TEXT

Top Abstract Text References

The interaction of the gram-negative bacterium Xanthomonas campestris pv. vesicatoria (hereafter X. campestris) with its host plants pepper and tomato is controlled by hrp (hypersensitive reaction and pathogenicity) genes (3), some of which are predicted to encode components of a type III protein secretion system (1, 5). hrp gene expression is suppressed in complex medium (NYG [5 g of peptone, 3 g of yeast extract, and 20 g of glycerol per liter]), but is induced in planta and in the synthetic XVM2 medium (16, 19) and is regulated by two genes, hrpX and hrpG. hrpX encodes a protein of the AraC family and activates transcription of the operons hrpB to hrpF (18). Expression of hrpX and hrpA depends on hrpG, which encodes a putative response regulator protein of two-component signal transduction systems (20). The predicted HrpG protein is most similar to proteins of a subclass containing Escherichia coli OmpR (4, 22) and Agrobacterium tumefaciens VirG (21). The X. campestris hrpG gene is expressed at a low level in NYG medium but requires hrp gene-inducing conditions for the activation of downstream genes (20). To achieve hrp gene expression in noninducing media, we performed a random mutagenesis of hrpG and selected for mutations rendering expression of all hrp genes constitutive in NYG medium.

First, a suitable reporter plasmid, pPCspec, was constructed by fusing a 320-bp fragment encompassing the hrpG-regulated X. campestris hrpC promoter to a promoterless aadA gene in pLAFR6 (2). The aadA gene, which confers resistance to spectinomycin, was amplified by PCR from the omega cassette (14) by using oligonucleotides 161 (5' GGGGAGCTCAAACAAAGTTAAACATC 3') and 162 (5' GGGGAATTCATTATTTGCCGACTACC 3') (restriction sites used for cloning are underlined). For mutagenesis of hrpG, plasmid pFG72, which carries hrpG on a 1.7-kb BamHI-KpnI fragment in the low-copy plasmid pUFR043 (gift from D. Gabriel, University of Florida, Gainesville), was transformed into the E. coli mutator strain Epicurian Coli XL1-Red (Stratagene, La Jolla, Calif.). Plasmid DNA was isolated at different time points and transformed into E. coli DH5. Transformants were conjugated en masse with X. campestris 85-10 (11) carrying pPCspec and plated on NYG agar containing 400 µg of spectinomycin per ml. Eleven spectinomycin-resistant transconjugants, obtained at a frequency of 10⁴, were analyzed further. To test the effect of the mutated pFG72 derivatives on hrp promoter activity, they were first introduced into X. campestris 85-10 carrying pPC490, an hrpC promoter-uidA fusion in pL6GUSB (2). -Glucuronidase (GUS) activities were determined after growth of bacteria in NYG and in hrp gene-inducing XVM2 medium. All 11 mutated hrpG plasmids (designated pFG72-1 to pFG72-11) activated the hrpC promoter in NYG. To determine the activity of all other known hrp promoters in presence of the mutated hrpG gene, pFG72-1 (hrpG herein referred to as hrpG*) was conjugated into X. campestris strains carrying hrp promoter-uidA fusion plasmids pPA2, pPB1, pPD3, pPFI42, pPG1, and pPX2 (2, 6, 18-20), which express the uidA gene under the control of the hrpA, hrpB, hrpD, hrpF, hrpG, and hrpX promoters, respectively. hrpE promoter activity was determined by using pXV4::E525 (18). As shown in Fig. 1A, additional copies of the wild-type hrpG plasmid pFG72 resulted in somewhat higher GUS activities of bacteria grown in NYG medium compared to 85-10. However, in the presence of hrpG*, all hrp promoters, except for the hrpG promoter, exhibited 20- to 1,000-fold higher activities in NYG compared to wild-type strain 85-10. In XVM2 medium, GUS activities in the presence of hrpG* were in the same order of magnitude as those with additional copies of the wild-type gene and were 10-fold higher than those for strains containing only the genomic copy (Fig. 1B).

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Activity of X. campestris hrp promoters in the presence of wild-type hrpG and hrpG*. Wild-type strain 85-10 carrying plasmid-borne transcriptional hrp promoter-uidA fusions and no additional hrpG, plasmid-borne hrpG (pFG72), and hrpG* (pFG72-1), respectively, were grown for 16 h in NYG (A) and XVM2 (B). For technical reasons, hrpE promoter activity was determined with hrpG* expressed in pDSK600. GUS assays were performed as described previously (16). Specific GUS activities are the average of two experiments and are graphically displayed with a logarithmic scale (1 U = 1 nmol of 4-methyl-umbelliferyl--D-glucuronide released per bacterium per min).

DNA sequencing of hrpG* in pFG72-1 revealed a single G-to-A mutation at position 130 of the hrpG coding sequence, leading to an E44K exchange in the translation product. Since this mutation destroys a SacI restriction site, the remaining 10 pFG72 derivatives were tested for presence or absence of this site. While eight of them had lost the SacI site and probably carry the same mutation as pFG72-1, the SacI site was present in two plasmids, pFG72-2 and pFG72-3. Sequence analysis revealed a G-to-A exchange at position 595 of hrpG in pFG72-2, leading to a D199N substitution, and an A-to-G mutation at position 581 in pFG72-3, resulting in an H194R exchange. All three hrpG mutations had identical effects on the activity of the hrpG-dependent promoters (data not shown).

Only a few known mutations in OmpR-type transcriptional activators render the protein constitutively active. These mutations have all been mapped to the N-terminal domain of a given protein (Table 1 and Fig. 2). For example, in the Agrobacterium irG protein, N54E or I106L substitutions (N80 or I132 in Fig. 2) resulted in constitutive vir gene expression (12). A D55E substitution in the E. coli OmpR protein rendered it active independent of the sensor kinase EnvZ (8). Whether the E44K mutation in the N-terminal domain of HrpG mimics the predicted phosphorylation of the D60 residue in HrpG (20) is a matter of speculation. Interestingly, the other two mutations in X. campestris HrpG identified here, H194R and D199N, are located in the C-terminal domain of the protein. In contrast to HrpG, substitutions at corresponding positions in E. coli OmpR (E193K and E198K) and PhoB (D192G) led to loss of function (9, 13, 15). It has been shown for OmpR that the region from amino acid 192 to amino acid 199 forms a loop exposed at the protein surface, which has been proposed to control RNA polymerase activity by direct interaction with its -subunit (10). The H194R and D199N substitutions in HrpG might, therefore, lead to transcriptional activation due to alteration of contact with RNA polymerase.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Protein sequence alignment of X. campestris HrpG and selected homologs. (A) Only the relevant region in the N terminus of proteins is shown. Highly conserved amino acids are boxed in black. At, A. tumefaciens; Ec, E. coli; Xcv, X. campestris pv. vesicatoria. (B) Only the relevant part of the C-terminal region of proteins is shown. Structural and functional elements are indicated above the sequence. The positions of -sheets, recognition helices, and -loops are based on structural data for OmpR (10). The asterisks mark the positions of three independent substitutions in HrpG; the dot indicates the position of the conserved aspartate residue, which is predicted to be phosphorylated (10). Database accession numbers: A. tumefaciens VirG, SWISS-PROT P06664; E. coli OmpR, SWISS-PROT P03025; E. coli PhoB, SWISS-PROT P08402; X. campestris pv. vesicatoria HrpG, GENBANK U57625.

Because hrpG is a key regulator for the interaction of X. campestris with plants, we wondered whether the three mutations in hrpG alter the plant reaction after infection. To avoid interference of mutated hrpG with the wild-type gene, a 950-bp SacI fragment spanning most of the hrpG coding region was deleted. The mutation was introduced into the chromosome of X. campestris wild-type strain 85-10 using suicide plasmid pOK1 (7), resulting in mutant 85G. Transconjugants of 85G harboring pFG72, pFG72-1, pFG72-2, or pFG72-3 were inoculated into leaves of susceptible (ECW) and resistant (ECW-10R) pepper plants (11). Symptoms caused by wild-type and mutated hrpG bacteria in ECW were indistinguishable. On pepper ECW-10R, strain 85-10 induces the hypersensitive reaction (HR), a rapid, localized cell death reaction, approximately 8 to 10 h after inoculation. Bacteria expressing hrpG* induced a complete, confluent collapse (earliest macroscopically visible reaction) on ECW-10R as soon as 4 h after inoculation (Fig. 3A). No difference was observed between hrpG* and mutated hrpG present in pFG72-2 and pFG72-3. The minimal concentration of hrpG* bacteria needed to induce the HR on ECW-10R was two- to fourfold lower than those for 85-10 and 85G(pFG72). Thus, hrpG* enhances the HR induction with respect to both time and efficiency. We then tested whether a stronger HR was also observed in the nonhost plant tobacco, which is resistant to X. campestris. Different densities of bacterial suspensions were inoculated into leaves of Nicotiana tabaci cv. Xanthi plants (Fig. 3B). Inoculation of the hrpG* strain induced a confluent HR 1 day after inoculation. In contrast, strains 85G(pFG72) and 85-10 induced only a weak chlorotic and necrotic reaction 2 to 3 days after infection.

View larger version (59K): [in this window] [in a new window]   FIG. 3.   Effect of hrpG* on HR induction on resistant plants. (A) HR induction on pepper line ECW-10R after inoculation with 85G(pFG72) and 85G(pFG72-1). Bacteria were grown overnight on NYG agar and resuspended in H2O. Serial twofold dilutions of bacterial suspensions, from 5 × 108 to 4 × 106 CFU/ml (no. 1 to 8), were infiltrated into the intercellular space of a fully expanded leaf of an 8-week-old plant. The photograph of the underside of the leaf was taken 6 h after inoculation. (B) Reaction of Nicotiana tabacum cv. Xanthi after inoculation with different X. campestris strains. Strains 85-10, 85G, 85G(pFG72), and 85G(pFG72-1) were inoculated at bacterial densities of 2 × 107 (A), 2 × 108 (B), 8 × 108 (C), and 2 × 109 (D) CFU/ml. The photograph was taken 24 h after inoculation.

To reexamine the effect of hrpG* on susceptible plants in a bacterial genetic background resembling the natural situation, the wild-type copy was replaced by hrpG* in strains 85-10 and 82-8 (11) with pOK1 (7) to generate 85* and 82*. Chromosomal hrpG* accelerated the HR induction, as described above for hrpG* expressed from a plasmid. However, there was a clear difference in timing and intensity of disease symptoms on the susceptible plant, ECW. Water-soaked lesions, typical early disease symptoms, appeared much earlier (49 h) and were stronger than with the wild-type strains (65 to 72 h). These effects were not due to increased bacterial growth in the presence of hrpG*, as determined by bacterial growth curves in planta (data not shown).

In conclusion, we identified three mutations in the key hrp regulatory gene hrpG rendering expression of downstream genes constitutive. The mutations residing in the C-terminal domain of HrpG are particularly interesting, because similar mutations in homologous proteins lead to inactivation. The amplifying effect of mutated hrpG on both disease symptoms and cell death is intriguing and might be indicative of more efficient Hrp type III protein delivery into the plant tissue.