A JUMBLED CODE: UNVEILING THE ENIGMATIC DNA BINDING PATTERNS OF THE MARR FAMILY

Transcription factors from the multiple antibiotic resistance regulator family (MarR) form a widespread group in Bacteria and Archaea. In Bacteria, MarR gene expression control is known for multidrug resistance, organic solvent resistance, virulence, stress responses and metabolism. The MarR family is, therefore, essential for understanding bacterial gene regulation, especially relevant in the current conjuncture of emergence and propagation of multidrug resistant bacterial strains. Usually, MarR proteins bind to palindromic sequences found in the intergenic region of regulated genes and the gene that encodes the protein, repressing or, more rarely, activating gene expression. They have variable DNA binding modes and affinities and, although most of them recognize specific sequences, many interact with degenerated sequences. The current study sought to identify evolutionary patterns in DNA recognition by MarR proteins, aiming to enable the prediction of which genes would be regulated by proteins of the family without a structure deposited on PDB or in which it is not in complex with DNA. Accordingly, identification of protein-DNA interactions in a subset of the MarR family, multiple sequence alignment (MSA) analysis in search of conserved and/or co-evolving positions in this family and molecular dynamics simulations investigating the protein-DNA interface at 2 positions of interest were employed. A Pfam alignment representative of the MarR family was used in coevolution analysis after filtering out fragments and redundancy. From the resulting proteins, those which possessed a structure in complex with DNA were chosen. Protein-DNA interactions were obtained by literature research and annotation of interactions. Ten positions involved in DNA binding in the 5 analyzed structures were evaluated for residue type variation and DNA interaction. All these positions showed some variation in their residues, with the highest degree of conservation observed in one position showing only positive residues (four arginines and one lysine). In most cases, a pattern in which the presence of a given amino acid appears to determine which specific base it binds to was not observed, except for an arginine with respect to three positions in the DNA chain. Our results corroborate the idea that, unlike other transcription factors, the MarR family does not seem to have clear specificity relationships between DNA contact regions and responsive elements. We can conclude that genes controlled by these factors cannot be predicted solely by amino acid sequences, and the effects of the residues found in this analysis needs further investigation. Moreover, the inspection of the electronic density of these residues revealed a poor fit to the modeled residue for the most important residue sets in one of the structures: the most conserved and the one involved in base determination. The DNA binding modes of these residues will be investigated further using molecular dynamics (distribution and evolution of hydrogen bonds, end-state free energy etc). We aim to accurately evaluate the flexibility, energetic impact and model quality for these regions presumed to affect DNA binding specificity.
This work was supported by Conselho Nac. Des. Cient. Tecnológico (CNPq) and by the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG).