether endogenously synthesized or exogenously applied--was associated with a significant decrease in sucrose sensing, and a small but significant increase in fructose sensing. Additionally, in vitro assays demonstrated that Pa299R initiates growth more rapidly when using glucose or fructose as a sole carbon source than when using sucrose. Together, this suggests that IAA biosynthesis may function as a resource conversion strategy to enhance carbon acquisition and rapid bacterial growth in an oligotrophic, environmentally dynamic phyllosphere.

In addition to its role in carbon acquisition, we also investigated several alternative hypotheses of bacterial IAA production in the phyllosphere (Chapter Three). We demonstrated that bacterial IAA exerts no measurable impact on resistance of Pa299R to environmental stress. We also found that bacterial IAA does not affect the autofluorescence of substrate leaf epidermal cells that support large bacterial aggregates of Pa299R. Finally, we describe unsuccessful efforts to determine the role of bacterial IAA in another foliar epiphyte and compatible bean pathogen, Pseudomonas syringae pv. syringae B728a.