Further Unraveling the Regulatory Twist by Elucidating Metabolic Coinducer-Mediated CbbR-cbb_I Promoter Interactions in Rhodopseudomonas palustris CGA010

Abstract

The cbb_I region of Rhodopseudomonas palustris (Rp. palustris) contains the cbbLS genes encoding form I ribulose-1,5-bisphosphate (RuBP) carboxylase oxygenase (RubisCO) along with a divergently transcribed regulator gene, cbbR. Juxtaposed between cbbR and cbbLS are the cbbRRS genes, encoding an unusual three-protein two-component (CbbRRS) system that modulates the ability of CbbR to influence cbbLS expression. The nature of the metabolic signals that Rp. palustris CbbR perceives to regulate cbbLS transcription is not known. Thus, in this study, the CbbR binding region was first mapped within the cbbLS promoter by the use of gel mobility shift assays and DNase I footprinting. In addition, potential metabolic coinducers (metabolites) were tested for their ability to alter the cbbLS promoter binding properties of CbbR. Gel mobility shift assays and surface plasmon resonance analyses together indicated that biosynthetic intermediates such as RuBP, ATP, fructose 1,6-bisphosphate, and NADPH enhanced DNA binding by CbbR. These coinducers did not yield identical CbbR-dependent DNase I footprints, indicating that the coinducers caused significant changes in DNA structure. These in vitro studies suggest that cellular signals such as fluctuating metabolite concentrations are perceived by and transduced to the cbbLS promoter via the master regulator CbbR.

INTRODUCTION

In Rhodopseudomonas palustris (Rp. palustris), CO₂ is primarily assimilated via the Calvin-Benson-Bassham (CBB) reductive pentose phosphate pathway, with most of the key enzymes encoded within the cbb_I and cbb_II operons (18–20). The cbb_I regulon of Rp. palustris contains the divergently transcribed cbbR gene, encoding a LysR-type transcriptional regulator (LTTR), CbbR, and the cbbLS genes encoding the large and small subunits of form I ribulose-1,5-bisphosphate (RuBP) carboxylase oxygenase (RubisCO). CbbR is the master and positive regulator of the cbb_I operon and specifically controls the expression of the cbbLS genes, encoding form I RubisCO, while the cbb_II operon, which contains other structural genes of the CBB pathway, including cbbM, encoding form II RubisCO, is constitutively expressed. Juxtaposed between the divergently transcribed cbbR and cbbLS genes of the cbb_I regulon is a unique three-protein two-component regulatory system (the CbbRRS system), encoded by the cbbRR1, cbbRR2, and cbbSR genes. The hybrid sensor kinase CbbSR contains an N-terminal sensor domain, a central transmitter domain, and a C-terminal receiver domain. The two atypical response regulators (CbbRR1 [response regulator 1 of the CbbRRS system] and CbbRR2) are unique in that they contain domains that are normally part of hybrid sensor kinases; response regulator 1 (CbbRR1) contains a phosphate receiver domain and a histidine phosphotransfer domain, while response regulator 2 (CbbRR2) contains two phosphate receiver domains (18, 19). In addition, these regulators do not contain traditional DNA binding domains and are not able to bind to any discernible target DNA sequences, and yet they are able to interact and modify the properties of CbbR (11) and influence CbbR-mediated cbbLS transcription under specialized growth conditions (10, 11, 18).

As a member of the well-studied LysR-type transcriptional regulator (LTTR) family of regulators, CbbR has been shown to positively regulate cbb operon expression in several photosynthetic and chemoautotrophic bacteria (7, 12, 15, 22). LTTRs usually bind a symmetric sequence motif; T-N₁₁-A, located anywhere between 65 and 100 bp upstream of the transcription start (8, 13, 21, 31). The functional domains on the protein comprise a conserved helix-turn-helix DNA binding domain at the amino terminus, a central coinducer recognition domain, and a C-terminal oligomerization domain. There is little sequence conservation within the coinducer binding domain, as different LTTRs encounter a variety of coinducer molecules, depending on the cellular process that is regulated (5, 13, 20–22). Often, binding of a coinducer molecule to the LTTR is essential for transcriptional activation of its target gene, believed to occur by virtue of changes in DNA bending and interactions with RNA polymerase (RNAP). The major coinducer is usually a unique metabolite of the pathway regulated by the specific LTTR. Coinducers confer altered DNA binding properties to the LTTR protein, possibly due to a conformation change to the tertiary structure of the protein (13).

Previous studies indicated that the CbbRRS system differentially regulates cbb_I operon transcription under different growth conditions and is specifically involved in the regulation of form I RubisCO synthesis in photoautotrophically grown cells (10, 18). Subsequent studies from this laboratory have provided direct evidence of specific protein-protein interactions between CbbR and CbbRR1, and both response regulators (CbbRR1 and CbbRR2) influence the interaction of CbbR at the cbb_I promoter (11). When these initial findings were established, it was apparent that a detailed and systematic probe of the interaction of CbbR with target sequences within the cbb_I region would be necessary in order to gain an understanding of the interplay between the metabolic and regulatory protein factors involved in the control of cbb_I operon transcription. In this study, after the CbbR binding region was mapped, the role of different coinducer molecules in influencing interactions of CbbR with specific binding sites within the cbbLS promoter region was ascertained, strongly suggesting that biosynthetic intermediates and other potential metabolic effectors play an important part in the regulatory mechanism.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The plasmids used in this study are listed in Table 1.

Table 1.

Plasmids used in this study

Plasmid	Descriptiona	Reference or source
pQE80	cis-repressed, IPTG-inducible, N-terminal His6-tagged recombinant protein expression vector; Ampr	Qiagen
pQE160	cbbR coding region cloned into the BamHI-HindIII sites of pQE80; Ampr	Laboratory stock
pGSJSP	Intergenic region between cbbSR and cbbL (199 bp) in pCR-Blunt	11
pGSJSP90	Intergenic region between cbbSR and cbbL (90 bp) in pCR-Blunt	11
pGSJFP	Intergenic region between cbbSR and cbbL (199 bp) extending 216 bp 5′ into cbbSR and 44 bp 3′ into cbbL	This study

^aIPTG, isopropyl-β-d-thiogalactopyranoside; Amp^r, ampicillin resistant.

DNA manipulations.

All DNA manipulations such as cloning, restriction enzyme digestion, DNA ligation, and agarose gel electrophoresis were performed in accordance with standard protocols as previously described (10, 11).

Synthesis and purification of CbbR.

N-terminal His₆-tagged recombinant CbbR was synthesized in Escherichia coli BL21(DE3) and purified by nickel affinity chromatography as previously described (11).

Gel mobility shift assays.

Gel mobility shift assays were performed as previously described (2, 11). The pGSJSP and pGSJSP90 plasmids, which contained the intergenic region between cbbSR and cbbL, were used for PCR amplification of the target DNA and probes SP and SP90. The primers for PCR amplification of SP and SP90 used for CbbR binding were primers 5′-GATCGGATCCTCGCGCCTGCCGCGATATAAG-3′ and 5′-GATCGGATCCGTCGTCCTCCTTGAAAGCCC-3′ and primers 5′-GATCGGATCCTCGCGCCTGCCGCGATATAAG-3′ and 5′-GATCGGATCCCGGGAGGACTTTGACCACAAC-3′, respectively (underlined sequences represent restriction sites for convenient subsequent labeling reactions). Target DNA regions (Fig. 1) SP2 (5′-GATCGGATCCCGGCGTTGTGGTCAAAGTCCTCC-3′ and 5′-GATCGGATCCGTCGTCCTCCTTGAAAGCCC-3′), SPΔ1 (106 bp; 5′-GATCGGATCCGCGATATCGAAG-3′ and 5′-GATCGGATCCTCGGTCGCGAAGTTC-3′), and SPΔ2 (96 bp; 5′-GATCGGATCCATGCCGCGCTCGACAATG-3′ and 5′-GATCGGATCCCGATATCGCAAAAG-3′) were PCR amplified and used for elucidating the CbbR binding sites within the cbb_I promoter region. The target DNA regions were labeled with [³²P]CTP in an end-filling reaction using Klenow DNA polymerase. Each of the binding reaction mixtures contained 0.3 nM labeled target. Binding reaction mixtures (50 μl) were set up containing CbbR (25 nM) radiolabeled target (∼10,000 cpm) and poly(dI-dC) (1.83 μg/μl) in a buffer containing 30 mM potassium glutamate, 10 mM Tris-Cl (pH 8.5), 30% glycerol, and 1 mM dithiothreitol (DTT). CbbR was incubated in the presence of competitor poly(dI-dC) DNA for 5 min at room temperature prior to the addition of the radiolabeled target DNA. Coinducers were included in the reaction mixture to achieve a final concentration ranging between 5 and 500 μM and were added prior to the addition of the target. Controls containing similar ionic strengths of inert salts were routinely employed and run beside the experimental reaction mixtures. The reaction mixture was incubated for 6 min at room temperature after addition of the target. The samples were separated in a Tris-glycine buffer system (3). Following drying, the gel was analyzed by autoradiography and visualized with a Storm 840 PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Fig 1.

Gene organization of the Rp. palustris cbb_I operon. (A) The nucleotide sequence of the cbbSR-cbbL intergenic region (SP) containing LTTR consensus binding motif T-N₁₁-A is indicated. Putative CbbR binding sites (1 and 2) are highlighted in red. This region was used to generate DNA targets SP, SP90, SP2, SPΔ1, and SPΔ2, sequences of which are indicated, for gel mobility shift analyses. (B) Comparison of the cbbSR-cbbL intergenic region of Rp. palustris (R. pal) with those of the Rb. sphaeroides (R. sph) and Rb. capsulatus (R. cap) cbb_I operons. Presumptive CbbR binding sites (T-N₁₁-A) in the Rp. palustris sequence are indicated.

DNase I footprinting.

To map the binding site(s) of CbbR within the cbb_I promoter region, DNase I footprinting was performed using fluorescently labeled DNA and an automated fluorescent DNA analyzer to resolve the digestion products, as described by Zianni et al. (33). Plasmid pGSJFP, which contains the intergenic region between cbbSR and cbbL (199 bp) and extends 216 bp into cbbSR and 44 bp into cbbL, was used as the template to generate the 459-bp probe (FP). The probe was generated by PCR amplification with the primers Fpcbb1-VIC [5′-(VIC)-AAGCTCGATCACGAGGCG-3′] and Fpcbb2-FAM [5′-6-carboxyfluorescein (FAM)-GACTTGTAGCGTTCCTTGC-3′] from Applied Biosystems. PCR was performed for 30 cycles under the following conditions: 94 C for 60 s, 54 C for 30 s, and 72 C for 30 s. PCR products were gel purified and quantified by UV spectrophotometry. CbbR (100 nM) was then incubated with 1.83 μg of poly(dI-dC) for 10 min at room temperature in binding buffer (30 mM potassium glutamate, 1 mM DTT, 5 mM magnesium acetate, 2 mM calcium chloride, 0.125 mg/ml bovine serum albumin [BSA], 30% glycerol, 10 mM Tris-Cl [pH 8.5]). After this incubation, 100 to 120 ng of fluorescently labeled probe was added to the reaction mixture (final volume, 20 μl), which was then incubated for 20 min at room temperature. DNase I (Worthington Biochemicals, Lakewood, NJ) digestion was performed with 0.125 μg of DNase I per 20 μl of reaction mixture for 5 min at room temperature. The reaction was then stopped by incubation at 75°C for 10 min. Control digestions with the probe were performed in the absence of the protein or in the presence of BSA. The DNA fragments were purified with a QIAquick PCR purification kit (Qiagen, Valencia, CA) and eluted in 40 μl of water. The digested DNA was combined with a Genescan 600 LIZ size standard (Applied Biosystems). Samples were analyzed with a 3730 DNA analyzer (Applied Biosystems). The electropherograms were aligned using GeneMapper 4.0 software (Applied Biosystems), and the signal was normalized for total peak area per electropherogram. The resulting DNA fragments were organized into bins (sets of equivalent peaks from each sample) through a combination of software automation and manual selection to account for fragment migration variation. The differences between the peak heights corresponding to the different conditions and the negative control (BSA) were calculated, and the differences were exported to Sigma Plot 11.0 for display as a histogram. The 3′ base of each fragment was determined by dideoxy/Sanger DNA sequencing utilizing the same plasmid and primers as described above in combination with a Thermo Sequenase dye primer kit (USB, Cleveland, OH) according to the manufacturer's recommended protocol. The resulting primer extension products were analyzed and aligned to the same bin set as used in the procedure described above.

Surface plasmon resonance (SPR) detection of protein-DNA interactions.

CbbR-DNA interactions were determined as described previously (11). CbbR-DNA interactions and the effects of coinducers on CbbR binding were determined with a fully automated BIAcore T100 system (Uppsala, Sweden). For binding, the biotinylated DNA (in HEPES-buffered saline [HBS] with 0.5 M NaCl) was injected at a flow rate of 5 μl/min (for 540 s). This resulted in a response of 270 response units (RU). CbbR was purified as described previously (11). In tests of the interaction of CbbR with DNA, 50 mM Tris-Cl (pH 7.5)–200 mM NaCl–5 mM MgCl₂–1 mM DTT was used as the running buffer. To determine the effect of the coinducers on CbbR binding, a 61-μl sample of CbbR (0.065 nM) containing the coinducer at the desired micromolar concentration in running buffer was injected over the sensor surface. Binding analysis was performed after the CbbR samples were incubated for 5 min at room temperature after the addition of the coinducer. Kinetic data analysis of CbbR binding reactions at the cbb_I promoter was done using BIAevaluation software version 4.0 and a 1:1 kinetic binding model. This model describes a 1:1 interaction at the sensor surface as follows: A + B = AB, where A represents the concentration of the analyte (CbbR), B represents the total ligand (DNA) on the sensor surface, and AB represents the initial binding response. The parameters reported by the model include the association rate constant (k_a [per mole per second]), the dissociation rate constant (k_d [per second]), and the affinity constant (K_D = k_d/k_a [in nanomoles]).

RESULTS

Gel mobility shift assays to map the region of CbbR that binds to the cbbLS promoter.

CbbR was purified as described previously (11). Inspection of the intergenic region between cbbSR and cbbL revealed the presence of two AT-rich sites (reference 11 and Fig. 1). These putative CbbR binding sites, homologous to other CbbR binding sites (3, 4, 29, 30), prompted the use of this region as a target in subsequent binding experiments. After verifying specific binding of CbbR to this region in gel mobility shift assays, to further narrow down the CbbR binding sites, a second, shorter target DNA region was generated (SP90; 90 bp), which contained the first 90 bp between cbbSR and cbbL. There was efficient CbbR binding to this region, as it encompassed the putative consensus binding sites (reference 11 and Fig. 1). It is known that CbbRR1 (response regulator 1 of the CbbRRS system) enhances the binding affinity of CbbR to this target 3- to 5-fold, and various control experiments verified the specificity of this interaction (11; see also Fig. S1 in the supplemental material). Clearly, the two AT-rich sites, as expected, contained the CbbR binding sites, as no detectable CbbR binding to the target DNA was obtained with probe SP2 (134 bp) due to deletion of the two AT-rich sites. Consequently, response regulator CbbRR1 had no effect on binding of this probe (Fig. 2A), as previously shown (11). From these results, the binding region was narrowed to the first 65 bp 3′ of cbbSR (Fig. 1A). The presence of two subsites within the binding region seems to be a common feature of LTTR-controlled promoters (21). At concentrations ranging between 25 and 100 nM, and in the absence of coinducer, it was observed that CbbR binding was more severely affected by the absence of site 1 than site 2. In addition, the binding of CbbR to site 2 was substantially improved in the presence of CbbRR1 (Fig. 2B and C).

Fig 2.

Deletion analysis to map the Rp. palustris CbbR binding region. (A) Phosphorimage of a gel mobility shift assay, demonstrating specific binding of CbbR (50 nM) to [³²P]-labeled target SP90 DNA (0.3 nM) (lane 2). The presence of CbbRR1 increased CbbR binding at the cbb_I promoter (lane 3). Deletion of the putative CbbR binding sites (SP2) resulted in loss of CbbR binding (lanes 5 and 7), and therefore there was no effect of CbbRR1 on the interaction (lanes 6 and 8). SP90 and SP2 free-target DNA data are presented in lanes 1 and 4, respectively. (B and C) The CbbR binding region is comprised of 2 subsites. Phosphorimages from a gel mobility shift assay demonstrating CbbR binding (5 to 100 nM) in the absence of site 1, probe SPΔ1 (B; see Fig. 1), or site 2, probe SPΔ2 (C; see Fig. 1), are shown.

DNase I footprinting to map the region of CbbR binding to the cbbLS promoter.

DNase I footprint analysis was performed to confirm the results of the gel mobility shift analyses as well as obtain information of greater precision concerning the CbbR binding sites. The fluorescently labeled FP probe (459 bp), consisting of the entire intergenic region between cbbSR and cbbL and extending 216 bp into cbbSR and 44 bp into cbbL on either side of the intergenic region, was incubated with 100 nM His-tagged CbbR and then digested with DNase I. The digested DNA fragments were analyzed by capillary electrophoresis using an automated sequencer. The digested DNA fragment pattern after analysis revealed protected regions within the probe extending between −202 bp and −120 bp (relative to ATG translation start of cbbL), with strong hypersensitive or unprotected sites at positions −175, −162, −142, and −125 bp (Fig. 3B; see also Fig. 7A). Negative controls consisted of binding reactions in which CbbR was replaced by BSA (Fig. 3B) and no protein. The negative-control reactions (BSA or no protein) demonstrated a uniform peak pattern across the entire DNA probe and no protected regions.

Fig 3.

Binding of CbbR to the Rp. palustris cbbLS promoter. (A) Comparison of the cbb_I promoter region of Rb. capsulatus (R. cap) with that of Rp. palustris (R. pal). The sequences beneath the solid line represent the −35 and −10 regions of the Rb. capsulatus cbb promoter. The putative −35 and −10 regions of the Rp. palustris cbb promoter are indicated by brackets. Hyphens indicate gaps inserted to maximize sequence identity. +1, transcription start of Rb. capsulatus cbbL. (B) DNase I footprinting of the CbbR binding sites on the cbb_I region of Rp. palustris. The DNase I-digested reactions were prepared and analyzed by capillary electrophoresis with a 3730 DNA analyzer (Applied Biosystems). The differences in the peak heights of CbbR and the negative control (BSA) were calculated for each nucleotide position after digestion of the probe incubated with either CbbR or BSA as a negative control. Protection is represented by negative peak heights, while positive peaks are signs of hypersensitivity. Nucleotide positions refer to the region of CbbR binding (−202 to −120) on the probe DNA relative to the translation start site (ATG) of cbbL.

Fig 7.

Summary of DNase I protection results in the absence (A) or presence (B to E) of coinducers, including (B) RuBP, (C) NADPH, (D) ATP, and (E) FBP, each at 500 μM. Results are compared to a BSA control. Putative CbbR binding sites are indicated in bold. The nucleotides protected by CbbR binding in the cbbSR-cbbL intergenic region are highlighted in yellow, the moderately protected sites in green, the strongly hypersensitive sites in turquoise, the moderately hypersensitive sites in pink, and the stop (TGA) and start (ATG) codons of cbbSR and cbbL, respectively, in gray. The putative −35 and −10 regions are underlined.

The total protected region was large, extending over 55 bp in length, with moderately hypersensitive sites observed at positions −168, −160, −140, and −131 bp. Thus, the CbbR protected region lies about 30 to 35 bp upstream of the putative transcription start and is in close proximity to or partly overlaps the putative −35 region, suggesting potential interactions with RNAP that are important for transcriptional activation (Fig. 3A; see also Fig. 7A). The mapped CbbR binding sites comprise two distinct and nonoverlapping regions to which Rp. palustris CbbR binds, which is similar to what was previously observed for CbbR from Rhodobacter sphaeroides (Rb. sphaeroides) (4, 33). The two protected regions span the putative AT-rich sites identified to be essential for CbbR binding in gel mobility shift assays via deletion analysis. Thus, the results of the DNase I footprint analysis confirmed the gel mobility shift experiment findings (Fig. 2A).

CbbR binding in gel mobility shift assays in the presence of coinducers.

A general mechanism of LTTR-mediated activation involves coinducer binding, resulting in enhanced transcriptional activation. Binding of the coinducer usually influences the DNA bend angle or mobility of the protein-DNA complex (13). The signals that influence the interaction of CbbR with its target promoter are thought to arise from perturbations in the concentrations of metabolites of the central carbon (biosynthetic) and/or energy pathways (23, 26). Thus, the role of potential physiologically relevant coinducer molecules in alteration of the DNA binding ability of Rp. palustris CbbR was tested by gel mobility shift assays (Fig. 4; see also Fig. S1 in the supplemental material). The potential coinducers examined were phosphorylated CBB cycle and glycolytic intermediates, as well as redox/energy indicators present in the cell that might likely influence biosynthetic processes. For these experiments, the amount of CbbR used in mobility shift assays was just enough to show minimum binding in positive-control reaction mixtures containing no added coinducers. The binding of Rp. palustris CbbR to the cbbLS promoter was visually and considerably enhanced in the presence of ATP (2.5-fold), FBP (2-fold), RuBP (1.6-fold), and NADPH (4-fold). This was evident as a large increase in the intensity of the single CbbR-DNA complex, seen only in the presence of the coinducer (Fig. 4). There was a 2- to 3-fold increase in CbbR binding in the presence of ATP, FBP, and RuBP, indicating that these coinducers may act cooperatively on CbbR (see Fig. S1 in the supplemental material).

Fig 4.

Binding of CbbR to the cbbLS promoter in the presence of various coinducers at 500 μM. Phosphorimages of gel shift assays performed with Rp. palustris CbbR (25 nM) and the cbb_I target SP90 DNA are pictured. (A) Lanes: target DNA only (lane 1), CbbR and target DNA, no coinducer (lane 2), plus phosphoenolpyruvate (PEP) (lane 3), plus 3-phosphoglycerate (3-PGA) (lane 4), plus ATP (lane 5), plus fructose-1,6-bisphosphate (FBP) (lane 6), plus NADP (lane 7), plus ribose-5-phosphate (R5P) (lane 8), plus RuBP (lane 9). (B) Lanes: target DNA only (lane 1), CbbR and target DNA, no coinducer (lane 2), plus NADPH (lane 3), plus ribulose-5-phosphate (Ru5P) (lane 4). The free DNA and CbbR-DNA complexes are indicated.

CbbR DNA binding was not affected by the presence of 3-phosphoglycerate (3-PGA), NADP (Fig. 4A), ADP, fructose-6-phosphate, glucose-6-phosphate, or dihydroxyacetone phosphate (DHAP) (see Fig. S2 in the supplemental material). It appeared that CbbR DNA binding was inhibited in the presence of phosphoenolpyruvate (PEP; Fig. 4A), ribose-5-phosphate, ribulose-5-phosphate (Fig. 4A and B), and KH₂PO₄ (see Fig. S2 in the supplemental material). The latter compounds may act as negative effectors, and their accumulation during cellular metabolism might serve to decrease cbbLS transcription. However, KH₂PO₄ had no appreciable effect on CbbR binding in the presence of ATP, FBP, and RuBP (see Fig. S1 in the supplemental material).

RuBP is a metabolite totally specific to RubisCO and the CBB pathway, as this compound has no known metabolic fate beyond carboxylation or oxygenation catalyzed by RubisCO (24). Thus, it was particularly relevant to determine whether this compound might affect the binding of CbbR to the cbbLS promoter. The RuBP concentration dependence of the gel mobility shift assays revealed that there was a direct correlation of CbbR with DNA binding as the concentration increased from 1 to 100 μM (Fig. 5B); the binding did not vary significantly at concentrations greater than 100 μM (Fig. 5A). As a consequence of these results, additional gel mobility shift assays were performed to examine whether CbbR binding was affected by different levels of the other coinducers. Thus, a range of different concentrations (0 to 250 μM) of ATP, FBP, and NADPH were tested for their ability to influence binding of CbbR to the DNA. For all the three positive coinducers (ATP, FBP, and NADPH), it was observed that a concentration of 250 μM stimulated CbbR binding. At a lower concentration (50 μM), CbbR binding either was reduced (NADPH, ATP) or remained similar to that seen with CbbR alone (RuBP, FBP) (see Fig. S3 in the supplemental material).

Fig 5.

Binding of CbbR to target SP90 DNA in the presence of RuBP. The figure presents phosphorimages of gel mobility shift assays showing CbbR binding in the presence of increasing concentrations of RuBP from 100 to 1,000 μM (A) and concentrations from 1 to 100 μM (B).

In summary, various coinducer compounds resulted in increased CbbR-DNA binding and produced a single predominant protein-DNA complex, which is similar to results achieved in the absence of the coinducer. The fact that the presence of ADP, fructose-6-phosphate (see Fig. S2 in the supplemental material), ribulose-5-phosphate (Fig. 4B), and NADP (Fig. 4A) did not increase the ability of CbbR to bind to the DNA suggested that the observed results obtained with other compounds were not due to nonspecific effects of phosphorylated compounds. The presence of the coinducer did not alter the electrophoretic mobility of the complex or the number of protein-DNA complexes, as has been observed in the Rb. sphaeroides and Rb. capsulatus systems (2, 6). Thus, the four coinducers (ATP, RuBP, FBP, and NADPH) were selected for further CbbR-DNA binding experiments, as well as DNase I protection and SPR determinations.

DNase I protection assays to determine coinducer effects on CbbR binding.

Coinducer binding to LTTR proteins often influences the footprint at the target promoter (1, 2, 6, 13, 14, 17, 27, 32). In the absence of a coinducer, it was shown that Rp. palustris CbbR (100 nM) protected regions within the probe FP extending between −202 bp and −120 bp (relative to the ATG translation start of cbbL) (Fig. 3B; see also Fig. 7A). The two CbbR binding sites are located within the cbbSR-cbbL intergenic region.

In the presence of RuBP (500 μM), there was no change in the length of the footprint; however, there was increased protection of the 2 AT-rich sites in the promoter region from DNase I cleavage due to CbbR binding (Fig. 6A). This protection was also accompanied by the disappearance of the hypersensitivity at −142 and the appearance of a new and strongly hypersensitive site at position −140 in addition to the ones at −175, −162, and −125 resulting from CbbR binding in the absence of coinducer (Fig. 6A). This footprint experiment suggested that RuBP influenced CbbR binding to favor the formation of a CbbR-DNA complex. An almost identical protection pattern was observed in the presence of NADPH (500 μM) (Fig. 6A, Fig. 7B and C), e.g., the appearance of a new hypersensitive site at position −140 as well as retention of other hypersensitive sites compared to CbbR alone.

Fig 6.

DNase I footprinting of the CbbR binding sites on the cbbLS region in the presence of coinducer. The DNase I-digested reactions were prepared and analyzed by capillary electrophoresis with a 3730 DNA analyzer (Applied Biosystems). The differences in the peak heights of CbbR and the negative control (BSA) were calculated for each nucleotide position after digestion of the probe incubated with either CbbR plus coinducer (RuBP/NADPH [A] and ATP/FBP [B], each at 500 μM) or BSA as a negative control. Protection is represented by negative peak heights, while positive peaks are signs of hypersensitivity. Nucleotide positions refer to the region of CbbR binding (−202 to −120) on the probe DNA relative to the translation start site (ATG) of cbbL.

Unlike RuBP and NADPH, there were severe changes in the footprint upon inclusion of ATP (500 μM) in the binding reaction with CbbR (Fig. 6B). This resulted in an extensive region of hypersensitivity within the region spanning −202 bp to −120 bp (relative to ATG translation start of cbbL), and new bases were protected by CbbR binding at positions −197, −192, −187, −186 and −184, −169, −158, −152, and −135. Most of the bases that now appeared strongly hypersensitive in the presence of ATP were originally protected (−153) or minimally hypersensitive (−172, −163, −155, −151, −143, −136, and −129 to −126) to moderately hypersensitive (−168, −160, and −131) in the absence of ATP. There was, however, increased protection 5′ of the hypersensitive region and significant hypersensitivity in the putative −35 region of the promoter. Along similar lines, the presence of FBP also resulted in conversion of the status of the CbbR binding region (between −202 bp and −120 bp) from protected to hypersensitive (Fig. 6B). This was also accompanied by a “shift” in the region of protection on the 5′ side of the hypersensitive region. Thus, the presence of an ATP or FBP coinducer resulted in very similar footprints as well as significant and extensive changes to the promoter compared to CbbR alone. There was CbbR binding on the 5′ end of the extensive hypersensitive region, indicating extreme DNA structural changes brought about by protein binding. These results are summarized in Fig. 7.

SPR determination of coinducer effects on CbbR binding.

Surface plasmon resonance (SPR) analysis was used to examine the kinetics and quantify the binding interaction between CbbR and the cbbLS promoter. The affinity constant (K_D = 32 ± 3.7 nM) and the apparent association (k_a = 3.6 ± 0.6 × 10⁶ M⁻¹ s⁻¹) and dissociation rate (k_d = 0.12 ± 0.01 s⁻¹) constants were measured for CbbR in the absence of coinducers (Fig. 8 and Table 2). To confirm previous gel mobility shift and DNase protection study results and to provide a potential quantitative perspective with respect to the ability of the coinducers to influence CbbR binding to specific DNA sequences, additional SPR studies were initiated. In this approach, the ligand (5′ biotin-labeled DNA probe SP90; 270 RU) was immobilized on the Biacore SA sensor chip via a tight biotin-streptavidin interaction. The analyte (purified CbbR alone or CbbR plus coinducer) was passed over the sensor chip, and changes in the binding of CbbR to the cbbLS promoter were monitored by SPR detection over time (to create a sensorgram). CbbR bound specifically to the cbbLS promoter, as revealed by an increase in the resonance signal (approximately 125 RU; Fig. 9). A change in mass at the sensor surface as a result of protein binding alters its refractive index, thus accounting for the changes in the measured resonance signal.

Fig 8.

Surface plasmon resonance sensorgram depicting the real-time interaction of CbbR with the cbbLS promoter. Target SP90 DNA was immobilized on a Biacore SA sensor chip. Binding analysis was carried out by injecting different concentrations of CbbR (0 to 0.25 nM) over the sensor chip. All samples were injected for 400 s at a flow rate of 5 μl/min. The results are expressed as response units (RU) over time (seconds).

Table 2.

Kinetic constants and affinities calculated for the interaction of CbbR with the Rp. palustris cbbLS promoter in the absence or presence of coinducers

Condition (μM)	ka (× 106 M−1 s−1)	kd (s−1)	KD (kd/ka) (nM)
CbbR	3.6 ± 0.6	0.12 ± 0.01	32 ± 3.7
CbbR + RuBP (500)	14 ± 2.7	0.08 ± 0.01	5.9 ± 0.1
CbbR + ATP (500)	23.4 ± 2.3	0.02 ± 0.004	0.9 ± 0.07
CbbR + FBP (500)	11.8 ± 0.15	0.07 ± 0.01	5.7 ± 0.09

Fig 9.

Effect of coinducers on the DNA binding activity of CbbR. A surface plasmon resonance sensorgram depicting the real-time interaction of CbbR with the cbbLS promoter in the presence of coinducers (RuBP, ATP, FBP, and NADPH at 500 μM) is shown. Target SP90 DNA was immobilized on a Biacore SA sensor chip. Binding analysis was carried out by injecting CbbR (0.065 nM) alone or in combination with each coinducer. All samples were injected for 400 s at a flow rate of 5 μl/min. The results are expressed as response units (RU) over time (seconds).

Passage of CbbR over the sensor chip in the presence of the coinducer RuBP, ATP, or FBP (at a final concentration of 500 μM) resulted in an increased CbbR binding response compared to CbbR alone. The highest response was obtained with RuBP (∼275 RU), which is the substrate molecule for RubisCO. The presence of ATP or FBP with CbbR resulted in identical yet intermediate responses of approximately 175 RU (Fig. 9). Thus, the results of this experiment suggested that CbbR responds to 500 μM RuBP, ATP, and FBP to different degrees and confirms the results of the gel mobility shift assays. Surprisingly, in contrast to the results of the gel mobility shift assays and unlike the other coinducers, NADPH (500 μM) did not change the binding of CbbR at the promoter. This was reflected by a binding response that was almost identical to or slightly lower than that seen with CbbR alone (Fig. 9). The kinetic data for CbbR binding in the presence and absence of coinducers are summarized in Table 2. The apparent affinity for CbbR binding in the absence of coinducer was in the nanomolar range, as expected for a site-specific DNA binding protein; the binding data appeared to fit a 1:1 interaction model. The affinity increased considerably in the presence of the coinducers RuBP, ATP, and FBP, as reflected by the significantly higher apparent association rates (Table 2).

DISCUSSION

CbbR, an LTTR family member, regulates transcription of the CO₂ fixation (cbb) operons of chemo- and photoautotrophic bacteria (7, 12, 15, 22, 29). Prior physiological and genetic studies showed that form I RubisCO production in Rp. palustris (encoded by the cbbLS genes that comprise the cbb_I operon) absolutely requires a functional cbbR gene (10, 18). Thus, in order to gain additional insight into the regulatory mechanism by which CbbR influences cbb_I gene expression in Rp. palustris, we have further probed CbbR-cbbLS promoter interactions. The CbbR binding site was mapped via deletion analysis of the cbbLS control region by the use of gel mobility shift assays, and this binding site was confirmed after DNase I footprinting. Using probe FP, the binding site was shown to be centered between bp −202 and −120 within the cbbSR-cbbL intergenic region. Four DNA bases (at positions −175, −162, −142, and −125) within probe FP were strongly hypersensitive to DNase I digestion, which is suggestive of the presence of a potential DNA bending region upon CbbR binding. Both gel mobility shift assays and footprint analyses indicated that the CbbR binding region consisted of 2 binding sites with sequences characteristic of classic LTTR binding motifs. Deletion of site 1 had a more severe effect on binding compared to removal of site 2, suggesting that site 1 might play a more important role in controlling gene expression by influencing the overall DNA structure.

The nature of the metabolic signals that Rp. palustris CbbR perceives for the regulation of transcription is not known. Therefore, this study was also designed to consider the effects of likely physiologically significant coinducer or effector molecules on in vitro binding of CbbR to its target. Four different coinducers, namely, RuBP, ATP, FBP, and NADPH, were identified that altered the promoter binding properties of CbbR. While CbbR was shown to bind to the cbb_I regulatory region, yielding a single DNA-protein complex in gel shift assays, the addition of RuBP, FBP, ATP, and NADPH (500 μM each) visibly enhanced the binding of CbbR in a coinducer concentration-dependent manner. Other related metabolites, such as Ru5P, ADP, AMP (see Fig. S4 in the supplemental material), NADP, and fructose-6-phosphate did not elicit similar responses, ruling out the possibility of nonspecific effects of phosphorylated compounds. Interestingly, phosphoenolpyruvate (PEP) reduced the DNA binding of CbbR, while 3-phos-phoglycerate (3-PGA), the RubisCO end product, had no effect. PEP, derived from 3-PGA, may serve as an indicator of the organic carbon status of the cell and function as a negative effector for CbbR, as previously suggested in a Ralstonia eutropha study (9). This diverse recognition of coinducers by a single regulatory protein is indeed intriguing and would most likely enable the rapid integration of multiple metabolic signals. The ability of a LysR regulator to bind two separate coinducer molecules has been reported. BenM, the benzoate degradation regulator in Acinetobacter sp. strain ADP1, is known to bind two components of the pathway (benzoate and cis,cis-muconate), resulting in synergistic transcriptional activation of benA. The resulting DNase I footprint patterns were dramatically different from those seen with either coinducer alone (1).

Similarly, the increased binding affinity of CbbR to the cbbLS promoter in the presence of all four coinducers raises the possibility of discrete cooperative binding sites on the protein. Clearly, the stoichiometry of coinducer binding to CbbR needs additional and thorough investigation. The synergistic effect of the coinducers on cbbLS transcription would likely involve unique conformations of the DNA-bound CbbR tetramer while also affecting interactions with RNAP.

The finding that RuBP is a coinducer of CbbR in Rp. palustris was not surprising. RuBP enhanced CbbR DNA binding over a wide concentration range (100 to 1,000 μM) (Fig. 5 and Table 2) but had no effect at concentrations below 100 μM. This was also confirmed by surface plasmon resonance analysis. Previous physiological and genetic studies using cbbP (phosphoribulokinase) knockout strains of Rb. capsulatus and Rb. sphaeroides had indicated that RuBP was a likely coinducer for CbbR (23, 26). Subsequent in vitro studies with both Rhodobacter strains confirmed that RuBP is a cofactor for CbbR; however, different binding responses were observed (2, 6). Given the complexity presented by Rp. palustris and the potential additional involvement of the CbbSSR two-component system (11), it was important to better define the interaction of CbbR with its cognate binding sites on the DNA. For Rp. palustris CbbR, RuBP did not change the length of the DNase protected sequence; however, there was enhanced protection of the DNA from DNase I cleavage due to the appearance of additional protected bases (Fig. 6 and 7). Thus, as previously suggested (2), RuBP bound to CbbR appears to favor the formation of a CbbR-cbb_I promoter complex in Rp. palustris. The fact that RuBP affects CbbR binding in vitro for three CBB-dependent organisms strengthens previous physiological and genetic studies suggesting that RuBP and other intermediates of the CBB cycle act as positive coinducers of CbbR-mediated cbb gene expression (15, 23, 26).

It should be noted that coinducers ATP and FBP had identical effects on CbbR-DNA binding in Rp. palustris. Both compounds enhanced the ability of CbbR to bind to DNA, as observed with both gel mobility shift and SPR analyses. Moreover, the interaction of ATP and FBP with CbbR significantly affected the CbbR footprint. This interaction increased the accessibility of the CbbR binding region to DNase I cleavage. Consequently, there was extensive hypersensitivity within this region as well as CbbR-induced protection on the 5′ end of the hypersensitive region. Previous studies in Rb. capsulatus indicated that ATP enhances DNA binding of CbbR_I, resulting in a single protein-DNA complex. On the other hand, FBP not only enhanced the binding of the second CbbR protein of this organism (CbbR_II) but also induced the formation of higher-order CbbR-DNA complexes, reflecting the potential of oligomerization or various degrees of DNA bending. FBP also caused a concentration-dependent reduction in the DNase I footprint of CbbR_II from Rb. capsulatus (6). However, with Rp. palustris CbbR, the increased binding in gel mobility shift experiments and the presence of an extensive and strong hypersensitive region in the footprint as well as in the putative −35 region are unusual. These results suggest that CbbR bends or alters the conformation of DNA in the presence of ATP or FBP. Promoter activity is strongly influenced by DNA bending induced either by structural changes to the DNA helix or by facilitating interaction with RNAP (16). Coinducer binding usually relaxes the DNA bend induced by protein binding (28); in the Rp. palustris scenario, however, it appears that ATP and FBP induce DNA bending. Thus, at the concentration tested (500 μM), the presence of coinducers ATP (an intermediate whose increased intracellular presence typically favors biosynthetic metabolism) and FBP (a CBB cycle biosynthetic intermediate) may reflect intracellular conditions that are conducive for the CBB cycle. At lower intracellular biosynthesis-specific coinducer metabolite concentrations, e.g., under conditions where organic carbon rather than CO₂ is the preferred carbon source, it is conceivable that some CbbR coinducers behave as negative effectors and diminish cbb_I transcription and the subsequent progress of the CBB cycle.

NADPH, another intermediate required for biosynthetic metabolism, is also a coinducer of CbbR in Rp. palustris, as it increased the affinity of CbbR to its promoter in gel mobility shift assays. The CbbR footprint in the presence of NADPH resembled the RuBP footprint very closely, and no changes in the extent of protection were observed. However, the occurrence of hypersensitive sites within the protected region indicated an altered DNA conformation induced by protein binding. It is unclear at this time why there were no concomitant changes in the SPR response in the presence of NADPH, and this lack of a response in SPR measurements, especially with respect to potential stability issues of this compound under the conditions of SPR assay, needs further investigation. It is also apparent that the response of CbbR to NADPH is organism specific, as NADPH did not influence CbbR binding in Rb. capsulatus and Rb. sphaeroides (2, 6) but had a pronounced effect and relaxed the bend induced in Xanthobacter flavus CbbR (28). Similarly, in Hydrogenophilus thermoluteolus, NADPH altered the electrophoretic mobility of CbbR DNA complex 1 accompanied by the disappearance of complex 2. The altered mobility was speculated to be a result of relaxation of DNA bending induced by CbbR, in similarity to that seen with X. flavus (25).

To summarize, protection of DNA binding by Rp. palustris CbbR was enhanced in the presence of both RuBP and NADPH. Similarly, ATP and FBP affected CbbR-DNA binding by inducing significant conformational changes in the DNA helix, as observed by the increased hypersensitivity. However, none of these coinducers changed the length of the CbbR footprint in the manner observed with coinducers of other LTTRs (1, 2, 6, 13, 14, 17, 21, 32). It can be speculated that the observed structural changes in the cbbLS promoter may be a prerequisite for increased transcriptional activation. Clearly, the results of this in vitro study indicate a complex mechanism of transcriptional activation of the cbb_I operon in Rp. palustris. There is a hierarchy of signals transduced to the cbbLS control region via the master regulator CbbR. It could be envisioned that the concentrations of RuBP and NADPH influence cbbLS expression, for example, during transition from photoheterotrophic to photoautotrophic growth. Maximal cbbLS transcription would then proceed based on the cellular ATP and NADPH concentrations, resulting in productive completion of the CBB cycle. The concentration of FBP, a biosynthetic CBB cycle intermediate, would then indicate the regeneration and availability of RuBP for another turn of the CBB cycle. This would also help explain why the footprint responses of CbbR to RuBP and NADPH are identical and yet the responses to ATP and FBP are dramatically different. Under such cellular conditions, there are likely to be additional requirements for one or both response regulators of the CbbRRS system to stabilize the CbbR-DNA interaction (11); further experiments relative to understanding the combined effect of coinducers and response regulators on CbbR-cbbLS promoter interactions are clearly warranted, as are studies investigating whether the transcription start site is affected.