In prokaryotes, two-component signal transduction systems, consisting of a histidine kinase and a response regulator, play a critical role in regulating a range of cellular functions. A recent study suggests that XCC3315, a response regulator with a CheY-like receiver domain attached to an uncharacterized HD-related output domain (HDOD domain), plays a role in the general stress response of the Gram-negative bacterium Xanthomonas campestris pv. campestris (Xcc), the causal agent of black rot in cruciferous plants. Here, we demonstrated genetically that XCC3315, designated as gsmR (general stress and motility regulator), is involved in the expression of genes responsible for flagellum synthesis, including rpoN2, flhF, flhB, and fliC. Site-directed mutagenesis revealed that Glu9 and Arg100 in the receiver domain and Gly205, Asp263, His287, Trp298 and His311 in the HDOD are critical amino acids for GsmR function in cell motility regulation. The gsmR transcription initiation site was mapped. Promoter analysis and gel retardation assay revealed that the expression of gsmR is positively controlled by the global transcriptional regulator Clp in a direct manner, and is subject to catabolite repression. Our findings not only extend the previous work on Clp regulation to show that it influences the expression of gsmR in Xcc, but are also the first to characterize the expression of this response regulator gene in this phytopathogen. Furthermore, GsmR is the first HDOD-containing protein of bacteria in which key amino acids have been experimentally identified and characterized.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
In prokaryotes, two-component signal transduction systems (TCSTSs) represent the dominant sense–response mechanisms for the regulation of a wide range of biological functions, such as cell–cell signaling, chemotaxis, sporulation, osmolarity, nutrient assimilation, and virulence [1, 2]. A typical TCSTS is composed of a transmembrane sensor histidine kinase and a cytoplasmic response regulator with a CheY-like receiver (REC) domain that is attached to different classes of output domain. The environmental signal is detected by the sensor, which can be autophosphorylated at a conserved histidine, and then transfers the phosphoryl group to a conserved aspartate in the N-terminal REC domain of a cognate response regulator. The phosphorylated response regulator can activate its C-terminal output domain and trigger the adaptive response by modulation of gene expression or cellular machinery [3, 4].
Xanthomonas campestris pv. campestris (Xcc) is the causative agent of black rot in crucifers, a disease that results in tremendous agricultural losses . The virulence of Xcc towards plants depends on a number of factors, including the ability to secrete several extracellular enzymes and produce extracellular polysaccharide xanthan [6-8], as well as cell motility . The expression of these virulence determinants is upregulated by Clp, a homolog of the cAMP receptor protein [7, 10-15]. To date, the complete genomic sequences of three Xcc strains (ATCC33913, 8004, and B100) have been deposited in a public database [16-18]. Comparative genomic analysis has predicted that these three Xcc strains encode 106 TCSTS genes, which constitute ∼ 3% of the nucleotide sequences of these genomes [1, 19]. Only eight TCSTS genes have been characterized and experimentally confirmed to be associated with virulence in Xcc, accounting for ∼ 8% of the TCSTS genes in the Xcc genome. They are genes encoding the typical TCSTSs RpfC/RpfG , VgrS/VgrR (ColSXC1050/ColRXC1049) [21, 22], and RavS/RavR , as well as genes encoding the orphan response regulators HrpG  and VemR .
This study was inspired by the fact that insertional inactivation of genes encoding 51 response regulators in Xcc strain ATCC33913, followed by testing of the survival of these mutants under various stresses, reveals a response regulator (gene ID XCC3315) with a role in the general stress response . Domain organization analysis of XCC3315 suggests that it contains a REC domain (PF00072) at the N-terminus and an HD-related output domain (HDOD, PF08668) of unknown function at the C-terminus. The HDOD was first described by Galperin . Proteins containing the HDOD are widespread in diverse bacteria; it can be present as a stand-alone domain, and also associated with other domains, including REC, GGDEF and EAL domains, suggesting a role in regulation and signaling. At present, there are 2070 sequences with the HDOD listed in the Pfam protein family database . Among them, 1288 sequences contain the stand-alone domain, and 131 sequences are putative response regulators with the REC–HDOD architecture. Virulence factor CJ0248, a stand-alone HDOD protein from Campylobacter jejuni, is the first HDOD-containing protein to be reported . It has been indicated that the CJ0248 mutant of C. jejuni exhibits an altered flagellar motility phenotype and is attenuated for cecal colonization . Response regulators with HDOD as the output domain are found in several proteobacteria, such as Pseudomonas, Vibrio, and Xanthomonas. However, no experimental studies on these response regulators were found in the available literature, and their functions remain unknown, except for XCC3315. Because of the genome-wide research focusing on the general picture rather than individual genes, it is only known that inactivation of XCC3315 results in decreased survival under heat shock, high osmotic stress, and SDS stresses . The downstream target(s) that are regulated by XCC3315 and the regulation of XCC3315 expression have not yet been studied. The aim of the current work was to further characterize XCC3315 and to gain insights into its additional biological functions in Xcc. To this end, we first examined the regulation of XCC3315 expression in Xcc, and found that XCC3315 is under the positive control of the global transcription regulator Clp. Second, we constructed an XCC3315 mutant and subjected it to phenotypic evaluation and proteomic analysis. We observed that several protein amounts are altered in the XCC3315 mutant and that XCC3315 has a role in cell motility. We also tested the effects of XCC3315 mutation on the expression of several motility-related genes. The analysis revealed that XCC3315 is implicated in regulating the expression of motility-related genes. Finally, we identified the conserved amino acids that are essential for XCC3315 function. Considering that XCC3315 has important roles in the general stress response and cell motility, the gene was named gsmR (general stress and motility regulator) for convenience.
Xcc GsmR homologs are found in several Xanthomonas species
In Xcc strain ATCC33913, the gsmR ORF is 1113 bp in length and is located in the genome sequence at position 3 932 871–3 933 983 . It encodes a protein of 370 amino acids with a calculated molecular mass of 39 605 Da and a pI of 4.94. Domain organization analysis revealed that GsmR contains a REC domain (amino acids 3–116, E value is 3.70e-04) at the N-terminus and an HDOD (amino acids 140–314, E value is 6.00e-33) at the C-terminus. A database search revealed that GsmR is highly conserved in other sequenced Xanthomonas species, such as Xanthomonas oryzae, Xanthomonas axonopodis, and Xanthomonas albilineans, with > 70% amino acid identity (Table S1). All of these homologs are putative response regulators with unknown functions.
The gsmR gene possesses its own promoter and is likely to be monocistronic
The genes encompassing gsmR are ppa (537 bp, encoding an inorganic pyrophosphatase) and btuB (2892 bp, encoding a TonB-dependent receptor). The former is located upstream of gsmR in the same direction with a 146-bp intergenic space, whereas the latter is located downstream of gsmR in the opposite orientation with a separation of 743 bp. Owing to its orientation and intergenic regions with flanking genes, gsmR is likely to be a monocistronic gene possessing its own promoter.
In 5′-RACE experiments, sequencing of the fragment generated by nested PCR (231 bp) showed that nucleotide A, 45 nucleotides upstream from the start codon, was the transcription initiation site of gsmR (Fig. 1A,B). A putative ribosomal binding site (GAAG) was present five nucleotides upstream of the start codon (Fig. 1B). A possible σ70 promoter with a −10 box (TACATT) and a −35 box (GTGACC) were located at −7 and −30 (with a spacer of 17 nucleotides) relative to the transcription initiation site, respectively. A predicted Clp-binding site GGCGAN6TCACG, with seven of 10 bases matching (underlined bases) the consensus sequence , was located at −68/−53 relative to the gsmR transcription initiation point (Fig. 1B).
Gel retardation assay shows that Clp binds to the gsmR upstream region
The finding that there is a predicted Clp-binding site upstream of gsmR (Fig. 1B) suggests that Clp binds directly to the Clp-binding site. To demonstrate this binding, a gel retardation assay was performed with biotinylated probes: regions −156/+80 (probe a), −132/+80 (probe b), −103/+80 (probe c), −63/+80 (probe d) and −27/+80 (probe e) relative to the gsmR transcription initiation point (Fig. 1C). As a positive control, an assay with the engA promoter, which is directly regulated by Clp , was performed in parallel. As shown in Fig. 1D, Clp bound to probes a, b, and c, but not to probes d and e, indicating that region −103/+80 possesses the complete sequence needed for Clp to bind directly and is where the predicted Clp-binding site (−68/−53) is located.
Expression of gsmR requires Clp and is subject to catabolite repression
The findings that there is a Clp-binding site upstream of gsmR (Fig. 1B) and that the Clp and gsmR upstream region can form a Clp·DNA complex in a gel retardation assay (Fig. 1D) suggest that gsmR transcription is regulated by Clp directly. To demonstrate this transcriptional regulation, five PgsmR–lacZ reporter constructs (pFY−156+80, pFY−132+80, pFY−103+80, pFY−63+80, and pFY−27+80) were generated and introduced into Xc17 (wild type) and AU56E (clp mutant). Numbers with the pFY prefix are nucleotide positions relative to the gsmR transcription initiation site. Then, the resulting strains were subjected to β-galactosidase assays in LB medium. As shown in Fig. 2A, the β-galactosidase levels in AU56E carrying different constructs were lower than those in Xc17, indicating that Clp is required for transcription of the gsmR gene. Furthermore, the highest levels of enzyme, 1448 and 1344 U, were detected in Xc17(pFY−156+80) and Xc17(pFY−132+80), respectively. The levels of β-galactosidase expressed by Xc17(pFY−103+80) and Xc17(pFY−63+80) were lower (849 and 228 U, respectively) than those expressed by Xc17(pFY−156+80) and Xc17(pFY−132+80) (Fig. 2A). The remaining transformant, Xc17(pFY−27+80), showed the same levels of enzyme as those of Xc17(pFY13–9) without the gsmR promoter.
To better understand the regulation of gsmR expression, the promoter activities of Xc17 containing different fusion constructs were determined when cells were cultured in XOLN basal medium containing glycerol or glucose as a carbon source. The promoter activities of gsmR were lower (∼ 50%) when glucose was used as a carbon source than during glycerol supplementation (Fig. 2B). These observations, showing repression by glucose, suggest that expression of gsmR in Xcc is subject to catabolite repression.
Proteomic analysis indicates that the amounts of several proteins are altered after gsmR mutation
Prior to this study, nothing was known about the regulatory targets of GsmR. It was only known that GsmR might have a role in the general stress response of Xcc . To examine the effect of XCC3315 deletion on gene expression and to identify candidate target genes under GsmR control, 2D gel electrophoresis was performed to compare the protein profiles between Xc17 (wild type) and WL171 (gsmR mutant). It was found that at least 16 protein spots differed in amount between the samples prepared from Xc17 and those prepared from WL171 (Fig. 3). Three protein spots were identified. They were TonB-dependent receptor (IroN, accession number NP_635514), 2-deoxy-d-gluconate-3-dehydrogenase (KduD, NP_635547), and flagellin (FliC, NP_637306). The gene IDs of IroN, KduD and FliC in Xcc strain ATCC33913 are XCC0119, XCC0152, and XCC1941, respectively. As shown in Fig. 3, protein amounts of IroN (spot 1) and KduD (spot 2) increased in the gsmR mutant, whereas FliC (spot 3) production decreased in the gsmR mutant. Reporter assays were used to confirm the effects of gsmR mutation on iroN, kduD (Doc. S1; Fig. S1) and fliC expression (see below).
Cell motility is reduced following gsmR mutation
In Xcc, mutation in fliC causes loss of cell motility [27, 28]. The observation that gsmR plays a role in FliC expression at the protein level suggests that GsmR is implicated in Xcc cell motility. To address this possibility, 0.3% agar plate, in which presented the flagellum-dependent swimming motility, was used to evaluate cell motility. It was found that the diameters of the motility zones of WL171(pRK415) were significantly decreased (0.66 cm) as compared with Xc17(pRK415) (1.16 cm) (Fig. S2). In WL171 with cloned gsmR, WL171(pRKgsmR), wild-type-level motility (1.14 cm) was restored (Fig. S2).
Expression of genes involved in flagellar synthesis is reduced in the gsmR mutant
Xcc possesses a single polar flagellum that is essential for swimming motility . In addition to fliC, the genes rpoN2, fleQ, fliA, flhA, flhB and flgM are essential for motility and normal flagellar biogenesis [27, 29]. To test the involvement of GsmR in the expression of these motility-related flagellar genes, a reporter assay was performed to analyze their expression in WL171, with Xc17 for comparison. The fliC gene is flanked by the upstream XCC1940 (1206 bp, encoding a flagellar hook-associated protein FlgL) and the downstream XCC1942 (1329 bp, encoding the flagellar protein FliD), both in the same direction as fliC, with intergenic regions of 318 and 220 bp, respectively (Fig. 4A, left). According to microbesonlineoperonpredictions (http://www.microbesonline.org/operons/gnc190485.html), the other experimentally identified motility-related flagellar genes are arranged in four separate clusters and predicted to be formed as operons with their flanking genes: (a) rpoN2–vemR–fleQ; (b) flhF–fleN–fliA; (c) flhB–flhA; and (d) flgA–flgM (Fig. 4B–D and 4E, left part). Therefore, we set out to clone the upstream regions of fliC, rpoN2, flhF, flhB and flgA into pFY13–9 to create the reporter constructs pFYfliC, pFYrpoN2, pFYflhF, pFYflhB, and pFYflgA (Table 1). Then, the regulation of these motility-related flagellar biosynthesis genes was examined by measuring the β-galactosidase activities of these fusion constructs in WL171 and Xc17. The β-galactosidase levels expressed by Xc17(pFYfliC) and WL171(pFYfliC) were 138.0 and 84.4, respectively (Fig. 4A, right). Thus, the fliC promoter activity in the gsmR mutant was ∼ 60% that of the wild type. These transcription activity data are in agreement with the observation that the amount of FliC is reduced in WL171, as revealed on 2D gels (Fig. 3). The promoter activities of pFYrpoN2, pFYflhF and pFYfhlB were lower in WL171 (11601, 1233 and 155 U) than in Xc17 (16506, 2115 and 182 U) (Fig. 4B–D), indicating that GsmR is involved in the regulation of these motility-related genes. In addition, mutation in gsmR did not affect the expression of the flagellar protein gene flgA (Fig. 4E).
Table 1. Bacterial strains and plasmids used in this study. Apr, ampicillin-resistant; Cmr, chloramphenicol-resistant; Gmr, gentamycin-resistant; Kmr, kanamycin-resistant; Tcr, tetracycline-resistant, TIS, transcription initiation site, TSS, translation start site
The 236-bp fragment, −156/+80 relative to gsmR TIS, cloned into the XhoI and XbaI sites of pFY13–9
The 212-bp fragment, −132/+80 relative to gsmR TIS, cloned into the XhoI and XbaI sites of pFY13–9
The 183-bp fragment, −103/+80 relative to gsmR TIS, cloned into the XhoI and XbaI sites of pFY13–9
The 143-bp fragment, −63/+80 relative to gsmR TIS, cloned into the XhoI and XbaI sites of pFY13–9
The 107-bp fragment, −27/+80 relative to gsmR TIS, cloned into the XhoI and XbaI sites of pFY13–9
The 286-bp fragment, −286/−1 relative to iroN TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 449-bp fragment, −449/−1 relative to kduI TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 240-bp fragment, −240/−1 relative to kduD TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 318-bp fragment, −318/−1 relative to fliC TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 240-bp fragment, −238/+2 relative to rpoN2 TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 551-bp fragment, −553/−3 relative to flhF TSS, cloned into the XhoI and XbaI sites of pFY13–9
The 428-bp fragment, −438/−11 relative to flhB TSS, cloned into the PstI and NotI sites of pFY13–9
The 295-bp fragment, −297/−3 relative to flgA TSS, cloned into the XhoI and XbaI sites of pFY13–9
Conserved amino acids of the REC domain implicated in phosphorylation and putative metal-binding amino acids of the HDOD are essential for cell motility regulation
So far, the crystal structures of 193 sequences with the REC domain have been determined and deposited in the Protein Data Bank (PDB) database. Several crystallized response regulator REC domains have been studied extensively and characterized by mutational analysis, namely the transcription factor PhoB from Escherichia coli (PDB ID 1B00) , the sporulation initiation response regulator Spo0F from Bacillus subtilis (PDB ID 1SRR) , the chemotaxis regulator CheY from E. coli (PDB ID 3CHY) , and the response regulator NarL from E. coli (PDB ID 1RNL) . Sequence alignment of the REC domains of GsmR, PhoB, Spo0F, CheY and NarL implies that the amino acids corresponding to the phosphorylation site are highly conserved (Fig. 5A, left). In Xcc GsmR, it is suggested that: (a) Asp8 and Glu9 coordinate the metal ion needed for phosphorylation; and (b) Asp48 is the phosphorylation site that also establishes a salt-bridge with Arg100. The predicted 3D structure of the REC domain of GsmR (Fig. 5C, left) displays a (β/α)5-scaffold consisting of five parallel β-strands surrounded by five α-helices and a general topology in the REC domains . The side chains of Glu9, Asp48 and Arg100 are in close proximity to the putative phosphorylation site of the predicted 3D structure of the GsmR REC domain (Fig. 5C, left). To test the impact of these amino acids in the REC domain on the regulation of cell motility of GsmR, we generated four gsmR derivatives (D8A, E9A, D48A, and R100A) by site-directed mutagenesis. The mutated gsmR was cloned separately into the expression vector pRK415 for in trans expression in the mutant WL171. For comparison, strains carrying the expression vector pRK415 were analyzed in parallel, including Xc17(pRK415) (wild type) and WL171(pRK415) (gsmR mutant). Consistent with the results shown in Fig. S2, cell motility was reduced (57%) in WL171(pRK415) as compared with Xc17(pRK415) (Fig. 6). In trans expression of wild-type GsmR in WL171 resulted in cell motility being increased to the wild-type level (98%). Whereas expression of the mutated GsmR (D8A and E9A) in WL171 partially restored cell motility (91% and 66%), expression of another GsmR derivative (R100A) failed to restore cell motility (47%) (Fig. 6). Although D48A was successfully constructed, it may have a toxic effect on Xcc, so no viable transformant was obtained.
The HDOD of GsmR shares ∼ 30% identity with various crystallized HDOD proteins from other organisms, such as the HDOD-containing protein GSU2296 from Geobacter sulfurreducens, the putative signal transduction protein Sama_3455 from Shewanella amazonensis, and the virulence factor CJ0248 from C. jejuni. The 3D structures of GSU2296 and Sama_3455 have been determined by the Joint Center for Structural Genomics, and submitted to the PDB data bank with PDB codes 3HC1 and 3M1T, respectively. Only the 3D structure of CJ0248 from C. jejuni has been determined (PDB code 1VQR) and published . Sequence alignments of the HDODs of GsmR, GSU2296, Sama_3455 and CJ02498 are shown in the right-hand section of Fig. 5A. Note that several metal ion-coordinating amino acids found in GSU2296 are well conserved: (a) His151 and Asp152 in GSU2296 correspond to Ala262 and Asp263 in GsmR; (b) His192 in GSU2296 corresponds to His287 in GsmR; and (c) His216 in GSU2296 corresponds to His311 in GsmR. Crystal structures of GSU2296, CJ0248 and Sama_3455 show compact, helical molecules composed of 20, 18 and 19 α-helices, respectively. The predicted 3D structure of the GsmR HDOD is similar (Fig. 5C, right). According to the iron-bound GSU2296 structure, Asp263, His287 and His311 coordinate metal binding, and the side chains of these residues are in close proximity, forming a His–Asp–His triad active center in the predicted 3D structure of the GsmR HDOD (Fig. 5C, right). To determine the effect of these amino acids on GsmR function, site-directed mutagenesis was performed. In addition, two amino acids (Gly205 and Trp298 in GsmR) that are invariant in the PF08668 family were included. The cell motility of WL171 complemented with different mutated GsmRs was examined and compared with that of the strain complemented with the wild-type version. As shown in Fig. 6: (a) expression of the GsmR variants (G205A, D263A, and H311A) in gsmR mutants only partially restored cell motility (83%, 81%, and 69%); (b) the ability to restote cell motility was eliminated when WL171 was complemented with other mutated derivatives, including pRKgsmR-H287A and pRKgsmR-W298A; and (c) there were no differences between WL171(pRKGsmR) and WL171(pRKgsmR-A262H).
GsmR does not exhibit phosphohydrolase activity with p-nitrophenyl phosphate (pNPP) and bis-p-nitrophenyl phosphate (bis-pNPP) as substrates
The HDOD belongs to the HD superfamily of metal-dependent phosphohydrolases . As the highly conserved amino acids in the HD superfamily are histidine and aspartate, coordination of divalent cations is essential for the activity of these proteins . Sequence alignment revealed that the HD motif (His151-Asp152) in GSU2296 is replaced by Ala262-Asp263 in GsmR (Fig. 5A, right). The lack of conservation of the key metal-binding amino acids of the typical HD superfamily prevented us from predicting whether GsmR has phosphohydrolase activity. To investigate the enzymatic activity of GsmR, the coding region of gsmR was cloned into the expression plasmid pET30b (Novagen, Darmstadt, Germany) and overexpressed in E. coli BL21(DE3), as outlined in Experimental procedures. For comparison, E. coli BL21(DE3) carrying a mutated version of pETgsmR-A262H in which Ala262 is replaced with histidine was investigated in parallel. The pET-expressed proteins (GsmR and GsmR-A262H), which had been purified by passing through a His-binding affinity column, were tested for the presence of phosphohydrolase activity towards pNPP and bis-pNPP. The purified proteins formed one major protein band in SDS/PAGE with staining with Coomassie blue (data not shown). The mutated version, GsmR-A262H, showed low phosphohydrolase activity. The enzymatic activities of GsmR-A262H were 10.6 and 6.9 U·mg−1, respectively, when pNPP and bis-pNPP were used as substrates. Under the same conditions, negligible activity was detected when the purified wild-type GsmR was incubated with pNPP and bis-pNPP. These results indicate that the product encoded by gsmR does not possess phosphohydrolase activity. As shown in Fig. 6, WL171 carrying pRKgsmR-A262H showed similar cell motility to that of WL171 carrying wild-type GsmR, indicating that Ala262 is not an essential amino acid in GsmR and that its functionality is not enhanced, even though the conserved HD motif is incorporated in the protein. Together with the enzymatic assays, this suggests that reintroduction of the histidine makes GsmR enzymatically active, but that this modification does not influence its role in motility regulation.
In this study, we first characterized the upstream region of gsmR. The determined gsmR transcription initiation site was located 53 nucleotides downstream of a Clp-binding site (5′-GGCGA-tcggca-TCACG-3′) (Fig. 1A). It is intriguing that, although the GTG motif essential for Clp binding  is changed to GCG in the left arm (bold and underlined), a more conserved right arm (underlined) and GC-rich content in the central region (lower case) may compensate for the lack of conservation in the left arm . Several experimentally identified Xcc promoters directly regulated by Clp also have similar sequence properties, such as gumB (site I, 5′-GACGA-cccggg-TCACG-3′; site II, 5′-GGCCA-acgccg-TCAAA-3′) , XCC1294 (5′-CGCGA-cgtgct-TCGCA-3′) , and XCC2731 (5′-TGAGC-ggtcag-TCACA-3′) .
Deletion mapping analysis suggested that the −132/+80 region contains the complete promoter sequence and is capable of maximal expression under the assay conditions (Fig. 2A). The observation that expression from pFY−103+80 and pFY−63+80 gave lower activity (63% and 17% of that of pFY−132+80) suggests that region −132/−104 can somehow enhance promoter activity. Our reporter assay also suggested that gsmR expression is positively controlled by Clp (Fig. 2A) and is subject to catabolite repression (Fig. 2B). Recently, microarray analyses have revealed that transcription of 299 Xcc genes in strain XC1 is affected by clp mutation . Among these genes, 86 possess a predicted Clp-binding site upstream of their predicted ORFs . However, no TCSTS genes were included except for XCC2085 and XCC3349, the expression levels of which are positively regulated by Clp . The transcriptional activation mechanism of Clp for these two genes remains to be evaluated. Both XCC2085 and XCC3349 are response regulators with an REC domain attached to a DNA-binding output domain, and they belong to the LuxR/UhpA family of regulatory proteins. Our study not only extended previous findings on Clp regulon, but also revealed that Clp plays a role in regulating TCSTS proteins with different domain organizations.
Second, to identify genes and new functions regulated by GsmR, a gsmR-disrupted mutant was constructed and subjected to proteomic analysis. The three LC/MS/MS-identified GsmR-regulated proteins belong to different functional categories, indicating that GsmR regulates diverse functions in Xcc in addition to the general stress response. IroN (spot 1) is an outer membrane protein that is mainly known for the active transport of iron siderophore complexes in Gram-negative bacteria . Inactivation of iroN in Xcc ATCC33913 does not significantly affect Xcc virulence in the Arabidopsis thaliana Sf-2 ecotype , whereas mutation of XC_0123, the ortholog of iroN in Xcc strain 8004, results in impaired virulence in cabbage (Brassica oleracae cultivar Jingfeng 1) . KduD (spot 2) and KduI play a role in pectin metabolism in Erwinia . However, their biological function has not yet been documented in Xcc. FliC (spot 3) is a flagellin protein involved in bacterial motility [27, 28]. Expression of fliC is upregulated by Clp , and mutations of clp and fliC result in reduced motility [12, 28]. Clp upregulates gsmR expression directly (Figs 1D and 2A), and fliC is upregulated by Clp  and GsmR (Figs 3 and 4A). Thus, it is predicted that Clp regulates fliC expression through GsmR.
The finding that GsmR plays a role in cell motility (Fig. S2) prompted us to carry out further studies, because > 40 genes in the complete genome sequence of Xcc are predicted to be involved in flagellar biogenesis and motility . The flagellar biogenesis of Xcc requires the alternative sigma factors RpoN2 and FliA, and is temporally regulated by FlhA, FlhB, and FlgM . Within the rpoN2–vemR–fleQ gene cluster (Fig. 4B), RpoN2 and FleQ are important factors regulating the expression of flagellar biosynthesis genes in Xcc strain Xc17 . VemR positively regulates cell motility in another Xcc strain, 8004 . In Xcc, the transcriptional regulation of the gene clusters rpoN2–vemR–fleQ (Fig. 4B), flhB–flhA (Fig. 4D) and flgA–flgM (Fig. 4E) has not been previously documented. Within the flhF–fleN–fliA gene cluster (Fig. 4C), expression of flhF, fleN and fliA depends on the promoter region upstream of flhF . Here, we have provided evidence that GsmR plays a role in cell motility by controlling the expression of several flagellar gene clusters, including rpoN2–vemR–fleQ, flhF–fleN–fliA, and flhB–flhA (Fig. 4). It is worth mentioning that transcription analysis cannot distinguish direct regulation from these indirect or cascade regulatory events (see discussion below).
Finally, the amino acids essential for GsmR function were investigated with site-directed mutagenesis. Our mutagenesis experiments suggested that the predicted phosphorylation site amino acids Asp8, Glu9 and Arg100 in the REC domain are critical to the function of GsmR (Fig. 6). However, we were not able to isolate viable transformants harboring mutated GsmR with an alanine substitution at Asp48. We also failed to obtain wild-type Xcc carrying pRKgsmR-D48A. Whether the protein product encoded by such a mutated version has a toxic effect on Xcc remains to be investigated. We also observed that the putative metal-binding amino acids Asp263, His287 and His311 in the HDOD are associated with the full activity of GsmR (Fig. 6), although their actual role in metal binding requires further evaluation.
The observations that the conserved HD sequence motif is replaced by Ala262-Asp263 in GsmR, that substitution at Ala262 to His262 to form a conserved HD motif (His262-Asp263) in GsmR does not have any significant influence on its function (Fig. 6) and that recombinant GsmR produced by E. coli BL21(DE3) shows no phosphatase and phosphodiesterase activities when pNPP and bis-pNPP are used as substrates indicate that GsmR does not function as a phosphohydrolase. Alternatively, the phosphohydrolase activity of GsmR relies on phosphorylation of its REC domain or another, unknown, activation pathway. As the HD superfamily includes a Zn2+-dependent cyanamide hydratase (urea hydrolyase), it might possess additional catalytic activities [36, 42]. The possibility that GsmR has other enzymatic activities and uses other unidentified substrates cannot be excluded.
Although most response regulators contain DNA-binding output domains and serve as transcriptional regulators , GsmR is not a transcription factor. In addition to DNA-binding and enzymatic output domains, some response regulators combine the REC domain with RNA-binding, ligand-binding or protein-binding domains . Thus, GsmR may exert its regulatory effect at the post-transcriptional or post-translational level, or by an unknown biochemical mechanism. The role of GsmR in the expression of flagellar synthesis genes is most likely indirect. For example, there may be physical interactions with a regulatory protein, or a transcription factor, that directly regulates the expression of flagellar synthesis genes. The critical amino acids in the HDOD might be required for the interaction of GsmR with its target(s). Further studies are needed to address these possibilities.
In conclusion, in the present study, we performed transcriptional regulation and functional characterization of gsmR. The results provided experimental clues for further investigations of the functions of genes encoding proteins harboring the HDOD in bacteria. To the best of our knowledge, GsmR from Xcc strain Xc17, reported in the current study, is the first response regulator using the HDOD as an output domain to be characterized at the molecular level.
Bacterial strains, plasmids, media, and culture conditions
The bacterial strains and plasmids used in this study are listed in Table 1. LB broth and LB agar  were the general-purpose media used for cultivating E. coli and Xcc at 37 °C and 28 °C, respectively. XOLN was a basal salt medium containing 0.0625% tryptone and 0.0625% yeast extract . Glycerol or glucose was added (2%) as required. Antibiotics were added when necessary: ampicillin (50 μg·mL−1), kanamycin (50 μg·mL−1), gentamycin (15 μg·mL−1), and tetracycline (15 μg·mL−1). Liquid cultures were shaken at 220 r.p.m. Solid media contained 1.5% agar.
Recombinant DNA techniques
Enzymes were purchased from Promega and Roche. All DNA manipulations were carried out according to standard protocols . PCR was carried out as previously described , with the primers listed in Table S2. DNA sequences were determined by Mission Biotech Co. Ltd (Taipei, Taiwan). Transformation of E. coli was performed with the standard method , and transformation of Xcc was performed with electroporation .
Mapping of the 5′-end of the gsmR mRNA
The 5′-RACE system was used to determine the transcription initiation site with the Invitrogen Version 2.0 kit (Invitrogen, Carlsbad, CA, USA). Total RNA was isolated from Xc17 (mid-exponential phase) with the Qiagen RNA extraction system (Qiagen, Valencia, CA, USA). The abridged anchor primer (5′-GGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIG-3′) and abridged universal amplification primer (5′-GGCCACGCGTCGACTAGTAC-3′) were used in combination with the gene-specific primers. The gene-specific primers for RT-PCR and nested PCR were 297R (complementary to gsmR nucleotides 278–297, 5′-CTGGGCATCGATCAGGGCAA-3′) and 195R (complementary to gsmR nucleotides 176–195, 5′-GACGGCATCGAAGGGCGACA-3′), respectively. The PCR products were ligated into the yT&A vector (Yeastern, Taipei, Taiwan) for sequence verification.
Gel retardation assay
The DNA probes used for the gel retardation assay were prepared by PCR amplification of the desired gsmR upstream regions, with 5′-end biotinylated synthetic oligonucleotides as the primers (Table S2). The recombinant Clp protein, the binding conditions and detection procedures were as previously described .
Promoter activity assay
The upstream region of each gene was obtained by PCR amplification with Xc17 chromosomal DNA as the template, with the primers listed in Table S2. Then, the PCR fragments were cloned into the promoter probing vector pFY13–9  to create several promoter–β-galactosidase reporter constructs (Table 1). Xcc strains harboring these constructs were grown overnight and inoculated into fresh media to obtain an initial D550 nm of 0.35. β-Galactosidase activity was assayed following cell growth as described previously, and the enzyme levels are expressed in Miller units .
Construction of gsmR mutant
Procedures for construction of the gsmR mutant WL171 were as follows: The 1314-bp fragment encompassing the upstream 201-bp fragment plus the entire coding region of the Xc17 gsmR was PCR-amplified with primer pairs −156XhoI/+1158EcoRI (Table S2) and ligated into yT&A, giving pTgsmR. After sequence confirmation, the fragment was excised from pTgsmR, and cloned in pOK12 , giving pOKgsmR. A Gmr cartridge from pUCGM  was inserted into the PstI site within the pOKgsmR insert. The resultant plasmid, pOKgsmRG, was electroporated into Xc17, allowing for double crossover. Insertion of the Gmr cartridge into the target gene was confirmed by PCR.
Complementation of gsmR mutant
For construction of the WL171 complementation plasmid, the 1314-bp PCR-amplified fragment from pTgsmR was cloned in pRK415 , giving pRKgsmR. For complementation of the gsmR mutant, pRKgsmR was electroporated into the mutant WL171.
The site-directed mutagenesis of GsmR was performed with the QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA, USA), according to the manufacturer's instructions. The mutations were constructed in the predicted phosphorylation site residues in the REC domain (D8A, E9A, or R100A) and the highly conserved residues in the HDOD (G205A, A262H, D263A, H287A, W298A, or H311A) of GsmR, with pTgsmR as a template and the primers listed in Table S2. After DNA sequencing verification, the mutated gsmR was cloned into pRK415 to give pRKgsmR-D8A, pRKgsmR-E9A, pRKgsmR-R100A, pRKgsmR-G205A, pRKgsmR-A262H, pRKgsmR-D263A, pRKgsmR-H287A, pRKgsmR-W298A, and pRKgsmR-H311A. These constructs were then separately transferred to strain WL171 by electroporation.
Cell extract preparation
Total protein was extracted from bacterial cells as previously described , with some modifications. Briefly, bacteria were grown overnight in LB medium and diluted with fresh XOLN medium plus 2% glycerol to a D550 nm of 0.1, after which growth was allowed to continue. Cells were harvested from the culture at a D550 nm of 1.0–1.1 (∼ 22 h) by centrifugation (12 000 g at 4 °C for 2 min). The cells were then rinsed and resuspended in lysis buffer (10 mm Tris/HCl, pH 8.0, 1 mm EDTA, 0.1% Triton X-100, 1 mm phenylmethanesulfonyl fluoride). Cell extracts were obtained by sonication (cycles of 10-s pulses and 10-s rest periods on ice for 4 min). Following sonication, the cell debris and intact cells were removed by centrifugation (12 000 g at 4 °C for 2 min), and the supernatant fraction was treated with a 2-D CLEAN-UP kit (GE Healthcare Life Sciences, Uppsala, Sweden) and subjected to two-dimensional gel electrophoresis.
Two-dimensional electrophoretic analysis and protein identification
The protein concentration was adjusted to 100 μg in 200 μL of rehydration buffer (8 m urea, 2% Chaps, 40 mm dithiothreitol, 0.5% IPG Buffer, pH 3–10, 0.002% bromophenol blue). The samples were applied to an 11-cm, linear pH 3–10 IPG strip (GE) for in-gel rehydration. After rehydration (20 °C for 16 h), the first-dimension IEF was carried out on an Ettan IPGphorII (Amersham, Piscataway, NJ, USA) system, with the following program: (a) step, 500 V, 500 Vh (Volt-hour); (b) gradient, 500 V to 1000 V, 800 Vh; (c) gradient, 1000 V to 6000 V, 7000 Vh; and (d) step, 6000 V, 3700 Vh. The IPG strips were subsequently placed in equilibration buffer (50 mm Tris/HCl, pH 8.8, 6 m urea, 30% glycerol, 2% SDS, 0.002% bromophenol blue) containing 1% dithiothreitol for 30 min, and soaked in the same buffer containing 2.5% iodoacetamide for 30 min. Following equilibration, the strips were then placed on an 10% SDS/PAGE gel, and mounted with agarose sealing solution (0.5% low-melting agarose, 25 mm Tris/HCl, 192 mm glycine, 0.1% SDS, 0.002% bromophenol blue). The second-dimension electrophoresis was performed with the following program: 15 mA per gel, 15 min; and 50 mA per gel, 3 h. The gels were rinsed with deionized water for 5 min, fixed in fixing buffer (40% ethanol and 10% acetic acid) for 30 min, and stained in Sypro Ruby Protein Gel Stain (Bio-Rad, Hercules, CA, USA) overnight. Finally, the gels were washed in wash buffer (10% ethanol and 7% acetic acid) for 30 min and in deionized water three times for 5 min. The protein spots were isolated for MS analysis with a ThermoFinnigan LC/MS/MS (Schaumburg, IL, USA). The resulting data were analyzed with turbosequest (Bioworks Browser 3.2; Thermo-Finnigan).
To test the motility, 3 μL of overnight culture (D550 nm of 1) was deposited onto the surface of XOLN plates containing glycerol (2%) and agar (0.3%). The diameter of migrating rings was measured after 2 days of incubation at 28 °C.
Recombinant GsmR production and enzymatic assay
Plasmids pETgsmR and pETgsmR-A262H were constructed by inserting the 1110-bp PCR-amplified Xc17 gsmR coding region into pET30b (Novagen), with Xc17 genomic DNA and pRKgsmR-A262H, respectively, as templates. E. coli BL21(DE3) carrying pETgsmR or pETgsmR-A262H was grown in LB broth until a D595 nm of 0.5–0.6 was reached. At this point, protein production was induced by the addition of isopropyl thio-β-d-galactoside to a final concentration of 1 mm. The recombinant proteins were purified from the crude extracts prepared from a 50-mL culture with Bug-Buster protein extraction reagent and the His-Bind affinity column as recommended by the manufacturer (Novagen). Protein purity was checked by SDS/PAGE followed by Coomassie blue staining of the protein bands. Phosphatase or phosphodiesterase activities were measured in a reaction mixture containing 50 mm Tris/HCl (pH 8.0), 1 mm MnCl2, and 5 mm pNPP or bis-pNPP, by following the absorption increase at 410 nm. The amount of p-nitrophenol formed was estimated on the basis of its molar extinction coefficient of 18 000 m−1·cm−1. One unit of activity was defined as the amount of enzyme needed to produce 1 nmol of p-nitrophenol per min at 37 °C under the assay condition.
Sequence alignment and homology modeling
Multiple sequence alignment was generated by use of the clustalx package, with a standard point accepted mutation series protein weight matrix. Four and three related similar structural protein sequences were obtained from the RCSB PDB database, and the key residues were identified on the basis on alignment, as previously described for the REC domain and HDOD, respectively . The 3D structural models of the REC domain and HDOD of Xcc GsmR were based on the E. coli response regulator NarL (PDB ID 1RNL) and the G. sulfurreducens HDOD protein (PDB ID 3HC1), respectively, from the RCSB PDB database. These homology modeling structures of the conserved domains used the alignment mode from the swiss-model workspace .
Values are the means of three technical replicates per experiment, and each experiment was performed at least three times. Student's t-test was used to determine the statistical significance of differences between means.
This work was supported by grants No. NSC 97-2313-B-166-005-MY3 and No. NSC 101-2313-B-166-001-MY3 from the National Science Council, Republic of China.