Molecular and Functional Insights into the Regulation of d-Galactonate Metabolism by the Transcriptional Regulator DgoR in Escherichia coli

d-Galactonate is a widely prevalent aldonic sugar acid. Despite the proposed significance of the d-galactonate metabolic pathway in the interaction of enteric bacteria with their hosts, there are no details on its regulation even in Escherichia coli, which has been known to utilize d-galactonate since the 1970s. Here, using multiple methodologies, we identified the promoter, operator, and effector of DgoR, the transcriptional repressor of d-galactonate metabolism in E. coli. We establish DgoR as a GntR family transcriptional regulator. Recently, a human urinary tract isolate of E. coli introduced in the mouse gut was found to accumulate missense mutations in dgoR. Our results show these mutants to be DNA binding defective, hence emphasizing the role of the d-galactonate metabolic pathway in bacterial colonization of the mammalian gut.

ABSTRACT

d-Galactonate, an aldonic sugar acid, is used as a carbon source by Escherichia coli, and the structural dgo genes involved in its metabolism have previously been investigated. Here, using genetic, biochemical and bioinformatics approaches, we present the first detailed molecular and functional insights into the regulation of d-galactonate metabolism in E. coli K-12 by the transcriptional regulator DgoR. We found that dgoR deletion accelerates the growth of E. coli in d-galactonate concomitant with the strong constitutive expression of dgo genes. In the dgo locus, sequence upstream of dgoR alone harbors the d-galactonate-inducible promoter that likely drives the expression of all dgo genes. DgoR exerts repression on the dgo operon by binding two inverted repeats overlapping the dgo promoter. Binding of d-galactonate induces a conformational change in DgoR to derepress the dgo operon. The findings from our work firmly place DgoR in the GntR family of transcriptional regulators: DgoR binds an operator sequence [5′-TTGTA(G/C)TACA(A/T)-3′] matching the signature of GntR family members that recognize inverted repeats [5′-(N)_yGT(N)_xAC(N)_y-3′, where x and y indicate the number of nucleotides, which varies], and it shares critical protein-DNA contacts. We also identified features in DgoR that are otherwise less conserved in the GntR family. Recently, missense mutations in dgoR were recovered in a natural E. coli isolate adapted to the mammalian gut. Our results show these mutants to be DNA binding defective, emphasizing that mutations in the dgo-regulatory elements are selected in the host to allow simultaneous induction of dgo genes. The present study sets the basis to explore the regulation of dgo genes in additional enterobacterial strains where they have been implicated in host-bacterium interactions.

IMPORTANCE d-Galactonate is a widely prevalent aldonic sugar acid. Despite the proposed significance of the d-galactonate metabolic pathway in the interaction of enteric bacteria with their hosts, there are no details on its regulation even in Escherichia coli, which has been known to utilize d-galactonate since the 1970s. Here, using multiple methodologies, we identified the promoter, operator, and effector of DgoR, the transcriptional repressor of d-galactonate metabolism in E. coli. We establish DgoR as a GntR family transcriptional regulator. Recently, a human urinary tract isolate of E. coli introduced in the mouse gut was found to accumulate missense mutations in dgoR. Our results show these mutants to be DNA binding defective, hence emphasizing the role of the d-galactonate metabolic pathway in bacterial colonization of the mammalian gut.

INTRODUCTION

The common gut bacterium Escherichia coli can utilize a variety of sugar acids, i.e., hexonates, hexuronates, hexuronides, and aldarates, as carbon and energy sources (1). Utilization of sugar acids has been implicated in the colonization of E. coli in the mammalian gut. The gut microbiota liberates sugar acids from polysaccharides present in nutrients ingested by the host and the mucosal layer that lines the intestinal epithelial cells (2–4). Certain gut microbes also produce sugar acids from simple sugars as catabolic intermediates of metabolism (5, 6). Antibiotic treatment has also been reported to induce the host to oxidize sugars present in the gut into sugar acids (7). E. coli degrades sugar acids via the Entner-Doudoroff or Ashwell pathway into glyceraldehyde 3-phosphate and pyruvate, which further enter the central metabolism through glycolysis and the tricarboxylic acid cycle, respectively (3, 8). The sugar acid metabolic pathways are regulated by specific transcriptional regulators whose DNA-binding properties are influenced by binding to effectors, which could be either the sugar acid itself, its catabolic intermediate, or both (1, 9–11).

d-Galactonate, a hexonate sugar acid, was first reported as a carbon source for E. coli in studies conducted in the 1970s (12). Through classical mutagenesis and biochemical approaches, it was shown that E. coli metabolizes d-galactonate through a modified form of the Entner-Doudoroff pathway by a set of genes arranged in a putative d-galactonate operon (dgo) (Fig. 1A) (12, 13). Briefly, d-galactonate is transported into the cytoplasm by a putative transporter, DgoT, followed by dehydration by the dehydratase DgoD to form 2-dehydro-3-deoxy-d-galactonate. This intermediate is phosphorylated by the kinase DgoK to form 2-dehydro-3-deoxy-d-galactonate 6-phosphate, which is further cleaved by the aldolase DgoA to d-glyceraldehyde 3-phosphate and pyruvate (Fig. 1B). Further, a mutant that led to the constitutive expression of Dgo enzymes was isolated, and the mutation was mapped very close to the predicted operon. Thus, a regulator of the dgo operon, DgoR, was proposed (Fig. 1A) (13). In later studies, bioinformatics analysis of transcriptional regulators harboring N-terminal winged helix-turn-helix (wHTH) DNA-binding domains predicted DgoR to be a member of the GntR family (14, 15). On the basis of the similarity of the C-terminal effector-binding and oligomerization (E-O) domain, DgoR has been placed in the FadR subfamily within the GntR family (15, 16).

FIG 1.

DgoR negatively regulates d-galactonate metabolism. (A) Schematic of the chromosomal organization of the putative dgo operon in E. coli K-12. Filled arrows (not drawn to scale) show the direction of dgo genes, and the bent arrow indicates the direction of transcription. (B) Pathway of d-galactonate transport and degradation in E. coli K-12. (C) The dgo operon is involved in the metabolism of d-galactonate. Dilutions of the cultures were spotted on M9 minimal medium containing either d-glucose or d-galactonate as the carbon source. The experiment was repeated 3 times. A representative data set is shown. (D) Deletion of dgoR leads to faster growth of E. coli in d-galactonate. WT and dgoR::kan strains were grown in shake flasks in minimal medium containing one of the indicated carbon sources, and the OD₄₅₀ was measured. The experiment was performed 3 times. A representative data set is shown. (E) dgoR cloned in plasmid complements the faster growth phenotype of the dgoR::kan strain in d-galactonate. The WT and dgoR::kan strains carrying either empty plasmid (pACYC177) or pACYC177 with dgoR (pdgoR: pBS13) were grown in 96-well plates in minimal medium containing either d-glucose or d-galactonate as the carbon source, and OD₄₅₀ was measured. The experiment was done 3 times; each experiment had 3 technical replicates. A representative data set, with average and standard deviation (SD) of technical replicates, is shown. (F) DgoR negatively regulates transcription of dgo operon. WT and dgoR::kan strains were grown in minimal medium containing glycerol as the carbon source until exponential phase. RNA was isolated, and transcript abundance of dgo genes was assayed by qRT-PCR. Data were normalized to the transcript levels of dgo genes in WT and represent the average (±SD) from 3 independent experiments.

Several studies suggest the widespread occurrence of d-galactonate in nature. Bacteria such as Stenotrophomonas maltophilia (isolated from intestinal crypts and mucous membranes of mammalian hosts) (17), Gluconobacter liquefaciens (present in sugar-rich habitats) (18), and Pseudomonas saccharophila (a mud bacterium) (19) produce d-galactonate as an intermediate in d-galactose metabolism (6, 20, 21). Humans also produce small amounts of d-galactonate, and its levels increase in galactosemic patients, who lack the enzyme(s) of d-galactose metabolism. Various reports have demonstrated the presence of d-galactonate in mammalian tissues as well as body secretions (22–25). Genome-scale studies in the last couple of decades illustrate the importance of the d-galactonate metabolic pathway in the physiology of enteric bacteria. For example, the transcriptional profiles of Salmonella enterica strains grown in macrophages, soft-rotted leaves, or egg white and of an asymptomatic E. coli strain cultured in human urine showed upregulation of dgo genes (26–29), metabolic models of 55 fully sequenced E. coli and Shigella strains predicted 40 strains to be capable of growing on d-galactonate (30), and a mutant library screen of S. enterica serovar. Choleraesuis identified dgoT as a virulence determinant in pigs (31). Finally, in a recent in vivo evolution experiment where an E. coli strain isolated from a patient suffering from urinary tract infection was colonized in mouse gut for a year, a maximum number of mutations were found to accumulate in dgoR, indicating a role of dgo genes in bacterial colonization of the mammalian gut (32).

The above-mentioned examples of d-galactonate prevalence in nature and the proposed significance of dgo genes emphasize the importance of investigating d-galactonate metabolism in enteric bacteria. Although there are a few biochemical studies on Dgo enzymes (33–35), the molecular and functional characterization of DgoR is still lacking. Here, we show that deletion of dgoR leads to faster growth of E. coli K-12 in d-galactonate and results in a considerable increase in the expression of dgo genes. DgoR represses dgo genes by binding two closely spaced inverted repeats (IRs) in the cis-acting element that notably overlaps with the d-galactonate-inducible promoter. Mutational analysis of the predicted wHTH domain identified amino acid residues critical for DgoR-operator interactions. A comparison of DgoR and its operator with other characterized GntR family members revealed conserved and semiconserved features of protein-DNA recognition in the GntR family. Finally, we provide evidence that d-galactonate is a specific effector of DgoR.

RESULTS

The putative dgo operon is required for d-galactonate metabolism and is negatively regulated by DgoR.

Earlier studies showed the involvement of the putative dgo operon (Fig. 1A) in d-galactonate metabolism by isolating mutants unable to grow on d-galactonate as the sole carbon source (13). However, the growth phenotype of strains carrying a clean deletion of dgo genes has not been reported. Before performing a detailed investigation on the role of DgoR, we validated the involvement of the predicted dgo operon in metabolizing d-galactonate. We examined the growth of dgo single-gene deletion strains obtained from the Keio library (dgoR::kan, dgoK::kan, dgoA::kan, dgoD::kan, and dgoT::kan strains) (36) and of a strain carrying deletion of the entire dgo operon (dgo::kan, constructed in this study) in solid M9 minimal medium containing either d-glucose or d-galactonate as the carbon source. In contrast to the normal growth of strains in d-glucose, none of the dgo strains except the dgoR::kan strain grew in d-galactonate (Fig. 1C). We therefore assessed the growth of the dgoR deletion strain in a liquid medium. The dgoR::kan strain exhibited faster growth in d-galactonate (Fig. 1D). This phenotype was specific to d-galactonate; the dgoR::kan strain showed a growth profile similar to that of the wild type (WT) in other tested carbon sources, i.e., d-glucose, d-galactose, and d-gluconate (Fig. 1D). Importantly, the accelerated growth phenotype was complemented by dgoR cloned with its putative promoter in pACYC177 (Fig. 1E). The faster growth of the dgoR::kan strain suggested that DgoR is a negative regulator of d-galactonate metabolism. Consistent with this, we observed constitutive expression of various dgo genes in the dgoR::kan strain grown in minimal medium supplemented with the non-catabolite-repressing carbon source glycerol (Fig. 1F).

dgo genes are transcriptionally induced by d-galactonate.

In an earlier study, enzymatic activities corresponding to d-galactonate dehydratase (DgoD), 2-dehydro-3-deoxygalactonate kinase (DgoK), and 2-dehydro-3-deoxygalactonate 6-phosphate aldolase (DgoA) were observed in extracts prepared from WT cells cultured in d-galactonate but were not detected in extracts prepared from glycerol-, d-gluconate-, or d-galactose-grown cells (12). In another study, the expression of a dgo promoter-reporter construct was induced in the presence of d-galactonate (32). Here we show that of all tested carbon sources, the dgo genes were significantly induced at a transcriptional level particularly by d-galactonate (Fig. 2A). Further, compared to cells grown in d-galactose or glycerol, cells grown in d-glucose and d-gluconate exhibited much lower dgo transcript levels, which can be attributed to the higher carbon catabolite repression exerted by the latter carbon sources (37, 38). The result with d-glucose is in accordance with an earlier finding that d-glucose represses the expression of d-galactonate catabolic enzymes (13). Consistent with the strong activation of dgo genes by d-galactonate, we observed higher levels of DgoR protein (expressed as a 3×FLAG-tagged fusion protein from the native chromosomal locus) in d-galactonate-grown cells (Fig. 2B).

FIG 2.

The dgo operon is significantly induced by d-galactonate. (A) The dgo operon is induced at a transcriptional level by d-galactonate. The WT strain was grown in minimal medium supplemented with carbon sources indicated. RNA was isolated in the exponential phase and processed for qRT-PCR. Data were normalized to the transcript levels of dgo genes in the WT grown in minimal medium containing d-glucose. Data represent the average (±SD) for three independent experiments. (B) DgoR protein levels are higher in d-galactonate-grown cells. The strain expressing C-terminally 3×FLAG-tagged DgoR from the chromosome was grown in minimal medium supplemented with one of the indicated carbon sources. Cells were harvested in the exponential phase and processed for Western blotting using anti-FLAG antibody. WT cells grown in d-glucose were used as negative control. *, nonspecific band detected by anti-FLAG antibody which served as a control for equal loading of samples.

dgo genes are transcribed from a d-galactonate-inducible promoter located upstream of dgoR.

Because promoters in E. coli are usually located within ∼200 bp from the protein-coding sequence (39), to determine functional promoters in the putative dgo operon, we fused ∼200 to 300 bp of the upstream region of each individual dgo gene (up-dgoR, up-dgoK, up-dgoA, up-dgoD, and up-dgoT) with the fluorescent Venus reporter in the pVenus plasmid, generating constructs pAS1 and pNK3 to pNK6 (Fig. 3A). These transcriptional reporter constructs were integrated into the chromosomes of WT and ΔdgoR strains at the attλ site. The resulting reporter strains were grown either in noninducing (M9 minimal medium containing glycerol) or inducing (M9 minimal medium containing glycerol and d-galactonate) medium, and fluorescence was measured in the exponential phase. In the presence of d-galactonate, the increased expression of the reporter was observed only for the transcriptional fusion bearing the upstream region of dgoR (up-dgoR). Further, only the same reporter construct displayed constitutive fluorescence in noninducing medium in the ΔdgoR strain (Fig. 3B). We conclude that in the dgo locus, sequences upstream of dgoR alone harbor the d-galactonate-inducible promoter, which is normally repressed by DgoR. Because all dgo genes show increased expression in WT cells in the inducing medium and in a dgoR deletion strain in the noninducing medium (Fig. 1F and Fig. 2A), their expression must be from the dgo promoter located upstream of dgoR, indicating that the five dgo genes constitute an operon. These results are consistent with an earlier study where a mutation that simultaneously prevented the synthesis of d-galactonate catabolic enzymes was isolated, leading the authors to speculate that the mutation was in the common promoter of dgo genes (13).

FIG 3.

Sequence upstream of dgoR alone harbors the d-galactonate-inducible promoter. (A) Schematic of the transcriptional reporter fusions analyzed to identify functional promoters in the putative dgo operon. Reporter plasmids (pAS1 and pNK1 to pNK6) carrying ∼200 to 300 bp upstream of each individual dgo gene (up-dgoR, up-dgoK, up-dgoA, up-dgoD, and up-dgoT) and the up-dgoR truncated variants (up-dgoR Δ−35 and up-dgoR Δ−35/−10) were introduced as single-copy fusions in WT and ΔdgoR strains. The sequences of promoter elements (−35 and −10 boxes) in the upstream region of dgoR, predicted by both BPROM and bTSSfinder, are shown. Transcriptional fusions are not drawn to scale. (B) up-dgoR harbors the functional promoter of the dgo operon. The reporter strains described for panel A were grown in M9 minimal medium supplemented either with glycerol or with glycerol and d-galactonate to exponential phase. Fluorescence was measured and normalized to the OD₄₅₀ of the samples. Data represent the average (±SD) from 3 independent experiments.

Next, we predicted −10 and −35 promoter elements in the sequence upstream of dgoR using the promoter prediction programs BPROM and bTSSfinder (39, 40) (Fig. 3A). To investigate whether the common promoter elements predicted by both the programs are indeed required for activity in vivo, the up-dgoR element was truncated to give rise to two variants, one that lacks the −35 element (up-dgoR Δ−35) and another that lacks both the −10 and −35 elements (up-dgoR Δ−35/−10) (Fig. 3A). The truncated variants were cloned in the pVenus plasmid, resulting in the constructs pNK1 and pNK2, which were integrated into the chromosome. Compared to the up-dgoR reporter construct, which displayed significant fluorescence in the presence of d-galactonate in the WT strain and constitutive fluorescence in the ΔdgoR strain, the Δ−35 and Δ−35/−10 reporter constructs showed near-background fluorescence under similar conditions (Fig. 3B). These data validate that Δ−35 and Δ−35/−10 constructs lack essential promoter elements.

DgoR binds inverted repeats IR1 and IR2, which overlap the dgo promoter.

A majority of GntR family regulators bind inverted repeat sequences on DNA (15, 16, 41). A search for DgoR in the RegPrecise database (41, 42) showed three inverted repeats (IR1, IR2 and IR3) in the nucleotide sequence upstream of the dgo operon to be the probable DgoR-binding sites (Fig. 4A), suggesting that DgoR binds the up-dgoR element (hereafter referred to as the cis-acting element) to regulate itself and other dgo genes in the operon. IR1 is a perfect inverted repeat, whereas both IR2 and IR3 have one mismatch. To determine the binding of DgoR to the cis-acting element, an ∼200-bp DNA fragment encompassing the three inverted repeats (Fig. 4A) was PCR amplified from plasmid pAS1 using ³²P-labeled primer and incubated with different concentrations of purified DgoR-6×His (Fig. 4B). DgoR shifted the labeled cis-acting element in a concentration-dependent manner (Fig. 4C). The addition of a 200-fold excess of the unlabeled cis-acting element completely abolished the interaction of DgoR with labeled DNA (Fig. 4C).

FIG 4.

DgoR binds its cis-acting element. (A) Sequence of the ∼200-bp cis-acting element used in EMSAs. Lowercase letters denote the start codon of dgoR. The predicted promoter elements (−35 and −10 boxes) are highlighted in gray. The three inverted repeats (IR1, IR2, and IR3) are underlined. Arrows depict the left and right halves of the repeats. (B) SDS-PAGE of purified DgoR-6×His. The WT strain carrying pBS2 was induced with IPTG. Cells were harvested and processed for protein purification using Co-NTA resin. Lane M, molecular mass markers (in kilodaltons). The tagged protein could be purified to a final concentration of ∼15 µM. (C) EMSA shows the binding of purified DgoR to the cis-acting element. A 20 nM concentration of the ³²P-labeled DNA fragment was incubated with different concentrations of purified DgoR-6×His with or without the unlabeled cis-acting element. Samples were resolved by native PAGE and subjected to autoradiography.

In order to investigate the importance of three inverted repeats in the cis-acting element for binding DgoR, the symmetry of each inverted repeat was broken by substitution or deletion mutations in the left and right halves of the repeat (Fig. 5A), and their binding to DgoR was examined in an electrophoretic mobility shift assay (EMSA). For this, similar to the case for the WT cis-acting element, we cloned mutated fragments upstream of the fluorescent Venus reporter in the pVenus plasmid, resulting in constructs pAS2 to pAS8. When required, cis-acting elements were amplified from the respective constructs using radiolabeled primer and incubated with DgoR-6×His. Whereas DgoR shifted the WT cis-acting element, the protein was unable to shift the mutant DNA fragments IR1L, IR1R, IR2L, IR2R, and IR2L'. On the contrary, both the IR3L and IR3R fragments formed protein-DNA complexes similar to that observed for the WT cis-acting element (Fig. 5B). Collectively, these data show that of the three inverted repeats, IR1 and IR2 are critical for DgoR to bind the cis-acting element.

FIG 5.

Inverted repeats IR1 and IR2 overlap the promoter and are critical for binding of DgoR to its cis-acting element. (A) Sequence of a region of the dgo cis-acting element, showing mutations created within the three inverted repeats. The sequences of a region of the WT cis-acting element and its various mutants (IR1L, IR1R, IR2L, IR2R, IR2L′, IR3L, and IR3R) are shown, where the predicted promoter core elements (−35 and −10 boxes) are highlighted in gray, the three inverted repeats (IR1, IR2, and IR3) are underlined, the left and right halves of the repeats are shown by arrows, nucleotide substitutions in the inverted repeats are shown in bold, and nucleotide deletions in the inverted repeats are shown by a line. L and R in the names of the inverted repeat mutants denote that the mutation is in the left or right half of the inverted repeat. For example, IR1L and IR1R denote that the mutations are in the left and right halves of IR1, respectively. (B) EMSA shows that IR1 and IR2 are critical for binding of DgoR to its cis-acting element in vitro. A 20 nM concentration of ³²P-labeled WT and various mutant fragments was incubated with (+) or without (−) 1.2 µM purified DgoR-6×His. Samples were resolved by native PAGE and subjected to autoradiography. (C) Fluorescence assay reveals the requirement of IR1 and IR2 in the cis-acting element for binding of DgoR in vivo. The WT and various mutant fragments transcriptionally fused to the Venus reporter were integrated into the chromosome. Strains were grown in M9 minimal medium supplemented either with glycerol or with glycerol and d-galactonate to exponential phase. Fluorescence was measured and normalized to the OD₄₅₀ of the samples. Data represent the average (±SD) from 3 independent experiments.

Next, we investigated the importance of inverted repeats in vivo. For this, we integrated the Venus reporter constructs, pAS2 to pAS8, into the chromosomes of the WT and ΔdgoR strains. In order to interpret the data from fluorescence assays for the inverted repeat mutants, it was important to consider that IR1 and IR2 sites overlap the predicted −35 and −10 promoter elements, respectively (Fig. 5A). The left half of IR1 overlaps the −35 element, whereas the right half of IR1 overlaps a few base pairs of the spacer region between the −35 and −10 elements. In the ΔdgoR background, both the IR1L and IR1R constructs exhibited fluorescence similar to that of the reporter construct harboring the WT cis-acting element (Fig. 5C). These data show that mutations in IR1 do not abrogate promoter activity. Importantly, both constructs showed constitutive fluorescence in the WT strain, suggesting a complete loss of DgoR repression (Fig. 5C). This result correlates with the defective binding of purified DgoR with the IR1L and IR1R fragments observed in EMSA (Fig. 5B). Together these data reveal that IR1 is essential for DgoR to bind the cis-acting element to repress transcription of the dgo operon.

The right half of IR2 overlaps the −10 element (Fig. 5A). In the ΔdgoR strain, the IR2R construct exhibited lower fluorescence than the construct harboring the WT cis-acting element, suggesting that mutations in the −10 box compromise promoter activity (Fig. 5C). However, the promoter was still active, because the fluorescence of IR2R was higher than that of the Δ−35 and Δ−35/−10 constructs, which lack essential promoter elements (compare Fig. 3B and Fig. 5C). Although the left half of IR2 lies outside the −10 box (Fig. 5A), its mutation in the IR2L and IR2L′ constructs resulted in low fluorescence in the ΔdgoR strain, suggesting that sequence adjacent to the −10 element also contributes to promoter activity. Further, in WT cells grown under noninducing conditions, all IR2 mutants exhibited higher fluorescence than the WT cis-acting element, indicating that mutations in IR2 compromise DgoR binding (Fig. 5C).

Consistent with the observation that IR3 does not overlap the promoter, in the ΔdgoR background, the IR3L and IR3R reporter constructs exhibited fluorescence comparable to that of the construct carrying the WT cis-acting element. In the WT strain as well, IR3L and IR3R showed regulatory behavior similar to that of the WT fragment (Fig. 5C). These data correlate with the ability of purified DgoR to bind IR3L and IR3R fragments in EMSA (Fig. 5B). Thus, the IR3 site is not critical for the interaction of DgoR with its cis-acting element.

Collectively, from our EMSAs and fluorescence assays, we conclude that DgoR binds IR1 and IR2 sites in the cis-acting element for repressing dgo genes. The information that DgoR-binding sites overlap the promoter suggests that DgoR represses transcription of the dgo operon by occluding the binding of RNA polymerase (RNAP) (43).

Mutations in the DNA-binding domain abrogate interaction of DgoR with its cis-acting element.

The structure of DgoR predicted by the Robetta server showed the presence of an N-terminal wHTH DNA-binding domain and a C-terminal seven-helix domain (see Fig. S1 in the supplemental material). Thus, based on its topology DgoR belongs to the FadR subgroup within the FadR subfamily of transcriptional regulators (15, 16). The far-UV circular dichroism (CD) spectrum of purified DgoR also showed a signature of a substantially helical protein (see Fig. S2 in the supplemental material).

The wHTH domain is composed of a tri-helical core followed by two beta strands connected through a small loop, designated the “wing” motif (Fig. S1). To ascertain whether secondary structure elements in the predicted DNA-binding domain of DgoR are involved in protein-DNA interaction, we created a missense mutation in each of the three α-helices (D7A in α1, L34A in α2, and R46A in α3), which are either conserved across the GntR family or reported to be functionally important in other characterized members (15) (see below). In addition, in a recent study where a human urinary tract isolate of E. coli was introduced into streptomycin-treated mouse gut and mutations in the evolved lineages were followed over a span of a year, three missense mutations were independently recovered in the predicted α2-α3 region of DgoR. These data suggested the importance of derepression of the dgo operon in bacterial adaptation in the mouse gut. Indeed, the evolved lineages grew faster than the parental strain in d-galactonate, implying a potential benefit to the evolved strains in utilizing this carbon source (32). However, because there were mutations at additional loci in the evolved strains, a reexamination of dgoR mutations in a clean background coupled with testing their DNA-binding ability is imperative to assess their true contribution to the upregulation of the dgo operon and adaptation in the host. To this end, we created mutations in dgoR corresponding to the three missense mutations (T40I, R42C, and S51L) isolated in the previous study.

We purified mutant DgoR as C-terminally His-tagged proteins. Far-UV CD showed that the mutants are folded (Fig. S2). As observed in EMSA, the interaction of mutants with the cis-acting element was completely abolished (Fig. 6A). Consistent with the loss of DNA-binding ability, the mutants could not complement the growth phenotype of the dgoR::kan strain in d-galactonate (Fig. 6B). Further, the expression of DgoR variants in the complementation strains in the noninducing medium largely correlated with their repressor ability (Fig. 6C). Whereas WT DgoR could not be detected due to autorepression, the D7A, T40I, R42C, R46A, and S51L mutants exhibited significant expression. Although the L34A mutant could not be detected, likely due to its rapid degradation, its low-level expression inside the cell is evident from the ability of this mutant to exhibit a weak dominant negative phenotype (see below [Fig. 7A]). We further tested the ability of a few of these mutants to repress expression from the dgo promoter in vivo. For this, we expressed either WT or mutant DgoR proteins from pACYC177 in a ΔdgoR reporter strain. In the noninducing medium, in contrast to the significant repression of the reporter in the ΔdgoR strain expressing WT DgoR from the plasmid, the mutants behaved similarly to the ΔdgoR strain carrying an empty plasmid; i.e., they exhibited constitutive expression of the reporter (Fig. 6D). In the inducing medium, as expected, all strains exhibited comparable reporter expression (Fig. 6D). Taken together, our results emphasize that the D7, L34, T40, R42, R46, and S51 amino acid residues are important for DgoR to bind its cis-acting element to repress the dgo operon.

FIG 6.

Mutations in the DNA-binding domain abrogate interaction of DgoR with its cis-acting element. (A) DNA-binding domain mutants do not bind the cis-acting element in vitro. A 20 nM concentration of ³²P-labeled DNA fragment was incubated with either 1.2 µM WT DgoR-6×His or its various mutants. Samples were resolved by native PAGE and subjected to autoradiography. (B) DNA-binding domain mutants fail to complement the faster growth phenotype of dgoR deletion strain in d-galactonate. The WT and various dgoR mutants cloned in pACYC177 [pdgoR(WT), pBS13; pdgoR(D7A), pBS22; pdgoR(L34A), pBS14; pdgoR(R46A), pBS15; pdgoR(T40I), pBS37; pdgoR(R42C), pBS36; and pdgoR(S51L), pBS35] were individually transformed in a dgoR::kan strain. Cultures were grown in 96-well plates in minimal medium containing d-galactonate as the carbon source, and the OD₄₅₀ was measured. The experiment was performed 3 times; each experiment had 3 technical replicates. A representative data set, with average and SD from technical replicates, is shown. (C) The expression of WT DgoR and various DNA-binding mutants from the native promoter largely correlates with their repressor ability. The strains described for panel B were grown in minimal medium supplemented with glycerol. Cells were harvested and processed for Western blotting using anti-His antibody. (D) Fluorescence assay shows that DNA-binding domain mutants are unable to repress expression from the dgo promoter in vivo. Plasmids [pdgoR(WT), pBS13; pdgoR(D7A), pBS22; pdgoR(L34A), pBS14; and pdgoR(R46A), pBS15] were individually transformed in a ΔdgoR strain carrying the fluorescent Venus reporter on the chromosome under the control of the dgo promoter. Strains were grown in minimal medium supplemented either with glycerol or with glycerol and d-galactonate to exponential phase. Fluorescence was measured and normalized to the OD₄₅₀ of the samples. Data represent the average (±SD) from 3 independent experiments.

FIG 7.

DgoR forms oligomers. (A) DNA-binding-defective mutants exhibit a dominant negative phenotype. The empty plasmid pACYC177 and various constructs, described in the legend to Fig. 6D, were individually transformed in the WT strain carrying the Venus reporter (under the control of the native dgo promoter) on the chromosome. Strains were grown in minimal medium supplemented either with glycerol or with glycerol and d-galactonate to exponential phase. Fluorescence was measured and normalized to the OD₄₅₀ of the samples. Data represent the average (±SD) from 3 independent experiments. (B) DgoR oligomerizes in vitro. DgoR-6×His from cell lysate was immobilized on Co-NTA beads to which either purified MBP-DgoR or MBP-βgal was added. A portion of these suspensions was saved as the input. The remaining suspension was incubated, and proteins were finally eluted as the pulldown. Samples were subjected to SDS-PAGE and processed for Western blotting using anti-MBP (upper panel) or anti-His (lower panel) antibody. *, band corresponding to MBP which likely arises due to spontaneous degradation of MBP-βgal.

DgoR forms oligomers.

Most GntR family proteins exist as oligomers (16). We made use of the DNA-binding-defective mutants, D7A, L34A, and R46A, to investigate whether DgoR oligomerizes in vivo. For this, we expressed either WT or mutant DgoR proteins from pACYC177 in a WT reporter strain. In the noninducing medium, whereas the WT protein from plasmid did not affect the expression of the Venus reporter, mutant DgoR proteins increased reporter expression (Fig. 7A). These results suggest that the DNA-binding-defective mutants expressed in trans oligomerize with WT DgoR from the chromosome, interfering with its binding to the cis-acting element and thereby exhibiting the dominant negative phenotype. On the other hand, as expected, under inducing conditions, all strains exhibited similar expression of the Venus reporter (Fig. 7A).

To determine whether DgoR forms oligomers in vitro, DgoR-6×His was immobilized on cobalt-nitrilotriacetic acid (Co-NTA) beads and incubated either with maltose-binding protein–β-galactosidase (MBP-βgal) or MBP-DgoR. DgoR-6×His was eluted using imidazole. Whereas MBP-βgal did not coelute with DgoR-6×His, MBP-DgoR was pulled down by DgoR-6×His, suggesting that DgoR molecules interact with each other to form oligomers (Fig. 7B).

d-Galactonate is a specific effector of DgoR.

An earlier study reported the enzymatic activities of DgoK and DgoA to be significantly induced in cell extracts prepared from a dgoD mutant grown in d-galactonate-supplemented medium. Because DgoD catalyzes the first step in d-galactonate metabolism and hence no catabolic intermediate is expected to accumulate in a dgoD mutant, it was suggested that d-galactonate is the inducer of the dgo operon (13). However, since the catabolic pathways of sugar acids cross talk with each other either due to their enzymatic interconversion or shared transcriptional regulation (1, 9, 10), it is important to examine whether d-galactonate directly binds DgoR. To test this, we incubated DgoR with its cis-acting element in the presence of various concentrations of d-galactonate. To determine if d-galactonate is a specific effector of DgoR, additional sugars (d-glucose and d-galactose), a sugar acid (d-gluconate), and intermediates of the d-galactonate degradation pathway (Fig. 1B) were also examined. Whereas d-galactonate relieved DNA bound by DgoR in a concentration-dependent manner (K_D [apparent dissociation constant] of 55 ± 6 µM), other tested carbohydrates did not affect the binding of DgoR to its target DNA (Fig. 8A and B; see Fig. S3 in the supplemental material).

FIG 8.

d-Galactonate directly interacts with DgoR. (A and B) d-Galactonate releases DgoR from its target DNA. A 1.2 µM concentration of DgoR-6×His was incubated with the indicated concentrations of various carbohydrates for 20 min, and then 20 nM ³²P-labeled dgo cis-acting element was added to the samples and incubated for 30 min. Samples were resolved by native PAGE and subjected to autoradiography. (C) d-Galactonate induces a conformational change in DgoR. SDS-polyacrylamide gels show the results of trypsin-mediated digestion of DgoR-6×His alone (left panel), DgoR-6×His incubated with d-gluconate (middle panel), and DgoR-6×His incubated with d-galactonate (right panel). Lane M, molecular mass markers (in kilodaltons). O/N, the reaction mix was incubated overnight.

Limited proteolysis is widely used to monitor conformational changes induced in regulators upon interaction with their cognate effectors (44–47). To determine whether the interaction of d-galactonate with DgoR is specific and results in a conformational change in the regulator, we performed limited proteolysis of the DgoR protein alone and of the protein incubated with either d-gluconate or d-galactonate (Fig. 8C). Compared to DgoR alone or DgoR incubated with d-gluconate, the protein incubated with d-galactonate was less accessible to trypsin digestion and showed a different digestion pattern. These data show that the presence of d-galactonate protects DgoR against trypsin cleavage and suggest a conformational change in DgoR induced by its interaction with d-galactonate.

DISCUSSION

Here we investigated interactions between the dgo repressor, operator, and effector. The combined results provide evidence that DgoR represses the d-galactonate metabolic pathway by binding inverted repeats in the dgo cis-acting element and employs a derepression mechanism using d-galactonate as a specific effector molecule.

Interaction of the transcriptional repressor, DgoR, with its operator.

Our work shows that DgoR forms a complex at the dgo operator comprising of two inverted repeats that are likely on the same face of the DNA (the distance between the inverted repeats is 10 bp, i.e., 1 helix turn) (Fig. 4). Importantly, the inverted repeat sequences recognized by DgoR [5′-TTGTA(G/C)TACA(A/T)-3′] (Fig. 5), match the signature of GntR family regulators that also bind inverted repeats [5′-(N)_y GT(N)_x AC(N)_y-3′, where x and y indicate the number of nucleotides, which varies] (16, 41, 48). The DgoR-binding sites overlap the promoter, indicating that DgoR represses transcription of the operon by occluding the binding of RNA polymerase (Fig. 5). Shortly after the submission of the manuscript for this article, a study using a high-throughput plasmid-based reporter assay reported a region in the dgo cis-acting element to harbor putative overlapping DgoR- and RNAP-binding sites (49). Importantly, the region identified in that study encompasses the sites determined by our single-copy reporter assays and EMSAs, providing strong support to our findings.

The mechanistic details of the interaction between the wHTH domain and the operator are derived mainly from the DNA-bound structures of GntR family regulators belonging to different subfamilies (AraR subfamily, Bacillus subtilis AraR [50]; FadR subfamily, E. coli FadR [51, 52] and Vibrio cholerae FadR [53]; and HutC subfamily, B. subtilis NagR [54]) and genetic/biochemical characterization of these and additional GntR family members (55–61). To understand how amino acid residues in DgoR mutated in this study determine its DNA-binding ability, we obtained the DNA-bound state of DgoR by superimposing the wHTH domain of a DgoR monomer onto the wHTH domain of E. coli FadR bound to its operator (PDB identification [ID] 1H9T) (see Fig. S4A in the supplemental material). In the characterized GntR family proteins, amino acid residues in α1 make nonspecific contacts with the phosphate backbone and support the specific contacts made by the downstream residues (50–54, 59). In the DgoR-DNA model, although the side chain of D7, an α1 residue, points toward the DNA, it does not make any contact (Fig. S4A). However, similar to the DNA-binding defect of the D7A mutant of DgoR, a previous study on another FadR subfamily member, AphS, reported a loss of repression by a mutant harboring a mutation in the corresponding α1 residue, E17 (56). A Glu or Asp residue is present at the analogous position in several GntR family members (15), and thus, an acidic residue at the start of α1 may play an important role in the repressor ability of these regulators.

Residues in α2 and α3 form specific contacts in the major groove, where α3 defines a larger part of the specificity and hence termed the “recognition helix” (16). The information that amino acid substitution of several residues in α2-α3 abrogated the interaction of DgoR with its operator (Fig. 6) further underscores the importance of the wHTH motif in DNA recognition by GntR family regulators. Corresponding to the α2 residue, L34 in DgoR, a hydrophobic amino acid is present in a majority of the GntR family members (Fig. 9A) (15). Mutation of an equivalent Leu in FadR (L37A) and AraR (L33S) led to a weak dominant negative phenotype and poor expression of the protein, respectively, suggesting that this hydrophobic amino acid is important for protein stability (55, 57, 58). Along similar lines, we observed a weak dominant negative phenotype of the L34A mutant of DgoR and were unable to detect its expression from the native dgo promoter (Fig. 6C and Fig. 7A). In the DgoR-DNA model, we find that L34 is part of a hydrophobic core, further emphasizing that a hydrophobic amino acid at this position is structurally important in GntR family regulators (Fig. S4B, left panel). Additionally, the importance of the L34 residue in DNA binding is evident from both the inability of L34A to form protein-DNA complexes in vitro and negative dominance in vivo (Fig. 6A and Fig. 7A). In the DgoR-DNA model, L34 interacts with two hydrophobic residues, I45 and F49, located in α3 (Fig. S4B, left panel). Because mutating L34 to A will reduce hydrophobicity and increase the distance of I45 and F49 from α2, we suggest that the recognition helix might become reoriented, affecting its contact with DNA. Similar to the DNA-binding defect of R42C and R46A in DgoR, mutation of Arg residues at equivalent positions in α3 in several other GntR family members compromises their repressor ability (55, 57–59). From the available DNA-bound structures, it is evident that both Arg residues interact with a common Glu residue in α2, which orients them such that arginine corresponding to R42 makes a specific contact with guanine (G-3), while arginine equivalent to R46 interacts nonspecifically with the phosphate backbone (50–54, 59). In the DgoR-DNA model, we observe the interaction of R42 and R46 with an equivalent Glu residue, E31 in α2 (Fig. S4B, right panel), which leads us to speculate that these Arg residues make contacts with DNA similar to those observed in other characterized members (Fig. 9). GntR family regulators that recognize the TGGT motif in the operator make a specific contact with G-4 through an Arg residue conserved in these members (Fig. 9) (15, 51–54, 59). However, GntR family proteins, including DgoR, that recognize the TTGT motif lack the analogous Arg (Fig. 9) (15, 50), emphasizing that differences in the operator sequence are reflected in the amino acid sequences of their cognate regulators.

FIG 9.

Structure-based sequence alignment of the N-terminal wHTH DNA-binding domains of GntR family members and comparison of their operator sequences. (A) The amino acid sequence of the N-terminal wHTH domain of DgoR is aligned with those of GntR family members for which either the DNA-bound structures are solved (BsAraR, PDB ID 4EGY; EcFadR, 1H9T; VcFadR, 4P9U; and BsNagR, 4WWC) or the structure of their apo form is modeled on DNA (CgLldR, 2DI3), using the DALI server (79). Arrows indicate amino acid residues in the wHTH domain of DgoR mutated in this study. Secondary structure elements are indicated by bars (α-helices) and horizontal arrows (β-strands). *, presence of an Arg residue at this position in GntR family members that recognize the TGGT motif in the operator (see panel B). Abbreviations: Ec, E. coli; Bs, B. subtilis; Vc, V. cholerae; and Cg, Corynebacterium glutamicum. (B) Comparison of the binding sites of DgoR (IR1 and IR2) with the operator sequences of GntR family members used to either solve the DNA-bound structures of their wHTH domains (BsAraR, EcFadR, VcFadR, and BsNagR) or biochemically test the binding to its transcriptional regulator (CgLldR). Nucleotides in the inverted repeat that are conserved and/or involved in making specific contacts with the protein are highlighted in gray. Abbreviations: Ec, E. coli; Bs, B. subtilis; Vc, V. cholerae; and Cg, C. glutamicum.

The DNA-binding defect of the T40I and S51L mutants cannot be explained from the DgoR-DNA model. Although the side chain of T40 points toward the DNA, there is no interaction with the DNA (Fig. S4A). In a majority of the GntR family regulators, a hydrophobic amino acid is present at a similar position (Fig. 9A) (15). The residue S51 does not face the DNA (Fig. S4A), and the nature of the amino acid at a similar position is highly variable across the GntR family (Fig. 9A) (15). Future studies aimed at solving the structure of DgoR alone and in complex with its operator might explain the importance of these amino acids in the functioning of the repressor.

d-Galactonate is the physiological effector of DgoR.

Besides DgoR, which regulates d-galactonate metabolism in E. coli, FadR subfamily members also control the metabolism of sugar acids such as d-gluconate, d-glucuronate, d-galacturonate, and d-glucarate (1, 10, 62, 63). In addition, FadR subfamily regulators also govern the metabolism of a variety of other carbon sources, such as fatty acids, sialic acids, l-lactate, and substrates emerging from central metabolism (59, 64–66). These regulators employ substrates (e.g., E. coli NanR, N-acetylneuraminic acid; B. subtilis GntR, d-gluconate), metabolic intermediates (e.g., E. coli FadR, long-chain fatty acyl coenzyme A [acyl-CoA]; Polaromonas sp. strain JS666 GguR, 5-keto-4-deoxy-d-glucarate/galactarate), or both (e.g., Rhizobium leguminosarum MatR, malonate and citrate; E. coli UxuR, d-glucuronate and d-fructuronate) as their effectors (9, 10, 62, 63, 66–69). We find that similar to the case for B. subtilis GntR, the substrate of the metabolic pathway, d-galactonate, serves as the effector of DgoR. d-Galactonate binds DgoR, brings about a conformational change in the protein, and interferes with its binding to the target DNA (Fig. 8). To date, within the FadR subfamily, the details of effector-induced allosteric changes are known for E. coli, V. cholerae, and Vibrio alginolyticus FadR proteins through a comparison of their apo, DNA-bound, and effector-bound structures (51, 53, 70). Such studies on DgoR will be important in the future, especially because the structural and molecular details of effector-regulator interactions have not been investigated for any FadR subfamily member that regulates sugar acid metabolism.

A bioinformatics study has predicted the presence of DgoR and its binding sites in the genomes of 13 different bacterial species, with the majority of members belonging to the family Enterobacteriaceae (41). The recovery of missense mutations in dgoR in a human urinary tract isolate of E. coli that inactivate the repressor, enabling faster growth of bacteria in d-galactonate and its adaptation in the mouse gut (Fig. 6A and B) (32), suggests that mutations in the regulatory elements are preferably selected to allow simultaneous induction of structural dgo genes. Considering this, it will be interesting to investigate the similarities/differences in the regulation of the d-galactonate metabolic pathway in enteric bacteria and understand their implications in the relative colonization of these bacteria in their environmental niches.

MATERIALS AND METHODS

Media and growth conditions.

The media had the following compositions: lysogeny broth (LB), 5 g/liter yeast extract, 10 g/liter Bacto tryptone, and 5 g/liter NaCl; and M9 minimal medium, 5.3 g/liter Na₂HPO₄, 3 g/liter KH₂PO₄, 0.5 g/liter NaCl, 1 g/liter NH₄Cl, 0.12 g/liter MgSO₄, 2 mg/liter biotin, 2 mg/liter nicotinamide, 0.2 mg/liter riboflavin, and 2 mg/liter thiamine. Where required, M9 minimal medium was supplemented with one of the following carbon sources: d-glucose (10 mM), sodium d-gluconate (10 mM), d-galactose (10 mM), d-galactonate (10 mM), glycerol (0.4%, vol/vol), or both glycerol and d-galactonate. Except for d-galactonate, which was prepared from calcium d-galactonate (MP Biomedicals) (see below), all carbon sources were used as procured from commercial sources (d-glucose from Fisher Scientific and sodium d-gluconate, d-galactose, and glycerol from Sigma). 2-Dehydro-3-deoxy-d-galactonate, 2-dehydro-3-deoxy-d-galactonate 6-phosphate, d-glyceraldehyde 3-phosphate, and pyruvate were procured from Sigma. Media were solidified using 1.5% (wt/vol) Bacto agar. When required, ampicillin (100 µg/ml) or kanamycin (30 µg/ml) was used. Cultures were incubated at 37°C unless otherwise specified. Primary cultures were grown in 3 ml LB liquid medium. Secondary cultures were set up either in LB or in M9 minimal medium containing the desired carbon source with an initial optical density (OD) of ∼0.01, unless indicated otherwise.

Preparation of d-galactonate.

d-Galactonate was prepared from its calcium salt as described previously (71). Briefly, equivalent amounts of calcium d-galactonate and oxalic acid were mixed in boiling water and vortexed for 3 min. The milky solution was filtered through a 0.2-µm filter. The filtrate was immediately transferred to 4°C for 15 min. Crystals were collected and dried overnight at room temperature on a Whatman filter paper. Crystals of d-galactonate were stored at room temperature.

Strains, plasmids, and primers.

Table 1 provides a list of bacterial strains and plasmids, and Table S1 in the supplemental material provides a list of primers used in the study. E. coli BW25113 and various deletion strains in this background were used for dilution spotting experiments, growth curve assays, quantitative reverse transcription-PCR (qRT-PCR), and fluorescence reporter assays. For deletion strains obtained from the Keio collection (36), the two independent clones and/or their fresh transductants (made by P1 transduction) were analyzed to rule out genetic errors. Strain DH5α was used for cloning in plasmids pACYC177, pRC10, and pMAL-c2, while strain BW25142 (pir⁺) was used for cloning in the pVenus plasmid. BL21(DE3) was used for protein expression and purification.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype	Source and/or reference
Strains
DH5α	F− Δ(argF-lac)169 ϕ80dlacZ58(M15) glnX44(AS) λ− rfbC1 gyrA96(Nalr) recA1 endA1 spoT1 thiE1 hsdR17	New England Biolabs
BL21(DE3)	F− lon-11 Δ(ompT-nfrA)885 Δ(galM-ybhJ)884 λDE3 [lacI lacUV5-T7 gene 1 ind1 sam7 nin5] Δ46 [mal+]K-12(λs) hsdS10	New England Biolabs
BW25142	lacIq rrnB3 ΔlacZ4787 hsdR514 Δ(araBAD)567 Δ(rhaBAD)568 ΔphoBR580 rph-1 galU95 ΔendA9 uidA(ΔMluI)::pir-116 recA1	Rao lab (73)
BW25113	F− Δ(araD-araB)567 ΔlacZ4787(::rrnB-3) λ− rph-1 Δ(rhaD-rhaB)568 hsdR514	CGSCa (36)
dgoR::kan	BW25113 dgoR::kan Kanr	Keio collection (36)
dgoK::kan	BW25113 dgoK::kan Kanr	Keio collection (36)
dgoA::kan	BW25113 dgoA::kan Kanr	Keio collection (36)
dgoD::kan	BW25113 dgoD::kan Kanr	Keio collection (36)
dgoT::kan	BW25113 dgoT::kan Kanr	Keio collection (36)
RC2132	BW25113 dgo::kan Kanr	This work
RC3067	BW25113 dgoR:3×FLAG-kan Kanr	This work
RC2069	ΔdgoR (Kan cassette flipped from BW25113 dgoR::kan)	This work
RC12022	BW25113 attλ::[Kan promoterless-venus oriR6K] Kanr	This work
RC12023	ΔdgoR attλ::[Kan promoterless-venus oriR6K] Kanr	This work
RC12018	BW25113 attλ::[Kan Pdgo-venus oriR6K] Kanr	This work
RC12020	ΔdgoR attλ::[Kan Pdgo-venus oriR6K] Kanr	This work
RC14029	BW25113 attλ::[Kan Pdgo(Δ-35)-venus oriR6K] Kanr	This work
RC14030	ΔdgoR attλ::[Kan Pdgo(Δ-35)-venus oriR6K] Kanr	This work
RC14031	BW25113 attλ::[Kan Pdgo(Δ-35/-10)-venus oriR6K] Kanr	This work
RC14032	ΔdgoR attλ::[Kan Pdgo(Δ-35/-10)-venus oriR6K] Kanr	This work
RC14033	BW25113 attλ::[Kan Pup-dgoK-venus oriR6K] Kanr	This work
RC14034	ΔdgoR attλ::[Kan Pup-dgoK-venus oriR6K] Kanr	This work
RC14036	BW25113 attλ::[Kan Pup-dgoA-venus oriR6K] Kanr	This work
RC14037	ΔdgoR attλ::[Kan Pup-dgoA-venus oriR6K] Kanr	This work
RC14038	BW25113 attλ::[Kan Pup-dgoD-venus oriR6K] Kanr	This work
RC14039	ΔdgoR attλ::[Kan Pup-dgoD-venus oriR6K] Kanr	This work
RC14040	BW25113 attλ::[Kan Pup-dgoT-venus oriR6K] Kanr	This work
RC14041	ΔdgoR attλ::[Kan Pup-dgoT-venus oriR6K] Kanr	This work
RC12035	BW25113 attλ::[Kan Pdgo(IR1L)-venus oriR6K] Kanr	This work
RC12036	ΔdgoR attλ::[Kan Pdgo(IR1L)-venus oriR6K] Kanr	This work
RC12040	BW25113 attλ::[Kan Pdgo(IR1R)-venus oriR6K] Kanr	This work
RC12041	ΔdgoR attλ::[Kan Pdgo(IR1R)-venus oriR6K] Kanr	This work
RC12028	BW25113 attλ::[Kan Pdgo(IR2L)-venus oriR6K] Kanr	This work
RC12029	ΔdgoR attλ::[Kan Pdgo(IR2L)-venus oriR6K] Kanr	This work
RC12031	BW25113 attλ::[Kan Pdgo(IR2R)-venus oriR6K] Kanr	This work
RC12032	ΔdgoR attλ::[Kan Pdgo(IR2R)-venus oriR6K] Kanr	This work
RC12043	BW25113 attλ::[Kan Pdgo(IR2L')-venus oriR6K] Kanr	This work
RC12047	ΔdgoR attλ::[Kan Pdgo(IR2L')-venus oriR6K] Kanr	This work
RC14053	BW25113 attλ::[Kan Pdgo(IR3L)-venus oriR6K] Kanr	This work
RC14056	ΔdgoR attλ::[Kan Pdgo(IR3L)-venus oriR6K] Kanr	This work
RC12025	BW25113 attλ::[Kan Pdgo(IR3R)-venus oriR6K] Kanr	This work
RC12026	ΔdgoR attλ::[Kan Pdgo(IR3R)-venus oriR6K] Kanr	This work
Plasmids
pKD13	oriR6K, FRT-flanked Kanr, pANTSγ PS1 PS4 Kanr	72
pKD46	pSC101 ori araC repA101(Ts) ParaBAD-λred Ampr	72
pCP20	pSC101 ori cI857 λ-PR flp ts Ampr Camr	72
pACYC177	p15A ori Ampr Kanr	New England Biolabs
pBS13	dgo promoter and dgoR-6×His in pACYC177, Ampr Kanr	This work
pBS14	dgo promoter and dgoR(L34A)-6×His in pACYC177, Ampr Kanr	This work
pBS15	dgo promoter and dgoR(R46A)-6×His in pACYC177, Ampr Kanr	This work
pBS22	dgo promoter and dgoR(D7A)-6×His in pACYC177, Ampr Kanr	This work
pBS35	dgo promoter and dgoR(S51L)-6×His in pACYC177, Ampr Kanr	This work
pBS36	dgo promoter and dgoR(R42C)-6×His in pACYC177, Ampr Kanr	This work
pBS37	dgo promoter and dgoR(T40I)-6×His in pACYC177, Ampr Kanr	This work
pRC10	pBR322 ori, −10 box of Ptrc changed to Plac in pTrc99a, ΔNcoI Ampr	80
pBS2	dgoR-6×His in pRC10, Ampr	This work
pBS19	dgoR(L34A)-6×His in pRC10, Ampr	This work
pBS20	dgoR(R46A)-6×His in pRC10, Ampr	This work
pBS25	dgoR(D7A)-6×His in pRC10, Ampr	This work
pBS32	dgoR(S51L)-6×His in pRC10, Ampr	This work
pBS33	dgoR(R42C)-6×His in pRC10, Ampr	This work
pBS34	dgoR(T40I)-6×His in pRC10, Ampr	This work
pMAL-c2	pBR322 ori, leaderless malE MCS,b Ampr	New England Biolabs
pBS11	dgoR cloned in pMAL-c2, Ampr	This work
pINT-ts	oriR6K int Ampr	Rao lab (73)
pVenus	oriR6K, MCS venus t0 attλ, Kanr	Rao lab (74)
pAS1	oriR6K, MCS Pdgo-venus t0 attλ, Kanr	This work
pAS2	oriR6K, MCS Pdgo(IR1L)-venus t0 attλ, Kanr	This work
pAS3	oriR6K, MCS Pdgo(IR1R)-venus t0 attλ, Kanr	This work
pAS4	oriR6K, MCS Pdgo(IR2L)-venus t0 attλ, Kanr	This work
pAS5	oriR6K, MCS Pdgo(IR2R)-venus t0 attλ, Kanr	This work
pAS6	oriR6K, MCS Pdgo(IR2L')-venus t0 attλ, Kanr	This work
pAS7	oriR6K, MCS Pdgo(IR3L)-venus t0 attλ, Kanr	This work
pAS8	oriR6K, MCS Pdgo(IR3L')-venus t0 attλ, Kanr	This work
pNK1	oriR6K, MCS Pdgo(Δ-35)-venus t0 attλ, Kanr	This work
pNK2	oriR6K, MCS Pdgo(Δ-35/-10)-venus t0 attλ, Kanr	This work
pNK3	oriR6K, MCS Pup-dgoK-venus t0 attλ, Kanr	This work
pNK4	oriR6K, MCS Pup-dgoA-venus t0 attλ, Kanr	This work
pNK5	oriR6K, MCS Pup-dgoD-venus t0 attλ, Kanr	This work
pNK6	oriR6K, MCS Pup-dgoT-venus t0 attλ, Kanr	This work

^aCGSC, E. coli Genetic Stock Center.

^bMCS, multiple cloning site.

Strains carrying a deletion of the entire dgo operon or expressing C-terminally 3×FLAG-tagged DgoR (DgoR-3×FLAG) from the chromosome were constructed using the λ Red recombinase method (72). For the deletion of the dgo operon, a kanamycin cassette was PCR amplified from the plasmid pKD13, while for tagging the C terminus-encoding end of dgoR on the chromosome with 3×FLAG, the sequence of 3×FLAG and the kanamycin cassette was PCR amplified from a sequential peptide affinity (SPA)-tagged strain obtained from the Keio collection (36). The amplified cassettes flanked by regions homologous to the target sites were transformed into BW25113 expressing λ Red recombinase enzymes from the pKD46 helper plasmid. Recombinants were selected on LB-kanamycin plates. Strains were PCR verified. The dgo operon deletion and dgoR:3×FLAG-kan were subsequently transduced into a clean BW25113 background using P1 phage, yielding strains RC2132 and RC3067, respectively.

For complementation experiments, dgoR encoding a C-terminally 6×His-tagged protein was cloned in pACYC177. The coding region of dgoR along with its putative promoter (446 bp upstream of the dgoR start codon) was PCR amplified and cloned in AatII and BamHI sites of pACYC177 to generate plasmid pBS13. Various dgoR mutations in pACYC177 were created by overlap extension PCR using pBS13 as the template.

Plasmid pBS2 was constructed for high-level expression and purification of C-terminally 6×His-tagged DgoR (DgoR-6×His). Briefly, the coding region of dgoR was PCR amplified and cloned in the EcoRI and BamHI sites of IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible plasmid pRC10. For expression and purification of various DgoR mutant proteins, mutations were generated by overlap extension PCR using pBS2 as the template. Plasmid pBS11 was constructed for expression and purification of MBP-DgoR by amplifying the coding region of dgoR and cloning in the EcoRI and SalI sites of pMAL-c2 to create an in-frame fusion with MBP.

To determine functional promoters and DgoR-binding sites in the dgo region, single-copy transcriptional fusions were constructed in the pVenus integration vector by a previously described method (73, 74). Briefly, to identify functional promoters in the putative dgo operon, sequences upstream of dgoR and its truncated variants were PCR amplified from pBS13, while sequences upstream of other dgo genes (dgoK, dgoA, dgoD, and dgoT) were amplified from genomic DNA. To identify DgoR-binding sites, mutations were created in the dgo cis-acting element by overlap extension PCR using pBS13 as the template. The DNA fragments were cloned upstream of the fluorescent Venus reporter at the KpnI and EcoRI sites of the pVenus plasmid. The reporter constructs were integrated at the attλ site of BW25113 with the help of integrase expressed from the pINT-ts helper plasmid. Strains were PCR verified for single-copy integration. Reporter integrations were P1 transduced into clean BW25113 or ΔdgoR strains. The ΔdgoR strain (derived by flipping out the kanamycin cassette from the dgoR::kan strain) was used since the Venus reporter construct carried a kanamycin marker.

Dilution spotting.

Overnight cultures were pelleted and resuspended in M9 minimal medium without any carbon source to a similar OD at 450 nm (OD₄₅₀). Serial dilutions of resuspended cells were spotted on M9 minimal agar plates containing the desired carbon source. Plates were incubated at 37°C and imaged at definite time intervals using the Gel Doc XR⁺ imaging system from Bio-Rad.

RNA isolation, cDNA preparation, and qRT-PCR.

Secondary cultures were grown in 25 ml M9 minimal medium with the desired carbon source to an OD₄₅₀ of ∼0.5, and 20-ml samples were collected and processed for RNA isolation by the hot-phenol method as described previously (75). RNA suspended in nuclease-free water was treated with DNA-free Turbo DNase (Applied Biosystems, USA) to avoid genomic DNA contamination, and 4 µg RNA was used to prepare cDNA using SuperScript III reverse transcriptase (Applied Biosystems, USA) using the protocol described previously (76). qRT-PCR was carried out with Power Sybr green PCR master mix (Applied Biosystems, USA) in accordance with the manufacturer’s recommended protocol using 50 ng cDNA as the template and 5 pmol of each forward and reverse primer listed in Table S1. qRT-PCR was performed using the Quant Studio 6 Flex system (Applied Biosystems, USA). Data were analyzed as described previously (77) using recA and gyrA as internal controls.

Western blotting.

The expression of DgoR-3×FLAG, DgoR-6×His, and MBP-DgoR was monitored by Western blot analysis. Samples were subjected to SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blocked with 5% (wt/vol) skimmed milk at 4°C overnight and probed with anti-FLAG (1:1,000; Sigma) or anti-His (1:1,000; Thermo Fisher Scientific) primary antibody and horseradish peroxidase (HRP)-conjugated anti-mouse (1:5,000; Sigma) secondary antibody. Where required, anti-MBP conjugated to HRP (1:5,000; New England Biolabs) was used. Blots were developed with the SuperSignal West Dura extended-duration substrate (Pierce). Signal was captured with an LAS4000 instrument (GE Health Care) or on X-ray film.

Growth curve experiments in shake flasks and 96-well plates.

Overnight cultures were pelleted, washed, and resuspended in M9 minimal medium. For growth curves in shake flasks, cells were reinoculated in 25 ml M9 minimal medium supplemented with the desired carbon source (in 125-ml flasks) to an initial OD₄₅₀ of ∼0.01. Cultures were grown at 37°C, and the OD₄₅₀ was measured at defined time intervals. For growth curves determined in 96-well plates, cells were reinoculated in 200 µl medium to a starting OD₄₅₀ of ∼0.03, using a robotic liquid handling system (Tecan). Plates were incubated in a shaker at 37°C, and the OD₄₅₀ of the cultures was measured at hourly intervals (Tecan Infinite M200 monochromator). The incubator shaker and microplate reader were integrated with the liquid handling system, and the transfer of plates between shaker and reader was automated.

Overexpression and purification of DgoR-6×His, MBP-βgal, and MBP-DgoR.

For the purification of WT and mutant DgoR-6×His proteins, cultures were grown in 400 ml LB at 37°C to an OD₆₀₀ of ∼0.5. Cultures were induced with 0.5 mM IPTG and grown overnight at 18°C. The cell pellet was resuspended in 20 ml lysis buffer containing 50 mM sodium phosphate (pH 7.5), 300 mM NaCl, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 20 mM imidazole and disrupted by sonication. Cell debris was removed by centrifugation. The supernatant was incubated with Co-NTA beads (Pierce) on ice for 1 to 1.5 h with continuous shaking. Beads were washed with 40 ml wash buffer A containing 50 mM sodium phosphate (pH 7.5), 1 M NaCl, and 20 mM imidazole, followed by washing with 40 ml wash buffer B containing 50 mM sodium phosphate (pH 7.5), 300 mM NaCl, and 60 mM imidazole, and finally the protein was eluted in elution buffer containing 50 mM sodium phosphate (pH 7.5), 300 mM NaCl, and 500 mM imidazole. The eluted protein was dialyzed against buffer containing 20 mM Tris (pH 8.0), 300 mM NaCl, 1 mM dithiothreitol (DTT), and 10% (wt/vol) glycerol.

For the purification of MBP-βgal and MBP-DgoR, cultures were grown in 400 ml LB supplemented with 2 g/liter d-glucose at 37°C until the OD₆₀₀ reached 0.5. Cultures were induced with 1 mM IPTG for 4 h. The cell pellet was resuspended in column buffer containing 20 mM HEPES (pH 7.5), 300 mM NaCl, and 1 mM EDTA (5 ml/g cell pellet) and lysed by sonication. The lysate was centrifuged to remove cell debris, and the supernatant was incubated with amylose resin for 1.5 to 2 h at 4°C. Amylose beads were washed with five bed volumes of column buffer and eluted with one bed volume of the same buffer containing 10 mM maltose. Proteins were analyzed by SDS-PAGE to assess purity and stored at −80°C until use. The concentrations of the proteins were determined by the Bradford assay using bovine serum albumin as the standard, as per the manufacturer’s protocol (Sigma).

EMSA.

For electrophoretic mobility shift assay (EMSA), the dgo cis-acting elements were PCR amplified with primers BS104 and ³²P-labeled BS105, using appropriate pVenus constructs as the templates, and gel purified. A 20-µl reaction mixture was set up in protein dialysis buffer supplemented with 10 mM MgCl₂ and 1 mM PMSF, to which 20 nM labeled DNA, 1 µg herring sperm DNA, and either WT or mutant DgoR-6×His proteins were added. The reaction mix was incubated at 27°C for 30 min, and samples were subjected to 8% native PAGE at 60 V for 6 h. The signal was visualized using a PhosphorImager (Fuji or Bio-Rad). For experiments involving effectors, DgoR-6×His was incubated with effectors at 37°C for 20 min before incubation with DNA.

Fluorescence assay.

Overnight cultures of reporter strains in LB were pelleted and resuspended in M9 minimal medium. Secondary cultures were set up in 96-well, black, clear-bottom plates (Costar, Corning) in M9 minimal medium supplemented with either glycerol or glycerol and d-galactonate, followed by incubation at 37°C with shaking. Fluorescence was measured in exponential phase using a microplate reader (Tecan Infinite M200 monochromator) in the top mode with excitation and emission wavelengths of 498 nm and 568 nm, respectively. Fluorescence normalized to the OD₄₅₀ of the samples was plotted in a bar graph.

Pulldown experiments.

Culture expressing DgoR-6×His was harvested, and the cell pellet was resuspended in lysis buffer (50 mM sodium phosphate [pH 7.5], 300 mM NaCl, 1 mM PMSF, and 20 mM imidazole). Cells were lysed by sonication and centrifuged to remove the debris. The supernatant was incubated with 60 µl Co-NTA beads at 4°C for 90 min. Beads were washed with 35 bed volumes of lysis buffer. The beads were equally split into two fractions. Purified MBP-DgoR was added to one fraction, while to the other fraction purified MBP-βgal was added, and immediately a portion of these suspensions was saved as “input.” The remaining suspension was incubated for 20 min at 4°C. Beads were washed with 25 bed volumes of column buffer (20 mM Tris [pH 7.4], 200 mM NaCl, and 1 mM EDTA). Proteins were finally eluted with 120 µl elution buffer (50 mM sodium phosphate [pH 7.5], 300 mM NaCl, and 300 mM imidazole). The eluted sample was saved as the “pulldown.”

Limited proteolysis assay.

Trypsin digestion was performed following the procedure described previously (45, 47) with modifications. A 200-µl reaction mixture was set up at 37°C in protein dialysis buffer supplemented with 500 mM NaCl, 10 mM MgCl₂, and 0.02 mg/ml trypsin (Sigma), to which either DgoR-6×His (10 µM) alone or both protein and effectors (2 mM) were added. Twenty-microliter samples were withdrawn at various time intervals. The reaction was quenched by adding SDS sample buffer (4× stock: 10% [wt/vol] glycerol, 0.6% [wt/vol] Tris-HCl [pH 6.8], 2% [wt/vol] SDS, 0.02% [wt/vol] bromophenol blue, and 1.5% [wt/vol] DTT). Samples were boiled at 95°C for 5 min and subjected to 15% SDS-PAGE followed by staining with Coomassie brilliant blue.

Bioinformatics approaches.

The amino acid sequence of full-length DgoR was submitted to the Robetta server (78) for model building. The top model build by Robetta was used for analysis. The PDB database was searched for proteins with a structure similar to that of the N-terminal wHTH domain of DgoR (from amino acids 1 to 73) using the DALI server (79). The DALI server was also used to generate structure-based sequence alignment of DgoR with selected members of the GntR family. ICM-Browser (www.molsoft.com) was used for model visualization, structure comparison and analysis, and preparation of related figures.