Crystal structures of the DNA-binding domain of Escherichia coli proline utilization A flavoprotein and analysis of the role of Lys9 in DNA recognition

Abstract

PutA (proline utilization A) from Escherichia coli is a 1320-amino-acid residue protein that is both a bifunctional proline catabolic enzyme and an autogenous transcriptional repressor. Here, we report the first crystal structure of a PutA DNA-binding domain along with functional analysis of a mutant PutA defective in DNA binding. Crystals were grown using a polypeptide corresponding to residues 1-52 of E. coli PutA (PutA52). The 2.1 Angstrom resolution structure of PutA52 mutant Lys9Met was determined using Se-Met MAD phasing, and the structure of native PutA52 was solved at 1.9 Angstrom resolution using molecular replacement. Residues 3-46 form a ribbon-helix-helix (RHH) substructure, thus establishing PutA as the largest protein to contain an RHH domain. The PutA RHH domain forms the intertwined dimer with tightly packed hydrophobic core that is characteristic of the RHH family. The structures were used to examine the three-dimensional context of residues conserved in PutA RHH domains. Homology modeling suggests that Lys9 and Thr5 contact DNA bases through the major groove, while Arg15, Thr28, and His30 may interact with the phosphate backbone. Lys9 is shown to be essential for specific recognition of put control DNA using gel shift analysis of the Lys9Met mutant of full-length PutA. Lys9 is disordered in the PutA52 structure, which implies an induced-fit binding mechanism in which the side chain of Lys9 becomes ordered through interaction with DNA. These results provide new insights into the structural basis of DNA recognition by PutA and reveal three-dimensional structural details of the PutA dimer interface.

Figure 1.

Structure of the PutA RHH domain. (A) Stereoscopic view of one PutA52K9M chain, covered by a figure of merit weighted experimental electron density map (1 σ). The map was calculated using amplitudes from the remote energy data set and density-modified MAD phases from RESOLVE. This map was used to build the initial model of PutA52K9M. The model shown is the final, refined structure of the D chain of PutA52K9M. (B) Ribbon drawing of the PutA RHH dimer. The two subunits are colored red and green. Secondary structure elements are labeled β1, αA, and αB for the green subunit. The blue patches and side chains indicate locations of residues mentioned in the text: Thr5, Gly7, Lys9 (Met9), Arg15, Asp26, Pro29, His30, and Trp31.

Figure 2.

Amino acid sequence alignments of RHH and PutA proteins. (A) Structure-based sequence alignment of the E. coli PutA DNA-binding domain (PutA52) with other RHH domains. The numbers above the alignment correspond to E. coli PutA. The numbers on the left indicate the starting residue for each protein sequence. Boxes indicate conserved basic residues of β1, conserved glycine in the αA–αB turn, and the N_cap of αB. Stars below the alignment denote residues of β1 that typically interact with DNA. Ovals below the alignment denote conserved hydrophobic residues of the RHH family. Triangles above the alignment indicate residues unique to PutA RHH domain sequences. PDB accession codes for the sequences shown in this alignment are as follows: Arc (1BDT), Omega (1IRQ), ParG (1P94), NikR (1Q5V), MetJ (1CMC), and CopG (2CPG). (B) Sequence alignment of nine PutA RHH domains. Residues identically conserved in all nine sequences are indicated in black. Stars below the alignment denote residues of β1 that typically interact with DNA. Ovals below the alignment denote conserved hydrophobic residues of the RHH family. Triangles above the alignment indicate residues unique to PutA RHH domain sequences.

Figure 3.

Comparison of the αA–αB loops of PutA52 and Arc. PutA52 is magenta, and Arc is white. Residues are labeled as PutA52/Arc. The orientation of this figure is similar to that of the green subunit in ▶.

Figure 4.

Gel mobility shift assay of wild-type PutA and PutA mutant K9M. Wild-type PutA (0.3 μM dimer) and PutAK9M (0.2–4 μM dimer) were incubated with IRdye-700 labeled put control DNA (5 nM) for 20 min at 20°C in 50 mM Tris (pH 8.0) containing 10% glycerol and 100 μg/mL nonspecific calf thymus DNA. Protein–DNA complexes were separated using a nondenaturing polyacrylamide gel (4%).

Figure 5.

Model of the PutA RHH domain interacting with DNA. (A) Subunits of the PutA52 dimer are shown in green and yellow. DNA is shown in CPK mode. Helices of PutA52 are labeled αA, αB for the green subunit and αA′, αB′ for the yellow subunit. Note that the β-sheet is inserted into the major groove of DNA. N termini of the green and yellow subunits are labeled N and N′, respectively. Side chains are drawn for Thr4, Thr5, Lys9, Arg15, Thr28 (N_cap of αB), and His30. This model is based on homology with the Arc repressor as described in Materials and Methods. (B) Electrostatic potential of the DNA-binding surface of the PutA52 dimer. The potential ranges from −0.25 V (red) to +0.25 V (blue). The orientation of this panel is related to that of the panel A by a rotation of 90° around the horizontal axis, such that the DNA-binding surface faces the viewer. The arrows indicate the location of the β-sheet.