Approaches to study protein interactions

Abstract

This thesis presents studies on series of modified peptides to study protein-protein interactions. Specifically, the use of new fluorescent peptide modifications to influence secondary structure and secondly, development of a peptidomimetic scaffold to target Proliferating Cell Nuclear Antigen (PCNA). Chapter 1 introduces the importance of protein-protein interactions, their role in disease, and the difficulty in targeting these large-surface-area interactions. Peptides present as an ideal compromise between the large size of proteins and the drug-like properties of small molecules, to study such interactions. The importance of secondary structure and methods to stabilise a conformation favourable to binding specific proteins are outlined in order to address limitations associated with the use of peptides as therapeutics. Methods used to prepare such peptides, and determine the resulting conformations are outlined, along descriptions of techniques to characterise biological activity of peptides and their interaction with proteins. Chapter 2 details studies on the synthesis of peptide macrocycles that are constrained by a bimane group that covalently links two cysteine residues. The methodology allows preparation of nine macrocycles that range in size from 16 to 31 atoms. These peptides are cell permeable, where blue fluorescence corresponding to the bimane is present in the cell cytosol. CD and NMR secondary shift analysis revealed that the i-i+4 bimane-constrained pentapeptide is α-helical. Chapter 3 extends the investigation presented in Chapter 2 and demonstrates the i-i+4 bimane constraint introduces 310-helical structure into a 12 amino-acid sequence known to target Estrogen Receptor alpha (ERα). This same peptide adopts an α-helical geometry in the presence of 20% TFE, and when the peptide is bound to ERα as shown by computational modelling. This demonstrates sufficient flexibility in the bimane-constrained macrocycles to adopt α-helical conformation. Interestingly, helical structure is also adopted in acyclic peptides, where a bimane-modified cysteine is six amino-acids away from a tryptophan or phenylalanine residue. The fluorescence properties of the bimane-modified peptides are also presented. The methodology to introduce the bimane constraint into peptides and define helical structure is summarised in a mini-review-style Focus article in Appendix 1. Chapter 4 reviews the structure-activity relationship of peptides and molecules that bind PCNA, and summarises the current knowledge of features that allow specific and high affinity interaction with PCNA. Chapter 5 presents a series of 51 modified p21-derived peptides (141-155) with sequence modifications to the PCNA-interacting motif (known as the PIP-box). SPR analysis identified seven peptides that bind PCNA with higher affinity than the native p21 sequence (12.3 nM). The PCNA-binding affinity was correlated to the binding conformation of these peptides bound to PCNA, as studied by X-ray crystallography and computational modelling, and highlights a series of important hydrogen bonding networks that modulate PCNA binding affinity. Collectively, these data elucidate the rational design of a new PCNA-binding peptide that binds with the highest affinity reported to date (1.21 nM). A series of five macrocyclic p21 peptides is presented in Chapter 6, where a range of covalent constraints were installed by dithiol bis-alkylation in a p21 peptide (143-154) containing two cysteines at positions 146 and 149. The binding affinity of the resulting i-i+4 constrained macrocycles for PCNA was determined by SPR to reveal the bimane-constrained p21 peptide as the most potent at 570 nM. The secondary structure adopted by the macrocycles bound to PCNA was studied by X-ray crystallography and computational modelling, and indicated that the bimane-constrained peptides are the only macrocycle to adopt the key 310-helical binding conformation. Additionally, the bimane peptide was cell permeable, in comparison to a linear fluorescein-tagged p21 peptide of the same length, where confocal microscopy revealed blue fluorescence corresponding to the bimane in the cell cytosol. Chapter 7 examines the role of the peptide sequence flanking PCNA-binding motif of the p21 peptide, and reveals that a short p21 peptide of 12 amino-acids (143-154) can bind PCNA with 102 nM affinity. These short p21 peptides, however, are not cell permeable and a series of nuclear locating sequence (NLS)-tags were appended to a p21 peptide in an endeavour to improve cell and nuclear uptake, but were unsuccessful. Albeit, nuclear permeability was achieved when the NLS-tags were instead appended to the macrocyclic p21 bimane-constrained peptide presented in Chapter 6. Only the bimane-constrained p21 macrocycle appended to a SV40 NLS-tag via a thiol-maleimide linkage was nuclear permeable. Interestingly, nuclear entry was only permitted when a fluorescein was coupled to the N-terminus of the SV40-tag. Chapter 8 explores the use of solvatochromic amino-acids in a p21 sequence (141-155), at the positions known to insert into the hydrophobic cleft on the PCNA surface (147, 150 and 151), to determine whether a peptide could be utilised as a selective turn-on fluorescent PCNA sensor. Two different solvatochromic amino-acids were utilised, 4-DMNA and 4-DMAP, and the fluorescence properties and affinity of the six resulting peptides for PCNA was characterised. This revealed that only the peptides with 4-DMNA or 4-DMAP at position 150 maintained high affinity binding to PCNA. A 10-fold increase in fluorescence intensity, relative to the peptide alone, was achieved in the presence of 2.5 equivalents of PCNA for the 151-substituted 4-DMNA peptide. The 151-substituted 4-DMAP peptide only produced a 3.5-fold change in fluorescence under the same conditions. Chapter 9 provides an overall summary of this thesis, including the methodology to introduce a bimane modification into short peptides and the optimal configurations to introduce helical structure. It also details key mutations presented in Chapter 5 to improve binding affinity of peptides for PCNA, along with the importance of stabilising a 310-helical binding conformation (Chapter 6). Chapter 9 presents a new series of nine i-i+4 constrained macrocyclic peptidomimetics that include the high affinity sequence modifications, in order to consolidate these distinct achievements. The binding affinity of these macrocycles for PCNA, and corresponding linear analogues, was determined by SPR and revealed that in all except for one case, the binding affinity of the macrocycle was greater than the linear analogue. The position of a polar group (e.g. amide or acyl thioether) in asymmetric constraints alters PCNA binding affinity, where higher affinity for PCNA was observed when the polar group is nearer the C-terminal end of the constraint. The most potent peptide of this series is a Lys/Glu lactam macrocycle which binds PCNA with 8.16 nM affinity, the highest affinity peptidomimetic that binds PCNA to date. Finally, this chapter outlines future directions to build on this work and develop a peptidomimetic that targets PCNA for application as a cancer therapeutic.