In silico molecular understanding of opioid receptor activation

<p>This thesis explores the behaviour the <em>homo sapiens</em> δ opioid receptor (δ OR) in complex with a variety of morphinan (morphine derivatives) opioids, aiming to answer whether molecular dynamics (MD) tools can be used to determine the molecular underpinning of opioid struc...

Volledige beschrijving

Bibliografische gegevens
Hoofdauteur: Velgy, N
Andere auteurs: Biggin, P
Formaat: Thesis
Gepubliceerd in: 2018
Onderwerpen:
Omschrijving
Samenvatting:<p>This thesis explores the behaviour the <em>homo sapiens</em> δ opioid receptor (δ OR) in complex with a variety of morphinan (morphine derivatives) opioids, aiming to answer whether molecular dynamics (MD) tools can be used to determine the molecular underpinning of opioid structure-activity relationship.</p> <p>To do so, the results of conventional MD simulations are reported and analysed to shed light on the local changes to the binding pocket, and a novel method, N-Body Information Theory, is adapted to study concerted motions within the receptor as a result of morphinan binding. The reported simulations reveal an interesting connection between Tyr130 <sup>3.30</sup> and Trp173 <sup>4.50</sup> (superscripts correspond to the Ballesteros-Weinstein numbering scheme), the latter of which is critical in stabilising inactive states of the receptor.</p> <p>Due to sampling limitations of medium timescale (100 ns) conventional MD is not capable of exploring large conformational changes (necessary for receptor activation, e.g. outer movement of TM6), a novel collective variable (CV) based on unique contacts between two receptor states was developed. Using the μ OR, this new CV (termed FNC) was designed to explore the transition path from inactive to active states (and vice-versa) of the receptor. Umbrella sampling simulations showed that, despite the high free energy values, the presence of the agonist DIPP-NH2 lowered the energetic barrier for activation, agreeing with previously reported experimental results.</p> <p>Encouraged by this, FNC were generated for the δ OR. Enhanced sampling simulations of the δ OR bound to four ligands (the antagonists naltrindole and naltrexone, and the agonists morphine and oxycodone), however, failed to elucidate the differences between agonists and antagonists, likely due to poor phase space sampling borne from umbrella sampling. Improvements to the work performed are suggested, particularly with respect to the sampling of the FNC CV and orthogonal states.</p>