Nitrogen Oxyanion-dependent Dissociation of a Two-component Complex That Regulates Bacterial Nitrate Assimilation*

Background: Nitrogen oxyanion-responsive two-component regulators that control assimilatory nitrate metabolism in heterotrophic bacteria are poorly characterized. Results: The nasT and nasS genes encode a regulatory complex that dissociates upon sensing nitrate/nitrite, releasing the RNA-binding protein NasT. Conclusion: NasS-NasT is a two-component regulator for nitrate/nitrite perception. Significance: A nitrate/nitrite-sensitive interaction between NasS and NasT has been demonstrated for the first time.

A supply of bioavailable nitrogen can be a limiting factor for the growth of bacteria in both terrestrial and aquatic environments. Although these organisms readily assimilate inorganic nitrogen from ammonium (NH 4 ϩ ), many species have been shown to use nitrate (NO 3 Ϫ ) or nitrite (NO 2 Ϫ ) as their sole source of nitrogen. The ability to assimilate these readily water-soluble oxyanions is particularly widespread in heterotrophic bacteria and is associated with the expression of a cytoplasmic assimilatory NO 3 Ϫ /NO 2 Ϫ reductase system (NAS) 4 that performs the two-electron reduction of NO 3 Ϫ to NO 2 Ϫ , followed by the sixelectron reduction of NO 2 Ϫ to NH 4 ϩ (1-5). The NH 4 ϩ formed from the NO 3 Ϫ assimilation pathway can fuel reactions that yield L-glutamate, which plays a pivotal role in biosynthetic cellular metabolism. For example, under nitrogen-sufficient growth conditions, NH 4 ϩ may be used directly via a reaction with 2-oxoglutarate that is mediated by glutamate dehydrogenase. Alternatively, when the availability of NH 4 ϩ is limited, the bulk of L-glutamate is formed by the concerted action of the NH 4 ϩ -dependent glutamine synthetase and the glutamine:2oxoglutarate amidotransferase (also known as glutamate synthase) in the glutamine synthetase/glutamine:2-oxoglutarate amidotransferase cycle (6).
Paracoccus denitrificans PD1222 has recently been shown to assimilate inorganic nitrogen from NO 3 Ϫ or NO 2 Ϫ via an NADH-dependent NAS system encoded by the nasABGHC genes (hereafter termed the nas gene cluster) (5). NAS activity could be clearly measured in cytoplasmic extracts prepared from cells grown with NO 3 Ϫ as the sole nitrogen source. However, activity was not detected in extracts prepared from cells grown in NH 4 ϩ -sufficient culture medium (5). This is consistent with other studies involving Gram-negative bacteria, in which expression of the NO 3 Ϫ /NO 2 Ϫ assimilation pathway is subject to tight hierarchical control involving (i) primary induction by the general nitrogen regulatory system during NH 4 ϩ starvation (7) and (ii) additional system-specific NO 3 Ϫ /NO 2 Ϫ -responsive regulatory proteins typically encoded within, or in close proximity to, nas loci (2, 8 -10).
Pathway-specific control of bacterial NO 3 Ϫ assimilation has been extensively studied in Klebsiella pneumoniae M5al. Here, the key regulator NasR is an example of a single-component NO 3 Ϫ /NO 2 Ϫ -responsive transcription antiterminator protein for which the signal transduction mechanism has been studied in detail (8). NasR polypeptides comprise an N-terminal NO 3 Ϫ / NO 2 Ϫ -sensing NIT domain fused to a C-terminal ANTAR (AmiR and NasR transcription antitermination regulator) signaling domain. This arrangement has been recently confirmed by structural resolution of the NasR protein, which exists as a homodimer in the absence of inducer, i.e. the "inactive" state (11). The regulatory target for the NasR ANTAR signaling domain is a cis-acting regulatory element or "antiterminator" secondary structure within the leader region of the nasFED-CBA mRNA transcript (12,13).
In addition to NasR, a two-component regulatory system (NasS-NasT) has been proposed to be involved in the specific control of NO 3 Ϫ assimilation in the diazotrophs Azotobacter vinelandii (2) and Rhodobacter capsulatus (4) and in members of the Pseudomonas genus such as Pseudomonas aeruginosa (10) and Pseudomonas putida JLR11 (14). Bioinformatics analyses of bacterial genome sequences suggest that nasT and nasS are widely distributed in Gram-negative bacteria that assimilate NO 3 Ϫ and NO 2 Ϫ , including important symbionts, pathogens, and denitrifiers (5,9,10,15). In P. denitrificans, a putative NO 3 Ϫ -sensing two-component regulatory system is encoded by the nasT and nasS genes, which are located immediately upstream of the nas gene cluster on chromosome II (5,9).
NasT is a member of the ANTAR protein family (15). In contrast to NasR, NasT does not contain any recognized NO 3 Ϫ / NO 2 Ϫ -sensing domain. Instead, NasS belongs to the small molecule-binding protein superfamily, the members of which are typically present in ABC-type transport systems. One such example includes the cyanobacterial NO 3 Ϫ -binding protein NrtA from Synechocystis sp. PCC 6803, which has been structurally characterized, revealing a single NO 3 Ϫ anion bound at a defined site within the protein (16). In A. vinelandii, the phenotypes of nasS and nasT strains suggest that NasS and NasT proteins play negative and positive regulatory roles in assimila-tory NO 3 Ϫ /NO 2 Ϫ reductase gene expression, respectively (2). This is consistent with NasS and the ANTAR-type protein NasT being a two-component configuration for regulation of nas gene expression in which the sensor and signal transduction functions are segregated into different proteins, i.e. NasS and NasT, respectively. However, to our knowledge, neither a protein-protein interaction between NasS and NasT nor direct NO 3 Ϫ /NO 2 Ϫ sensing by NasS has yet been experimentally demonstrated. In this work, focused on the P. denitrificans NAS pathway, we present the first biochemical characterization of a NO 3 Ϫ /NO 2 Ϫ -responsive two-component system (NasS-NasT), in which binding of NO 3 Ϫ or NO 2 Ϫ by the sensor NasS triggers release of the positive RNA-binding regulator NasT.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Media, and Growth Conditions-P. denitrificans PD1222 was routinely cultured under aerobic conditions at 30°C in either LB medium or a defined mineral salts medium (5) supplemented with ammonium chloride (10 mM), potassium nitrate (20 mM), potassium nitrite (10 mM), or sodium L-glutamate (5 mM) as the sole nitrogen source as required. Escherichia coli strains were cultured aerobically in LB medium at 37°C unless stated otherwise. Cell growth was followed by measuring the absorbance of cultures at 600 nm (A 600 ). Antibiotics were used at the indicated final concentrations: ampicillin, 100 g/ml; gentamycin, 20 g/ml; kanamycin, 25 g/ml; rifampicin, 100 g/ml; spectinomycin, 25 g/ml; and streptomycin, 60 g/ml.
Construction of nasT and nasS Strains-P. denitrificans mutant strains were constructed by replacement of significant portions of the target gene essentially as described previously (5). To generate the nasT strain (nasT⌬::streptomycin), the front and rear sections of the nasT gene were amplified from genomic DNA isolated from P. denitrificans PD1222 in separate reactions using oligonucleotide primer sets T1/T2 and T3/T4 (Table 1), respectively. Reactions were performed using the Expand High Fidelity PCR system (Roche Applied Science). A BamHI restriction site was introduced into the end of each fragment, allowing ordered assembly of the gene sections within the multiple cloning site of the pGEM-T Easy vector (Promega). The resulting construct had a unique BamHI site at the interface of the front and rear sections of nasT, into which a streptomycin resistance cassette, obtained from pSRA2, was introduced (17). The nasT⌬::streptomycin fragment was then transferred to the mobilizable vector pSUP202* (5) as an EcoRI fragment. The nasS strain (nasS⌬::kanamycin) was constructed in a similar manner. PCR amplification of the front and rear gene sections was performed using primer sets S1/S2 and S3/S4 (Table 1), respectively, and the fragments were then cloned into pGEM-T Easy. A kanamycin resistance cassette derived from pSUP2021 was inserted into a unique BamHI site between the front and rear sections of nasS. The nasS⌬::kanamycin fragment was transferred to the mobilizable vector pSUP202* as an EcoRI fragment. All cloning steps were carried out using an E. coli DH5␣ host following standard transformation and ligation protocols (18). Conjugation, selection, and validation of mutants were performed as described previously (5).

Assay for Assimilatory NO 3
Ϫ /NO 2 Ϫ Reductase Activity in P. denitrificans Strains-NADH-dependent assimilatory NO 3 Ϫ /NO 2 Ϫ reductase activity was measured in cytoplasmic extracts as described previously (5). Given that NADH is consumed at a ratio of ϳ3:1 NO 2 Ϫ :NO 3 Ϫ , NAS activity assay was performed with NO 2 Ϫ as the electron acceptor to allow rapid reproducible initial rate determinations using cytoplasmic extracts prepared from relatively small cell volumes.
Cloning, Expression, and Purification of NasT and NasS-A 1.75-kb fragment containing the coding regions for nasT and nasS was amplified by PCR. Reactions containing 5% (v/v) Me 2 SO were performed essentially as described by Sambrook and Russell (18) using primers TS1 and TS2 ( Table 1). The purified product was cloned into pGEM-T Easy and then transferred to the pET-24a expression vector (Novagen) as an NheI-HindIII restriction fragment. The resulting construct, pET-24a/nasTS, was sequenced and transformed into E. coli BL21(DE3) for protein expression. Cells containing the expression plasmid were grown at 37°C in 500 ml of LB medium until cultures reached an A 600 reading of ϳ0.5. Expression was induced by addition of 1 mM isopropyl ␤-D-thiogalactopyranoside, after which the culture temperature was lowered to 28°C. Cells were harvested 3 h after induction by centrifugation at 12,000 ϫ g for 20 min at 4°C. Soluble cell extracts were prepared at 4°C. Pellets were resuspended in buffer A (20 mM sodium phosphate, 150 mM NaCl, 25 mM imidazole (Sigma-Aldrich), and 10% (v/v) glycerol, pH 7.0). The cell suspension was supplemented with a protease inhibitor mixture (cOmplete, EDTA-free, Roche Applied Science) and 1 mg/ml lysozyme (hen egg white, EC 3.2.1.17, Fluka) and incubated at 4°C for 30 min. Cell lysis was achieved following addition of 1% (v/v) Triton X-100 (Sigma-Aldrich) to the mixture and incubation on a rocking platform for 10 min. The lysate was supplemented with DNase I (bovine pancreas, EC 3.1.21.1, Sigma) and RNase A (bovine pancreas, EC 3.1.27.5, Sigma) and then incubated for a further 10 min, after which it was fluid. Insoluble cell debris was removed by ultracentrifugation at 260,000 ϫ g for 60 min at 4°C. NasS-NasT was purified by immobilized metal affinity chromatography (IMAC), followed by anion-exchange and size-ex-clusion chromatography. All purification steps were performed at 4°C with a flow rate of 1 ml/min unless stated otherwise. Soluble cell extract prepared from a 2-liter culture was loaded onto a 10-ml Ni 2ϩ IMAC column (HiTrap Chelating HP, GE Healthcare) that was precharged with nickel sulfate, washed with analytical reagent-grade water, and equilibrated with buffer A. Following loading, the column was then washed with a further 4 column volumes of buffer A to elute unbound protein. Bound protein was eluted with a linear gradient of 25-500 mM imidazole applied over 5 column volumes. Fractions containing NasS and NasT were pooled and buffer-exchanged into buffer B (50 mM NaHEPES, 1 mM EDTA, and 10% (v/v) glycerol, pH 7.0). This sample was loaded onto a 5-ml HiTrap Q HP anion-exchange column (GE Healthcare) that was pre-equilibrated with buffer B. Elution of bound protein was achieved by applying a linear gradient of 0 -2 M NaCl over five column volumes. Peak fractions containing NasS-NasT were then pooled, buffer-exchanged into buffer C (50 mM NaHEPES and 100 mM NaCl, pH 7.0), and concentrated by ultrafiltration. Samples were loaded onto a 70-ml preparative size-exclusion column (Sephacryl S-200 high resolution, GE Healthcare) that was preequilibrated with buffer C. Protein concentration was determined by bicinchoninic acid assay (19).
For identification of purified proteins, bands corresponding to the correct molecular mass of NasS (ϳ42 kDa) and NasT (ϳ22 kDa) were excised from denaturing SDS-polyacrylamide gels. Each gel slice was washed, reduced, alkylated, and treated with trypsin according to standard procedures adapted from Shevchenko et al. (20). The tryptic peptide fragments were analyzed by mass spectrometry using an ultraflex TM MALDI-TOF/ TOF spectrometer (Bruker). Briefly, 0.5-0.8 l of the peptide samples was applied to a Prespotted AnchorChip TM MALDI target plate (Bruker), and the spots were washed with 10 -15 l of 10 mM ammonium phosphate and 0.1% trifluoroacetic acid according to the manufacturer's protocol. The instrument was then calibrated using the prespotted standards. Samples were analyzed using a flexControl TM method (version 3.0, Bruker) optimized for peptide detection. Acquired spectra were processed using flexAnalysis TM (version 3.0, Bruker). The resulting peak lists were used for a database search using an in-house Mascot 2.4 server (Matrix Science, London, United Kingdom). The search was performed on the UniProt Swiss-Prot/ TrEMBL database (release 20121031) with taxonomy set to bacteria and on a common contaminants database using the trypsin/P enzyme with a maximum of one missed cleavage, a peptide mass tolerance of 50 ppm, carbamidomethylation as fixed, and oxidation and acetylation (protein N terminus) as variable modifications. Using those parameters, Mascot protein scores Ͼ85 were significant (p Ͻ 0.05). NasS and NasT peptides were identified with significance scores of 187 (sequence coverage of 45%, expect value of 6.7 ϫ 10 Ϫ13 ) and 123 (sequence coverage of 65%, expect value of 1.79 ϫ 10 Ϫ6 ), respectively.
UV-visible Electronic Absorbance and Fluorescence Spectroscopy-Absorbance spectra were recorded for purified protein (ϳ1.5 mg/ml) using a Hitachi U-3000 spectrophotometer. An extinction coefficient of 47,100 M Ϫ1 cm Ϫ1 at 280 nm was estimated for the NasS-NasT complex. Emission spectra were recorded at 295 nm using a Varian Cary Eclipse fluorescence spectrophotometer. Curve fitting was performed using Origin 7.0 (OriginLab Corp.).
Analytical Ultracentrifugation and Size-exclusion Chromatography-Analytical ultracentrifugation sedimentation equilibrium experiments were performed at 20°C using a Beckman Optima XL-I analytical ultracentrifuge equipped with an integrated UV-visible absorbance optical system and an An-50 Ti analytical rotor (Beckman Instruments). Protein samples and buffer controls were loaded into the relevant sectors of twochannel EPON cells (1.2-cm path). Sedimentation equilibrium profiles were recorded at 280 nm at a range of protein concentrations (3-23 M) and rotation speeds (7.5, 12, and 16 krpm). The partial specific volume for NasS-NasT was calculated as 0.744 ml/g using the sedimentation interpretation program SEDNTERP (version 20120111 BETA, Biomolecular Interactions Technology Centre). Analytical ultracentrifugation experiments were performed according to published methods (21). Data analysis was performed using the UltraScan II software package (version 9.9) (22). Fitting of sedimentation profiles to an ideal one-component model was used to determine the apparent molecular masses of proteins.
Analytical size-exclusion chromatography was performed on a 24-ml Superdex 200 HR 10/30 column (Amersham Biosciences) that was equilibrated with buffer C. The column was loaded with 250 l of purified protein (1.5 mg/ml) and developed at a flow rate of 0.5 ml/min. Time elution of protein from the column was followed automatically at 280 nm using an ÄKTA fast protein liquid chromatograph (GE Healthcare). The low aromatic amino acid content of NasT did, however, complicate assigning the retention peak position at 280 nm. Instead, this was done manually by measuring the absorbance of column fractions at 260 nm. The apparent molecular masses of NasS-NasT and the isolated NasS and NasT proteins were estimated by comparison with known protein standards supplied in a gelfiltration calibration kit (high molecular weight kit, GE Healthcare), which were run individually under the relevant buffer conditions.
Protein-RNA Binding Monitored by Electrophoretic Mobility Shift Assay-RNA molecules for the nasA leader and regions of nasB and sdhA were prepared by in vitro transcription. Primer sets SA1/SA2, AB1/AB2, and SDH1/SDH2 (Table 1) were used in separate PCRs to generate DNA (ϳ300 bp) for the putative control region upstream of the nasA gene and regions of nasB and sdhA genes as controls. The T7 promoter sequence was included in forward primers SA1, AB1, and SDH1 for subsequent RNA transcription. The in vitro synthesis of singlestranded RNA was performed with the HiScribe TM T7 in vitro transcription kit (New England Biolabs). Overnight reactions were performed at 42°C. DNA template was then removed by DNase treatment at 37°C for 30 min. RNA products were visualized on 1.5% agarose gels stained with ethidium bromide.
For electrophoretic mobility shift assays, ϳ20 M purified NasS-NasT was incubated with different concentrations of RNA (50 -90 nM) for 15 min at room temperature in 10-l binding reaction volumes. The buffer contained 10 mM Tris, pH 7.4, 150 mM KCl, 0.1 mM EDTA, 0.1 mM dithiothreitol, and 1 mM NaNO 3 . Samples were loaded onto native 5% polyacryl-amide gels (37.5:1 acrylamide:bisacrylamide) in 45 mM Tris, 45 mM boric acid, and 1 mM EDTA, pH 8.3. Gels were stained for RNA using a SYBR Green EMSA kit (Molecular Probes) and visualized in an FX scanner (Bio-Rad). After electrophoresis, the shifted band present in lane 4 of Fig. 3A was excised and prepared for MALDI-TOF-MS analysis for protein identification as described above. The NasT peptide was identified with a significance score of 87.

Role of NasT and NasS in NO 3
Ϫ /NO 2 Ϫ Assimilation-In the absence of NH 4 ϩ , P. denitrificans may use NO 3 Ϫ or NO 2 Ϫ as the sole nitrogen source for growth (Fig. 1A), an ability that has been directly linked to expression of the nas gene cluster (5). To explore the role of the NasS-NasT two-component regulator in expression of the NAS system, P. denitrificans strains that were deficient in either nasT or nasS were constructed. A nasT strain was unable to grow with either NO 3 Ϫ or NO 2 Ϫ as the sole nitrogen source, but growth of this strain was unaffected when cells were grown in the presence of NH 4 ϩ (Fig. 1B) or L-glutamate  OCTOBER 11, 2013 • VOLUME 288 • NUMBER 41

JOURNAL OF BIOLOGICAL CHEMISTRY 29695
(data not shown). Growth on NO 3 Ϫ and NO 2 Ϫ could be restored to WT levels upon complementation with a functional gene copy that was expressed in trans from the expression vector pEG276-nasT.
NAS expression is induced in WT cells by NO 3 Ϫ when NH 4 ϩ is absent and can be monitored by assaying NADH-dependent NO 2 Ϫ reductase activity in cytoplasmic extracts (hereafter termed NAS activity). The level of enzyme activity detected in induced WT cells (14.8 Ϯ 1.2 units) was similar to that described previously (5). However, this NAS activity was not detectable in cytoplasmic extracts prepared from the nasT strain grown on L-glutamate with or without the additional inclusion of NO 3 Ϫ (Table 2). NAS activity was restored to WT levels in the nasT strain when the deletion was complemented with a functional plasmid-borne gene copy. This is consistent with NasT being a positive regulator of the NO 3 Ϫ /NO 2 Ϫ assimilation pathway.
Using NO 3 Ϫ or NO 2 Ϫ as the nitrogen source, a strain in which the nasS gene was mutated showed no clear growth defect with respect to the WT (Fig. 1C). However, in contrast to the WT, NAS activity could be readily detected in cytoplasmic extracts prepared from cells grown on L-glutamate (13.5 Ϯ 2.7 units) despite omission of NO 3 Ϫ as inducer for nas expression. The level of NAS activity present in the nasS strain was comparable to that observed in NO 3 Ϫ -induced WT cells (Table 2). That disruption of nasS did not have any pronounced effect on growth but instead led to the deregulation of NAS activity (such that it became constitutive irrespective of the presence of NO 3 Ϫ ) is consistent with NasS normally having an inhibitory role in expression of the nas gene cluster.
Coexpression and Purification of NasS-NasT-The mechanism by which the putative NO 3 Ϫ sensor NasS and the ANTARtype protein NasT cooperate to control nas gene expression in bacteria is unclear but may involve a protein-protein interaction (2). To investigate whether NasS and NasT interact, the expression construct pET-24a/nasTS was produced, which would not only yield high levels of recombinant forms of P. denitrificans NasS and NasT proteins but would also provide a means of immobilizing NasT via a polyhistidine tag during affinity purification. SDS-PAGE analysis of soluble protein extracts prepared from E. coli host cells containing the pET-24a/nasTS construct showed clear overexpression of two proteins at ϳ22 and 42 kDa when isopropyl ␤-D-thiogalactopyranoside was added to the cell cultures ( Fig. 2A).
Soluble extracts containing NasS and NasT were subjected to Ni 2ϩ IMAC, which revealed that both proteins bound tightly to the affinity matrix despite only NasT being His-tagged (Fig. 2B). At this stage, protein bands were extracted from SDS-polyacrylamide gels and identified by mass spectrometry, which confirmed the 22-and 42-kDa bands as the NasT and NasS proteins from P. denitrificans, respectively. Co-purification of NasS and NasT was also observed during the subsequent polishing steps of the purification, which included Q-Sepharose anion-exchange (Fig. 2C) followed by size-exclusion (Fig. 2D) chromatography, consistent with a strong protein-protein interaction.
Nitrate/Nitrite Binding and Dissociation of the NasS-NasT Complex-Co-purification of approximately equivalent amounts of NasS and NasT (as assessed visually by SDS-PAGE) was found to be dependent on the concentration of NO 3 Ϫ present in solution. To explore the impact of NO 3 Ϫ further, the purified NasS-NasT complex was re-immobilized on an IMAC column pre-equilibrated with buffer containing 50 mM NaHEPES and 100 mM NaCl, pH 7.5, that was additionally supplemented with 1 mM NO 3 Ϫ as indicated (Fig. 3). In the absence of NO 3 Ϫ , NasS and NasT readily co-eluted upon washing the column with 500 mM imidazole (Fig. 3A, lane 2). In stark contrast, nearly complete separate elution of NasS was observed when a column containing freshly immobilized NasS-NasT was washed with binding buffer containing NO 3 Ϫ (Fig. 3A, lane 3).
Step elution of the remaining bound protein, the overwhelming majority being His-tagged NasT, was achieved by washing the column with imidazole (Fig. 3A, lane 4). Dissociation of NasS from the bound NasS-NasT complex was also observed in similar experiments in which NO 2 Ϫ was used in place of NO 3 Ϫ , but dissociation was minimal in buffers supplemented with sulfate (SO 4 2Ϫ ) (data not shown). Inspection of the NasS polypeptide sequence reveals that this putative NO 3 Ϫ sensor contains seven tryptophan residues. Accordingly, a clear shoulder at ϳ288 nm was present in the UV-visible electronic absorbance spectrum of the purified NasS-NasT complex and the isolated NasS protein (Fig. 3B). By contrast, NasT contains no tryptophan residues. This is consistent with the relatively weak absorbance at 288 nm in the UV-visible spectrum of the isolated NasT protein (Fig. 3B).
The fluorescence emission spectrum of NasS-NasT that resulted from excitation at 295 nm revealed a clear peak at ϳ334 nm, consistent with fluorescence of buried tryptophan residues (Fig. 4A). The magnitude of peak fluorescence was sensitive to NO 3 Ϫ and was "quenched" to a resting value of ϳ35% at concentrations above ϳ250 M. NasS-NasT fluorescence was also quenched by NO 2 Ϫ , indicating binding, but was insensitive to SO 4 2Ϫ and NH 4 ϩ (Fig. 4B) and a range of other ionic compounds, including chloride (Cl Ϫ ), chlorate (ClO 3 Ϫ ), azide (N 3 Ϫ ), and bicarbonate (HCO 3 Ϫ ). That the other ions tested had little effect on the intrinsic protein fluorescence of NasS implies that the small molecule-binding site present in the NasS-NasT regulatory complex is specific to NO 3 Ϫ but may also accommodate the smaller chemically similar anion NO 2 Ϫ . The change in protein fluorescence (⌬F) observed in response to the solution concentration of ligand (L) can be explained by the following minimal binding model: NasS-NasT ϩ L 7 NasS-L ϩ NasT. Here, the apparent equilibrium constant (K D app ) for ligand binding and dissociation of the NasS- Ϫ and NO 2 Ϫ , respectively. Solution State Properties of the NasS-NasT Complex-To establish the composition of the NasS-NasT complex in solution, analytical ultracentrifugation experiments and size-exclusion chromatography were performed. Fig. 5 shows the sedimentation profile of the purified NasS-NasT complex. In the absence of NO 3 Ϫ , the sedimentation equilibrium profile of this complex fitted well to a single component with an apparent molecular mass of 132 Ϯ 5 kDa (Fig. 5A). Given that equivalent amounts of NasS and NasT were observed by SDS-PAGE, the experimental value determined is consistent with the expected value of 128 kDa for a heterotetrameric solution state complex consisting of two NasS proteins and two NasT proteins. Similar experiments performed in the presence of 1 mM NO 3 Ϫ resulted in a decrease in the apparent molecular mass for the complex (Fig. 5B). Given the low aromatic residue content of NasT in comparison with NasS, the imbalance of extinction coefficients  OCTOBER 11, 2013 • VOLUME 288 • NUMBER 41 precludes accurate discrimination of the proteins in the analytical ultracentrifugation experiment. Therefore, the ϳ2-fold decrease in the observed apparent molecular mass for NasS-NasT in the presence of NO 3 Ϫ is qualitative but consistent with NO 3

A Regulatory Complex for Assimilatory Nitrate Reduction
Ϫ -mediated dissociation of the larger NasS-NasT complex to lower apparent molecular mass species.
Additional experiments involving analytical size-exclusion chromatography were performed to investigate the result of NO 3 Ϫ -dependent dissociation of the NasS-NasT complex in more detail (Fig. 6). In the absence of NO 3 Ϫ , NasS-NasT eluted at ϳ14 ml as a single symmetrical peak. SDS-PAGE analysis revealed equivalent amounts of NasS and NasT in all fractions across this peak. An apparent molecular mass of 134 Ϯ 10 kDa could be assigned to NasS-NasT by comparison with various protein standards applied to the same column under identical conditions. When the column equilibration buffer was supplemented with NO 3 Ϫ , an asymmetric protein elution profile was observed that was clearly different from that observed for the protein in the absence of NO 3 Ϫ . SDS-PAGE analysis of eluted protein revealed differential elution of NasS and NasT. Peak NasS elution was observed at ϳ16 ml, corresponding to an apparent molecular mass of 38 Ϯ 5 kDa, consistent with a monomeric solution state for NasS. By contrast, the bulk of NasT eluted at ϳ14 ml, corresponding to an apparent molecular mass of ϳ130 kDa. Given that retention of NasT was within experimental error of that observed for the NasS-NasT complex prior to NO 3 Ϫ exposure, the solution state of the isolated protein is considerably larger than would be  expected for monomeric NasT (22 kDa) and implies that a substantial population of NasT can form a homo-oligomeric solution state when separated from NasS. Such behavior was observed at a range of pH values and salt concentrations and persisted despite the inclusion of the reducing agent dithiothreitol at 2 mM in all purification buffers. Thus, the multimeric state of NasT would likely consist of approximately six monomers and is unlikely to be the result of an adventitious proteinprotein interaction. In summary, the combination of analytical ultracentrifugation and gel-filtration data provides compelling evidence that not only do NasS and NasT dissociate in the presence of NO 3 Ϫ , but that once separated, the NasT protein may also form a homo-oligomeric state in solution.
Specific Interaction of NasT with the Leader RNA of the nasA Gene-Analysis of the region upstream of nasA revealed repeated inverted sequence tracts for a series of regulatory hairpins similar to those present in the leader RNA of genes regulated by related RNA-binding proteins of the ANTAR signaling family (data not shown). The capacity of the NasS-NasT regulatory proteins to bind the leader RNA of the nasA gene from P. denitrificans was assessed in a series of electrophoretic mobility shift assays. Here, the purified NasS-NasT complex was "activated" by addition of 1 mM NO 3 Ϫ and then incubated with RNA molecules produced by in vitro transcription. The RNAs tested included the predicted leader region of the nasA gene and two control sequences that included regions of the nasB and the sdhA genes, also from P. denitrificans, which did not contain similar hairpin structures associated with transcription antitermination.
The results presented in Fig. 7 reveal that the migration of the nasA RNA was significantly slower in the presence of the activated NasS-NasT protein relative to the migration of the same RNA when the protein was absent (compare lanes 1 and 4). In contrast, migration of either the nasB (Fig. 7, compare lanes 2  and 5) or sdhA (compare lanes 3 and 6) RNA molecules was essentially unaltered upon addition of activated NasS-NasT.  When RNA was absent, the NasS and NasT proteins were not resolved by the staining procedure (Fig. 7, lane 7). Given that protein-nucleic acid complexes migrate more slowly than free linear nucleic acid fragments, the "mobility shift" observed for the leader RNA of the nasA gene in the presence of activated NasS-NasT is indicative of a specific protein-RNA interaction. The gel band in lane 4 of Fig. 7 (denoted by an asterisk) was excised, and the protein component was identified as P. denitrificans NasT by mass spectrometry. This confirmed that the ANTAR protein NasT was responsible for RNA binding.

DISCUSSION
Transcription antitermination is a control mechanism for gene expression that regulates a growing number of systems in bacteria, including those responsible for nitrogen metabolism (13,23,24). Specifically, one-and two-component systems (NasR and NasS-NasT, respectively) have been shown to regulate NO 3 Ϫ assimilation. However, despite wide distribution among bacterial heterotrophs that assimilate NO 3 Ϫ and/or NO 2 Ϫ , the biochemical properties of NasS-NasT two-component systems have been scarcely explored (2,10,14).
In this study, we have demonstrated that nasT is essential for growth of P. denitrificans with NO 3 Ϫ or NO 2 Ϫ as the sole nitrogen source. Deletion of nasT removes the capacity for NO 3 Ϫ / NO 2 Ϫ induction of NAS expression. NasT polypeptides are predicted to contain an N-terminal CheY-like receiver domain (termed the REC domain) in addition to the C-terminal ANTAR domain similar to that present in NasR (26). The REC domain is found in a range of prokaryotic proteins that undergo distinctive conformational modulation during signal transduction as a consequence of covalent (e.g. phosphorylation) and/or physical (e.g. protein-protein interaction) modification by their cognate sensors (26,27).
In this case, NasS is the cognate sensor. Sequence analysis reveals that NasS shares ϳ44% sequence similarity with the cyanobacterial periplasmic NO 3 Ϫ -binding protein NrtA (16). Notably, NasS conserves all residues required for NO 3 Ϫ coordination but lacks the N-terminal signal sequence and transmembrane helix present in NrtA required for periplasmic export and membrane localization, respectively (data not shown) (16). Accordingly, NasS is predicted to be a soluble cytoplasmic NO 3 Ϫ -binding protein. Given the subtlety of the distinguishing sequence features between NasS and related periplasmic NO 3 Ϫbinding proteins, the nasS regulatory gene may have been incorrectly annotated in a number of bacterial genomes. Thus, the importance of the NasS-NasT system may have been underestimated.
Without nasS, P. denitrificans is unable to perceive the presence of the inducer (either NO 3 Ϫ or NO 2 Ϫ ), which results in the deregulation of gene expression such that the NAS system is expressed constitutively. That loss of NasS does not appear to significantly attenuate growth on NO 3 Ϫ or NO 2 Ϫ but instead leaves the bacterium unable to regulate expression of the NAS system suggests that NasS plays an inhibitory regulatory role in NAS expression when the inducer is absent. These results are consistent with the published phenotypes of nasT and nasS mutants in other bacteria (2,10,14) and imply that both NasS and NasT act together as a NO 3 Ϫ /NO 2 Ϫ -responsive two-component regulatory system to control nas gene expression.
To transmit the induction signal from NasS to NasT, a putative regulatory interaction between these two proteins is necessary. Such an interaction has been inferred for the A. vinelandii and P. aeruginosa NasS-NasT systems, but to our knowledge, no experimental evidence has yet been presented. In this work, in vitro experiments revealed that the NasS and NasT proteins from P. denitrificans co-purify as a stable heterotetrameric complex. This complex comprises the NasS and NasT proteins in a 1:1 ratio and persists during a wide range of purification methods, including affinity, strong anion-exchange, and sizeexclusion chromatography. This robust NasS-NasT proteinprotein interaction is, however, sensitive to NO 3 Ϫ and NO 2 Ϫ but not other anions.
The relative affinity of the NasS-NasT regulatory complex for NO 3 Ϫ was found to be in the low micromolar range (K D app ϳ 15 M), which is in good agreement with the K m value of ϳ17 M reported for the NasC NO 3 Ϫ reductase from P. denitrificans (5). Notably, this value is also consistent with that (ϳ5 M) reported by Chai and Stewart (12) for the NasR-nasF leader RNA complex with NO 3 Ϫ . In contrast, the K D app value of NasS-NasT for NO 2 Ϫ (ϳ94 M) was an order of magnitude higher than the K m value of ϳ5 M reported for the NasB NO 2 Ϫ reductase (5) and thus may reflect the inability of NasS to discriminate between NO 3 Ϫ and the smaller chemically similar NO 2 Ϫ anion. The ability of NasS-NasT to bind NO 2 Ϫ , albeit with lower affinity than NO 3 Ϫ , is consistent with P. denitrificans being able to grow with millimolar levels of NO 2 Ϫ as the sole nitrogen source. Given that NO 3 Ϫ and NO 2 Ϫ are assimilated via a common pathway, there may be a selective advantage for some bacteria to express a sensor with dual specificity that is capable of detecting the presence of both inorganic nitrogen sources.
Significantly, exposure to NO 3 Ϫ or NO 2 Ϫ triggered dissociation of the heterotetrameric NasS-NasT complex (apparent molecular mass of ϳ134 kDa) into monomeric NasS (apparent molecular mass of ϳ38 kDa) and a homo-oligomeric state of NasT (apparent molecular mass of ϳ130 kDa). Given the error of the gel-filtration experiment and that NasT is a low molecu- RNAs that were preincubated with 20 M purified NasS-NasT in the presence of 1 mM NaNO 3 prior to loading. Lane 7 was loaded with a control incubation of the NasS-NasT protein in binding buffer without RNA. RNA was resolved on native polyacrylamide gels and visualized using SYBR Green stain. The asterisk denotes the shifted band excised for identification by mass spectrometry. lar mass protein, the multimeric state was broadly consistent with a hexamer.
In P. aeruginosa, the AmiC and AmiR proteins act to mediate inducer-responsive regulation of the amiEBCRS operon, which encodes the necessary genes for catabolic degradation of aliphatic amides (30). Notably, the ANTAR protein AmiR has been structurally resolved with its cognate small moleculebinding partner AmiC in a heterotetrameric ligand-responsive regulatory complex, (AmiC-AmiR) 2 (29). When considered with the genetic results presented for nasS and nasT strains, the biochemical properties of NasS-NasT suggest that, prior to NO 3 Ϫ /NO 2 Ϫ -dependent induction of nas gene expression, a ligand-free NasS-NasT complex exists in which NasT is inactive. Thus, the NasS-NasT and AmiC-AmiR regulatory systems may share mechanistic similarities.
The regulatory mechanism of NasR, whose target is a hairpin in the leader RNA of nasF, the promoter-proximal gene of the nas operon in Klebsiella sp., has been extensively studied (11,15,25). Transcription antitermination control mechanisms mediated by NasT have also been postulated to regulate nas gene expression (10,28). In support, we present evidence from electrophoretic mobility shift assays that, in the presence of NO 3 Ϫ , NasT is able to bind the leader RNA of the nasA gene, which contains putative regulatory elements. Formation of this NasT-nasA RNA complex is consistent with the proposed regulatory interaction required for ANTAR-type signaling proteins.
The data presented herein for NasS-NasT suggest that, following inducer perception by this NO 3 Ϫ /NO 2 Ϫ -responsive regulatory complex, the ANTAR-type protein NasT is released from the complex with NasS. Once free, NasT can activate transcription of the P. denitrificans nasABGHC gene cluster necessary for the reductive assimilation of this nitrogen source (Fig. 8).
Finally, the structural basis of the protein-RNA interaction remains poorly understood, but an oligomeric form of an ANTAR-type protein, as suggested here for NasT, may be functionally relevant. In this respect, it is notable that AmiR has also been shown to form oligomers of a similar size range after inducer-mediated dissociation of the (AmiC-AmiR) 2 regulatory complex (29) and that other recognized RNA-binding proteins such as Hfq are functional as homohexamers (31).