Which way up? Recognition of homologous DNA segments in parallel and antiparallel alignment

TL;DR

This study investigates the recognition energy between homologous DNA segments in parallel and antiparallel alignments, revealing a shallow recognition well for antiparallel DNA and validating theoretical models with simulations.

Contribution

It is the first to analyze homologous DNA recognition in antiparallel alignment and compares theoretical approximations with Monte Carlo simulations considering DNA flexibility.

Findings

Recognition well exists for antiparallel homologous DNA but is very shallow.

Theoretical approximations are validated against MC simulations, especially when DNA flexibility is included.

DNA torsional flexibility improves agreement between models and simulations.

Abstract

Homologous gene shuffling between DNA promotes genetic diversity and is an important pathway for DNA repair. For this to occur, homologous genes need to find and recognize each other. However, despite its central role in homologous recombination, the mechanism of homology recognition is still an unsolved puzzle. While specific proteins are known to play a role at later stages of recombination, an initial coarse grained recognition step has been proposed. This relies on the sequence dependence of the DNA structural parameters, such as twist and rise, mediated by intermolecular interactions, in particular electrostatic ones. In this proposed mechanism, sequences having the same base pair text, or are homologous, have lower interaction energy than those sequences with uncorrelated base pair texts; the difference termed the recognition energy. Here, we probe how the recognition energy changes when one DNA fragment slides past another, and consider, for the first time, homologous sequences in antiparallel alignment. This dependence on sliding was termed the recognition well. We find that there is recognition well for anti-parallel, homologous DNA tracts, but only a very shallow one, so that their interaction will differ little from the interaction between two nonhomologous tracts. This fact may be utilized in single molecule experiments specially targeted to test the theory. As well as this, we test previous theoretical approximations in calculating the recognition well for parallel molecules against MC simulations, and consider more rigorously the optimization of the orientations of the fragments about their long axes. The more rigorous treatment affects the recognition energy a little, when the molecules are considered rigid. However when torsional flexibility of the DNA molecules is introduced, we find excellent agreement between analytical approximation and simulation.