INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Agrobacterium tumefaciens is a soil bacterium which infects plant wound sites and induces tumor formation. The bacterium harbors a large tumor inducing plasmid (Ti plasmid) encoding virulence genes and transfer DNA (T-DNA). The function of the virulence genes is the processing and transfer of the T-DNA from the Ti plasmid into susceptible plant cells, with subsequent integration into the host genome (for recent reviews, see references 27 and 56). Located on the T-DNA are genes that direct the biosynthesis of the plant growth regulators auxin and cytokinin in the infected cells (1, 49). The synthesis of these plant growth regulators results in a rapid, uncontrolled cell division leading to production of a characteristic tumor at the site of infection. In addition, the T-DNA contains genes for the synthesis of a unique class of compounds called opines, which A. tumefaciens can utilize as a carbon and energy source (36).

In A. tumefaciens, the expression of virulence genes is under the control of a two-component regulatory system comprised of VirA and VirG (45, 52). VirA is an inner membrane histidine kinase (34, 54) which autophosphorylates in response to certain phenolic compounds released from wounded plants (44) with subsequent transfer of the phosphate moiety to the response regulator VirG (16, 22, 24). Once phosphorylated, VirG activates transcription from promoters containing a specific 12-bp sequence called the vir box, which is present in the promoters of all vir genes (25, 40). This expression is augmented by the presence of certain monosaccharides (5, 43) and an acidic pH (32), which is characteristic of plant wound sites. A periplasmic sugar-binding protein, ChvE, which is highly homologous to glucose-binding protein of Escherichia coli, interacts with the periplasmic portion of the VirA molecule in the presence of certain monosaccharides, including glucose and arabinose (2, 15). This interaction alone does not induce vir gene expression, but it sensitizes the VirA molecule to the phenolic inducers.

While A. tumefaciens is a potentially serious plant pathogen, the main interest in this organism is due primarily to its ability to transform plant cells. Researchers have developed delivery systems based on T-DNA transfer to engineer new traits into selected plant species (14). However, the exact mechanism of T-DNA transfer is still not well understood. The ability to use E. coli as a heterologous host in which to study the regulation of A. tumefaciens virulence genes and the mechanism of T-DNA transfer would constitute an excellent model system given the degree of characterization at both the biochemical and the genetic levels and the relative ease of genetic manipulation. However, all previous attempts to reconstitute inducer-mediated vir gene expression in E. coli have not been successful. Previously, we reported the identification of the rpoA gene, encoding the alpha subunit of RNA polymerase from A. tumefaciens, and that it was required for transcription of a virB promoter fused to lacZ (virBp::lacZ) fusion in E. coli using a constitutively active VirG mutant [VirG(Con)] which can activate vir gene expression independent of VirA and inducers (31). In this study, we report the successful reconstitution of inducer-dependent expression of a virBp::lacZ fusion in E. coli utilizing lac promoter-driven virA and virG. Effects of various environmental conditions on virulence gene activation in E. coli are also described. Furthermore, by the gene fusion approach, the C-terminal domain of RpoA from A. tumefaciens is shown to be required for interaction with the transcriptional activator VirG.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and growth conditions. The bacterial strains used in this study are listed in Table 1. Bacterial strains were grown in either Luria-Bertani (LB) medium, mannitol glutamate-Luria salts (MG/L) medium (50), or induction medium containing either 55 mM glucose, mannitol, or glycerol (53). When appropriate, the medium was supplemented with ampicillin (100 µg/ml), kanamycin (50 µg/ml), gentamicin (5 µg/ml), and tetracycline (20 µg/ml). For the growth of DH5 in induction medium, Casamino Acids (Difco) and thiamine were added to final concentrations of 0.1% (wt/vol) and 12 µg/ml, respectively. For M182 and M182crp grown in induction medium, arginine and thiamine were added to final concentrations of 20 and 12 µg/ml, respectively. For visualization of expression of the virBp::lacZ fusion on plates, 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) and isopropyl--D-thiogalactopyranoside (IPTG) were included at final concentrations of 75 µg/ml and 200 µM, respectively. The virulence gene inducers acetosyringone, acetovanillone, hydroxyacetophenone, and syringealdehyde (Sigma) were added where indicated at a final concentration of 200 µM.

Plasmid constructs and DNA manipulations. The plasmids used in this study are listed in Table 1. Plasmids pZL2 and pECH are wild-type rpoA constructs from A. tumefaciens A136 (ropA_At) and E. coli MC4100 (rpoA_Ec), respectively (31). Plasmid pAD8 encodes N-terminal His-tagged RpoA_At with the eight C-terminal amino acids deleted and was constructed by PCR amplification from pPS1.3 using the primers 5'-GGA AGG ATC CAA GAT GAT TCA GAA GA-3' and 5'-TTA GAG ATC TTC GAT GTT CTC AGG CGG-5', followed by digestion with BamHI and ligation into pQE30 that was digested with BamHI and SmaI. Plasmid pEA8 encodes N-terminal His-tagged RpoA_Ec plus the eight C-terminal amino acids from RpoA_At fused to the C terminus. This plasmid was generated by inserting 24 nucleotides encoding the eight amino acids into the pECH by site-directed mutagenesis using oligonucleotide 5'-TAG AGG ATC GGG TTA GTA TTG GTC TTC GTA ACG CTT TGC CTC GTC AGC GAT GCT-3'. Plasmid pADC encodes N-terminal His-tagged RpoA_At with the C-terminal 94 amino acids deleted and was generated by PCR amplification of pPS1.3 using primers 5'-GGA AGG ATC CAA GAT GAT TCA GAA GA-3' and 5'-TTA AGA TCT CGA TTC TTC TGC TTC CTT CTG-3'. The PCR product was digested with BamHI and ligated into pQE30 that was digested with BamHI and SmaI. The C-terminal domain of rpoA_Ec was amplified with primers 5'-GAA AGG ATC CAA ACC AGA GTT CGA TCC-3' and 5'-CCT GGG ATC CGG TTA CTC GTC AGC-3' using pECH as a template, and the PCR product was cleaved with BamHI and cloned into the BglII site of pADC, generating pANEC. To generate a hybrid RpoA containing the N-terminal domain from E. coli and the C-terminal domain from A. tumefaciens, each domain was PCR amplified. For PCR amplification of the N-terminal domain of RpoA_Ec, we used pECH as a template and primers that contain a BamHI site (5'-CCA AAG AGA GGA TCC AAT GCA GGG-3') and a KpnI site (5'-CGG GGT ACC CTC TTC TTT CAC TTC AGG CTG ACG-3'), followed by digestion with BamHI and KpnI and ligation into pQE31, to generate pEN. To PCR amplify the C-terminal domain of RpoA_At, pPS1.3 was used as a template and oligonucleotides (5'-CGG GGT ACC GAA CTC GCG TTC AAC CCG GCG-3' and 5'-CCT GGA TCC TGC AGA TGA CTT ATC TG-3') were used as primers. The PCR product was digested with KpnI and ligated with pEN that was digested with KpnI and SmaI to generate pENACN.

Virulence gene induction assays. For reconstitution of virulence gene expression in E. coli, plasmid construct pSL204 was introduced into various E. coli strains by electroporation. Constructs containing either lacp-driven A. tumefaciens rpoA (pPS1.3 or pHO98) or non-lacp-driven rpoA (pPS1.3R) were then introduced into these strains and initially screened on MG/L medium (pH 6.0) containing appropriate antibiotics, 200 µM IPTG, 75 µg of X-Gal per ml, and 200 µM acetosyringone.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

C-terminal RpoA of A. tumefaciens is required to interact with the VirG protein. RpoA of E. coli is known to be composed of N- and C-terminal domains (12, 55). RpoA of A. tumefaciens shares 62% sequence similarity and 51% sequence identity with that of E. coli (31). An obvious difference between them is an extra eight amino acids in the C-terminal tail in the RpoA of A. tumefaciens. To determine which domain of the A. tumefaciens RpoA is required for interaction with the VirG protein, a series of protein fusion constructs were generated, and the abilities of the hybrid rpoA constructs to activate transcription of virBp::lacZ in a VirG(Con)-dependent manner in E. coli MC4100 and pSY215 were evaluated (Fig. 1A). When introduced into DH5, all of the fusion constructs clearly overproduced protein bands with predicted sizes (data not shown), but only pENACN was able to provide significant expression of the virBp::lacZ fusion (Fig. 1B). The wild-type His-tagged A. tumefaciens RpoA construct pZL2 exhibited reduced expression of virBp::lacZ relative to both pPS1.3 (59% reduction) and pENACN (48% reduction). Deletion of 8 and 94 amino acids from the C terminus of A. tumefaciens RpoA (pAD8 and pADC, respectively) resulted in a complete loss of virBp::lacZ expression. Additionally, addition of the eight amino acids from the A. tumefaciens RpoA C terminus to the C terminus of E. coli RpoA (pEA8) also failed to activate transcription of virBp::lacZ. These results indicate that the C-terminal domain of the A. tumefaciens RpoA is required to interact with the VirG protein.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Abilities of various rpoA fusion constructs to activate vir gene expression in E. coli. (A) Schematic representation of gene fusions between rpoA of A. tumefaciens and E. coli. Hatched boxes represent RpoAAt, while shaded boxes represent RpoAEc. (B) -Galactosidase activities of E. coli MC4100 harboring pSY215 and one of the rpoA constructs. Cultures were grown for 18 h in LB medium at pH 7.0 without acetosyringone. -Galactosidase activity values are averages from three replicates.

Influence of rpoA copy number on expression of virBp::lacZ in E. coli. We have constructed pSL204, which contains lac promoter-driven virA and virG, as well as a virBp::lacZ reporter gene fusion. This construct is compatible with the rpoA_At clones pHO98 and pPS1.3 (31). Transformation of DH5 containing pSL204 with pHO98 or pPS1.3 resulted in a blue color on MG/L and induction medium plates containing X-Gal and acetosyringone. No blue color was observed on identical medium in the absence of acetosyringone, indicating inducer-dependent expression of the virBp::lacZ fusion. In contrast, transformants containing the non-lacp-driven A. tumefaciens rpoA construct pPS1.3R did not exhibit any blue color on MG/L or induction medium containing X-Gal and acetosyringone.

View this table: [in this window] [in a new window]   TABLE 2.   Effect of rpoA copy number on wild-type expression of virBp::lacZ in E. coli DH5

lacZ'

View larger version (15K): [in this window] [in a new window]   FIG. 2.   Effect of cAMP on the expression of virulence genes in E. coli. Strain DH5 harboring both pSL204 and pPS1.3 was grown in induction medium containing 55 mM glucose. The medium was supplemented with 200 µM acetosyringone, 30 mM cAMP, or both. -Galactosidase activity values are the average of three replicates.

Effect of other virulence gene inducers on the expression of virBp::lacZ in DH5. It is known that vir genes of different A. tumefaciens strains respond to different sets of phenolic compounds depending on the origin of the VirA molecules. The virA of pSL204 is derived from strain A348, whose vir gene expression is activated by acetosyringone, acetovanillone, and syringealdehyde but not by hydroxyacetophenone (29). The ability of DH5 containing pSL204 and pPS1.3 to express virBp::lacZ in response to these inducing compounds was evaluated. DH5 containing pSL204 with or without pPS1.3 were grown in induction medium amended with glycerol and with one of the following inducers: acetosyringone, acetovanillone, hydroxyacetophenone, or syringealdehyde. Significant levels of virBp::lacZ expression were obtained with DH5 containing pPS1.3 when acetosyringone and syringealdehyde were present. Low-level expression was detected in the presence of acetovanillone, and no induction was seen when hydroxyacetophenone was present (Table 4). Similarly, a Ti plasmidless A. tumefaciens strain, A136, harboring pSL204 responded strongly to acetosyringone and syringealdehyde and weakly to acetovanillone, but not at all to hydroxyacetophenone (Table 4).

View this table: [in this window] [in a new window]   TABLE 4.   Wild-type expression of virBp::lacZ in DH5 in response to other inducers

Inducer-dependent virBp::lacZ expression in DH5 is affected by pH, temperature, and choice of media. To assess whether the expression of the virBp::lacZ in E. coli is responding to the same environmental signals that affect vir expression in A. tumefaciens, the effect of variations in temperature and pH were examined. When DH5 cells containing pSL204 with or without pPS1.3 were grown in induction medium containing acetosyringone and glycerol at 28 or 37°C, high-level expression was observed at 28°C but not at 37°C (Fig. 3). When the same strains were assayed for virBp::lacZ expression in induction medium at pH 6.0 and 7.0, there was approximately a 36% reduction in -galactosidase activity when the bacterium was grown at pH 7.0 compared to when it was grown at pH 6.0 (Fig. 4).

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Effect of temperature on the expression of the virB::lacZ fusion in E. coli. Strain DH5 harboring both pSL204 and pPS1.3 was grown for 18 h in induction medium (pH 6.0) containing 55 mM glycerol at either 28 or 37°C. Acetosyringone (AS) was added where indicated at a final concentration of 200 µM. -Galactosidase activity values are average from three replicates.

View larger version (15K): [in this window] [in a new window]   FIG. 4.   Effect of pH on the expression of virB::lacZ expression in E. coli. Strain DH5 containing both pSL204 and pPS1.3 was grown for 18 h in induction medium, at pH 6.0 or 7.0, containing 55 mM glycerol with (+) or without () 200 µM acetosyringone (AS). -Galactosidase activity values are averages from three replicates.

View larger version (12K): [in this window] [in a new window]   FIG. 5.   Effect of growth media on the expression of virB::lacZ expression in E. coli. Strain DH5 containing both pSL204 and pPS1.3 was grown for 18 h in induction medium containing 55 mM glycerol (IM) or MG/L or LB medium at pH 6.0. Acetosyringone (AS) was added where indicated at a 200 µM final concentration. -Galactosidase activity values are averages from three replicates.

Effect of chvE on virBp::lacZ expression. To evaluate whether chvE from A. tumefaciens can increase virBp::lacZ expression in E. coli, we introduced pTCR2.4, which contains lacp-driven chvE and lacp-driven rpoA of A. tumefaciens into DH5 containing pSL204 and compared virBp::lacZ expression with that of DH5 containing pSL204 and pPS1.3. To avoid the inhibitory effect of the glucose on the lac promoter driving the expression of virA, virG, and rpoA, 30 mM cAMP was added to the induction medium. As shown in Fig. 6, the virBp::lacZ expression level was highest in the presence of both ChvE and glucose, suggesting that similar interactions between VirA and sugar-bound ChvE also occur in E. coli, which enhances the sensitivity of VirA to the acetosyringone signal. As in A. tumefaciens, the presence of glucose and ChvE sensitizes the VirA molecule, resulting in a high level of virBp::lacZ expression in the presence of a low concentration of acetosyringone (5 µM), whereas in the absence of ChvE, virBp::lacZ responded only to a high level of acetosyringone (200 µM) (Fig. 6). The fact that no glucose effect was detected in the absence of chvE indicates that the glucose-binding protein of E. coli is incapable of interacting with the VirA, although there is significant amino acid sequence homology between the ChvE and the E. coli glucose-binding protein (5).

View larger version (13K): [in this window] [in a new window]   FIG. 6.   Effect of sugar and ChvE on the expression of virB::lacZ in E. coli. Strain DH5 harboring pSL204 and either pZL2, which contains rpoA of A. tumefaciens, or pTC2.4, which contains both lac-promoter driven rpoA and chvE. Bacterial cultures were grown for 18 h in induction medium supplemented with 30 mM cAMP and either 55 mM glucose or glycerol. Acetosyringone (AS) was added at either a 5 or a 200 µM final concentration. -Galactosidase activity values are averages from three replicates.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

For many years it was believed that the specificity of RNA polymerase for promoters was determined by various sigma factors. Beginning in the early 1990s it was demonstrated that rpoA, encoding the alpha subunit of RNA polymerase, plays an essential role in the transcription of many operons in E. coli controlled by transcriptional regulators such as FNR (51), GalR (11), MarA (20), MerR (9, 26), MetR (18), OxyR (48), OmpR (17), Rob (21), SoxS (19), UhpA (39), and TyrR (28). Recently, it has become clear that RpoA also plays an essential role in transcription in other bacterial systems such as Agrobacterium tumefaciens (31), Bacillus subtilis (35, 38), Bordetella pertussis (6, 46), Rhodospirullum rubrum (13), Pseudomonas putida (33), and Vibrio fischeri (47).

Similar to results obtained for other transcriptional regulators in other bacterial systems, the results of the domain swap indicate the importance of the C-terminal domain of A. tumefaciens rpoA in VirG-mediated transcription from virB promoters (35, 46). The ability of pENACN to allow virBp::lacZ expression indicates that the residues essential for transcription are located in the C-terminal 89 amino acids of A. tumefaciens RpoA. Interestingly, deletion of eight residues from the C terminus of A. tumefaciens RpoA abolished virB::lacZ transcription, indicating an essential role for these residues. However, when these eight residues were added to the C terminus of E. coli RpoA, we did not obtain any transcription, indicating that while this region is essential for transcription, other determinants in the C-terminal region are also required. As expected, deletion of the C-terminal 94 amino acids from A. tumefaciens RpoA (pADC) or replacement with the C-terminal domain of E. coli RpoA (pANEC) failed to result in any expression of the virBp::lacZ fusion. The reduced virBp::lacZ expression obtained with the wild-type His-tagged A. tumefaciens RpoA indicates that the N-terminal histidine residues are interfering with RpoA function. Interestingly, pENACN gave intermediate levels of virBp::lacZ expression even though it contains the N-terminal domain of E. coli RpoA from pECH, which contains an N-terminal His tag plus the C-terminal domain of A. tumefaciens. This suggests that an N-terminal His tag negatively affects A. tumefaciens RpoA function to a greater degree than for E. coli RpoA, at least at the virB promoter.

Another purpose of this study was to determine if we could reconstitute wild-type, inducer-dependent A. tumefaciens virulence gene expression in the heterologous host E. coli. In a previous study (31), we were unable to detect acetosyringone-mediated expression of virBp::lacZ in E. coli MC4100 using the constructs pHO98 (lacp::rpoA) and pSL107 (lacp::virA/G, virBp::lacZ). The pSL107 plasmid used in that study was not stable and resulted in high-frequency spontaneous mutations, probably due to the presence of two replicons (ColE1 and IncP origin) (data not shown). In response to this finding, we constructed pSL204, which also contains lacp-driven virA and virG, virBp::lacZ, and in contrast to pSL107, only the IncP origin. This allowed us to utilize the high-copy-number plasmid pPS1.3 in our attempts to reconstitute inducer-dependent virBp::lacZ expression in E. coli. Introduction of either pHO98 or pPS1.3 into DH5 containing pSL204 resulted in significant expression of the virBp::lacZ fusion in the presence of acetosyringone. The higher level of expression observed with pPS1.3 most likely is the result of an increase in the copy number of rpoA compared to that in pHO98.

The expression of virulence genes in A. tumefaciens is known to be affected by changes in temperature (23) and pH (32), with 28°C and pH 5.5 being optimal. The thermosensitive nature of virulence expression in A. tumefaciens is due to a reversible inactivation of VirA protein (23). Similar to what is observed in A. tumefaciens, we have also seen virBp::lacZ expression at 28°C, while no expression was observed at 37°C. In addition, we demonstrated that virBp::lacZ expression in DH5 is also affected by pH, with a 36% decrease in expression at pH 7.0 compared with that seen at pH 6.0. We did not use pH 5.5, since the growth of DH5 was adversely affected by this pH. Taken together, these results reinforce our conclusion that we have obtained inducer-dependent virBp::lacZ expression in E. coli and that the VirA-VirG signaling mechanism appears to be functioning.

Our observation that the choice of sugar greatly influences virBp::lacZ expression in DH5 is indicative of catabolite repression. To confirm this, we utilized the isogenic E. coli strains M182 and M182crp (3, 4, 8), which differ only in the presence of the cAMP receptor gene crp. Our observation that virBp::lacZ expression in M182crp was dramatically reduced in induction medium amended with glycerol confirms the requirement of CRP. Since the addition of cAMP alone in the induction medium did not induce virBp::lacZ expression, the CRP may not directly activate the virB promoter. The fact that both cAMP and acetosyringone are needed for virBp::lacZ expression suggests that the CRP is required for the expression of virA, virG, and rpoA that are under the control of the lac promoter, a well-characterized promoter that requires activation by CRP. An alternative promoter that is not influenced by CRP or any other factors might be needed to further clarify this.

One unresolved question in virulence gene expression is the exact mechanism of sensing of phenolic inducers by the VirA-VirG system. The two possibilities are either a direct binding of the inducer by VirA or an intermediate receptor protein that binds the inducer and then interacts with VirA. Although the genetic evidence supporting direct binding of inducers by VirA has been reported (29, 30), all attempts to demonstrate direct binding by VirA have been unsuccessful. We were able to demonstrate significant expression of virBp::lacZ in response to acetosyringone, syringealdehyde and, to a lesser extent, acetovanillone but not to hydroxyacetophenone. This result provides further evidence that inducers may be recognized directly by VirA. However, we cannot rule out the possibility that E. coli may contain homologues of A. tumefaciens receptors for the phenolic inducers. In the case of sugar effect, given the fact that E. coli encodes a ChvE homologue, the profound effect of ChvE and sugar on virBp::lacZ expression was somewhat unexpected. Apparently, VirA can be sensitized only by sugar- bound ChvE but not by the glucose-binding protein of E. coli which shares significant amino acid homology with ChvE (5).

Taken together, these results provide conclusive evidence that we have indeed reconstituted inducer-dependent vir gene expression in E. coli. However, the level of expression in E. coli is still significantly less than what is usually observed in A. tumefaciens (29, 30). One possible explanation for the relatively low induction in E. coli may be the E. coli sigma factors, which are inefficient in recognizing the vir gene promoters. It is conceivable that E. coli sigma factors may have a lower affinity for the virB promoter than sigma factors from A. tumefaciens. Although the vegetative sigma factor from A. tumefaciens has been identified (42), it is unclear whether this or an alternative sigma factor is involved in vir gene transcription. It is also likely that the relatively low expression is a consequence, at least in part, of RNA polymerase-containing endogenous E. coli RpoA subunits. One approach to resolve this would be to engineer E. coli RpoA such that it is able to activate virBp::lacZ expression. The results of our domain swap experiments have demonstrated that it is possible to obtain a hybrid RpoA molecule that will function at vir promoters, although the effect of such hybrids on the transcription of E. coli genes is unclear. In a recent study, Carbonetti et al. (6) reported that, in B. pertussis, overexpression of RpoA_Bp reduced the transcription of Bvg-activated virulence genes. It was suggested that excess RpoA_Bp was able to interact with Bvg and prevent interaction with virulence promoters. Since we previously reported that RpoA from A. tumefaciens, but not from E. coli, is able to interact with VirG (31), it may be that the constructs we are using result in cellular levels of A. tumefaciens RpoA that bind VirG and prevent optimal transcription. Although the current level of expression in E. coli is relatively low, it may still be sufficient to begin studies on A. tumefaciens processes in E. coli such as T-DNA transfer. Studies are under way to address these possibilities.