Plant gene expression response to Agrobacterium tumefaciens.

To elucidate the nature of plant response to infection and transformation by Agrobacterium tumefaciens, we compared the cDNA-amplified fragment length polymorphism (AFLP) pattern of Agrobacterium- and mock-inoculated Ageratum conyzoides plant cell cultures. From 16,000 cDNA fragments analyzed, 251 (1.6%) were differentially regulated (0.5% down-regulated) 48 h after cocultivation with Agrobacterium. From 75 strongly regulated fragments, 56 were already regulated 24 h after cocultivation. Sequence similarities were obtained for 20 of these fragments, and reverse transcription-PCR analysis was carried out with seven to confirm their cDNA-AFLP differential pattern. Their sequence similarities suggest a role for these genes in signal perception, transduction, and plant defense. Reverse transcription-PCR analysis indicated that four genes involved in defense response are regulated in a similar manner by nonpathogenic bacteria, whereas one gene putatively involved in signal transduction appeared to respond more strongly to Agrobacterium. A nodulin-like gene was regulated only by Agrobacterium. These results demonstrate a rapid plant cell response to Agrobacterium infection, which overlaps a general response to bacteria but also has Agrobacterium-specific features.

Characterization of a putative RND-type efflux system in Agrobacterium tumefaciens.

Sequencing of a 7277 bp fragment adjacent to the chvH locus of Agrobacterium tumefaciens revealed four open reading frames (ORFs), designated ameR, ameA, ameB and ameC, respectively. These ORFs exhibit amino acid similarities to components of Resistance-Nodulation-Cell Division (RND) type efflux systems. AmeA and AmeB show high homology to membrane fusion proteins (MFP) and RND-type transporters, whereas AmeC shows similarity to NodT and other members of outer membrane factor families. Mutations of the ameA and ameB genes did not affect the susceptibility profile of the wild-type strain to several detergents and antibiotics. In contrast, mutations of the ameC gene dramatically affected the susceptibility of the strain to these same inhibitory compounds. RT-PCR analysis demonstrated that the ameABC genes form an operon. In addition, ameC gene has its own promoter gene located in the intergenic region between ameB and ameC. Mapping upstream of ameA is ameR, which encodes a protein that shows similarity especially at its N-terminal end to the TetR family of bacterial transcriptional regulators. AmeR negatively regulates expression of the ameABC operon. A mutation in ameR increased the resistance of the cells to several antimicrobial agents. This regulatory locus appears to be in the same operon as a gene located upstream which codes for a product that has high similarity to numerous 4-(N-succinocarboxamide)-5-aminoimidazole ribonucleotide (SAICAR) synthetases. The possible role of the putative efflux pump coded by the ame genes is discussed.

The chvH locus of Agrobacterium encodes a homologue of an elongation factor involved in protein synthesis.

The virulence of Agrobacterium tumefaciens depends on both chromosome- and Ti plasmid-encoded gene products. In this study, we characterize a chromosomal locus, chvH, previously identified by TnphoA mutagenesis and shown to be required for tumor formation. Through DNA sequencing and comparison of the sequence with identified sequences in the database, we show that this locus encodes a protein similar in sequence to elongation factor P, a protein thought to be involved in peptide bond synthesis in Escherichia coli. The analysis of vir-lacZ and vir-phoA translational fusions as well as Western immunoblotting revealed that the expression of Vir proteins such as VirE2 was significantly reduced in the chvH mutant compared with the wild-type strain. The E. coli efp gene complemented detergent sensitivity, virulence, and expression of VirE2 in the chvH mutant, suggesting that chvH and efp are functionally homologous. As expected, ChvH exerts its activity at the posttranscriptional level. Southern analysis suggests that the gene encoding this elongation factor is present as a single copy in A. tumefaciens. We constructed a chvH deletion mutant in which a 445-bp fragment within its coding sequence was deleted and replaced with an omega fragment. On complex medium, this mutant grew more slowly than the wild-type strain, indicating that elongation factor P is important but not essential for the growth of Agrobacterium.

Transferred DNA (T-DNA)-associated proteins of Agrobacterium tumefaciens are exported independently of virB.

The transfer of T-DNA from Agrobacterium to plant cells is mediated by a system which involves the virB operon of the Ti plasmid. We report that VirE2 and VirD2, two T-DNA-associated proteins, as well as VirF, a protein known to be secreted into plant cells, are present in the periplasm and supernatant fractions of growing cells of Agrobacterium as are VirJ and ChvE, two known periplasmic proteins. Two cytoplasmic proteins, Ros and chloramphenicol acetyl transferase, and a VirE2green fluorescent protein construct were not detected in the above fraction. Export of VirE2 into the culture supernatant did not require any Ti plasmid genes, except for VirE1, a specific chaperone for VirE2. The levels of the VirE2 and VirD2 proteins in the supernatant increased significantly when cells were grown at 19 degrees C as compared with 28 degrees C. When Agrobacterium expressed the oncogenic suppressive activity protein (Osa), VirE2 and VirF proteins could not be detected in the supernatant or the periplasm and the level of VirD2 was greatly reduced. However, oncogenic suppressive activity protein did not block the accumulation of VirJ and ChvE in the periplasm. Our data suggest that VirD2, VirE2, and VirF are transported across the cytoplasmic membrane by a specific pathway, independent of virB. Thus, transfer of the T-complex of Agrobacterium may take place in two steps, the first mediated by an unidentified pathway and the second by the virB system.

DNA and protein transfer from bacteria to eukaryotes- the agrobacterium story.

Abstract Agrobacterium is a well-studied plant pathogen, which has the unique ability to transfer DNA and protein into a number of eukaryotes. The DNA is integrated randomly into the plant genome where it is expressed, thereby leading to the disease crown gall. This system is a paradigm for the interaction of a number of plant and animal pathogens which transfer proteins into their host cells. In Agrobacterium, the tumour inducing (Ti) plasmid codes for the functions specifically required for the transfer process. These genes, termed virulence or vir genes, are activated by plant signal molecules acting through a two component regulatory system. A key structure coded by 11 genes of the vir B operon is a pilus, synthesized at 20 degrees C, but poorly at 25 degrees C. How this pilus functions in DNA and protein transfer is unclear, but homologous genes are found in many animal pathogens. In addition to Ti plasmid-encoded vir genes, chromosomal virulence genes have also been identified. However, these mutations are often pleiotropic because they involve both the normal physiology of Agrobacterium as well as the metabolism of Agrobacterium when it is associated with plant cells. Based on 16S ribosomal RNA sequencing, Agrobacterium is closely related to the intracellular pathogen of animals, Brucella. Several chromosomal mutations of Agrobacterium required for virulence in plants are also required for invasion of animal host cells by Brucella.

VirE1 is a specific molecular chaperone for the exported single-stranded-DNA-binding protein VirE2 in Agrobacterium.

Agrobacterium tumefaciens induces tumours on plants by transferring a nucleoprotein complex, the T-complex, from the bacterium to the plant cell. The T-complex consists of a single-stranded DNA (ssDNA) segment, the T-DNA, and VirD2, an endonuclease covalently attached to the 5' end of the T-DNA. A type IV secretion system encoded by the virB operon and virD4 is required for the entry of the T-complex and VirE2, a ssDNA-binding protein, into plant cells. The VirE1 protein is specifically required for the export of the VirE2 protein, as demonstrated by extracellular complementation and tumour formation. In this report, using a yeast two-hybrid system, we demonstrated that the VirE1 and VirE2 proteins interact and confirmed this interaction by in vitro binding assays. Although VirE2 is a ssDNA-binding protein, addition of ssDNA into the binding buffer did not interfere with the interaction of VirE1 and VirE2. VirE2 also interacts with itself, but the interaction between VirE1 and VirE2 is stronger than the VirE2 self-interaction, as measured in a lacZ reporter gene assay. In addition, the interaction of VirE2 with itself is inhibited by VirE1, indicating that VirE2 binds VirE1 preferentially. Analysis of various virE2 deletions indicated that the VirE1 interaction domain of VirE2 overlaps the VirE2 self-interaction domain. Incubation of extracts from Escherichia coli overexpressing His-VirE1 with the extracts of E. coli overexpressing His-VirE2 increased the yield of His-VirE2 in the soluble fraction. In a similar purified protein solubility assay, His-VirE1 increased the amount of His-VirE2 partitioning into the soluble fraction. In Agrobacterium, VirE2 was undetectable in the soluble protein fraction unless VirE1 was co-expressed. When urea was added to solubilize any large protein aggregates, a low level of VirE2 was detected. These results indicate that VirE1 prevents VirE2 from aggregating, enhances the stability of VirE2 and, perhaps, maintains VirE2 in an export-competent state. Analysis of the deduced amino acid sequence of the VirE1 protein revealed that the VirE1 protein shares a number of properties with molecular chaperones that are involved in the transport of specific proteins into animal and plant cells using type III secretion systems. We suggest that VirE1 functions as a specific molecular chaperone for VirE2, the first such chaperone linked to the presumed type IV secretion system.

The phenolic recognition profiles of the Agrobacterium tumefaciens VirA protein are broadened by a high level of the sugar binding protein ChvE.

The formation of crown gall tumors by Agrobacterium tumefaciens requires that the virulence (vir) genes be induced by chemical signals which consist of specific phenolic compounds and monosaccharides, synthesized at plant wound sites. Signal transduction in the activation of these genes is mediated by the VirA-VirG two-component regulatory system, together with ChvE, a glucose-galactose binding protein which interacts with VirA. We have previously presented genetic evidence that virA senses phenolic compounds directly (Y.-W. Lee, S. Jin, W.-S. Sim, and E. W. Nester, Proc. Natl. Acad. Sci. USA 92:12245-12249, 1995). The vir genes of strain KU12 can be induced by 4-hydroxyacetophenone, p-coumaric acid, and phenol, whereas these same phenolic compounds are weak inducers of the vir genes of strain A6. In this report, we show that a specific inducing sugar can broaden the specificity of the phenolic compound which VirA senses. 4-Hydroxyacetophenone and other related phenolic compounds function as inducing phenolic compounds with the virA gene of A6 if arabinose replaces glucose as the inducing sugar. We further demonstrate that this broadened specificity for phenolic inducers results from the increased level of ChvE through induction by arabinose via the regulatory protein GbpR. If high levels of ChvE are present, then poorly inducing phenolic compounds can induce the vir genes to high levels in combination with glucose. Comparing the induction response of the wild type and that of a VirA mutant with a mutation in its receiver domain revealed that the activity of the receiver domain is controlled by the periplasmic domain. We discuss these observations in terms of how VirA senses and transduces signals elicited by the two classes of plant signal molecules.

Agrobacterium VirD2 protein interacts with plant host cyclophilins.

Agrobacterium tumefaciens induces crown gall tumors on plants by transferring a nucleoprotein complex, the T-complex, from the bacterium to the plant cell. The T-complex consists of T-DNA, a single-stranded DNA segment of the tumor-inducing plasmid, VirD2, an endonuclease covalently bound to the 5' end of the T-DNA, and perhaps VirE2, a single-stranded DNA binding protein. The yeast two-hybrid system was used to screen for proteins interacting with VirD2 and VirE2 to identify components in Arabidopsis thaliana that interact with the T-complex. Three VirD2- and two VirE2-interacting proteins were identified. Here we characterize the interactions of VirD2 with two isoforms of Arabidopsis cyclophilins identified by using this analysis. The VirD2 domain interacting with the cyclophilins is distinct from the endonuclease, omega, and the nuclear localization signal domains. The VirD2-cyclophilin interaction is disrupted in vitro by cyclosporin A, which also inhibits Agrobacterium-mediated transformation of Arabidopsis and tobacco. These data strongly suggest that host cyclophilins play a role in T-DNA transfer.

Characterization of an unusual sensor gene (virA) of Agrobacterium.

Previous studies have shown that the virulence(vir) genes of Agrobacterium tumefaciens strain KU12 are induced by a unique set of phenolic compounds that are non-functional in most strains of Agrobacterium. Further, strain KU12 is not induced by phenolic compounds that induce the vir genes in other strains. Previous studies have shown that these differences in inducing activity result from differences in the sensor protein for these signal molecules, the VirA protein. To gain some understanding of the basis for these differences in sensing ability, we sequenced the entire virA locus of pTiKU12, including its promoter region and compared this sequence with five different published virA sequences that respond in different ways to inducing compounds. The virA gene of KU12 is composed of an open single reading frame coding for 851 aa. At the aa level, the VirA protein of pTiKU12 is 45, 45, 49, 49 and 64% identical to the VirA proteins from pTiA6, pTi15955, pRiA4, pTiC58 and pTiAg162, respectively. The transcription start sites of pTiKU12 and pTiA6 virA genes differ significantly when mapped by primer extension. Unlike all other vir genes, except the virA gene of pTiAg162, pTiKU12 virA is constitutively expressed, and its synthesis is not induced by phenolic compounds. The lack of induction is accounted for by the fact that the promoter region does not have the conserved VirG-binding dodecadeoxynucleotide sequence (vir-box) that was previously identified in all promoter regions of inducible vir genes.

Variable efficiency of a Ti plasmid-encoded VirA protein in different agrobacterial hosts.

The transconjugant CB100, harboring the Ti plasmid from the Agrobacterium tumefaciens biovar 2 strain D10B/87 in the chromosomal background of the biovar 1 strain C58, was defective in vir gene induction. This defect was corrected in the presence of virA from pTiA6. Based on this complementation result and an analysis of the induction requirements of the transconjugant CB100 and its parent strains, it was hypothesized that the defective vir gene induction in CB100 was related to a dysfunctional interaction between the pTi-encoded D10B/87 VirA and the chromosome-encoded C58 ChvE. To verify this hypothesis, D10B/87 and C58 virA were compared, and conclusions from this first set of analyses were then corroborated by comparing D10B/87 and C58 chvE. Whereas only a few nucleotide differences were identified in the promoters and 5' ends of the coding regions of D10B/87 and C58 virA, analysis of hybrid virA genes showed that these differences collectively accounted for the poor vir gene induction of strain CB100. In contrast with the sequence similarity of the VirA proteins, extensive divergence was seen between the chromosome-encoded D10B/87 and C58 ChvE. Although D10B/87 chvE introduced in trans had little effect on vir gene induction of CB100, it enhanced the induction response of a strain CB100 derivative in which the chromosomal C58 chvE had been inactivated by marker exchange. These results suggest that chromosomal backgrounds provided by different strains of A. tumefaciens are not equivalent for VirA function. Following conjugative transfer of certain Ti plasmids to a new agrobacterial host, evolution of the newly introduced virA, or coevolution of chvE and virA, may lead to optimization of ChvE-VirA interaction and vir gene induction levels.

The Agrobacterium tumefaciens virulence gene chvE is part of a putative ABC-type sugar transport operon.

The Agrobacterium tumefaciens virulence determinant ChvE is a periplasmic binding protein which participates in chemotaxis and virulence gene induction in response to monosaccharides which occur in the plant wound environment. The region downstream of the A. tumefaciens chvE gene was cloned and sequenced for nucleotide and expression analysis. Three open reading frames transcribed in the same direction as chvE were revealed. The first two, together with chvE, encode putative proteins of a periplasmic binding protein-dependent sugar uptake system, or ABC-type (ATP binding cassette) transporter. The third open reading frame encodes a protein of unknown function. The deduced transporter gene products are related on the amino acid level to bacterial sugar transporters and probably function in glucose and galactose uptake. We have named these genes gguA, -B, and -C, for glucose galactose uptake. Mutations in gguA, gguB, or gguC do not affect virulence of A. tumefaciens on Kalanchoe diagremontiana; growth on 1 mM galactose, glucose, xylose, ribose, arabinose, fucose, or sucrose; or chemotaxis toward glucose, galactose, xylose, or arabinose.

Discrete regions of the sensor protein virA determine the strain-specific ability of Agrobacterium to agroinfect maize.

The ability of Agrobacterium strains to infect transformation-recalcitrant maize plants has been shown to be determined mainly by the virA locus, implicating vir gene induction as the major factor influencing maize infection. In this report, we further explore the roles of vir induction-associated bacterial factors in maize infection using the technique of agroinfection. The Ti plasmid and virA source are shown to be important in determining the ability of a strain to infect maize, and the monosaccharide binding protein ChvE is absolutely required for maize agroinfection. The linker domain of VirAC58 from an agroinfection-competent strain, C58, is sufficient to convert VirAA6 of a nonagroinfecting strain, A348,to agroinfection competence. The periplasmic domain of VirAC58 is also able to confer a moderate level of agroinfection competence to VirAA6. In addition, the VirAA6 protein from A348 is agroinfection competent when removed from its cognate Ti plasmid background and placed in a pTiC58 background. The presence of a pTiA6-encoded, VirAA6-specific inhibitor is hypothesized and examined.

The sensing of plant signal molecules by Agrobacterium: genetic evidence for direct recognition of phenolic inducers by the VirA protein.

The virulence (vir) genes of Agrobacterium tumefaciens are induced by low-molecular-weight phenolic compounds and monosaccharides through a two-component regulatory system consisting of the VirA and VirG proteins. Although it is clear that the monosaccharides require binding to a periplasmic binding protein before they can interact with the sensor VirA protein, it is not certain whether the phenolic compounds also interact with a binding protein or directly interact with the sensor protein. To shed light on this question, we tested the vir-inducing abilities of several different phenolic compounds using two wild-type strains of A. tumefaciens, KU12 and A6. We found that several compounds such as 4-hydroxyacetophone and p-coumaric acid induced the vir of KU12, but not A6. On the other hand, acetosyringone and several other phenolic compounds induced the vir of A6, but not KU12. By transferring different Ti plasmids into isogenic chromosomal backgrounds, we showed that the phenolic sensing determinant is associated with the Ti plasmid. Subcloning of the Ti plasmid indicated that the virA locus determines which phenolic compounds can function as vir inducers. These results suggest that VirA directly senses the phenolic compounds for vir activation.

Pilus assembly by Agrobacterium T-DNA transfer genes.

Agrobacterium tumefaciens can genetically transform eukaryotic cells. In many bacteria, pili are required for interbacterial DNA transfer. The formation of pili by Agrobacterium required induction of tumor-inducing (Ti) plasmid-encoded virulence genes and growth at low temperature. A genetic analysis demonstrated that virA, virG, virB1 through virB11, and virD4 are the only Ti plasmid genes necessary for pilus assembly. The loss and gain of pili in various mutants correlated with the loss and gain of transferred DNA (T-DNA) transfer functions, which is consistent with the view that Agrobacterium pili are required for transfer of DNA to plant cells in a process similar to that of conjugation.

A family of lysozyme-like virulence factors in bacterial pathogens of plants and animals.

We describe a conserved family of bacterial gene products that includes the VirB1 virulence factor encoded by tumor-inducing plasmids of Agrobacterium spp., proteins involved in conjugative DNA transfer of broad-host-range bacterial plasmids, and gene products that may be involved in invasion by Shigella spp. and Salmonella enterica. Sequence analysis and structural modeling show that the proteins in this group are related to chicken egg white lysozyme and are likely to adopt a lysozyme-like structural fold. Based on their similarity to lysozyme, we predict that these proteins have glycosidase activity. Iterative data base searches with three conserved sequence motifs from this protein family detect a more distant relationship to bacterial and bacteriophage lytic transglycosylases, and goose egg white lysozyme. Two acidic residues in the VirB1 protein of Agrobacterium tumefaciens form a putative catalytic dyad, Each of these residues was changed into the corresponding amide by site-directed mutagenesis. Strains of A. tumefaciens that express mutated VirB1 proteins have a significantly reduced virulence. We hypothesize that many bacterial proteins involved in export of macromolecules belong to a widespread class of hydrolases and cleave beta-1,4-glycosidic bonds as part of their function.

Temperature affects the T-DNA transfer machinery of Agrobacterium tumefaciens.

Early studies on Agrobacterium tumefaciens showed that development of tumors on plants following infection by A. tumefaciens was optimal at temperatures around 22 degrees C and did not occur at temperatures above 29 degrees C. To assess whether this inability to induce tumors is due to a defect in the T-DNA transfer machinery, mobilization of an incompatibility group Q (IncQ) plasmid by the T-DNA transfer machinery of A. tumefaciens was tested at various temperatures. Optimal transfer occurred when matings were performed at 19 degrees C, and transfer was not seen when matings were incubated above 28 degrees C. Transfer of the IncQ plasmid was dependent upon induction of the virB and virD operons by acetosyringone but was not dependent upon induction of the tra genes by octopine. However, alterations in the level of vir gene induction could not account for the decrease in transfer with increasing temperature. A. tumefaciens did successfully mobilize IncQ plasmids at higher temperatures when alternative transfer machineries were provided. Thus, the defect in transfer at high temperature is apparently in the T-DNA transfer machinery itself. As these data correlate with earlier tumorigenesis studies, we propose that tumor suppression at higher temperatures results from a T-DNA transfer machinery which does not function properly.

Mutational analysis of the input domain of the VirA protein of Agrobacterium tumefaciens.

The transmembrane sensor protein VirA activates VirG in response to high levels of acetosyringone (AS). In order to respond to low levels of AS, VirA requires the periplasmic sugar-binding protein ChvE and monosaccharides released from plant wound sites. To better understand how VirA senses these inducers, the C58 virA gene was randomly mutagenized, and 14 mutants defective in vir gene induction and containing mutations which mapped to the input domain of VirA were isolated. Six mutants had single missense mutatiions in three widely separated areas of the periplasmic domain. Eight mutants had mutations in or near an amphipathic helix, TM1, or TM2. Four of the mutations in the periplasmic domain, when introduced into the corresponding A6 virA sequence, caused a specific defect in the vir gene response to glucose. This suggests that most of the periplasmic domain is required for the interaction with, or response to, ChvE. Three of the mutations from outside the periplasmic domain, one from each transmembrane domain and one from the amphiphathic helix, were made in A6 virA. These mutants were defective in the vir gene response to AS. These mutations did not affect the stability or topology of VirA or prevent dimerization; therefore, they may interfere with detection of AS or transmission of the signals to the kinase domain. Characterization of C58 chvE mutants revealed that, unlike A6 VirA, C58 VirA requires ChvE for activation of the vir genes.

Agrobacterium tumefaciens-mediated transformation of yeast.

Agrobacterium tumefaciens transfers a piece of its Ti plasmid DNA (transferred DNA or T-DNA) into plant cells during crown gall tumorigenesis. A. tumefaciens can transfer its T-DNA to a wide variety of hosts, including both dicotyledonous and monocotyledonous plants. We show that the host range of A. tumefaciens can be extended to include Saccharomyces cerevisiae. Additionally, we demonstrate that while T-DNA transfer into S. cerevisiae is very similar to T-DNA transfer into plants, the requirements are not entirely conserved. The Ti plasmid-encoded vir genes of A. tumefaciens that are required for T-DNA transfer into plants are also required for T-DNA transfer into S. cerevisiae, as is vir gene induction. However, mutations in the chromosomal virulence genes of A. tumefaciens involved in attachment to plant cells have no effect on the efficiency of T-DNA transfer into S. cerevisiae. We also demonstrate that transformation efficiency is improved 500-fold by the addition of yeast telomeric sequences within the T-DNA sequence.

Genetic evidence for direct sensing of phenolic compounds by the VirA protein of Agrobacterium tumefaciens.

The virulence (vir) genes of Agrobacterium tumefaciens are induced by low-molecular-weight phenolic compounds and monosaccharides through a two-component regulatory system consisting of the VirA and VirG proteins. However, it is not clear how the phenolic compounds are sensed by the VirA/VirG system. We tested the vir-inducing abilities of 15 different phenolic compounds using four wild-type strains of A. tumefaciens--KU12, C58, A6, and Bo542. We analyzed the relationship between structures of the phenolic compounds and levels of vir gene expression in these strains. In strain KU12, vir genes were not induced by phenolic compounds containing 4'-hydroxy, 3'-methoxy, and 5'-methoxy groups, such as acetosyringone, which strongly induced vir genes of the other three strains. On the other hand, vir genes of strain KU12 were induced by phenolic compounds containing only a 4'-hydroxy group, such as 4-hydroxyacetophenone, which did not induce vir genes of the other three strains. The vir genes of strains KU12, A6, and Bo542 were all induced by phenolic compounds containing 4'-hydroxy and 3'-methoxy groups, such as acetovanillone. By transferring different Ti plasmids into isogenic chromosomal backgrounds, we showed that the phenolic-sensing determinant is associated with Ti plasmid. Subcloning of Ti plasmid indicates that the virA locus determines which phenolic compounds can function as vir gene inducers. These results suggest that the VirA protein directly senses the phenolic compounds for vir gene activation.