An autoinhibitory mechanism controls RNA-binding activity of the nitrate-sensing protein NasR

Abstract

The ANTAR domain harnesses RNA-binding activity to promote transcription attenuation. Although several ANTAR proteins have been analyzed by high-resolution structural analyses, the residues involved in RNA-recognition and transcription attenuation have not been identified. Nor is it clear how signal-responsive domains are allosterically coupled with ANTAR domains for control of gene expression. Herein, we examined the sequence conservation of ANTAR domains to find residues that may associate with RNA. We subjected the corresponding positions of Klebsiella oxytoca NasR to site-directed alanine substitutions and measured RNA-binding activity. This revealed a functionally important patch of residues that forms amino acid pairing interactions with residues from NasR's nitrate-sensing NIT domain. We hypothesize these amino acid pairing interactions are part of an autoinhibitory mechanism that holds the structure in an "off" state in the absence of nitrate signal. Indeed, mutational disruption of these interactions resulted in constitutively active proteins, freed from autoinhibition and no longer influenced by nitrate. Moreover, sequence analyses suggested the autoinhibitory mechanism has been evolutionarily maintained by NasR proteins. These data reveal a molecular mechanism for how NasR couples its nitrate signal to RNA-binding activity, and generally show how signal-responsive domains of one-component regulatory proteins have evolved to exert control over RNA-binding ANTAR domains.

Keywords: RNA-binding proteins; antiterminator proteins; bacteria; gene expression; nucleic acid; regulatory sequences; transcription attenuation.

FIGURE 1

Klebsiella oxytoca NasR specifically binds a tandem stem‐loop RNA in the presence of nitrate or nitrite. (a) Purified NasR protein was subjected to systematic evolution of ligands by exponential enrichment (SELEX) and protein‐binding RNA aptamers were analyzed by high‐throughput sequencing. After four rounds of selection, a majority of protein‐associated RNAs included a common consensus motif (shown as a sequence logo) forming the closing base pair and terminal loops of two consecutive hairpins. (b) Cy3‐labeled RNA molecules were synthesized for the two‐hairpin portion of the Klebsiella oxytoca nasF leader (“P1P2”), along with two different unrelated RNAs (“UR1” and “UR2”). These RNA molecules were used for fluorescence anisotropy binding assays. (c) Quantification of fluorescence anisotropy of RNA in binding buffer at different concentrations. (d) Equilibrium saturation binding curves of RNA molecules from (b) bound to different concentrations of purified His6‐MBP‐NasR protein in the presence of 1 mM nitrate. (e) Equilibrium binding curves of P1P2 RNA bound to 8 μM His6‐MBP‐NasR protein in the presence of varying concentration of different ionic species. For all panels, error bars represent the standard deviation of the fluorescence anisotropy (b) or change in anisotropy (d‐e) of four replicate wells relative to the no‐protein (d) or no‐ion sample (e). All datasets were fit to a Hill equation model

FIGURE 2

Alignment of representative ANTAR protein sequences. (a) Multiple sequence alignment of nineteen ANTAR domains from ANTAR regulators previously described in the literature (Klebsiella oxytoca NasR, Pseudomonas aeruginosa AmiR, Mycobacterium tuberculosis Rv1626, Nakamurella multipartita PAL, Enterococcus faecalis EutV, Listeria monocytogenes EutV, and a Streptomyces coelicolor response regulator protein) in addition to 12 additional representatives of the ANTAR domain selected from Pfam PF03861 listed by UniProt accession ID. Conserved residues selected for mutagenesis are highlighted with magenta. Highlighted in gray are residues that exhibited an increased level of conservation but were not chosen for mutagenesis due to their hydrophobicity and the strong likelihood they participate in interhelical structural interactions. (b) NasR is composed of eight alpha helices in the NIT domain, one connecting helix, and three alpha helices in the ANTAR domain. The location of the ANTAR helices (10‐12) in NasR (PDB: 4AKK) are shown in green. (b) A ribbon depiction of the ANTAR domain region of NasR (PDB: 4AKK) is shown in green, superimposed with structures of the ANTAR domain from Pseudomonas aeruginosa AmiR (PDB: 1QO0), Mycobacterium tuberculosis Rv1626 (PDB: 1S8N), and Nakamurella multipartita PAL (PDB: 6HMJ)

FIGURE 3

Site‐directed mutagenesis of conserved ANTAR residues affects RNA binding. (a) Equilibrium binding measurements for alanine substitution mutations for His10‐MBP‐NasR in the presence of 1 mM nitrate (full binding curves are shown in Figure S2). Binding measurements were colored according to those exhibiting an affinity better than or comparable to wild‐type NasR (green), those exhibiting greater than 2‐fold reduction in binding affinity (yellow), and those with negligible binding activity (red). (b) Ribbon depictions of the ANTAR domain (PDB: 4AKK) with residues at positions corresponding to the alanine substitutions highlighted in color

FIGURE 4

Compositional bias of NIT and ANTAR residues at an interdomain interface. (a) A sequence logo shows conservation of residues across all NIT domain‐containing proteins corresponding to residues in the helix 5‐6 region of Klebsiella oxytoca NasR. Letter height represents relative entropy above the background residue distribution frequencies. A modified sequence logo showing the differential information content of NasR proteins and non‐NasR NIT domain proteins identifies residues differentially conserved in the two groups. Letter height represents the relative entropy or KL divergence of the sequence distributions between each group. (b) A sequence logo shows conservation of residues across all ANTAR domain‐containing proteins corresponding to residues in the helix 10‐11 region of Klebsiella oxytoca NasR. Letter height represents relative entropy above the background residue distribution frequencies. A modified sequence logo showing the differential information content of NasR proteins and non‐NasR ANTAR domain proteins identifies residues differentially conserved in the two groups. Letter height represents the relative entropy or KL divergence of the sequence distributions between each group. (c) Residues showing conservation bias for NasR‐like proteins cluster at the interdomain interface, as shown in the schematic. Some of these residues appear to participate in amino acid pairing interactions between the NIT and ANTAR domains of NasR as well as part of the hydrogen bonding network connecting this interface with the proposed nitrate binding pocket. Individual residues in helices 1, 5 and 6 (NIT domain) and 10 and 11 (ANTAR domain) are shown in circles. Wide dashed lines represent likely hydrogen bonding interactions between associated residues and narrow dashed lines represent proposed interactions between arginine residues and nitrate

FIGURE 5

The identity of ANTAR residues in a surface‐associated positive patch can affect RNA binding. (a) A predicted electrostatic surface potential map of the NasR structure (PDB: 4AKK) shows a region of the ANTAR domain dimer with positive surface potential (bottom center). The positions of residues R340, K345 and K347 are demarcated with black circles. K345 is exposed near the center of the positive patch, while R340 and K347 are sequestered by nearby negatively charged residues. (b) Equilibrium saturation binding curves for three charge‐conservative substitution mutant proteins of His10‐MBP‐NasR in the presence of 1 mM nitrate. Error bars represent the standard deviation of the anisotropy change of four replicate wells relative to the no‐ligand sample. The fit lines represent a Hill equation model

FIGURE 6

Amino acid interactions at the NIT‐ANTAR interface are required for control of RNA‐binding activity. (a‐k) Equilibrium saturation binding curves for alanine substitution mutants of residues near the NIT‐ANTAR interdomain interface. Blue data points represent binding curves performed in the absence of nitrate, while red data points represent binding curves with 1 mM nitrate. Error bars represent the standard deviation of the anisotropy change of four experiments relative to the no‐ligand sample. The fit lines represent a Hill equation model. (l) Site‐directed mutations that perturb interdomain interactions affect transcription attenuation by NasR. DNA templates encompassing the K. oxytoca nasF leader region were amplified by PCR and incubated with E. coli RNA polymerase (New England Biolabs) and P³²‐radiolabeled UTP. These reactions also contained 8 μM purified His10‐MBP‐NasR proteins and were incubated in the presence or absence of 1 mM nitrate. The products of these transcription reactions were resolved by 8% urea‐denaturing polyacrylamide gel electrophoresis (PAGE). This resulted in two primary transcription products, one that corresponds to premature transcription termination (“T”) and a longer transcript corresponding to run‐off transcription (“RO”). The fraction of RO was determined for wild‐type protein and each of the mutant proteins

FIGURE 7

Summary model for one‐ and two‐component ANTAR regulatory proteins. (a) The default state of response regulator ANTAR proteins is presumed to be that of a monomer. These monomers bind either one‐ or two‐hairpin RNA with a similar low affinity. (b) Upon phosphorylation by the cognate histidine kinase, the protein dimerizes—binding both hairpins of a two‐hairpin RNA with an increased apparent affinity—but retains low‐affinity for single‐hairpin RNA. (c) NasR‐like proteins (containing NIT and ANTAR domains) form a constitutive dimer incapable of binding RNA. (d) Upon ligand binding, these one‐component regulators adopt an alternate configuration. Autoinhibitory interactions are disrupted, thereby releasing the RNA‐binding face of the ANTAR domain to bind RNA with apparent affinity comparable to dimerized response regulator ANTAR proteins