Structural basis for higher-order DNA binding by a bacterial transcriptional regulator

Abstract

Transcriptional regulation by binding of transcription factors to palindromic sequences in promoter regions is a fundamental process in bacteria. Some transcription factors have multiple dimeric DNA-binding domains, in principle enabling interaction with higher-order DNA structures; however, mechanistic and structural insights into this phenomenon remain limited. The Pseudomonas putida toxin-antitoxin (TA) system Xre-RES has an unusual 4:2 stoichiometry including two potential DNA-binding sites, compatible with a complex mechanism of transcriptional autoregulation. Here, we show that the Xre-RES complex interacts specifically with a palindromic DNA repeat in the promoter in a 1:1 molar ratio, leading to transcriptional repression. We determine the 2.7 Å crystal structure of the protein-DNA complex, revealing an unexpected asymmetry in the interaction and suggesting the presence of a secondary binding site, which is supported by structural prediction of the binding to the intact promoter region. Additionally, we show that the antitoxin can be partially dislodged from the Xre-RES complex, resulting in Xre monomers and a 2:2 Xre-RES complex, neither of which repress transcription. These findings highlight a dynamic, concentration-dependent model of transcriptional autoregulation, in which the Xre-RES complex transitions between a non-binding (2:2) and a DNA-binding (4:2) form.

Fig 1. Organisation, activity, and regulation of the xre-res^Pp operon.

(A) Schematic representation the P. putida KT2440 xre-res locus (xre-res^Pp) including the 172-bp promoter region (P_XR) and the upstream gene (nfsB). Top, the xre^Pp gene (orange) encodes a protein antitoxin with an N-terminal helix-turn-helix (HTH) domain and a C-terminal domain of unknown function (DUF2384), known to interact with the RES toxin, while res^Pp (blue) encodes a NADase toxin with a RES domain. Below, close-up of the P_XR promoter region showing the three inverted repeats (S1, S2, and S4) and one directional repeat (S3), of which S4 is imperfect and contains one mismatch (yellow box, see also S1A Fig). (B) Overview of the GFP-based assay to measure regulation of P_XR in E. coli MG1655 in vivo. The pBAD33 vector allows for arabinose induction of the xre-res^Pp genes while P_XR coupled to GFP allows for read-out of the transcriptional activity of the promoter. (C) and (D) Promoter activity assays in E. coli MG1655. Activity is measured as GFP signal in relative fluorescence units (RFU) during growth without ((C), -Ara) or with ((D), + Ara) arabinose using the promoter reporter pGH254Kgfp::P_XR (P_XR), as well as empty pBAD33 (-, green curves) or pBAD33::xre-res^Pp vector (xre-res, orange curves). All curves represent mean-of-mean values (line, n = 3) with standard error of mean (SEM) shown as a shadow. The dotted vertical line indicates the 350 min time point.

Fig 2. Determination of the minimal promoter requirements for transcription and regulation of P_XR.

(A) Detailed overview of the P_XR promoter (top) and schematic representation of promoter variants (below). Full lines indicate the included parts while dotted lines represent regions that have been removed. Complete growth curves and details of the constructs are found in S1 Fig and S1 Table. (B) Promoter activity assays in E. coli MG1655. The reporter plasmid pGH254Kgfp encoding promoter variations was combined with pBAD33::xre-res^Pp. For each variant, the GFP signal (RFU) measured at t = 350 min from non-induced cultures (-Ara, green bars) is compared to that from arabinose-induced cultures (+Ara, orange bars). Mean RFU values extracted at the same time point from three independent experiments were treated as biological replicates (dots, n = 3) and error bars show SEM. Complete growth curves are included in S1 Fig. Statistical analysis was performed using Welch's t-test (with equal variance). ns, not significant; **-, p < 0.01; ***, p < 0.001. (C) Schematic representation of the repeat variants of the imperfect inverted repeat in S4 of the P_XR promoter region. The DNA sequence of the imperfect inverted repeat is highlighted (light grey), including the mismatch in the first repeat (red), the perfect part of the repeat (dark grey) and the alterations (green). The mismatch was removed by a substitution either in the first (5’-perf) or second (3’-perf) half of the repeat. (D) As b, but using promoter variations from c. ns, not significant; *, p < 0.05.

Fig 3. Xre-RES^Pp forms a 4:2 hexameric complex in solution and interacts specifically with the

S4 element of the promoter. (A) SEC-MALS analysis of purified Xre-RES^Pp_His6 complex. The chromatogram shows the elution profile as measured by absorption at 280 nm (blue) and the molecular mass measured by MALS (red). The average mass corresponding to the middle of the peak is 92 ± 4 kDa (dotted line). (B) The crystal structure of heterohexameric Xre-RES^Pp (PDB: 6GW6) with total mass (103 kDa) and HTH domains indicated [14]. The RES toxins are coloured in shades of blue, while the Xre antitoxins are in shades of green. (C) Analytical SEC (a-SEC) analysis of Xre-RES^Pp_His6 (Xre-RES, blue), S4 (S4-DNA only, red), and DNA mixed with protein in a 1:1 ratio (Xre-RES + S4-DNA, purple). Black arrows represent known molecular weight standards (conalbumin, 75 kDa; carbonic anhydrase, 29 kDa). The full line shows 280 nm absorbance, while the dotted line is 260 nm absorbance. (D) Isothermal Titration Calorimetry (ITC) titration of S4 dsDNA into the Xre-RES^Pp complex. The top panel shows the raw thermogram (DP, differential power), while the bottom panel shows the integrated heat per injection (ΔH, kJ/mol). The solid is line is the fit using the “one set of sites” model.

Fig 4. Crystal structure of the DNA-bound Xre-RESPp complex.

(A) Overview of the P. putida Xre-RES complex (green/blue cartoon) is bound to a 30 bp dsDNA (red/orange) with both HTH recognition helices of the of a single Xre dimer interacting with DNA. (B) Schematic overview of the interactions between Xre and DNA including both sequence specific and unspecific contacts as well as the contacts from the symmetry related interaction. The shaded DNA region could not be build. (C) Close-up on the interaction between Xre and DNA with interactions between selected residues of both Xre molecules and the DNA highlighted. (D) Partial sequence alignment of P. putida Xre and a range of homologous proteins from other bacterial species. UniProt IDs are listed in parentheses for all proteins, along with PDB IDs for P. putida Xre (PDB ID 6GW6) [14], M. tuberculosis MbcA (PDB ID 6FKG) [12], V. parahaemolyticus VPA0770 (PDB ID 8GUG) [18], and P. aeruginosa NatR (PDB ID 8QNL) [19]. Residues conserved with P. putida Xre are shown in red with functionally similar residues in orange. Starting and ending residue numbers are listed at each end of the sequences. (E) Promoter activity in E. coli MG1655 using promoter reporter plasmid pGH254Kgfp::P_XR (P_XR) and either pBAD33 (-), pBAD33::xre-res^Pp (xre-res), or pBAD33::xre^R67A-res^Pp (xre^R67A-res), as indicated. Complete growth curves and details of the constructs are found in S4 Fig. (F) In the crystal, the Xre-RES complex (blue/green) interacts with a primary dsDNA molecule (DNA 1) as well as a symmetry-related DNA molecule (DNA 2).

Fig 5. DNA binding requires the Xre dimer, which is unstable at low concentrations.

(A) Promoter activity assays in E. coli MG1655, where the promoter reporter plasmid pGH254Kgfp::P_XR (P_XR) was combined with either empty pBAD33 (-), pBAD33::xre^Pp (xre), or pBAD33::xre-res^Pp vector (xre-res), as indicated. Complete growth curves and details of the constructs are found in S4 Fig. Statistical analysis was performed using Welch's t-test (with equal variances). ns, not significant; ***, p < 0.001. (B) SEC-MALS analysis showing that isolated Xre^Pp_CHis6 is a monomer with an approximate mass of 22 ± 1 kDa in solution. The chromatogram shows elution profile measured by 280 nm absorbance (blue) and calculated mass from MALS (red). (C) Single-molecule mass determination of Xre^Pp-RES^Pp_His6 at 2, 5, and 10 nM. The violin plots represent the density of all measured single-molecule events. The solid black horizontal line indicates the median mass, while the vertical line shows the standard deviation. The theoretical masses of the Xre^Pp-RES^Pp_His6 complex with 4:2, 3:2, and 2:2 stoichiometries are indicated by dotted lines. Significance was tested with the Wilcoxon signed rank test. *****, p < 5.55e-7. (D) Single-molecule mass determination for Xre_His6-RES^Pp on the 2:2 complex form at 5 and 10 nM, as in (C). (E) High resolution analytical SEC (a-SEC) analysis of Xre_NHis6-RES (Xre-RES, blue), S4 (DNA, red), and DNA mixed with protein in a 1:10 ratio (Xre-RES + DNA, purple). The full line shows 280 nm absorbance, and the dotted line is 260 nm absorbance. (F) AlphaFold3-predicted structure of the Xre-RES complex interacting with the full 172 bp promoter region (not all nucleotides shown). (G) Alignment of crystal structure of the DNA-bound Xre-RES complex, including the symmetry-related dsDNA molecule (blue/green/orange) with the AlphaFold 3 prediction (grey).

Fig 6. Model for transcriptional regulation of the xre-res^Pp operon.

Expression of xre-res^Pp leads to the formation of a 4:2 Xre-RES^Pp complex (A), in which one Xre dimer binds an imperfect inverted repeat in the promoter region (S4), followed binding of the second Xre dimer to the region between S3 and S4 and subsequently transcriptional repression (B). This lowers the concentration of the Xre antitoxin, leading to the dissociation of the Xre dimer (C) and formation of a toxin-neutralising, but non-DNA binding 2:2 complex (D), and thereby initiation of transcription by the RNA polymerase.