The biological function of large macromolecular assemblies depends on their structure and their dynamics over a broad range of timescales; for this reason its investigation poses significant challenges to conventional experimental techniques. A promising experimental technique is hydrogen-deuterium exchange detected by mass spectrometry (HDX-MS). I begin by presenting a new computational method for quantitative interpretation of deuterium exchange kinetics. The method is tested on a hexameric viral helicase φ12 P4 that pumps RNA into a virus capsid at the expense of ATP hydrolysis. Molecular dynamics simulations predict accurately the exchange kinetics of most peptide fragments and provide a residue-level interpretation of the low-resolution experimental results. This approach is also a powerful tool to probe mechanisms that cannot be observed by X-ray crystallography, or that occur over timescales longer than those that can be realistically simulated, such as the opening of the hexameric ring. Once validated, the method is applied on a homologous system, the packaging motor φ8 P4, for which RNA loading and translocation mechanisms remain elusive. Quantitative interpretation of HDX-MS data, as well as Förster resonance energy transfer (FRET) and computational observations, suggest that the C-terminal domain of the motor plays a crucial role. A new translocation model of φ8 P4 is proposed, for which the affinity between the motor and RNA is modulated by the C-termini. In the final result chapter, the amount of the structural information carried by HDX-MS data is quantitatively analysed. The impact of the averaging of the exchange over peptide fragments on the information content is investigated. The complementarity of data obtained from HDX-MS and data obtained from other techniques (such as NMR, FRET or SAXS) is also examined.