DNA Bends the Knee to Transcription Factors

Image result for dna blue background

Main Text

During gene expression, multiprotein-DNA complexes are formed to regulate transcription. Binding of transcription factors can locally deform DNA, for example, by bending its helical axis. Dimers or oligomers of transcription factors may interact with DNA at multiple sites and determine the bending angle collectively. Bending may have a role in transcription regulation in several ways (1). It can loop the DNA to bring distal DNA regions (such as promoters and enhancers) into close proximity. It can also mediate interactions between transcription factors by wrapping the DNA around the large interacting protein complexes. The release of elastic energy stored in the bent DNA may then favor processes leading to a straighter DNA conformation, such as the dissociation of an RNA polymerase during transition from initiation to elongation.

Protein-induced bending can be detected by molecular biology approaches (1) such as electrophoretic mobility assays and cyclization analysis, or biophysical techniques like small-angle x-ray scattering (SAXS), electron microscopy, atomic force microscopy (2), and Förster resonance energy transfer (FRET). Since its early introduction for the determination of the four-way DNA (Holliday) junction solution conformation (34), FRET has proven to be a powerful biophysical tool for measuring and comparing intramolecular distances and conformational changes in DNA and DNA-protein complexes. FRET has been used early on to detect sequence-induced DNA curvature due to adenine repeats (5) or DNA-bending induced by protein binding (6). More recently, the emergence of single-molecule FRET (smFRET) microscopy has provided unique opportunities to directly dissect subpopulations having different conformations of protein-DNA complexes, identify transient reaction intermediates, and quantitate DNA bending. smFRET has been used for, among other things, characterizing the conformational dynamics of the Holliday junction (7) or describing intermediates in the molecular mechanism of helicase-assisted DNA unwinding (8).

In their article appearing in this issue of Biophysical Journal, Rubio-Cosials et al. (9) elucidate the DNA-binding mechanism of human mitochondrial transcription factor A (TFAM) and the conformation of DNA bent by TFAM. TFAM regulates transcription and transcription-dependent replication of DNA in mitochondria. It possesses two high-mobility-group (HMG) domains; HMG proteins have been shown earlier to bend DNA and to bind preferentially to specific DNA structures such as four-way DNA junctions (10). The study elegantly combines smFRET, electrophoretic mobility shift assay, and SAXS experiments with atomistic molecular dynamics simulations. The work’s main objective is to decipher whether the light strand promoter DNA-TFAM complex in solution assumes a U-turn, as suggested by crystallographic data (11), or a V-shape, as proposed by other studies. The authors compare the end-to-end distances of two DNA constructs of 30 or 50 bp in length by smFRET measurements. The rationale is that if the DNA is kinked at only one site, it will have a V-shape, and the two ends of the helix will be farther apart for the 50 bp duplex as compared to the shorter, 30 bp one, whereas if the duplex is kinked at two sites and assumes a U-shape with approximately parallel duplex ends, the two duplexes will have similar end-to-end distances. The article convincingly proves that TFAM-bound DNA has a U-turn conformation due to the two kinks introduced by the progressive and cooperative binding of the two HMG-box domains having different affinities, and that this conformation is stabilized by the DNA binding of the linker region between the domains. SAXS reveals that the TFAM-DNA complex is dynamic and is flexible at the TFAM linker region. This is also confirmed by atomistic molecular dynamics simulations, which predict a butterfly-like motion of the entire complex, during which the DNA approaches its naked straightened conformation and the linker region of TFAM adapts this motion by unfolding and unbinding from the DNA, whereas the HMG domains maintain their DNA contacts. Molecular dynamics also adds to the overall picture that the DNA helix has extra intrinsic curvature at the TFAM contact points, which may also facilitate the stable binding of the protein. This exemplary combination of experimental and modeling methods sheds light on both the structural and dynamic properties of an important regulatory protein-DNA complex.


This work was supported by grants from Hungary’s National Scientific Research Fund Programs (OTKA) (K103965) and the Economic Development and Innovation Operational Program (GINOP) (GINOP-2.3.2-15-2016-00026 and GINOP-2.3.3-15-2016-00030 to G.V.). The Rueda Lab is funded by the MRC London Institute of Medical Sciences (RCUK MC-A658-5TY10), the Wellcome Trust (206292/C/17/Z), the Leverhulme Trust (RPG-2016-214), and Imperial College London.

See full article