What happens to DNA duplexes in the gas phase?

Massimiliano Porrini (1) , Frédéric Rosu (2) , Valérie Gabelica (1)

1 ARNA - Acides Nucléiques : Régulations Naturelle et Artificielle
2 Soutien à la Recherche de l'Institut Européen de Chimie Biologique

Introduction: DNA duplexes were among the first non-covalent complexes studied by mass spectrometry. Their gas-phase kinetic stability suggested that Watson-Crick base pairing and nearest-neighbor interactions are conserved in vacuo [1,2]. Ion mobility combined with modeling suggested that long duplexes keep a B-helix structure in vacuo, whereas shorter ones (8-16 base pairs) convert into an A-helix [3,4]. Here we challenge this proposal: collision cross sections (CCS) compatible with B-helix or A-helix are indeed obtained at high charge states, but not at the low charge states that predominate in ESI-MS spectra recorded from aqueous solutions at physiological ionic strength. They are much lower. We used extensive molecular dynamics simulations and semi-empirical calculations to understand what happens to DNA duplexes in the gas phase. Methods: Ion mobility spectrometry experiments were carried out on an Agilent 6560 IMS-Q-TOF instrument, and collision cross sections were determined with the drift tube in helium. All molecular dynamics simulations were performed with sander module of AmberTools14 (gas phase) and pmemd module of Amber12 (solution phase), using the parmbsc1 force field (M. Orozco, private communication) which is the latest refinement of parmbsc0, a force field specifically developed for nucleic acids [5]. Semi-empirical calculations were performed with MOPAC software implementing the PM7 method. Preliminary Data: Collision cross sections for [(12-mer)2]5- duplexes range from 700 Ų for a 100%-GC duplex to 750 Ų for a 33%-GC duplex. CCS distributions are broader than the instrumental peak width, indicating that multiple conformations co-exist. The experimental CCSs are all far smaller than that of a B-helix (~900 Ų), and also far smaller than the values reported on more densely charged duplexes of similar size [3,4]. Unbiased MD simulations of protonated duplexes lead to helices which retain some of their initial structural features, but these helices are distorted because hydrogen bonds form between phosphates across the minor groove. Similar gas-phase structures were proposed before [3,6]. Unfortunately, their computed CCS is above 850 Ų, meaning they are far larger than the experimental gas-phase structures of the 5- charge state (although they may represent the structures formed at higher charge states 6- and 7-). For the 5- charge state, the compact structures observed experimentally could be unstructured globular dimers, but this would not explain the persistence of a memory of the sequence in MS/MS experiments [1,2]. We therefore explored biased MD simulation to force the duplex exploring alternative regions on the potential energy surface. In particular, a compaction through hydrogen bond formation between phosphates across the major groove was found to lead the helix to collapse along the longitudinal axis, and the resulting CCSs are now compatible with the experimental ones (even more so when the structures are re-optimized at the semi-empirical level using the PM7 method). Novel Aspect: This work shows that DNA duplexes sprayed in native conditions are actually much more compact than previously thought.

Mots clés: Conformation changes; Nucleic acids; Mass spectrometry; Ion mobility spectrometry

Domaines: Chimie analytique; Chimie théorique et/ou physique; Biophysique

Massimiliano Porrini, Frédéric Rosu, Valérie Gabelica. What happens to DNA duplexes in the gas phase?. 63rd ASMS Conference on Mass Spectrometry and Allied Topics (ASMS 2015), May 2015, St Louis, United States. , Proceedings of the ASMS Annual Conference 2015 St. Louis, 2015. ⟨hal-01562195⟩