ABSTRACT
Monosaccharides capable of serving as nutrients for the soil bacterium Agrobacterium tumefaciens are also inducers of the vir regulon present in the tumor-inducing (Ti) plasmid of this plant pathogen. One such monosaccharide is galacturonate, the predominant monomer of pectin found in plant cell walls. This ligand is recognized by the periplasmic sugar binding protein ChvE, which interacts with the VirA histidine kinase that controls vir gene expression. Although ChvE is also a member of the ChvE-MmsAB ABC transporter involved in the utilization of many neutral sugars, it is not involved in galacturonate utilization. In this study, a putative tripartite ATP-independent periplasmic (TRAP) transporter, GaaPQM, is shown to be essential for the utilization of galacturonic acid; we show that residue R169 in the predicted sugar binding site of the GaaP is required for activity. The gene upstream of gaaPQM (gaaR) encodes a member of the GntR family of regulators. GaaR is shown to repress the expression of gaaPQM, and the repression is relieved in the presence of the substrate for GaaPQM. Moreover, GaaR is shown to bind putative promoter regions in the sequences required for galacturonic acid utilization. Finally, A. tumefaciens strains carrying a deletion of gaaPQM are more sensitive to galacturonate as an inducer of vir gene expression, while the overexpression of gaaPQM results in strains being less sensitive to this vir inducer. This supports a model in which transporter activity is crucial in ensuring that vir gene expression occurs only at sites of high ligand concentration, such as those at a plant wound site.
INTRODUCTION
Virulent strains of the soil bacterium Agrobacterium tumefaciens exhibit two distinctly different phases in their life history. When exposed to an appropriate plant environment, expression of the virulence (vir) genes on the tumor-inducing (Ti) plasmid is induced, ultimately leading to plant tumorigenesis and production by the tumor of carbon and nitrogen sources (the opines) that only the inciting bacteria can utilize (1). The molecular mechanism for this process has extensively been elucidated. A portion of the Ti plasmid DNA (T-DNA) and other effector Vir proteins are transferred into the plant cells by a type IV secretion system encoded by vir genes, ultimately resulting in T-DNA integration into the plant chromosomes, leading to tumor formation. Expression of the vir genes is controlled by the VirA-VirG two-component system (2), and plant-derived molecules that activate VirA are recognized either directly by VirA (phenolics) or indirectly by the periplasmic sugar binding protein ChvE (monosaccharides) (1, 2).
Besides its function in virulence, A. tumefaciens has also evolved a robust capacity to take up and catabolize a considerable array of nutrients available to support growth in the soil. The genome sequence of A. tumefaciens strain C58 suggests that there are 190 predicted ABC transporters (3), some of which have experimentally been characterized, including the sugar transporters ChvE-MmsAB (4–6), GxySBA (7), and PalEFGK (8); an octopine transporter, OccQMPJ (9); a choline transporter, ChoXWV (10); and a zinc transporter, ZnuABCT (11). Besides ABC transporters, there are also 3 predicted tripartite ATP-independent periplasmic (TRAP) transporter systems (Atu2742/Atu2743/Atu2744, Atu3366/Atu3367/Atu3368, and Atu3135/Atu3136/Atu3137). The relative richness of transporters makes A. tumefaciens significant in its competition and survival in its natural soil habitat.
Pectin is one of the main polymeric components of plant cell walls and is particularly abundant in citrus peel and sugar beet pulp (12, 13). Galacturonic acid, the primary monomer of pectin, is one of the important carbon sources for many microorganisms, including A. tumefaciens, which live in soil environments containing decaying plant material. Interestingly, galacturonic acid is utilized by A. tumefaciens as a carbon source and can also serve as an inducer of vir gene expression via binding to ChvE (4, 6, 14, 15). The catabolism of galacturonic acid by bacteria has been studied extensively and shown to occur through two different pathways referred to as the isomerase and oxidative pathways (16). The isomerase pathway has been described in Escherichia coli (17), Erwinia chrysanthemi (18), and Klebsiella pneumoniae (19). In this case, d-galacturonic acid is first converted to d-tagaturonate by an isomerase, which is ultimately degraded to pyruvate and d-glyceraldehyde-3-phosphate through the activity of an NADH-dependent reductase, a dehydratase, an ATP-requiring kinase, and an aldolase (12).
The oxidization pathway occurs in A. tumefaciens and Pseudomonas syringae (16) (Fig. 1A). In A. tumefaciens, the catabolism pathway of galacturonic acid was observed and characterized decades ago (20–22). More recently, most of the enzymes involved in this process have been cloned, purified, and biochemically characterized, and some were further structurally analyzed (Fig. 1B).
(A) Oxidization catabolism pathway of glucuronate in A. tumefaciens. d-Glucuronate is oxidized by the dehydrogenase encoded by udh, to galactaro-1,5-lactone, which is then isomerized to galactaro-1,4-lactone by the isomerase Gli. Through the subsequent functions of d-galactarolactone cycloisomerase, KDG dehydratase, and 2,5-dioxovalerate dehydrogenase, galactaro-1,4-lactone is converted to α-ketoglutarate, which can enter the tricarboxylic acid cycle. Most genes have been cloned and are shown at the left along with their respective gene loci. (B) Genetic map of the galacturonic acid utilization region in A. tumefaciens. The arrows indicate the direction of transcription, with the gene locus names and gene names shown above. The regions used in EMSA are shown below as P1 to P4. gaaM, gaaQ, and gaaP are the genes encoding the large membrane protein, the short membrane protein, and the binding component of the galacturonic acid transporter, respectively. gli encodes galactaro δ-lactone isomerase, gci encodes galactarolactone cycloisomerase, kdgD encodes KDG dehydratase, kduD encodes 2-deoxy-d-gluconate 3-dehydrogenase, kduI encodes 4-deoxy-l-threo-5-hexosulose-uronate, udh is the d-galacturonic acid dehydrogenase gene, and gaaR encodes the regulator for the expression of the whole operon. P1 to P4 are the putative promoters plus their respective vicinity regions with DNA sequences of as long as 200 bp, which were used in EMSA. P1 denotes the region in the promoter vicinity of gaaP (Atu3137). P2 denotes the region in the promoter vicinity of kdgD (Atu3140). P3 denotes the region of vicinity of intergenic region between kduI (Atu3142) and udh (Atu3143). P4 denotes the region in the promoter vicinity of gaaR (Atu3134). Control, the region in the gene gaaR, which is used as a negative control for EMSA.
In this pathway, d-galacturonate is first oxidized to galactaro-1,5-lactone by uronate dehydrogenase (UDH), encoded by gene locus Atu3143 (20, 21, 23–25), which is then isomerized to galactaro-1,4-lactone by a recently identified isomerase (encoded by gene locus Atu3138) (26). Galactarolactone cycloisomerase (encoded by gene locus Atu3139) (27) catalyzes the ring opening of galactaro-1,4-lactone to form 3-deoxy-2-keto-l-threo-hexarate, which is converted to α-ketoglutarate semialdehyde through the function of a keto-deoxy-d-galactarate (KDG) dehydratase (encoded by gene locus Atu3140) (28, 29). In the last step, a dehydrogenase oxidizes α-ketoglutarate semialdehyde to form α-ketoglutarate, which is an intermediate in the citric acid cycle (30).
Although the catabolism of galacturonate in A. tumefaciens is well elucidated, the mechanism of its uptake has not been reported. Studies in E. coli (31), Bacillus subtilis (32), Erwinia chrysanthemi (33–35), and Ralstonia solanacearum (36) showed that galacturonate uptake is mediated by the cytoplasmic membrane transporter ExuT, a major facilitator superfamily (MFT)-type transporter (37). Besides galacturonic acid, ExuT also transports glucuronic acid (35). While no homology to ExuT can be found in the A. tumefaciens genome (3), there are three genes adjacent to the galacturonic acid catabolism genes, with locus tags Atu3135, Atu3136, and Atu3137, whose predicted protein products collectively constitute a TRAP transporter (38). Based on the localization, we reasoned that this putative TRAP transporter may be involved in galacturonate uptake.
In this study, we first genetically verified the essential role of UDH in the utilization of galacturonate but not glucuronate, the other substrate for this enzyme, based on an in vitro assay (23, 24). Subsequently, through gene deletion and growth assays, we assigned the function of the TRAP transporter encoded by Atu3135/Atu3136/Atu3137 as a galacturonate transporter and named these genes gaaM, gaaQ, and gaaP. Through site-directed mutagenesis, the highly conserved residue Arg-169 was found to be critical for the function of GaaP, the galacturonate binding protein. Further, we demonstrated that the expression of the gaaPQM operon is induced by its substrate sugars and is regulated by the repressor GaaR, which can bind to multiple proposed promoter regions near the catabolic and transporter genes. We also showed that the expression of this operon is not regulated by the small-RNA binding protein Hfq, which has been shown to be important in the control of many sugar binding proteins in A. tumefaciens (39, 40). Finally, we demonstrated that the expression of this GaaPQM transporter also affected the expression of virulence genes, which further supports and broadens our previous model in which the sugar transporters are indirectly involved in the control of virulence gene expression (15).
MATERIALS AND METHODS
Bacteria, media, and growth conditions.E. coli strains were grown in Luria-Bertani (LB) medium (41) with appropriate antibiotics at 37°C. A. tumefaciens strains (Table 1) were routinely maintained on LB medium at 25°C. The defined Agrobacterium (AB) minimal medium (42) was used for the growth assays. For vir gene induction, AB induction medium (ABI) (43) was used along with 7.8 g/liter morpholineethanesulfonic acid (MES) (pH 5.5). Acetosyringone (AS) was included at 10 μM, and thiamine and Casamino Acids were omitted. Antibiotics were used at the following concentrations (in liquid and solid medium, respectively) for A. tumefaciens: gentamicin, 100 and 100 μg/ml; carbenicillin, 30 and 100 μg/ml; spectinomycin, 25 and 50 μg/ml; and kanamycin, 10 and 50 μg/ml.
Strains and plasmids used in this study
Genetic modification of bacterial strains and construction of plasmids.We used marker exchange eviction mutagenesis (44), as described by Zhao and Binns (6), to create strains carrying nonpolar deletions of gaaPQM in appropriate A. tumefaciens strains. The primer Datu313567.P1 (Table 2) was used with Datu313567.P2 to amplify an 800-bp fragment containing the downstream gaaPQM flanking sequence. PCR used to amplify the upstream gaaPQM flanking sequence used primers Datu313567.P3 and Datu313567.P4 to create an 800-bp product. The 5′ ends of the primers Datu313567.P2 and Datu313567.P3 had been designed to include complementary sequence, so that the products from the first two PCRs could function as a self-annealing template in a third PCR with primers Datu313567.P1 and Datu313567.P4 to generate a 1.6-kb fragment carrying the gaaPQM flanking sequences. This 1.6-kb PCR fragment was digested with BamHI and SalI and cloned into pK18mobsacB (44). The resulting construct was electroporated into strain A348 (45). The double-deletion strain was obtained and confirmed, as described by Zhao and Binns (6). The primers used for confirmation are outside the amplified gaaPQM flanking sequences, and they are Con313567.P1 (downstream) and Con313567.P2 (upstream). This gaaPQM deletion strain was named A. tumefaciens AB320.
Primers used in this study
The same deletion methods and confirmation strategies were used to make strains with in-frame deletions of udh, gaaP, gaaR, and hfq using the primers in Table 2. All strains were confirmed by PCR.
To monitor the expression of GaaP, a regulator of G protein signaling (RGS)-6×His tag was fused to the C terminus of gaaP, and this fusion was used to replace gaaP at its location in the chromosome. Two primer pairs, D3137.P1/3137RGS.P2 and 3137RGS.P3/Datu313567.P4, were used to perform two initial PCRs. Subsequently, the first two PCR products were used as a template to amplify a fragment carrying the RGS-6×His-tagged gaaP gene, along with additional upstream and downstream sequences. This third PCR used primers D3137.P1 and Datu313567.P4, and the resultant fragment was digested with BamHI and SalI and cloned into pK18mobsacB. The plasmid was introduced into gaaP deletion strain A. tumefaciens AB696. The resulting strain, containing the C-terminal RGS-6×His-tagged GaaP, was named A. tumefaciens AB900.
We used overlap extension PCR to create mutations (R148A/Q/K and R169A/Q/K) in the predicted galacturonic acid binding sites of GaaP, and we used marker-exchange eviction mutagenesis to replace the wild-type gene on the chromosome. In order to conveniently screen for the mutant strains, we first made partial gaaP in-frame deletions (N90 to R169) of A. tumefaciens strains AB900 and AB901 (both strains have an RGS-6×His-tagged GaaP), which were named A. tumefaciens AB341 and AB342, respectively. Next, AB341 and AB342 were used as the starting strains to make mutations at R148 and R169, as follows: for R148A mutagenesis, the primer pairs 2D3137RGS.P1/R148A.P1 and R148A.P2/2D3137RGS.P2 were used to perform two initial PCRs. Subsequently, the first two PCR products were used as the template for the third PCR with the primers 2D3137RGS.P1 and 2D3137RGS.P2, and the resultant fragment was digested with BamHI and XhoI and cloned into pK18mobsacB. The plasmid was introduced into the gaaP partial deletion strain AB341. The resulting strain, containing gaaPR148A, was named A. tumefaciens AB343. The same plasmid was also introduced into AB342, and the strain therefore obtained was named A. tumefaciens AB344. (The purpose of making a gaaP mutation in AB342 is to check the expression level of GaaP in a gaaR deletion strain grown in minimal medium with glycerol as the sole carbon source.) The same method was used to make R169A, R148Q, R169Q, R148K, and R169K in strain AB341, and these strains were named A. tumefaciens AB345, AB347, AB349, AB353, and AB355. These mutations were also made in strain AB342 with the strain names AB346, AB348, AB352, AB354, and AB356.
For complementation of the udh deletion, the primers Com3143.P1 and Com3143.P2 were used to amplify a fragment containing whole udh region and 140-bp sequence upstream of the udh start codon, which contains the putative promoter of udh. This PCR product was digested with KpnI and XbaI and cloned into pBBR1MCS-4 (46) to make pJZ41.
For complementation of the gaaPQM deletion, the primers com313567.P1 and com3137.P2 were used to amplify a fragment containing the entire gaaPQM region and 200-bp sequence upstream of the gaaP start codon, which contains the putative promoter of gaaPQM. This PCR product was digested with KpnI and XbaI and cloned into pBBR1MCS-4 (46) to make pJZ40.
For complementation of the hfq deletion, the primers comhfq.P1 and comhfq.P2 were used to amplify a fragment containing the entire hfq (Atu1450) region and 295-bp sequence upstream of the hfq start codon. This PCR product was digested with KpnI and XbaI and cloned into pBBR1MCS-5 (46) to make pJZ55.
Growth assays.The growth assays were carried out as described previously (6), except we used a concentration of 3 mM for sugars. Briefly, overnight cultures of A. tumefaciens in LB medium were centrifuged and diluted to an optical density at 600 nm (OD600) of 0.1 in AB minimal medium with glycerol. After approximately 20 h of growth, the cultures were pelleted, washed with AB medium (without any carbon source), and inoculated into AB medium with various carbon sources at a concentration of 3 mM and an OD600 of 0.06. Bacterial growth at 600 nm was measured with a spectrophotometer (DU-520; Beckman).
Immunoblot analysis for the expression of GaaP.Immunoblot analysis was performed as previously described (6). Briefly, cultures in AB medium were centrifuged and resuspended in SDS loading buffer (47), with an OD600 of 10. Samples were boiled for 5 min, and 5 μl was subjected to SDS-PAGE and subsequently transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore). For immunoblot analysis, RGS-His mouse monoclonal antibody (Qiagen) was used to detect RGS-6×His-tagged GaaP. The immunoblots were developed with an ECL Plus kit (Amersham), per the manufacturer's instructions.
β-Galactosidase assay.β-Galactosidase assays were performed as described previously (15), except in the ABI medium, 7.8 g/liter MES (pH 5.5) was used, and thiamine and Casamino Acids were omitted. In brief, strains carrying reporter plasmid pSW209Ω (48) were first grown in LB medium plus appropriate antibiotics overnight at 25°C and inoculated into ABI medium containing 0.25% glycerol plus appropriate concentrations of various sugars at 10 μM AS. After a 24-h induction, the strains were assayed for β-galactosidase using the method of Miller (41).
Protein overexpression and purification.The GaaR was overexpressed in E. coli BL21(DE3) as a C-terminal His fusion protein. In short, the gaaR gene was amplified by a PCR method using the primers 3144HIS.P1 and 3144HIS.P2. The resulting fragments were cloned into the NdeI/XhoI sites of pET22b plasmid to produce pJZ42. E. coli BL21(DE3) carrying plasmid pJZ42 was first grown at 37°C in LB medium to an optical density at 600 nm of 0.6. Next, 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) was added for induction. After cultivation for another 3 h at 37°C, cells were harvested by centrifugation and purified, according to the instructions depicted in The QIAexpressionist (49).
Analysis of GaaR binding by EMSAs.The binding of the GaaR protein to the putative promoter regions of the galacturonic acid utilization gene cluster was tested with an electrophoretic mobility shift assay (EMSA). Four gene fragments upstream of gaaP (P1), kdgD (P2), kduI/udh (P3), and gaaR (P4) and one fragment located in gaaR (control) were used in DNA-binding assays (Fig. 1B). The respective DNA fragments were generated by PCR using the primers described in Table 2, with sizes of around 200 bp. In the binding assays, different DNA fragments (60 to 100 ng) were incubated with purified GaaR (125 ng) and 20 μl of reaction buffer (20 mM Tris [pH 8.0], 30 mM KCl, 5 mM MgCl2, 1 mM EDTA, 10% glycerol) for 10 min at room temperature. Subsequently, the reaction solution was separated on a 4.5% nondenaturing polyacrylamide gel in 0.5× Tris-borate-EDTA (TBE) buffer, stained with ethidium bromide, and visualized with a UV transilluminator.
RESULTS
Search for a galacturonic acid transporter candidate.Previous studies (6, 15) showed that while ChvE can bind sugar acids and thereby mediate sugar acid-induced vir gene expression, the utilization of sugar acids as a carbon source was independent of the ChvE/MmsAB ABC transporter. To find the galacturonic acid transporter, the genome was scanned for genes known to be involved in galacturonic acid catabolism, searching for nearby putative transporter candidates. This search revealed a region on the linear chromosome encoding proteins previously identified as being involved in galacturonic acid catabolism (23–29). Nearby are three gene loci (Atu3135, Atu3136, and Atu3137) that encode proteins with homologs to the TRAP family transporters. Because such transporters are often involved in the transport of organic acids (38, 50, 51), they appeared to be good candidates for the galacturonic acid transporter. As a check on the activities of the catabolism genes, we first examined gene locus Atu3143 (udh), which encodes a galacturonic acid dehydrogenase that has been characterized biochemically (16, 23). These studies suggest that UDH can also use glucuronic acid as a substrate. To further define its function physiologically, udh was deleted in A348, as described in Materials and Methods, yielding strain AB321. Growth assays were performed comparing this strain with the wild-type strain A348. The udh deletion strain AB321 did not grow in minimal medium with galacturonic acid as the sole carbon source, whereas, surprisingly, it grew as well as wild-type strain A348 in glucuronic acid (Fig. 2), suggesting that at least another dehydrogenase must be involved in glucuronic acid catabolism. The introduction of the udh-containing plasmid pJZ41 restored the normal growth of AB321 in galacturonic acid (see Fig. S1 in the supplemental material). These results confirmed the previous in vitro study on the function of udh, demonstrated its essential function in galacturonic acid catabolism, and led us to test the hypothesis that the nearby gene loci Atu3135, Atu3136, and Atu3137 encode a transporter necessary for galacturonic acid utilization.
Growth of A. tumefaciens strains in minimal medium containing 3 mM galacturonic acid or glucuronic acid as the sole carbon source. At intervals, the optical density at 600 nm of the cultures was checked. The data are the means from the results with three samples, with the error bars representing the standard deviations. The strains used are wild-type strain A348, udh deletion strain AB321, and gaaPQM deletion strain AB320.
GaaPQM, encoded by Atu3135, Atu3136, and Atu3137, constitutes a TRAP-type galacturonic acid transporter.The gene loci Atu3135, Atu3136, and Atu3137 are predicted to encode a putative TRAP transporter, with Atu3135 encoding a large membrane protein, Atu3136 encoding a small membrane protein, and Atu3137 encoding a solute binding protein (Fig. 1B). Because of its confirmed involvement in galacturonic acid utilization (see below), this putative region was named gaa, and in accordance with the nomenclature for the first characterized TRAP transporter, DctPQM (38, 50), Atu3137, Atu3136, and Atu3135 are named gaaP, gaaQ, and gaaM, respectively.
The genes encoding GaaPQM were deleted from wild-type strain A348, and the resultant strain, AB320, was tested for growth in minimal AB medium containing various sugars as the sole carbon source. AB320 grew comparably to wild-type strain A348 in medium containing succinate, glycerol, arabinose, fucose, galactose, glucose, xylose, and glucosamine (data not shown), showed slightly slower growth on glucuronic acid than wild-type strain A348, and totally lost the ability to grow on galacturonic acid (Fig. 2). AB320 transformed with plasmid pJZ40 carrying gaaPQM and 200 bp of upstream sequence was able to utilize galacturonic acid as a sole carbon source, verifying that the growth defect of AB320 was caused by the deletion of gaaPQM (see Fig. S2 in the supplemental material). These data support the hypothesis that GaaPQM is the sole galacturonic acid transporter in A. tumefaciens and also that it slightly contributes to glucuronic acid utilization.
Residue R169 in the proposed binding pocket of GaaP is critical for its function.All solute binding proteins of TRAP transporters investigated to date that are involved in organic acid uptake contain conserved arginine residues that assist in ligand binding via salt bridges (52). To investigate the basis of GaaP binding specificity for galacturonic acid, we compared its sequence to that of other TRAP solute binding proteins and found two candidate residues, R148 and R169, which are conserved with those found in the binding pocket of galacturonate binding protein DctP from Bradyrhizobium sp. strain BTAi1 (PDB 4N8Y) (see Fig. S3 in the supplemental material). To test their possible role in galacturonic acid utilization by GaaP, the arginine residues R148 and R169 were each mutated to alanine, glutamine, and lysine. In order to monitor the expression of these mutant proteins, all the mutations were made in gaaP with an RGS-6×His tag and used as replacements in an otherwise wild-type strain (see Materials and Methods). Immunoblots showed that all the mutated proteins were expressed at the same level as the wild type (see Fig. S4 in the supplemental material). Growth assays were then performed to check the effects of these mutations on the utilization of galacturonate. The mutations at R169 (R169A, R169Q, and R169K) resulted in strains without the ability to utilize galacturonate as a sole carbon source, suggesting that R169 is essential for the function of GaaP in either the binding or transport process (Fig. 3). The mutations at R148 also partially affected the function of the protein, as shown by the slightly decreased or increased growth capability (Fig. 3).
Effect of mutations at R148 and R169 of GaaP on its function assessed by growth assay. wt* (AB900) is the wild-type strain with RGS-6×His-tagged GaaP. A partial deletion of gaaP was made first to construct the strain AB341 (wt*ΔgaaP). Based on strain AB341, gaaP site-directed mutants were made in the chromosome. The growth of A. tumefaciens strains was monitored in minimal medium containing 3 mM galacturonic acid as the sole carbon source. At intervals, the optical density at 600 nm of the cultures was checked. The data are the means of the results from three samples, with the error bars representing the standard deviations.
Regulation of gaaPQM expression.Because the genes encoding sugar transporters are usually induced by their transported substrates, we examined the expression of GaaP in bacteria grown in minimal medium with different carbon sources (53). Directly adjacent to the region downstream of the udh gene (Atu3143) is a gene locus, Atu3144, encoding a protein homologous to the GntR family of regulators. Since these are often adjacent to the genes that they control, we speculated that Atu3144 may regulate the expression of gaaPQM; therefore, we renamed it gaaR.
To check the role of gaaR on the expression of gaaPQM, we deleted gaaR and used strains AB900 (wild type with GaaP tagged with RGS-6×His) and AB901 (AB900 with gaaR deletion) and compared the expression of GaaP in these strains after growth in various carbon sources. An immunoblot assay showed that wild-type strain AB900 expressed GaaP only when grown in the presence of galacturonic acid or glucuronic acid (Fig. 4). In contrast, GaaP was expressed in AB901 regardless of the carbon source (Fig. 4), demonstrating that GaaR is a repressor of GaaP expression, and this repression can be relieved as a result of the growth in the sugar acids tested. Given that the expression of GaaPQM was depressed in the gaaR deletion strain, we speculate that the gaaR deletion strain would have a shorter lag phase when grown in galacturonic acid. Growth assays of the gaaR deletion strain AB335 confirmed this (see Fig. S5 in the supplemental material).
Immunoblot analysis of GaaP expression in strains grown at 30°C in minimal medium with 3 mM glycerol, arabinose, fucose, galactose, glucose, xylose, galacturonic acid (gal. acid), glucuronic acid (glu. acid), glucosamine, or N-acetylglucosamine (NAG) as the sole carbon source. The strain wt* (AB900) is similar to wild-type strain A348 except gaaP is tagged with an RGS-6×His tag. The strain AB901 (wt*ΔgaaR) has a gaaR deletion and is derived from AB900. For immunoblot analysis, RGS-His mouse monoclonal antibody (Qiagen) was used to detect RGS-6×His-tagged GaaP.
Our genetic study indicated that GaaR regulates the expression of the GaaPQM transporter. GaaR may also regulate the expression of the catabolism genes and itself. To identify potential promoters under its control, GaaR was overexpressed in E. coli as a His-tagged protein and used to perform an in vitro EMSA. A DNA sequence of around 200 bp, including the region upstream of gaaP, kdgD, and gaaR and the intergenic region of kduI and udh, was included in the assay (Fig. 1B). When GaaR was added together with the regions upstream of kdgD or the intergenic region of kduI and udh, an apparent shifted band was observed, indicating the binding of GaaR to these putative promoters (Fig. 5). Weak bands were observed when the regions upstream of gaaP or gaaR were incubated with GaaR, and these appear to be nonspecific interactions, because a similar shift was observed when an internal sequence of gaaR was used (Fig. 5).
EMSA to study the direct interaction between GaaR and the putative target DNA-binding regions. The binding reaction was in 20-μl systems with 125 ng of GaaR and 60 to 100 ng of DNA. P1 to P4 are the putative GaaR binding regions, and control DNA is from internal sequence of gaaR (see Fig. 1B for details). Note there are specific and strong shifted binding bands for P2 and P3 when GaaR was added (arrow with solid line). Unspecific and weak shifted bands appeared for P1, P4, and unrelated DNA (control) when GaaR was added (arrow with dotted line).
Known GntR family members bind the promoter recognition sequence as a dimer, and the binding may be counteracted by the binding of ligand to the C-terminal effector binding domain (54). However, galacturonic acid added to the EMSA reaction mixtures had no effect on the binding of GaaR to the putative promoter regions tested (data not shown). Thus, the identity of the effector ligand that is recognized by GaaR and controls the expression of the target genes remains unclear.
Expression of GaaP is not regulated by Hfq.Hfq is a small-RNA binding protein and has been shown to be involved in the control of numerous cellular processes, including regulation of the expression of ABC transporters. In Sinorhizobium meliloti, the deletion of hfq affected the expression of the mRNA of 140 transporter-related genes, including a 5.8-fold decrease in Smb21353, which is homologous to gaaP in A. tumefaciens (55). We therefore explored the effect of hfq on GaaP expression at the protein level. For this purpose, an hfq deletion was created in AB900 to make strain AB902. AB902 showed a severe growth defect in LB solid medium (see Fig. S6 in the supplemental material), which is consistent with previous observations (39). Subsequently, we checked the expression of GaaP in minimal medium supplemented with galacturonic acid as the sole carbon source. Samples were taken from two different growth periods, mid-log phase and stationary phase, and examined via immunoblot assay. The results indicate that there is no obvious difference in GaaP expression levels in the hfq deletion strain, either in log or stationary phase, when grown in galacturonic acid (Fig. 6).
Immunoblot analysis of GaaP expression in strains grown at 30°C in minimal medium with 3 mM galacturonic acid as the sole carbon source. The samples were taken at log phase and stationary phase. The strain wt* is AB900, which is similar to wild-type strain A348, except gaaP is tagged with an RGS-6×His tag. wt*Δ hfq (AB902) has an hfq deletion and is derived from AB900. For immunoblot analysis, an RGS-His mouse monoclonal antibody (Qiagen) was used to detect RGS-6×His tagged GaaP.
GaaPQM affected sensitivity to galacturonic acid in vir gene expression.In our previous studies, we observed that the deletion of sugar transporters can increase the sensitivity of A. tumefaciens to vir-inducing substrate sugars, and we suggested that the phenotype may be due to an increase in the effective concentration of the inducing sugar in the periplasm (7, 15). Here, we tested the sensitivity to galacturonic acid for vir gene expression in gaaPQM deletion strain AB320. As shown in Fig. 7A, this strain is more sensitive to galacturonic acid than wild-type strain A348 in terms of vir gene expression, as measured using a PvirB::lacZ reporter construct.
Effects of gaaPQM and gaaR deletion on galacturonic acid dose response for vir induction. Strains carrying plasmid pSW209Ω were grown in AB induction medium plus different concentrations of galacturonic acid. After growth at 25°C for 24 h, a β-galactosidase assay was performed as described in Materials and Methods. Note that 10 μM AS was used. All samples were assayed in triplicate, and results are plotted as the means with the standard deviations (SD). (A) wt, wild-type strain A348; ΔgaaPQM, strain AB320; ΔgaaR, strain AB335. (B) AB320 strain with ΔgaaPQM with vector pBBR1MCS-4 or the complementation plasmid pJZ40.
Given that the deletion of the repressor GaaR leads to the overexpression of GaaPQM (Fig. 4) (which might in turn lead to a lower galacturonic acid concentration in periplasm), we predicted that the gaaR deletion strain would be less sensitive to galacturonic acid than the wild-type strain A348. The result confirmed our prediction (Fig. 7A). A348 carrying the gaaPQM complementation plasmid pJZ40 grows much faster than the control (see Fig. S2 in the supplemental material) and may also decrease the galacturonic acid concentration in the periplasm. The results showed that the complementation strain is much less sensitive to galacturonic acid than the wild-type strain (Fig. 7B). All of the strains showed wild-type sensitivity to glucose, a nonsubstrate sugar for GaaPQM (see Fig. S7 in the supplemental material).
DISCUSSION
Through the sequence and functional analyses presented in this study, we have identified and characterized a novel TRAP-type galacturonate transporter, GaaPQM, in A. tumefaciens. gaaPQM is required for the utilization of galacturonate as a sole carbon source, and its expression is regulated via galacturonate-relieved repression of the operon by GaaR. TRAP transporters, distributed widely across the bacteria and archaea (38), are usually specific for the uptake of organic acids, including C4-dicarboxylates (50); keto-acids, such as pyruvate and alpha-ketobutyrate (56, 57); and the sugar acids, such as sialic acid (58). They are composed of the solute binding protein (the SBP), the small membrane protein, and the large membrane protein. The deletion of gaaPQM resulted in a strain (AB320) that could not use galacturonate as a sole carbon source but showed only a modest reduction in the capacity to utilize glucuronic acid (Fig. 2), suggesting that another transporter must exist for glucuronic acid.
Although the membrane components of TRAP transporters are relatively poorly studied, the SBP component has been well characterized structurally and functionally. For example, in Haemophilus influenzae, the N-acetylneuraminic acid (Neu5Ac) binding protein SiaP was the first TRAP SBP whose structure was solved (59). These studies showed that SiaP uses a highly conserved arginine residue to interact with the carboxylate group of the ligand through a salt bridge. In this study, we investigated the function of two arginine residues (R148 and R169) in the predicted ligand-binding site through site-directed mutagenesis. Each residue was changed to alanine (nonpolar R group), glutamine (same length, polar, but uncharged R group), and lysine (different length, positively charged R group). The mutants altered at R169 lost their capacity to use galacturonic acid as the sole carbon source (Fig. 3), implying that the residue R169 of GaaP is strictly required for the activity of GaaP, consistent with and supporting the previous structural studies in other TRAP SBPs (38, 51, 52). Furthermore, these data (Fig. 3) suggest that the positive charge and length of the R group of arginine are critical for function. The nearby arginine residue R148 had a slight effect on the function of GaaP, based on the growth assay. Interestingly, although R148A and R148Q had slight growth defects, the R148K mutant grew slightly better. Structural data will be needed to further decipher the role of residues R169 and R148 in ligand binding.
Our data also showed that the expression of the gaaPQM operon is regulated by GaaR. GaaR belongs to GntR family of regulators, which usually act as metabolite-responsive transcription factors (54). The DNA-binding HTH domain is located at the N terminus of the protein. At the C terminus, an effector-binding and/or oligomerization domain binds an effector molecule, and the subsequent conformational change in the protein will influence the DNA-binding properties of the regulator. Through immunoblot and EMSAs (Fig. 4 and 5), GaaR was found to repress the expression of gaaPQM by directly binding an upstream promoter (P2) region located slightly in the 5′ direction of kdgD and three genes upstream of gaaP. The expression of GaaP in the gaaR deletion mutant was derepressed under all the growth conditions used. The lack of specific binding to the region (P1) directly upstream of gaaP was surprising, given that this was the only putative promoter in the gaaPQM construct (pJZ40) that complemented the gaaPQM deletion. It is possible that the binding of GaaR to that region requires additional sequences beyond those tested in the EMSA or that a plasmid-born promoter activated the transcription of gaaPQM in our complementation assay. The finding that GaaR also binds to the promoters (P2 and P3) of galacturonate metabolism genes (kdgD, kduI, and udh) strongly suggests a role for this repressor in the control of these and their downstream genes (Fig. 5). In general, the GntR family regulators form dimers and bind to sequences with direct or inverted repeats in the promoter regions (54). A perfect palindromic sequence, PS1 (5′-TCATCATA-TAACTTG-TATGATGA-3′), is located 71 to 93 nucleotides upstream of the start codon of Atu3140 (kdgD). Moreover, 53 to 75 nucleotides upstream of the start codon of Atu3142 (kduI), there is also an imperfect palindromic sequence, PS2 (5′-TTGTCATA-CAGATCG-TACAAAAA-3′). Because both sequences are included in the DNA fragments used in our EMSA, we hypothesize that they can be recognized by GaaR. Interestingly, there is a transcription start site located 62 nucleotides upstream of the start codon of kdgD (60), placing PS1 in the −10 to −35 region and therefore further supporting our hypothesis.
When the substrate galacturonic acid, but not other sugars, is present, the expression of the transporter genes is derepressed (Fig. 4). This is likely caused by the binding of the effector ligand at the C-terminal domain of GaaR. Often, the effector molecule is a substrate or intermediate in the metabolic pathway controlled by the GntR transcription factor (54). To check the possibility that galacturonic acid is the effector, we added it in our reactions for the EMSA, but no effect was observed: the GaaR still bound to the promoter fragments P2 and P3 when galacturonic acid was present. It is conceivable that the ligand binding of GaaR was affected by the purification protocol or that the effector might be one of the intermediates in the galacturonic acid metabolism pathway.
In A. tumefaciens, the expression of many transporter genes is regulated by the small-RNA binding protein Hfq (39, 40). Hfq also regulates the expression of Smb21353, the homologue of GaaP in S. meliloti (55). In this study, we found that the expression of GaaP is not regulated by Hfq (Fig. 6), which further supports the previous proteomic studies that did not identify GaaP levels changing in an Agrobacterium strain harboring an hfq deletion (39, 40).
Our previous studies (7, 15) indicated that the sugar transporters MmsAB and GxySBA have an effect on the sensitivity to their substrate-active sugars in terms of virulence gene expression. In this study, we showed that deletion of the gaaPQM transporter increases the sensitivity of the strain to the vir-inducing activity of galacturonic acid, whereas the overexpression of GaaPQM, either via the deletion of gaaR or the use of a multicopy plasmid, decreases the sensitivity to galacturonic acid (Fig. 7). Together with our previous studies, these results strongly support a model in which the sugar transporters can affect virulence gene expression through changing the effective sugar concentration in the periplasm.
In a broader context, the GaaPQM transporter reported here is a novel TRAP transporter for galacturonic acid, and because galacturonic acid is the main monomer of pectin, this finding may be of interest to the biomass industry. Natural raw materials, especially those utilized in unexploited plant waste from the food industry, such as chitin, pectin, and others, are becoming recognized for their potential in the production of high-value fuels and chemicals (13, 61, 62). Currently, these raw materials containing pectin are mainly used as low-value cattle feed or treated as waste (16). To exploit such compounds, they would need to be transported into appropriate cells and catabolized there. For example, although yeast is a good system for utilizing galacturonate to produce biomass, this has been blocked by the absence of an efficient galacturonate uptake system. Recently, a galacturonate transporter found in Neurospora crassa was induced into Saccharomyces cerevisiae, which then exhibited significant galacturonate utilization (61, 63). The TRAP transport system described here might be of interest as a means of galacturonate transport in other settings.
To summarize, in this study, we found the novel TRAP transporter GaaPQM to be required for galacturonic acid utilization in A. tumefaciens. The expression of the transporter is regulated by the GntR family repressor GaaR but not by Hfq. Through site-directed mutagenesis combined with growth assays, we also showed that the highly conserved amino acid residue R169 is essential for the function of GaaP. Finally, virulence gene expression assays showed that the GaaPQM transporter is involved in mediating the sensitivity to galacturonic acid in vir gene expression. This provided further support for our previously proposed model (7, 15), in which the physiological activities of sugar transport systems affect the sensitivity of the vir-inducing sugars and allow Agrobacterium to distinguish between its usual soil environment and host plants that would be susceptible to its pathogenic effects.
ACKNOWLEDGMENTS
We thank the members of the Binns and Goulian labs for helpful discussion.
FOOTNOTES
- Received 4 September 2015.
- Accepted 1 December 2015.
- Accepted manuscript posted online 4 December 2015.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.02891-15.
- Copyright © 2016, American Society for Microbiology. All Rights Reserved.
