Kinetic modelling techniques for simultaneous mass and structural changes during peptide-membrane interactions
2017-02-27T00:15:29Z (GMT) by
Peptide-membrane interactions play an important role in many biological systems, drug delivery systems and therapeutics. Examples include antimicrobial peptides, which show promise as alternatives to conventional antibiotics, and cell-penetrating peptides, which can pass though the plasma membrane of cells and therefore have the potential for use in drug delivery. However, the precise mechanisms of these interactions is still uncertain and, in particular, there is limited understanding of the role that the membrane and its structure plays in these interactions. Dual-polarization interferometry (DPI) is a recently developed optical technique that can simultaneously measure multiple properties of a layer adsorbed to the surface of a chip. Here, the mass and birefringence (a measure of bilayer structure) of a supported lipid bilayer are measured simultaneously in real time during the binding and subsequent dissociation of certain peptides. A novel analysis technique is then used to fit a kinetic model simultaneously to the mass and birefringence data thus obtained. Depending on the data observed, the most appropriate model may be either two-state or three-state (in terms of distinct bound peptide states) with or without adjustments to account for lateral bilayer expansion, lag in birefringence changes from peptide binding, and mass thresholds for state transitions. These methods are used to investigate the membrane binding of several model peptides. For the binding mechanism of the antimicrobial peptide HPA3 it is found that a two-state model is sufficient for its binding to gel-phase membranes, while a three-state model with bilayer expansion is necessary for fluid-phase membranes. The binding of the magainin 2 analogue varied depending on the membrane type, but where a suitably accurate fit was obtained, a three-state model with bilayer expansion was needed, with birefringence lag in some cases. The cell-penetrating peptide penetratin and a truncated analogue R8K-biotin were compared, with R8K-biotin fitted by a two-state model while penetratin required a three-state model, which is indicative of a more complex and involved binding process. The binding of helix 8 of the angiotensin receptor was compared for membranes with or without phosphatidylinositol phosphates (PIPs); there was a dramatic change in the binding between membranes with and without PIPs, indicating a significant interaction between Helix 8 and PIPs. For each of these peptide types, a hypothesis about the binding mechanism is formed from the data obtained through modelling. Additionally, testing is performed to measure the level of variability of the modelling, both as a result of intrinsic uncertainties in the modelling itself, and variations in the data obtained between different experimental runs. This new modelling technique allows for powerful quantitative characterisation of DPI data, and provides new insights into peptide-membrane interactions that are likely to improve scientific understanding of these systems and assist the development of therapeutic products associated with them.