Towards Solving the Dopamine G Protein Coupled Receptor Modelling Problem

Wood, Christopher D (1994) Towards Solving the Dopamine G Protein Coupled Receptor Modelling Problem. PhD thesis, University of Glasgow.

Abstract

The overall aim of this work has been to furnish a model of the dopamine (DA) receptor D2. There are currently two sub-groups within the DA family of G protein coupled receptors (GPCRs): D1 sub-group (includes D1 and D5) and the D2 sub-group (includes D2, D3 and D4). Organon (UK) Ltd. supplied a disk containing the PDB atomic co-ordinates of the integral membrane protein bacteriorhodopsin (bRh; Henderson et al., 1975 and 1990) to use as a template to model D2 - the aim being to generate a model of D2 by simply mutating the side-residues of bRh. The assumption being that bRh had homology with members of the supergene class of GPCRs. However, using the GCG Wisconsin GAP algorithm (Devereux et al., 1984) no significant homology was detected between the primary structures of any member of the DA family of GPCRs and bRh. However, given the original brief to carry out homology modelling using bRh as a template (see appendix 1) I felt obliged to carry out further alignments using a shuffling technique and a standard statistical test to check for significant structural homology. The results clearly showed that there is no significant structural homology, on the basis of sequence similarity, between bRh and any member of the DA family of GPCRs. Indeed, the statistical analysis clearly demonstrated that while there is significant structural homology between every catecholamine binding GPCR, there is no structural homology what so ever between any catecholamine binding GPCR and bRh. Hydropathy analysis is frequently used to identify the location of putative transmembrane segments. However, is difficult to predict the end positions of each ptms. To this end a novel alignment algorithm (DH Scan) was coded to exploit transparallel supercomputer technology to provide a basis for identifying likely helix end points and to pinpoint areas of local homology between GPCRs. DH Scan clearly demonstrated characteristic transmembrane homology between different subtype DA GPCRs. Two further homology algorithms were coded (IH Scan and RH Scan) which provided evidence of internal homology. In particular IH Scan independently revealed a repeat region in the 3rd intracellular loop (iIII) of D4 and RH Scan revealed palindromic like short stretches of amino acids which were found to be particularly well represented in predicted ?-helices in each DA receptor subtype. In addition, the profile network prediction algorithm (PHD; Rost et al., 1994) predicted a short alpha-helix at greater than 80% probablility at each end of the third intracellular loop and between the carboxy terminal end of transmembrane VII and a conserved Cys residue in the forth intracellular loop. Fourier analysis of catecholamine binding GPCR primary structures in the form of a multiple-sequence file suggested that the consensus view that only those residues facing the protein interior are conserved is not entirely correct. In particular, transmembrane helices II and III do not exhibit residue conservancy characteristic of an amphipathic helix. It is proposed that these two helices undergo a form of helix interface shear to assist agonist binding to a Asp residue on helix II. This data in combination with information from a number of papers concerning helix shear interface mechanism and molecular dynamic studies of proline containing ?-helices suggested a physically plausible binding mechanism for agonists. While it was evident that homology modelling could not be scientifically justified, the combinatorial approach to protein modelling might be successfully applied to the transmembrane region of the D2 receptor. The probable arrangement of helices in the transmembrane region of GPCRs (Baldwin, 1993) which was based on a careful analysis of a low resolution projection map of rhodopsin (Gebhard et ah, 1993) was used as a guide to model the transmembrane region of D2. The backbone torsion angles of a helix with a middle Pro residue (Sankararamakrishnan et al., 1991) was used to model transmembrane helix V. Dopamine was successfully docked to the putative binding pocket of D2. Using this model as a template, models of D3 and D4 were produced. A separate model of Di was then produced and this in turn was used as a template to model D5.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Adviser: Neil Isaacs
• Keywords: Biochemistry
• Date of Award: 1994
• Depositing User: Enlighten Team
• Unique ID: glathesis:1994-75545
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 19 Nov 2019 19:30
• Last Modified: 19 Nov 2019 19:30
• URI: https://theses.gla.ac.uk/id/eprint/75545