![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Microbiology and Molecular Biology Reviews, March 2005, p. 155-194, Vol. 69, No. 1
1092-2172/05/$08.00+0 doi:10.1128/MMBR.69.1.155-194.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Microbiology, Cornell University, Ithaca, New York 148531
SUMMARY INTRODUCTION CHARACTERISTICS OF ENVIRONMENTS INHABITED BY PLANT-ASSOCIATED BACTERIA HOST DETECTION BY MEMBERS OF THE RHIZOBIACEAE DURING COMMENSAL PLANT COLONIZATION Chemotaxis toward Plant Root Exudates Binding to Host Surfaces Alteration of Gene Expression by Plant-Released Sugars Detection of Acidity HOST DETECTION DURING RHIZOBIUM-PLANT INTERACTIONS Host Detection during Nodule Formation Structure, function, and ecology of flavonoids. Nonflavonoid inducers of nod genes. Transcriptional regulators of nod genes. Metabolism of nod-inducing signal molecules. Production of Exopolysaccharides during Nodule Invasion Regulation of Symbiotic Nitrogen Fixation Regulation of nif and fix genes by intracellular oxygen tensions. Oxygen-controlled regulatory systems. (i) FixL-FixJ. (ii) FixK. (iii) NifA. Detection and metabolism of dicarboxylic acids Opine-Like Molecules in Sinorhizobium-Plant Interactions HOST DETECTION DURING AGROBACTERIUM-PLANT INTERACTIONS Host Detection and Expression of Genes Required for Infection vir gene inducers. VirA-VirG-ChvE regulators of vir genes. Metabolism of phenolic compounds. Other regulators of vir genes. Detection of Opines Released by Crown Gall Tumors PRODUCTION OF PECTINOLYTIC ENZYMES BY SOFT ROT ERWINIAS Detection of Pectin and Pectin Catabolic Products KdgR (RexZ). PecS and PecM. PecT. Catabolite repression and CRP. Regulation of the secondary or minor pel genes. Induction of Pectinolytic Enzymes by Plant Extracts Induction of Pectinolytic Enzymes by Iron Limitation Effect of pH on the Expression of Pectinolytic Enzymes Other Environmental Signals That Affect the Expression of Pectinolytic Enzymes REGULATION OF TYPE III SECRETION SYSTEMS AND ASSOCIATED EFFECTORS IN VARIOUS PLANT-PATHOGENIC BACTERIA Induction of Expression of Type III Secretion Systems by Host-Released Chemical Cues Regulatory Cascades Controlling the Expression of Type III Secretion System Genes Pseudomonas syringae. Erwinia amylovora and Pantoea spp. Xanthomonas campestris and Ralstonia solanacearum. Secretion of Effectors by the Type III Secretion System PRODUCTION OF PHYTOTOXINS BY PSEUDOMONAS SYRINGAE ANTIBIOTIC PRODUCTION BY ROOT-COLONIZING PSEUDOMONAS FLUORESCENS GLOBAL REGULATION OF PLANT-ASSOCIATED PHENOTYPES Role of Quorum Sensing in Regulation of Plant-Associated Phenotypes Role of GacS-GacA in Regulation of Plant-Associated Phenotypes CONCLUSIONS ACKNOWLEDGMENTS REFERENCES
| SUMMARY |
|---|
 Top NextReferences |
|---|
| INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
An interesting example is found with honeybees, which require pollen and flower nectar as their sole sources of protein, carbon, and energy. Bees are also the most important pollinating insects, and the interdependence between bees and plants makes them an excellent example of an animal-plant symbiosis. Behavioral ecologists have been interested in how various floral characteristics, such as color, shape, and size, enable a bee to choose appropriate flowers that will provide them with sufficient nutrients. These characteristics are detected by color receptors of the animal, and research is being conducted to understand how differences in these characteristics are computed on a neuronal level (54, 457).
Other interesting and well-studied examples of plant detection are found with plant-associated microorganisms. Many types of microbes live in close association with host plants and benefit from these associations by obtaining carbon and other nutrients from their hosts. Events that lead to establishment of these interactions are triggered by bacterial recognition of specific plant-associated signal molecules, which are detected by dedicated bacterial sensory proteins. Similar to what has been observed for honeybees, this recognition may play an important role in the host specificity or the host range of a bacterium. This is the case in the recognition of plant-released flavonoids by rhizobial NodD proteins and also in the recognition of translocated bacterial avirulence proteins by host-encoded resistance proteins (see below).
Over the course of a plant-microbe interaction, bacteria continue to monitor changes in the physiology of their host. These changes are often due to specific activities of the colonizing microbes, which in response continuously make adjustments to their own physiology. Thus, detection and response to various host signals in the plant-microbe interaction is a continuous process. In many cases, in addition to specific regulatory proteins, global regulators play a role in these interactions.
In this review, we describe the process of plant detection as it is known to occur in the best studied plant-microbe systems. These include the symbiosis between rhizobia and legumes and the pathogenesis between Agrobacterium tumefaciens and host plants that leads to crown gall tumors. We then turn to other plant pathogens, such as soft rot erwinias, Pseudomonas syringae, and biocontrol strains of Pseudomonas fluorescens. We describe plant-associated signals, bacterial proteins involved in their detection, and mechanisms and pathways of signal transduction leading to expression of specific bacterial genes that direct these interactions. We also discuss the role of various global regulators involved in the regulation of these processes.
| CHARACTERISTICS OF ENVIRONMENTS INHABITED BY PLANT-ASSOCIATED BACTERIA |
|---|
|
TopPreviousNextReferences |
|---|
The interior portions of leaves provide a somewhat more hospitable environment. Importantly, humidity is more carefully controlled within leaf tissues, due to a waxy cuticle on the plant surface that minimizes water loss. Leaf epidermis contains small openings called stomata, which allow the exchange of carbon dioxide and oxygen. Stomata are the main route by which water is lost from the plant; however, they can be closed during periods of dryness to conserve water. When stomata are open, they provide ready access to the intercellular spaces within leaves (the apoplast) and serve as an important entry point for many bacteria (27, 28).
Roots and the zone surrounding them (the rhizosphere) are also readily colonized by microbes (44, 48). The rhizosphere generally provides more protection than the phyllosphere from desiccation, temperature, and light stress. Furthermore, sources of carbon and minerals are more abundant in the rhizosphere (44, 99, 510). Plants exude high levels of nutrients from their roots, often in excess of 20% of all fixed carbon (309). Amino acids, organic acids, sugars, aromatics, and various other secondary metabolites comprise the majority of the low-molecular-weight root exudates, whereas high-molecular-weight exudates primarily include polysaccharides and proteins (309). This complex mixture of organic compounds results in much larger numbers of microbes in the rhizosphere than in the nearby bulk soil, where the microbial community is carbon limited. This phenomenon is referred to as the "rhizosphere effect" (44).
A large variety of bacteria, fungi, protozoa, and nematodes colonize the rhizosphere (44, 48). These organisms may exist as free-living organisms in the rhizosphere or may be attached to surfaces of roots. Colonization of root surfaces is characteristically nonuniform; some areas, including the extreme tip of the root, are relatively free of bacteria, whereas other areas can be heavily populated (151, 355). In studies with Pseudomonas spp., the heavily populated areas are usually found at junctions between epidermal root cells, indented parts of the epidermal surface, or sites of side root appearance, all of which are presumed sites of exudation (39, 303). Microbes can also gain access to the interior portions of roots through cracks in the epidermis made by the emergence of lateral roots or through wounds caused by various herbivores (18).
In recent years, there has been significant progress in our understanding of the environments that are experienced by the bacteria living in the rhizosphere or the phyllosphere. Much of this knowledge comes from studies that use bioreporter strains, in which an environmentally or metabolically responsive promoter is fused to a suitable reporter such as lacZ, gus, lux, inaZ, or gfp (286, 454). These studies allow monitoring of the spatial distribution and fluctuations of physicochemical factors that are relevant for the microbes inhabiting the plant-associated environments. The factors studied so far have included UV irradiation, temperature, water potential, and iron availability on surfaces of leaves (17, 251, 268, 499) and carbon, phosphate, nitrogen, iron, and oxygen availability in the soil (217, 241, 271, 275). Bioreporter strains have also been used to detect products of plant metabolism that are released into the surrounding environment and are used by the associated microbes as nutrients (53, 239, 285, 329). In general, these studies showed that there is substantial heterogeneity in the intensities of reporter gene expression in different microenvironments of a leaf or rhizosphere, suggesting that bacteria residing in different parts of these habitats may be exposed to remarkably different environments.
Studies performed with Pantoea agglomerans (previously known as Erwinia herbicola) harboring a sucrose- and fructose-responsive scrY promoter fused to a gfp or inaZ reporter revealed a high-level heterogeneity of apparent sucrose availability on surfaces of leaves (329). Workers performing a study in which the sucrose- and fructose-responsive fruB promoter was fused to a short-half-life variant of gfp came to similar conclusions by showing that, within 1 day after inoculation, only 1% of the bacteria expressed this fusion (285). The cells that continued to detect sugars were not randomly dispersed across the leaf surface but instead were localized to sites likely to release these nutrients, including stomata, trichomes, veins, and various crevices that are more likely to retain water (536).
Similar studies were performed to map the availability of sugars and amino acids along roots (239). A strain of P. agglomerans harboring an ice nucleation reporter gene, driven by either a sucrose- or tryptophan-responsive promoter, was used as a biosensor. When the strain was introduced into the rhizosphere of an annual grass, both tryptophan and sucrose were detected, but they showed different spatial patterns. Tryptophan was most abundant in soil around roots 12 to 16 cm from the tip, while sucrose was most abundant in soil near the root tip. High sucrose availability at the root tip is thought to be caused by its leakage from the immature, rapidly growing root tissues, while tryptophan loss from older root sections was proposed to result from lateral root perforation of the root epidermis (239). As might be expected, sites having the highest apparent sucrose or tryptophan exudation were the most heavily colonized parts of the root.
| HOST DETECTION BY MEMBERS OF THE RHIZOBIACEAE DURING COMMENSAL PLANT COLONIZATION |
|---|
|
TopPreviousNextReferences |
|---|
-proteobacteria include various nitrogen-fixing plant symbionts, Rhizobium, Sinorhizobium, Mesorhizobium, Azorhizobium, and Bradyrhizobium, collectively called rhizobia, and the plant pathogens Agrobacterium tumefaciens and A. rhizogenes, referred to as agrobacteria, which cause crown gall tumors and other neoplasias on a wide variety of plants. Species of rhizobia and agrobacteria are very closely related, and it was recently proposed that the Agrobacterium genus be abolished and that its members be referred to by the genus name Rhizobium (132, 528, 543). Major similarities between the two genera include metabolic, transport, and regulatory systems that may promote survival in the competitive rhizosphere, whereas the most striking differences lie in genes specifically required for interaction with a plant host (65, 158). These genes are carried on the tumor-inducing (Ti) plasmid in agrobacteria and on the symbiotic plasmids pSymA and pSymB in Sinorhizobium meliloti (158). In Bradyrhizobium japonicum and Mesorhizobium loti, the symbiosis-related genes are carried on a chromosomally located symbiotic island (258, 259). Although best known for forming nodules or crown gall tumors, rhizobia and agrobacteria also can colonize plants without causing either of these neoplasias and can utilize plant exudates to support growth and division. These associations are not dependent on the pathogenic or symbiotic determinants of the respective bacteria, and similar phenomena are therefore likely to occur in the various members of this family.
-proteobacterium. The protein was found in a search for A. brasilense plant-inducible genes and was designated SbpA (sugar binding protein A) (500). Enteric bacteria use similar sugar binding proteins for sugar chemotaxis, suggesting that sugar chemotaxis may occur by similar mechanisms (1, 204, 205). Rhizobia and agrobacteria also have several flagella located near one pole of the cell (201, 278), and mutations in genes encoding flagellin abolish motility and, in Agrobacterium, reduce tumorigenesis (35, 77, 201).
|
Binding of rhizobia and agrobacteria to plant surfaces is thought to take place in two steps (315). The first is a rather weak and reversible binding step that may involve a variety of bacterial polysaccharides. The products of the ndvA and ndvB genes in Sinorhizobium meliloti and of the homologous chvA and chvB genes in A. tumefaciens are involved in the synthesis of a cyclic glucan (64, 103, 236, 468), which could act as an adhesin via gelling interactions with host polysaccharides or could interact with plant lectins (215). Mutations in the chv genes reduced the binding of the agrobacteria to cultured cells and abolished tumorigenesis (117, 118, 395). However, these mutations are pleiotropic, and so it is difficult to know whether the cyclic glucan is a direct adhesin or whether its loss perturbs some other functions that are important in binding. Mutations in the ndv genes caused a moderate decrease in the binding of rhizobia to root hairs and had a strong defect in nodule invasion (123, 124). The invasion defect, however, was not due to the defect in adhesion, since revertants that were fully able to form nitrogen-fixing nodules remained defective in attachment (124).
A 30-kb cluster of A. tumefaciens att genes has also been described as being required for attachment and tumorigenesis (316, 317, 319). However, the recently published A. tumefaciens genome sequence revealed that the att genes are located on the cryptic plasmid pAtC58, which is not essential for virulence (233, 423). There have also been a few reports of a bacterial adhesin called rhicadhesin, although the gene encoding this adhesin has yet to be cloned or disrupted (475-477).
The second binding step requires the synthesis of bacterial cellulose, which causes a tight, irreversible binding and formation of bacterial aggregates on the host surface (314, 413). Mutants with mutations of the A. tumefaciens celABCDE operon no longer synthesize cellulose and can be readily dissociated from cultured plant cells by vortexing. However, these mutants are still tumorigenic (317, 318). To our knowledge, mutants with mutations in the orthologous genes of rhizobia have not been tested for binding.
Another A. tumefaciens gene was reported to be induced by plant sugar molecules. The picA (for Plant-Inducible Chromosomal) gene was first identified as being induced by the polygalacturonic acid fraction of carrot extracts (419-421). The gene resembles genes encoding a family of polygalacturonidases (also known as pectinases). It lies in a possible operon with a gene encoding a second pectinase and is located adjacent to an operon encoding an ABC-type uptake system for oligogalacturonic acids (176, 535).
In S. meliloti, the melA gene is required for utilization of -galactosides, and its transcription is induced by these substrates (53, 155). An S. meliloti strain expressing a melA-gfp reporter fusion was used for the detection of -galactosides around roots of several legumes and grasses. Bacteria expressed high levels of GFP in the areas around zones of lateral root initiation and around roots hairs but not around root tips. Other studies reported that vitamins, choline, stachydrine, trigonelline, and homoserine can also be secreted by plant roots and used by different species of rhizobia (42, 108, 385, 472).
In Rhizobium tropici, the atvA gene is transcriptionally upregulated by acid shock and is homologous to the acvB gene of A. tumefaciens (506). Both R. tropici atvA and A. tumefaciens acvB mutants are acid sensitive, indicating that the two genes are required for acid tolerance (506). The functions of these genes are unknown, although the AcvB protein (or its Ti-plasmid-encoded homolog, VirJ) is required for tumorigenesis (255, 260, 369, 533).
| HOST DETECTION DURING RHIZOBIUM-PLANT INTERACTIONS |
|---|
|
TopPreviousNextReferences |
|---|
Nod factors consist of a ß-1,4-linked N-acetyl-D-glucosamine backbone with four or five residues. Of these, the nonreducing-terminal residue is substituted at the C-2 position with an acyl chain, whose structure and saturation vary in different rhizobia. The oligosaccharide can also have acetyl, sulfonyl, or carbamoyl substitutions at defined positions (95, 119, 162, 248, 416). Nod factors are synthesized and exported from the bacteria by the products of nod genes (119). The nodABC genes are present in all rhizobia and are required for production of the basic Nod factor (162, 248, 416). Interspecific differences in Nod factors are in part due to allelic variations in nodABC genes. In addition, each rhizobium species possesses species-specific nod genes, which direct species-specific modifications of the basic Nod factor (95, 119). Each Rhizobium species therefore produces a different set of Nod factors, which play a critical role in host specificity (95, 119).
|
In general, LysR-type regulators contain a highly conserved N-terminal DNA binding domain and a less highly conserved C-terminal ligand binding domain (441). The structure of the full-length LysR-type regulator CbnR has been solved by X-ray crystallography (Fig. 2) (342), as have the structures of the C-terminal domains of CysB, DntR, and OxyR (79, 453, 496). The N-terminal domain of CbnR consists of three helices and is followed by two ß-strands, which are connected to the C-terminal domain by a long helix that mediates protein dimerization (Fig. 2). Helix 3 in the N-terminal domain is thought to lie perpendicular to the major groove of the nod box and to mediate sequence-specific DNA binding. The C-terminal domains of all four crystallized proteins are composed of two subdomains that close together on ligand binding and separate on release of the ligand. This domain has a strong structural resemblance to the family of periplasmic binding proteins that are components of ABC-type permeases. It is highly likely that NodD is structurally similar to the structures of these proteins (Fig. 2). Genetic studies have shown that the C-terminal domain of NodD is involved in the binding of flavonoids (460).
|
Many species of rhizobia possess more than one copy of the nodD gene, and the properties of different nodD genes vary within the same strain as well as from one Rhizobium species to another. Some strains possess two to five copies of nodD (133, 179, 444, 504) and may in addition possess one or two copies of another LysR-type regulator gene called syrM (symbiotic regulator) (193, 327, 328, 340, 474). SyrM is a NodD homolog and also acts as an activator of nod genes. Different NodD proteins may differ in their affinity for various nod boxes and may also have different flavonoid specificities. The nodD genotype therefore in part determines the host range of a given Rhizobium strain. Transfer of a nodD gene from one Rhizobium species to another can in some cases alter the host range of the recipient to that of the nodD donor strain (218, 461), whereas point mutations in nodD affect the recognition of inducing molecules and cause extension of the host range (60, 321). In S. meliloti, NodD3 and SyrM do not require flavonoids for nod gene activation and therefore act in a signal-independent fashion (474). B. japonicum possesses a two-component system, NodV-NodW, that is responsive to plant-produced isoflavone signals and functions as a positive regulator of nod genes (178, 437).
In addition to activating nod promoters, certain NodD proteins repress the expression of their own promoters. In several rhizobia, nodD genes are transcribed divergently from nearby nod operons, and by binding to the operator sequence between the two operons, these NodD proteins act as repressors of their own expression (444, 460). This autorepression occurs in the presence or absence of flavonoid signals.
Several nod regulons, in addition to being positively regulated by NodD proteins, are subject to negative regulation by NoIR, a 13-kDa protein that contains a DNA binding motif resembling those of other regulators of the LysR family (92, 273). NoIR binds as a dimer to conserved sequences found in the promoter regions of target nod genes and prevents their expression. Expression of noIR is negatively regulated by its own product and by the nod gene inducer luteolin (92).
Several environmental factors, such as calcium, ammonium, organic acids, and pH, also contribute to nod gene expression by unknown mechanisms (408, 409, 444). Low pH has a negative effect on the induction of nod genes as well as on rhizobial growth. In addition, plants produce smaller amounts of flavonoids in acidic soil, and flavonoids appear to accumulate in the bacterium in a pH-dependent manner (409, 410). Similarly, elevated levels of combined nitrogen have a negative effect on the production of aromatic compounds in the root, while in S. meliloti high levels of ammonia sensed by the Ntr system inhibit the induction of nod genes (121, 122). Expression of nod genes is also inhibited in the presence of dicarboxylic acids (see below) (544).
EPS production depends on the concentration of available phosphate, which might be sensed by the bacteria during the process of nodulation (324, 546). Phosphate concentration is very low in the soil (typically 1 to 10 µM) and considerably higher within plant tissues (10 to 20 mM). EPS II is produced preferentially under low-phosphate conditions, whereas succinoglycan synthesis is stimulated at high concentrations of phosphate (324). This suggests that inside the plant, bacteria produce succinoglycan, which is consistent with the observation that although both EPS can mediate nodule invasion, succinoglycan is much more efficient in this process (375). The mechanism by which phosphate concentration controls the production of EPS is unknown. Several S. meliloti regulatory proteins have been identified that are involved in the control of EPS synthesis, but most have not been matched with any signal. The ChvI-ExoS two-component system is involved in the control of both succinoglycan production and flagellum biosynthesis (76, 541). The system is homologous to the A. tumefaciens ChvI-ChvG system, which senses environmental acidity (Table 1) (287). It remains to be determined whether the ChvI-ExoS system is involved in the regulation of any acid-inducible genes in S. meliloti and whether pH plays a role in infection thread extension and nodule invasion.
|
FixL is directly responsible for detecting intracellular oxygen tensions. It is composed of a membrane-anchoring domain; as well as sensor and kinase domains, both of which are located in the cytoplasm (Fig. 3A) (298). The protein has only a few periplasmic residues (Fig. 3A), and a FixL fragment lacking these and the membrane-spanning residues is fully functional (165, 298, 299). FixL of B. japonicum is composed only of sensor and kinase domains and is thus fully cytoplasmic (168). In general, the most common physiological strategy for detecting gases occurs via heme-based sensors (415). Different structures are possible for the heme binding domains in these sensors, and of these, PAS domains are most commonly encountered. The sensor domain of FixL is a prototypical heme-binding PAS domain and is responsible for sensing oxygen. It is approximately 130 residues long and has a predicted /ß fold (Fig. 3B) (167). The heme-bound PAS domains can accomplish a ligand-dependent switching of a neighboring transmitter domain, which in the case of FixL is the histidine kinase domain.
The heme binding PAS domain of B. japonicum FixL has been crystallized in the presence and absence of oxygen (Fig. 3B) (175, 194). Oxygen binding is thought to cause the movement of a loop away from the heme center, accompanied by a switch in the H bonding of the heme with protein residues (167, 174, 175). It is not yet known whether the movement of the heme causes the conformational change or vice versa, and because of a lack of the three-dimensional structure of the full-length protein, the entire regulatory mechanism remains unknown.
The FixL-FixJ system is one of the few two-component systems whose signal-responsive autophosphorylation and phosphotransfer have been reconstituted in vitro (165, 299). Studies of this system showed that anoxic conditions enhance FixL autophosphorylation whereas phosphorylation of FixJ is independent of oxygen status. More recently, the rate and oxygen sensitivity of FixL autophosphorylation were reported to be greatly enhanced by the presence of the response regulator FixJ, and a model was proposed in which FixL forms a sensing complex with FixJ and ATP. Detection of oxygen and the consequent phosphorylation reactions occur within this complex, after which the phospho-FixJ and ADP are released (492, 493). FixL also possesses a phosphatase activity which is repressed under anoxic conditions (299). Therefore, the antagonistic effect of oxygen on kinase and phosphatase activity of FixL regulates transcriptional activity of FixJ.
FixJ is composed of an N-terminal receiver domain and a C-terminal DNA binding domain. A truncated FixJ containing just the C-terminal domain exhibits high-affinity binding to the nifA promoter, whereas unphosphorylated full-length FixJ is inactive (2, 159). Therefore, it seems that under high oxygen concentrations, the unphosphorylated receiver domain inhibits the C-terminal domain, which prevents the protein from binding to DNA. Phosphorylation of the FixJ receiver relieves this inhibition, resulting in the activation of the inherent DNA binding and the activation capacity of the C-terminal domain. At the fixK promoter, however, the N-terminal domain of FixJ was shown to contribute positively to transcriptional activation. The domain was required for the recruitment of RNA polymerase to the fixK promoter by phosphorylated FixJ (486). Apparently, the mechanism of action of FixJ can vary from promoter to promoter.
Dicarboxylic acids are imported into bacteroids through two transporters. One is of plant origin and is located in the so-called peribacteroid membrane, which encloses the bacteroids within the nodule (102, 497). The other is encoded by the dctA gene of rhizobia and is embedded in the inner membrane of the bacteroid (250, 422). In free-living rhizobia, transcription of dctA is activated by dicarboxylic acids via the DctB-DctD two-component system (164, 517). DctB was proposed to sense the presence of dicarboxylic acids in the periplasm and to phosphorylate DctD in a signal-dependent manner. DctD is an NtrC-like transcriptional regulator of the dctA promoter (164). The DctB-DctD system is, however, not required under symbiotic conditions, which suggests the existence of an alternative regulatory system operating in the bacteroids (257).
|
| HOST DETECTION DURING AGROBACTERIUM-PLANT INTERACTIONS |
|---|
|
TopPreviousNextReferences |
|---|
Proteins responsible for T-DNA processing and transfer are encoded in the vir region of the Ti plasmid (104, 249, 425, 549). Twenty-one genes in this region are essential for wild-type levels of pathogenesis and are expressed in six operons, virA, virB, virC, virD, virE, and virG. The proteins required for cleavage of the T-DNA borders are encoded by virD1 and virD2 (540). VirC1 and VirC2 bind to a site adjacent to T-DNA borders, called overdrive, and are required for efficient T-strand processing (487). The virB operon encodes the T-DNA transfer apparatus, which delivers the T-DNA strand with the VirD2 protein bound to its 5' terminus into the plant cell cytoplasm (66). The transfer process is very similar to bacterial conjugation, and the VirB channel closely resembles a type IV secretion system (80). Once in the plant cytoplasm, the VirE2 protein, which is also transferred to the host cell cytoplasm through the VirB pore, binds tightly and cooperatively to the single-stranded T-strand. VirE1 is required for transfer of VirE2 into the plant cytoplasm and probably acts as an export chaperone for VirE2. Both VirD2 and VirE2 contain nuclear localization sites that mediate transport of the T-strand from the cytoplasm to the nucleus, where the T-DNA is integrated into the plant genome. Other members of the vir regulon are not essential for tumorigenesis on all hosts and may be required only in specific hosts or may play other roles in pathogenesis. These include virD5, virE3, virF, virH, virJ, virK, virL, virM, virP, and virR. In addition to T-DNA and vir genes, the Ti plasmid harbors genes that are involved in the uptake and catabolism of opines, others that are required for replication of the Ti plasmid, and still others that direct the conjugal transfer of the plasmid (549).
Studies of the induction of vir genes began with the demonstration that cocultivation of strains carrying a vir-lacZ fusion with cultured plant cells or cultured roots caused elevated expression of ß-galactosidase (463, 465). Two phenolic compounds, acetosyringone (Fig. 1G) and -hydroxyacetosyringone, isolated from tobacco root cultures, were the first two compounds identified as specific vir gene inducers (464). Comprehensive analyses of many chemical derivatives of acetosyringone showed that A. tumefaciens detects numerous related compounds, many of which are ubiquitous or at least widespread among host plants (322, 462). The essential structural features required for vir-inducing activity of a compound are a benzene ring with a hydroxyl group at position 4 and a methoxy group at position 3. The presence of another methoxy group at position 5 enhances the activity of the inducer, while a wide variety of substituents at position 1 are tolerated (Fig. 1G and 1H) (322, 464). Relatively high concentrations of the inducing phenolic (5 to 500 µM, depending on the compound) are required for full activation, which may help to account for the relatively low signal specificity of this pathogen (322, 462, 464).
Phenolic-induced expression of vir genes is greatly enhanced by specific monosaccharides including arabinose, galactose, galacturonic acid, glucose, glucuronic acid, mannose, fucose, cellobiose, and xylose (11). Most of these sugars are monomers of plant cell wall polysaccharides or are otherwise involved in plant metabolism. Their effect on vir induction is especially pronounced at low concentrations of the inducing phenolic. Galacturonic and glucuronic acid were reported to have the strongest activity, having an effect at concentrations as low as 100 µM (11). Acidity and temperature are also important for vir induction; activation of vir gene transcription occurs only in acidic environments in the pH range of 5.2 to 5.7 (465). Similarly, tumor formation on several host species is optimal at 22°C and does not occur at temperatures above 29°C (49). The temperature sensitivity of tumor formation was correlated with that of vir gene expression, which does not occur at temperatures above 32°C (9, 322).
The combination of phenolics, monosaccharides, and acidity that is required for induction of the vir regulon is thought to reflect the chemical components of wound sap. It is thought that these metabolites are released in largest amounts from plant wound sites, specifically from cells that are undergoing lignin synthesis or cell wall repair. The vir-inducing phenolics accumulate at wound sites as precursors of lignin biosynthesis, which is required for wound healing (112, 116). In addition, they may play a role in protection against potential pathogens (112). The inducing sugars are present at a wound site as degradation products of plant cell wall polysaccharides and may be generated by both mechanical means and the enzymatic activity of the cell wall glycosidases. Wound sap also tends to be acidic due to acidic compounds (e.g., phenolic acids and acidic monosaccharides) that are released from plant cell vacuoles (116). Wounding may be caused by foreign agents, such as herbivores or frost, or may be the result of tissue damage occurring during normal plant growth, for example the cracks produced at the site of emergence of side roots.
The idea that release of vir gene inducers requires wounding is appealing, since wounds are generally thought to be required for tumorigenesis (253). However, this idea may be flawed in several ways. First, acetosyringone and -hydroxyacetosyringone were first isolated not from plant wounds but, rather, from cultured cells and cultured, unwounded roots (464). Their concentration was reported to increase on wounding (464). There are few if any other studies demonstrating that phenolics are released preferentially from wound sites, and these same compounds have been detected from unwounded tobacco seedlings (unpublished data). Second, monosaccharides are, as described above, released from unwounded roots (99, 309). Finally, the pH of apoplastic fluids is acidic and roots generally acidify the adjacent soil, possibly in order to increase the solubility of phosphate (309). It may therefore be necessary to question whether these signals are really "wound released" and to think of them instead as being "plant released." Furthermore, wounding may not be essential for tumorigenesis after all. A study in which vir-induced bacteria were sprayed onto tobacco plantlets demonstrated that cells in unwounded plants could also be efficiently transformed (129), suggesting that the proposed plant responses evoked by wounding (enhanced cell division and DNA replication) are not essential for transformation.
