Structure and conformational behaviour of peptoid peptidomimetics . Towards an understanding of the mechanism of action of substance P and NK1 antagonists

Abstract

Developing peptoid peptidomimetics is part of a strategy to come to a more rational development of ligands for therapeutically interesting macromolecular targets for which peptide-protein or protein-protein interactions play a role. While maintaining the functionalities and relative side-chain positions of the parent peptide, peptidomimetics are likely to have improved pharmacokinetics in comparison with peptides. In addition, the modular build-up of peptoids (oligomers of N-substituted glycine residues) facilitates the synthesis of compounds targeted at a plethora of biological systems. To validly implement this strategy for many potentially interesting targets, insight in the structure and conformational behaviour of peptoids, which forms the topic of this thesis, is of importance.

Here, the system of interest is the tachykinergic system with the undecapeptide substance P as its most prominent member. Substance P is a neuropeptide that is released from sensory nerve endings throughout the body and is involved in e.g. neurogenic inflammation and in the transmission of pain. As such, substance P plays a role in allergic airway diseases, such as allergic asthma, and in chronic pain. Antagonists of substance P can therefore perhaps be used as analgesics or anti-inflammatory compounds. Recently, peptoid analogues of substance P have been found to elicit full agonist activity at the (murine) NK1 receptor, which is the preferred receptor for substance P. The explanation for their activity is, however, still based on the topological similarity between the structural formula of the peptoids and the parent peptide. To gain insight in their molecular mechanism of action, the structure and conformational behaviour of peptoid peptidomimetics is studied in this thesis. Furthermore, intermolecular interactions in the crystal structures of non-peptide antagonists for the human NK1 receptor are described and serve as a basis for evaluating a method which can be employed in the study of mimicry.

In the absence of an experimentally determined NK1 receptor structure, or antagonist-receptor complex, the interaction geometries between antagonists and amino-acid residues that are important in binding are yet unknown. In Chapter 2, crystal structures of nine non-peptide NK1 antagonists are analyzed for the intermolecular interactions of their pharmacophoric groups with neighbouring molecules in the crystals. Based on these analyses we identified several N+aromatic and aromatic-aromatic interaction geometries for the positively charged quinuclidine rings and the pharmacophoric aromatic rings, respectively, which can explain the importance of aromatic amino-acid residues in the human NK1 antagonist binding-site. In addition, an interaction site for Gln-165 in the human NK1 receptor is explicitly proposed. This amino-acid residue has been reported to be involved in a hydrogen bond with the benzylamino nitrogen or benzylether oxygen of piperidine and quinuclidine based antagonists. A superposition of the crystal structure conformations of two prototypic NK1 antagonists, CP-96,345 and CP-99,994, based on pharmacophoric elements, revealed similarities in intermolecular interaction geometries in the crystal environment, that might be used as a starting point to explain the similar binding-sites at the human NK1 receptor that have been indicated for these two compounds.

The number and types of interactions encountered in crystal structures of low-molecular-weight compounds are severely restricted by the limited functionality present in the crystal, especially when compared to the functionality present in protein binding-sites. Also, although the crystal structure conformation of a compound is an energetically accessible conformation, it is in many cases not likely to be the bioactive one. To avoid these drawbacks the applicability of an empirical method developed to calculate favourable interaction sites and interaction geometries, irrespective of the molecular conformation or the nature of the interacting groups, was evaluated for a selection of non-peptide NK1 antagonists. This is reported in Chapter 3. Experimentally determined interaction sites for hydrogen bond acceptor groups and N+aromatic and aromatic-aromatic interaction geometries (as reported in Chapter 2) could be reproduced remarkably well. This method would facilitate the development of a pharmacophore based on interaction sites, and the construction of a model for the NK1 antagonist binding-site. Application of this method would also be highly valuable in studying mimicry of peptoids and peptides.

Experiences with crystallization attempts of substance P and its (retro)peptoid peptidomimetic are described in Chapter 4. Crystal structures of substance P and its retropeptoid would contribute to the fundamental understanding of their energetically accessible conformations and their intermolecular interactions in a highly structured molecular environment. There are, however, no precedents with respect to successful crystallizations of linear and otherwise unrestrained peptide oligomers of a comparable size. The crystallization strategy employed was based on the use of macromolecular crystallization screens in combination with a hanging drop-vapour diffusion set-up. For both substance P and its retropeptoid no initial crystallization conditions were identified. Crystallization of substance P was prevented by the formation of phase separations and amorphous precipitates, which may be related to the amphiphilic nature of substance P.

The consequences of the presence of tertiary amide bonds for the backbone geometry of peptoids, with emphasis on the geometry of the amide bonds itself, are described in Chapter 5. In in vacuo quantum chemical calculations tertiary amide bonds were found to be intrinsically non-planar. In the solid state, on the other hand, a preferred planar geometry was observed that still allows for considerable pyramidalization at the amide nitrogen in individual cases. In ligands in protein binding-sites a similar degree of pyramidalization at the tertiary amide nitrogen was observed. In trans amide bonds the monomer side chain is mainly responsible for a non-planar amide bond, whereas in cis amide bonds the peptoid backbone takes over this role. Pyramidalization at the tertiary amide nitrogen seems to be an additional feature for peptoids in accommodating their conformation to the requirements of a protein binding-site.

Chapter 6 turns to the relatively flexible parts of the peptoid backbone. The energetically accessible conformations for a fragment characteristic of the conformationally flexible methylene bridge connecting the relatively rigid amide bonds in a peptoid backbone were evaluated in small-molecule crystal structures and in ligands in crystallographically determined protein complexes. The methylene group was found to act as a ball-joint at which the amide bonds prefer a mutually perpendicular orientation. The most prominently present peptoid monomer conformation is shown to be compatible with the formation of a periodic structure.

Conformational analysis of tri- and hexapeptoid analogues derived from the C-terminal sequence of substance P, as described in Chapter 7, reveals that a mutually perpendicular orientation of adjacent amide bonds is also a characteristic of peptoid monomers conformations when they are part of a peptoid chain. The conformational behaviour of the peptoid backbone can be understood in terms of low-energy conformations of the model peptoid monomers Ac-NAla-NMe2, Ac-NAla-NHMe and Ac-Gly-NMe2. These model compounds represent in corresponding order (i) any peptoid monomer that is part of the peptoid chain by means of tertiary amide bonds on either end, (ii) a peptoid monomer that precedes a glycine residue in the chain and (iii) a glycine residue. In addition, the relative directions of side chains (when viewed from the position of the backbone) in prominent energetically accessible peptoid conformations and a number of secondary structure elements in peptides were found to correspond. Similar observations were made for minimum energy conformations of tripeptide and tri-retropeptoids derived from substance P. Mimicry between peptides and peptoids is found at this level, but is not sufficient to explain the biological activity of the substance P peptoids or any other peptoid. Though, it is a starting point for describing mimicry of other relevant molecular properties in energetically accessible peptoid and peptide conformations as well.

The discussion in Chapter 8 focuses on the insight gained in the structure and conformational behaviour of peptoid peptidomimetics as a result of the work described in this thesis. In summary, some idea about the characteristics and the sequence dependence of the backbone conformational behaviour of peptoids has emerged. These features also seem to describe the behaviour of peptoids in protein binding-sites. Molecular modelling studies aimed at directing synthetic efforts and to interpret pharmacological structure-activity studies with respect to the substance P peptoids and other peptoid peptidomimetics can be aided and validated by the methods and results presented in this thesis. It may form the start of a more rational development of peptidomimetic ligands for therapeutically interesting targets. Final validation can of course only be obtained by experiment.