INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Agrobacterium tumefaciens transfers at least three macromolecular substrates, oncogenic T-DNA, VirE2 single-stranded DNA-binding protein (SSB), and VirF protein, to plant cells during the course of infection (8). Substrate transfer is mediated by a type IV secretion system assembled from the products of the ~9.5-kbp virB operon and the virD4 gene (29). This transfer system is a bona fide conjugation apparatus, as suggested by its ancestral relatedness to transfer systems (Tra) of several broad-host-range plasmids and as demonstrated by its ability to transfer the mobilizable IncQ plasmid RSF1010 to bacterial and plant recipient cells. In recent years, other type IV systems have been shown to contribute to the virulence of several mammalian pathogens. The pathogens utilizing type IV systems during infection include Helicobacter pylori (12), Bordetella pertussis (3), Brucella spp. (31, 42), Bartonella henselae (35, 40), and Legionella pneumophila (47). A type IV system of H. pylori exports CagA to the cytosol of mammalian cells (32, 41, 44), whereas a related system exports pertussis toxin across the outer membrane of B. pertussis (3). Other type IV systems are thought to export effector molecules, whose identities are currently unknown, to the mammalian cell cytosol.

Mechanistic studies of the A. tumefaciens T-DNA transfer system and the related conjugal transfer systems of the IncP, IncN, and IncW broad-host-range plasmids have contributed important structure and function information about the type IV systems (8, 29). An area of special interest concerns the mechanism of substrate processing and presentation to the mating channel. Recent work has shown that these systems translocate two general classes of substrates: (i) conjugative plasmids and T-DNA that are translocated as transfer intermediates composed of a single-stranded molecule of DNA covalently bound at the 5' end by a nicking enzyme (4, 7) and (ii) proteins that are translocated independent of DNA (25, 46, 50). Interestingly, export of DNA and protein substrates via type IV machines appears to require common as well as distinct factors at the early stages of substrate processing. In A. tumefaciens, T-DNA translocation requires the VirD2 nicking enzyme and the VirD1 auxiliary factor for cleavage at the T-DNA borders, the VirD4 protein, which is thought to physically link the T-DNA processing system and the mating channel, and two membrane-associated proteins, VirC1 and VirC2, whose contributions to T-DNA transfer are largely undetermined (23, 53). By contrast, translocation of protein substrates, such as VirE2 SSB and VirF, does not require VirD1, VirD2, or the VirC proteins but is still dependent on VirD4 (7, 9, 46, 53). In addition, translocation of VirE2 requires a small upstream gene encoding the putative VirE1 secretion chaperone.

Definition of the role of VirE1 for VirE2 export has been the subject of recent investigations. VirE1 interacts with VirE2, as shown by the yeast two-hybrid screen (15, 45, 52) and pull-down (15) and immunoprecipitation (52) assays. An interaction domain was localized to the C-terminal half of VirE2, possibly overlapping with regions required for single-stranded DNA binding and for VirE2 self-association (15, 45, 52). An early study suggested that VirE1 stabilizes VirE2 when it is synthesized in Escherichia coli (30), although recent work has led to conflicting results regarding this activity in A. tumefaciens (15, 16). There is also some evidence that VirE1 prevents VirE2 from aggregating in solution (15). VirE1 is a small, acidic protein with an amphipathic -helix at its C terminus, the proposed site for interaction with VirE2 (45). Of considerable interest, these physical and functional properties of VirE1 are also features of the Syc chaperone or bodyguard proteins required for export of effector proteins via the type III secretion systems of plant and mammalian pathogens (6, 36).

In this study, we further defined the role of virE1 for VirE2 translocation. Our findings support a model in which virE1 acts at the level of translation to ensure efficient virE2 expression and VirE1 protein acts posttranslationally to stabilize VirE2 and prevent the formation of higher-order nonproductive complexes or aggregates in A. tumefaciens.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Enzymes, chemicals, and reagents. Restriction endonucleases were purchased from Promega (Madison, Wis.), New England Biolabs (Beverly, Mass.), or GIBCO-BRL (Grand Island, N.Y.). The Klenow fragment of E. coli DNA polymerase I and T4 DNA ligase were obtained from Promega. Isopropyl--D-thiogalactopyranoside (IPTG), phenylmethylsulfonyl fluoride (PMSF), carbenicillin, kanamycin, Triton X-100, Coomassie brilliant blue R250, p-nitroblue tetrazolium, and 5-bromo-4-chloro-3-indolylphosphate (BCIP) were purchased from Sigma Chemical Co. (St. Louis, Mo.). Acetosyringone (3',5'-dimethoxy-4'-hydoxyacetophenone) (AS) was obtained from Aldrich (Milwaukee, Wis.). Alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G and a Bradford protein assay kit were obtained from Bio-Rad Laboratories (Hercules, Calif.).

Bacterial strains, plasmids, and growth conditions. The bacterial strains and plasmids used in this study and their relevant characteristics are listed in Table 1. E. coli strains were maintained on Luria-Bertani medium, and A. tumefaciens strains were maintained on MG/L medium or on AB minimal salts medium (52). Media were supplemented with antibiotics as follows: for E. coli, 50 µg of chloramphenicol per ml; and for E. coli and A. tumefaciens, 100 µg of carbenicillin per ml, 100 µg of kanamycin per ml, and 5 µg of tetracycline per ml. For induction of vir genes, A. tumefaciens cells were grown in MG/L medium to an optical density at 600 nm (OD₆₀₀) of 0.6, harvested by centrifugation, and inoculated at an initial OD₆₀₀ of 0.25 into induction medium composed of AB minimal salts medium (pH 5.5), 1 mM phosphate, and 200 µM AS (52). Cultures were incubated with shaking at 22°C for 18 h and then harvested for protein analysis.

Recombinant DNA techniques. DNA manipulations and DNA electrophoresis were performed as described by Sambrook et al. (39). A. tumefaciens cells were transformed by electroporation, and plasmids were recovered for physical characterization as previously described (51). DNA sequencing was carried out at the DNA Core Facility of the Department of Microbiology and Molecular Genetics, University of Texas-Houston Medical School, Houston, Tex., with an ABI 373A DNA sequencer (Perkin-Elmer, Applied Biosystems Division) by using Taq polymerase in a thermal cycling reaction. PCR was performed with a Perkin-Elmer Cetus DNA thermocycler by using Pwo DNA polymerase from Boehringer (Mannheim, Germany). Oligonucleotides were purchased from Sigma-Genosys (Woodlands, Tex.).

Construction of virE1 deletion strain PC3001. To delete virE1 from pTiA6NC, a 2.5-kb XbaI-XhoI fragment (made blunt ended with the Klenow fragment) from pPC732 (carries P_virE::virE2) was introduced at the ScaI site of sacBR suicide plasmid pBB50. The new plasmid, pSF1, was recombined onto pTiA6NC by a single crossover event, and sucrose counterselection was used to enrich for double-crossover recombinants lacking the pSF1 plasmid. The virE1 deletion in strain PC3001 was confirmed by sequencing a PCR product resulting from amplification across the putative virE1 junction site with primers complementary to the sequences in the virE promoter (5'-TCAAGACCCGAGTATGGATG-3') and the virE2 gene (5'-TGCCAGAAAGATCCATCGTC-3').

Construction of virE expression plasmids. Plasmids expressing virE derivatives were constructed as follows. Both lacZ transcriptional and translational fusions to virE2 were made by inserting cassettes at the StuI site at bp 36 of virE2. For the lacZ transcriptional fusions, StuI-XhoI fragments of pPC731 and pPC732 were replaced by a 6.3-kbp SmaI-SalI fragment containing the lacZYA cassette from pRS551 to generate pZZ14 (P_virE::virE1 virE2' lacZYA) and pZZ15 (P_virE::virE2' lacZYA). For lacZ translational fusions, StuI-HindIII fragments of pPC731 and pPC732 were replaced by the SmaI-HindIII fragments containing the 'lacZ cassette from pMC1871 to generate pZZ16 (P_virE::virE1 virE2'-'lacZ) and pZZ17 (P_virE::virE2'-'lacZ). In these constructs, 12 N-terminal residues of VirE2 were fused to -galactosidase. For production of His-VirE1, the virE1 gene carried on a 1.0-kbp NdeI-XhoI fragment from pXZ426 was introduced into pET15b to make pZZ3. Plasmid pPC914KS⁺. NcoI was constructed as previously described for pPC914KS⁺. NdeI (2), except that an NcoI site was introduced at the virB1 start site by oligonucleotide-directed mutagenesis. For expression of his-virE1 from P_virB, a ~1.1-kbp NcoI-XhoI fragment from pZZ3 was introduced into pPC914KS⁺. NcoI, replacing virB1. For production of GST-VirE1, the virE1 gene carried on a ~1.0-kbp NdeI (blunt-ended)-EcoRI restriction fragment from pXZ426 was introduced into SmaI-EcoRI-digested pGEX-2T, resulting in pZZ6. For production of His-VirE2, the virE2 gene carried on a ~1.6-kbp NdeI-XhoI restriction fragment from pPC725 was introduced into pET15b to make pPC730. Plasmids with ColE1 replication origins that were ligated to IncP broad-host-range plasmid pSW172 or pXZ151 for introduction into A. tumefaciens were designated by using the ColE1 plasmid name plus a B to indicate ligation to the broad-host-range plasmid.

Protein analysis, immunoblotting, and cell fractionation. Proteins were resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) or with a Tricine-SDS-PAGE system as previously described (37). Vir proteins were visualized by SDS-PAGE, transfer of the proteins to nitrocellulose membranes, and development of immunoblots with goat anti-rabbit antibodies conjugated to alkaline phosphatase and histochemical substrates. For enhanced sensitivity, blots were developed with anti-rabbit antibodies conjugated to horseradish peroxidase, and antibody-antigen interactions were visualized by chemiluminescence (Amersham, Arlington Heights, Ill.). Proteins were loaded on a per-cell-equivalent basis to compare VirE protein abundance in different strains. Molecular size markers were obtained from GIBCO-BRL. Fractionation of A. tumefaciens into soluble material (cytoplasm and periplasm) and insoluble material (cytoplasmic and outer membranes) was carried out as previously described (19). Subcellular fractions were applied to SDS-polyacrylamide gels on a per-cell-equivalent basis.

Purification of tagged VirE proteins. Anti-VirE1 and anti-VirE2 antibodies were generated as follows. IPTG-induced BL21(DE3) cells transformed with pPC730 and pZZ6 were used as the starting materials for purification of His-VirE2 and GST-VirE1, respectively; this was done by performing immobilized Ni²⁺ metal affinity chromatography (IMAC) and glutathione affinity chromatography, respectively, as recommended by the manufacturers (Ni²⁺ resin from Novagen and glutathione resin from Pharmacia). Purified His-VirE2 and GST-VirE1 were sent to Cocalico Biologicals, Inc. (Reamstown, Pa.) to produce antibodies in New Zealand White rabbits. His-VirE1 was purified by IMAC with a Co²⁺ resin (Talon resin; Clontech) for analyses of His-VirE1 complexes.

Measurement of VirE2 turnover in AS-induced A. tumefaciens cells. A. tumefaciens strains were grown to an OD₆₀₀ of 0.5 in AB minimal medium (pH 7.0), washed, pelleted, and resuspended at a 1:5 dilution in induction medium. Cells were induced for 6 h (required because there was a long lag in induction of vir gene expression [5]), pelleted, and resuspended in 5 ml of induction medium supplemented with Pro-Mix ([³⁵S]methionine-[³⁵S]cysteine; Pharmacia) at a final concentration of 5 µCi/ml. Cells were pulse-labeled for 30 min at 22°C with shaking, and then excess cold methionine and cysteine were added at final concentrations of 25 mM. A 1-ml culture aliquot was removed (zero time), and aliquots were removed at various times during continued incubation of the cell cultures at 22°C with shaking. Washed cell pellets were resuspended in 600 µl of a solution containing 50 mM Tris (pH 8.0), 50 mM MgCl₂, 20% sucrose, 0.1 mg of DNase per ml, 0.1 mg of RNase per ml, and 1 mM PMSF. Cells were broken by sonication and centrifuged at 14,000 × g for 15 min to remove the cell debris. Anti-VirE2 antisera and protein A-Sepharose were used to precipitate VirE2 protein from the clarified culture supernatants (24). Briefly, 5 µl of VirE2 antisera was incubated with culture supernatants at 4°C for 3 h, and then 50 µl of 10% (wt/vol) protein A-Sepharose in Tris buffer (50 mM Tris [pH 8.0], 1 mM PMSF) was added and the slurry was incubated at 4°C for 2 h. Precipitable complexes were recovered by centrifugation at 10,000 × g for 1 min. The pellet was washed with 1 ml of Tris-EDTA buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 5 mM EDTA, 1 mM PMSF) and resuspended in 25 µl of 50 mM Tris (pH 8.0) and 25 µl of SDS loading buffer. The resuspended material was boiled and analyzed for the presence of VirE2 by SDS-PAGE and immunoblotting. VirE2 abundance was estimated by scanning densitometry of autoradiographs.

Gel filtration chromatography and native PAGE. Total soluble proteins from A. tumefaciens extracts or soluble His-VirE1 purified by Co²⁺ affinity chromatography was applied at a concentration of 5 mg of protein/ml of 50 mM Tris (pH 8)-50 mM MgCl₂ to a gel filtration column (30 by 1.5 cm) prepared with Toyopearl MW55 resin (TosoHaas, Montgomeryville, Pa.) according to the manufacturer's instructions. Material was fractionated with a Pharmacia Gradi-frac chromatography system at a flow rate of 0.6 ml/min, and 1.2-ml fractions were collected. The fractions were analyzed for the presence of VirE1 or VirE2 by SDS-PAGE and immunostaining as described above. The protein molecular size markers used included apoferritin (443 kDa), alcohol dehydrogenase (150 kDa), bovine serum albumin (BSA) (66 kDa), carbonic anhydrase (29 kDa), and cytochrome c (12 kDa). Size markers were electrophoresed both together with and independent of the test sample.

Enzyme assays. A. tumefaciens KE1 and At12516 cells expressing the lacZ gene fusions were grown to the mid-log phase in MG/L broth and induced for expression of the vir genes as described above. -Galactosidase activities (in Miller units) were determined as previously described (13). At least three independent assays were performed in triplicate for each strain.

Phage immunity assay. Transformants of E. coli AG1688 expressing wild-type or chimeric repressors were each cross-streaked against >10⁶ PFU of lytic phage KH54 spread on one-half of the surface of a Luria-Bertani agar plate. After incubation overnight at 37°C, the streaks were examined visually for bacterial growth in the presence of the phage. Immunity was defined by the ability of strains to grow in the presence of the phage (26, 37).

Virulence assays. Strains of A. tumefaciens were tested for virulence on uniformly wounded Kalanchoe daigremontiana leaves (51). The controls for the tumorigenesis assays included coinoculation of the same leaf with wild-type A348 and strain A136 lacking plasmid Ti. Virulence was scored in terms of tumor size and time course of tumor appearance. Assays were repeated at least four times for each strain on separate leaves. Tumors were photographed 4 to 5 weeks after inoculation.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of cis and trans expression of virE1 on virE2 expression. In a previous study, we determined that expression of virE2 from a P_virE promoter without cis coexpression of the upstream virE1 gene resulted in accumulation of low levels of VirE2 protein (52). To further define the specific contributions of virE1 to virE2 expression, we first compared steady-state levels of VirE2 in an A. tumefaciens virE1 mutant expressing various combinations of virE1 and virE2 from Ti or IncP plasmids (Fig. 1A).

View larger version (68K): [in this window] [in a new window]   FIG. 1.   Effect of virE1 on VirE2 protein accumulation. (A) VirE protein abundance as determined by immunoblot analysis. Lane 1, wild-type strain A348; lane 2, virE1 mutant PC3001; lane 3, PC3001(pPCB732) merodiploid with PvirE::virE2 carried on Ti and IncP replicons; lane 4, PC3001(pXZB46) merodiploid with PvirE::virE2 (Ti) and Plac::virE2 (IncP); lane 5, PC3001(pPCB731) merodiploid with PvirE::virE2 (Ti) and PvirE::virE1 virE2 (IncP); lane 6, PC3001(pZZB12) with PvirE::virE2 (Ti) and PvirB::his-virE1 (IncP); lane 7, PC3001 (pXZ426) with PvirE::virE2 (Ti) and PvirE::virE1 (IncP); lane 8, PC3001(pZZB68) with PvirE::virE2 (Ti) and both PvirE::virE1 and PvirE::virE2 (IncP); lane 9, PC3001(pZZB46) with PvirE::virE2 (Ti) and both PvirB::his-virE1 and PvirB::virE2 (IncP). (B) Virulence of virE mutants as determined by inoculation of wounded K. daigremontiana leaves. The numbers refer to the strains in the corresponding lanes in panel A. Ti, leaf wound site inoculated with negative control strain A136.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   -Galactosidase (-Gal) activities of the lacZ fusion strains expressed as percentages of the activity of strain A348(pSM358). Experiments were done in triplicate, and standard deviations are indicated by error bars. The virE genes expressed from the PvirE promoter on Ti and IncP plasmids are shown on the right.

VirE1 chaperone stabilizes VirE2 in vivo. Next, we examined the contribution of the VirE1 chaperone to the stability of VirE2. As shown above, PC3001 expressing virE1 or his-virE1 from an IncP replicon accumulated higher levels of VirE2 than parental PC3001 cells. Similar results were obtained when virE1 and virE2 were expressed from separate P_tac promoters (15). Furthermore, even though cells lacking virE1 accumulated high levels of VirE2 when virE2 was expressed from the P_virB promoter, the VirE2 levels were still appreciably higher in isogenic cells expressing virE1 (see below). A previous study showed that upon expression of virE2 from P_tac, detection of VirE2 in gels was enhanced by treatment of cell extracts with urea prior to electrophoresis (15). In our studies, urea treatment did not appreciably alter the VirE2 protein content upon electrophoresis of extracts from cells expressing virE2 from the P_virE or P_virB promoters (data not shown).

View larger version (46K): [in this window] [in a new window]   FIG. 3.   Effect of virE1 expression on VirE2 protein turnover in A. tumefaciens. (A) Autoradiographs showing VirE2 precipitated from extracts of radioactively labeled A. tumefaciens cells incubated for the chase times indicated. The strains expressing virE genes are indicated below the autoradiographs. (B) Graphic representation of the results expressed as the percentage of VirE2 detected at zero time for each strain. See text for experimental details.

VirE1 stabilization is mediated by an interaction with the N terminus of VirE2. VirE1 interacts with a domain in the C-terminal half of VirE2, as shown by yeast two-hybrid screens (15, 45, 52). It has been proposed that a second interaction domain is localized in the N-terminal half of VirE2, but the boundaries of this domain could not be delineated because the N terminus of VirE2 self-activates transcription at the GAL4 promoter when it is fused to the GAL4 DNA-binding domain (45). To localize the region of VirE2 necessary for VirE1-mediated stabilization, we monitored steady-state accumulation of VirE2 truncation and insertion derivatives in isogenic virE1⁺ and virE1 strains as an indicator of intrinsic resistance to endogenous proteases. In the following experiments, we expressed the virE2 derivatives from the P_virB promoter. This is because initial studies showed that native VirE2 accumulates at appreciably higher levels when virE2 is expressed from P_virB than when virE2 is expressed from P_virE in the absence of virE1 cis coexpression (Fig. 1A and 4). Moreover, VirE2 protein levels approach wild-type levels in strains expressing virE2 from P_virB and virE1 from a separate promoter (Fig. 4, bottom). Use of the P_virB promoter therefore enabled us to uncouple translational effects of the virE1 gene from stabilizing effects of the VirE1 chaperone.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   Abundance of VirE2 truncation and i31 peptide insertion derivatives upon introduction of expression constructs into wild-type strain A348 (virE1+ virE2+), At12516 (virE1+), and KE1 (virE). All derivatives were expressed from PvirB. The line at the top represents wild-type virE2 encoding the 533-residue SSB. The two nuclear localization sequences are indicated by white boxes; the VirE1 interaction domain was identified previously by yeast two-hybrid analyses (Y2H), and the second proposed interaction domain was identified by stabilization analyses (ST). Protein contents relative to wild-type VirE2 levels in A348(pXZB27 [PvirB::virE2]) cells are indicated by symbols ranging from +++ (wild-type levels) to  (no detectable protein); the content of the 39-84 mutant was detected by chemiluminescence and not by histochemical staining. At the bottom the numbers below the virE2 line representation correspond to the sites of the i31 peptide insertions. The level of each VirE2::i31 mutant in the KE1 host strain is shown in the lane below each number in the immunoblot. WT, pXZB27 [wild-type virE2 expressed from PvirB]. The immunoblots at the bottom show the protein levels in VirE2 and i31 derivatives. Lane 1, A348(pXZB27); lane 2, PC3001(pXZB27); lane 3, At12516(pXZB27); lane 4, KE1(pXZB27); lane 5, A348; lane 6, KE1(pXZB761 [VirE2.39::31]); lane 7, At12516(pXZB761); lane 8, KE1(pXZB762 [VirE2.84::i31]); lane 9, At12516(pXZB762); lane 10, KE1(pXZB768 [VirE2.147::i31]); lane 11, At12516(pXZB768); lane 12, KE1(pXZB763 [VirE2.184::i31]); lane 13, At12516(pXZB763). See reference 52 for levels of additional i31 mutant proteins in virE1+ strains.

VirE1 self-associates in vivo. VirE1 self-activates transcription from a yeast GAL4 promoter when it is fused to the GAL4 DNA-binding domain, which prevents the use of the yeast two-hybrid screening method to assay for VirE1 self-association (45). To test for VirE1 self-association, we used  cI repressor fusion and pull-down assays. cI repressor protein binds as a dimer to operator sites in  early lytic gene promoters, and each monomer consists of an N-terminal DNA-binding domain and a C-terminal dimerization domain. The N-terminal domain binds DNA efficiently and represses transcription only when it is fused to the C-terminal dimerization domain or a heterologous dimerizing protein or peptide (26, 27).

View larger version (56K): [in this window] [in a new window]   FIG. 5.   VirE1 self-association. (A) cI repressor fusion assay. Cells were grown in the absence (left side) or presence (right side) of KH54 phage. Dimerization of cI derivatives was assessed on the basis of immunity to phage infection. AG1688 cells expressing the proteins indicated carried the following plasmids: cI'::VirE2, pZZ1; cI'::VirE1, pZZ2; cI' only and His-VirE1, pKH101 and pZZB4; cI' only, pKH101; cI full length, pZ157; and vector only, pZ150. (B) His-tag pull-down assay. The first three lanes contained the total soluble proteins loaded onto the Co2+ columns; the other three lanes contained the proteins eluted from the columns. His-E1, KE1(pZZB12); E1, KE1(pXZ426); His-E1, E1, KE1(pZZB13). The arrow and the arrowhead indicate the positions of His-VirE1 and VirE1, respectively. The positions of molecular mass markers (in kilodaltons) are indicated on the right.

View larger version (54K): [in this window] [in a new window]   FIG. 6.   Formation of purified His-VirE1. (A) His-VirE1 purification from A. tumefaciens KE1(pZZB12) by IMAC. Lane 1, total soluble proteins loaded onto the Co2+ columns; lane 2, His-VirE1 eluted from the column and visualized by Coomassie blue staining of the gel (arrow). The positions of molecular mass markers (in kilodaltons) are indicated on the left. (B) His-VirE1 complexes analyzed by native denaturing PAGE as described in the text. His-VirE1 complexes were detected by immunoblot analysis (arrows); the positions of molecular mass markers used for electrophoresis in the first dimension (in kilodaltons) are indicated at the top, and the positions of the molecular mass markers used for the second dimension (in kilodaltons) are indicated on the right. (C) His-VirE1 complexes analyzed by gel filtration chromatography. Purified His-VirE1 (1 mg) was fractionated as described in the text, and fractions were analyzed for the presence of His-VirE1 by immunoblot analysis (arrow). The positions of fractions corresponding to elution peaks for molecular mass markers (in kilodaltons) (see Materials and Methods) are indicated at the top.

VirE1-VirE2 complex formation in A. tumefaciens. To further characterize VirE1 complex formation in A. tumefaciens, total soluble proteins from wild-type strain A348 and strains expressing either virE1 or virE2 were fractionated by gel filtration chromatography (Fig. 7). VirE1 from strain At12516 (virE2) eluted in fractions whose sizes corresponded to molecular masses of ~30 to 40 kDa, which were slightly larger than the masses expected for a homodimer. This species might correspond to an aberrantly migrating homodimeric species or to a higher-order homo- or heteromultimer.

View larger version (26K): [in this window] [in a new window]   FIG. 7.   VirE1 and VirE2 complex formation in A. tumefaciens as analyzed by gel filtration chromatography. Total soluble proteins (5 mg) from the strains indicated on the right were applied to a gel filtration column and fractionated under identical conditions. The VirE proteins indicated on the left were detected by immunoblot analysis. The positions of elution peaks for molecular mass markers (in kilodaltons) are indicated at the bottom.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

VirE2 contributes several functions to the A. tumefaciens infection process upon export from the bacterium (9, 46, 48). In a plant cell, VirE2 multimerizes with itself, the T-DNA-VirD2 transfer intermediate, and, presumably, plant proteins to yield a complex that is competent for intracellular trafficking to the site of T-DNA integration, the plant nuclear genome (9-11, 14, 22, 38). Of considerable further interest, purified VirE2 recently was shown to self-assemble as single-stranded DNA-specific channels in artificial membranes; it has been proposed that these channels facilitate translocation of substrates such as the T-DNA across the plant plasma membrane or through the plant cell (18). In a bacterium, VirE2 must be processed as a secretion-competent molecule. Minimally, this processing pathway must (i) configure VirE2 as a substrate recognizable by the type IV T-DNA transfer machine and (ii) prevent formation of VirE2 aggregates or pore-forming channels and nonproductive interactions with other exported effector molecules. In this study, we demonstrated that virE1 and its product play central roles in the second of these requisite processing activities.

Our findings indicate the virE1 and the VirE1 secretion chaperone contribute both translationally and posttranslationally to VirE2 export. A regulatory role for virE1 at the level of virE2 translation is supported by analyses of VirE2 abundance and stability in various A. tumefaciens mutants and by results of our lacZ fusion studies. The first line of investigation showed that in the context of the native P_virE promoter, virE1 and virE2 must be cis coexpressed for synthesis of abundant levels of VirE2 and for successful VirE2 export. In the absence of virE1 or when virE1 is expressed in trans to virE2, VirE2 accumulates at low levels and is very unstable. Our lacZ fusion studies further showed that virE1 mediates efficient translation of virE2 without affecting transcription. Based on our results, we propose that the virE1 gene plays an important role in production of export-competent VirE2 under native gene expression conditions. The fact that virE1 and virE2 are separated by only four nucleotides is compatible with a translational coupling model in which virE1 translation facilitates ribosomal loading at the virE2 Shine-Dalgarno sequence. Similar translational regulatory mechanisms have also been proposed for some of the chaperone genes associated with type III trafficking systems (28, 36).

A regulatory role for virE1 at the level of translation is appealing because a proposed biological consequence of translational coupling is that it facilitates formation of productive interactions between two or more subunits of a protein complex (17). In support of such an activity for production of export-competent VirE2, our pulse-chase studies showed that cis expression of virE1 and virE2 significantly favors formation of a stable VirE1-VirE2 complex, resulting in an appreciably longer half-life for VirE2 than when each gene is expressed from a separate promoter (Fig. 4). Furthermore, we present evidence that VirE1 exerts its stabilizing effect via an interaction with the N terminus of VirE2. Translational coupling might therefore serve as a mechanism to facilitate rapid association of newly synthesized VirE1 with the N terminus of VirE2, possibly as soon as this region of the protein emerges from the translational machinery. We propose that this association both stabilizes VirE2 and prevents N-terminally mediated protein misfolding and aggregation.

We identified one nonnative expression system in which trans expression of virE1 and virE2 can still result in synthesis of functional (i.e., export-competent) VirE2. In this case, virE2 must be expressed from the P_virB promoter, whereas virE1 can be expressed from any functional promoter. We suggested previously that virE2 expression from P_virB might simply yield a critical threshold level of VirE2 protein required for export (52). However, VirE2 abundance is not always correlated with virulence, and furthermore, we now have shown that VirE2 produced from P_virB is very unstable in the absence of VirE1. We therefore suggest an alternative possibility, that the P_virB promoter provides a cis-acting targeting function, perhaps an mRNA signal, that is normally provided by the 5' untranslated region of P_virE and/or virE1. This proposed targeting function would serve to localize the virE2 translation machinery to the base of the T-DNA transfer channel to temporally and spatially coordinate virE2 translation and VirE2 translocation. Further comparative studies of the P_virE and P_virB promoter activities are needed to explore this and other possibilities.

In general, we found that the VirE1 chaperone stabilized VirE2 derivatives with an intact N terminus. If VirE1 exerts its stabilizing effect via a direct interaction with an N-terminal domain of VirE2, disruption of that domain should eliminate VirE1-mediated stability. This phenotype was observed with our class III mutation, VirE239-84. In contrast, the Ty39::i31 insertion mutation conferred VirE1-independent stability. We do not know the molecular basis for this finding, but one explanation is that an i31 insertion at this site interferes with formation of an N-terminal structure that targets the protein for degradation. If this is so, the Tyr39::i31 mutation might correspond to a functional mimic of chaperone binding to the N-terminal interaction domain, with the result that VirE2.39::i31 adopts a stable configuration independent of VirE1 binding. Interestingly, VirE1 is still required for VirE2.39::i31 export (52; data not shown). Furthermore, results of our gel filtration (Fig. 7) and yeast two-hybrid studies (52) suggest that a VirE1-VirE2.39::i31 complex is still assembled. Thus, while a VirE1 interaction is not needed for stabilization of this mutant protein, an interaction probably is required to prevent formation of nonproductive complexes or aggregates or to expose a substrate signal.

Intriguingly, VirE1 displays 21% identity and 35% similarity with the N terminus of VirE2 (Fig. 8). This level of sequence relatedness is also evident for the VirE1 and VirE2 proteins encoded by the pTiC58 plasmid of A. tumefaciens, as well as the pa megaplasmid of Rhizobium etli (data not shown). In addition, VirE1 and this region of VirE2 share a couple of physical features that may have biological importance. The VirE1 chaperones from different Ti or pa plasmid sources are highly acidic, with predicted pIs ranging from 4.7 to 5.2. Similarly, N termini of VirE2 from different Ti or pa plasmids are also highly acidic, with pIs in the range from 4.3 to 5.0, whereas full-length VirE2 proteins have predicted pIs of 6.1 to 6.6. VirE1 and the N terminus of VirE2 also possess hydrophilic N- and C-terminal regions and a hydrophobic central domain. Of further interest is the finding that residues 39 to 84 of VirE2, suggested by our studies to be important for a stabilizing interaction with VirE1, align with the C-terminal half of VirE1. Conversely, the C-terminal half of VirE1, which is predicted to adopt an amphipathic -helix, was shown previously to mediate interactions with VirE2 (45). Taken together, these observations suggest that the VirE1 interaction with the N terminus of VirE2 might be mediated by conserved structural motifs present on both proteins.

View larger version (18K): [in this window] [in a new window]   FIG. 8.   Alignment of VirE1 with the N terminus of VirE2. The numbers correspond to positions in the VirE2 sequence. Identical residues are indicated by a black background; similar residues are indicated by a gray background. Alignment was performed with the ClustalX Multalign program.

Our studies showed that VirE1 forms at least two types of complexes. VirE1 can self-associate, at least in the absence of VirE2, and VirE1 also assembles as a VirE1-VirE2 complex with an apparent molecular mass of ~70 to 80 kDa. The latter finding agrees with results of an early study which identified an ~80-kDa complex with single-stranded DNA-binding activity in extracts of virE⁺ cells but not virE cells (22). Based on the predicted sizes of VirE1 (7.5 kDa) and VirE2 (60 kDa), the evidence for VirE1 self-association, and the estimated size of the putative VirE1-VirE2 complex, we suggest that the VirE1-VirE2 complex is composed of one molecule of VirE2 and two molecules of VirE1. Formation of such a VirE1-VirE2 complex clearly seems to be important for preventing assembly of higher-order VirE2 complexes or aggregates (Fig. 7) (15). Whether formation of this complex also is a prerequisite for delivery of the secretion substrate to the transfer channel awaits further study. It is intriguing to speculate that the VirE1 dimer interacts dynamically with VirE2, associating with SSB and then dissociating upon successful presentation of the substrate to the transfer channel. Whether a given VirE1 monomer or dimer can repetitively interact with newly synthesized VirE2 monomers destined for export is unknown; what minimally must occur is that VirE1 disengages from VirE2 prior to export.

Of final note, the multiple roles identified for virE1 and the VirE1 secretion chaperone for VirE2 secretion by this type IV transfer system are analogous to functions ascribed to the secretion chaperones of flagellar and type III secretion pathways (1, 6, 28, 36). Recent work on different members of the type III secretion family has demonstrated that there is considerable variation in the relative importance of mRNA and amino acid targeting signals, the contributions of chaperones for stabilization of effector proteins, and the number and locations of chaperone-binding domains on the effectors (36). Whether secretion chaperones are required for export of all type IV secretion substrates and whether these chaperones (or their cognate genes) provide targeting functions for delivery of the secretion substrates to the cognate type IV translocases are exciting areas for future study.