The kynurenine pathway: kinetics of a dioxygenase and inhibitors of a monooxygenase

Greater than 95% of circulating L-tryptophan is metabolised through the kynurenine pathway in mammals, with the remainder directed into protein synthesis and serotonin, melatonin, and tryptamine production. The pathway is initiated by the oxidative ring opening of L-tryptophan by one of three heme-based dioxygenases: the paralogues indoleamine 2,3-dioxygenase 1 or 2(IDO1, IDO2), or the otherwise unrelated tryptophan dioxygenase (TDO). The role of tryptophan and its metabolites, and of the kynurenine pathway in particular, is a widespread topic of research. IDO1, IDO2, and TDO, as the initiating enzymes of the kynurenine pathway, are drug targets of great interest due to their implication in a range of pathogenic states; either directly or through the action of the wider kynurenine pathway. The first stable product of the kynurenine pathway, kynurenine, represents a branch point as it acts as the substrate for three different enzymes (kynurenine monooxygenase (KMO), kynureninase, and kynurenine amino transferases (KATs). As such, these enzymes have received attention as candidate drug targets in order to modulate the relative flux through these branches. This study relates to the first two drug targetsof the kynurenine pathway: IDO1 and KMO. Chapter 1 provides a brief introduction to the biochemistry of the kynurenine pathway, and evidence for its role as a target for neurodegenerative and cancer related diseases. Particular focus has been given to the molecular mechanisms and evidence for IDO1 and KMO as pharmacologically relevant drug targets. Chapter 2 reviews the current knowledge surrounding KMO's structure, kinetics and mechanism, and inhibition structure-activity relationship (SAR). From this was generated a pharmacophore based on the common feature alignment of active KMO inhibitors (Chapter 3). The pharmacophore showed excellent shape complementarity to the active site features of KMO and was able to moderately describe the relationship between the structure of KMO inhibitors and inhibitory potency (R2 0.64, n = 39). New KMO lead molecules were screened in silico using the pharmacophore and docking, and 41 compounds were tested in vitro and in vivo. This yielded six molecules with an in vitro IC50 below 100 μM and one compound with an in vivo EC50 below 20μM. A study was also initiated in regards to the structure and function of IDO1, with emphasis on the possible roles that a large 20 amino acid loop might play in the binding of small molecules and protein partners. Sequence alignment also indicated that walrus IDO1 naturally possessed a truncated loop sequence and would be of interest. Mutant constructs of human and walrus IDO1s were designed, along with those of IDO1-interacting proteins cytochrome b5 and cytochrome P450:NAD(P)H reductase, as detailed in Chapter 4. The expression and characterisation of these proteins is presented in Chapter 5, including investigations into the impact of bacterial culturing conditions. A comparison was made between different growth temperatures and different induction procedures. A method utilising weaker induction through a smaller concentration of the molecule IPTG, combined with inducing at a later phase of the culture and with reduced agitation, gave the best results as judged by size exclusion chromatography. Whilst these proteins generally gave expected results in regards to the UV-vis profiles, wild-type walrus IDO1 gave a double Soret peak in multiple instances, compared to the single Soret peak of other proteins. The Q-band region of this walrus IDO1 also behaved in an inconsistent manner compared to other heme-containing proteins when exposed to redox reagents. This led to the hypothesis that walrus wild-type IDO1 possessed batch-specific variation in the isolated protein leading to some molecules having a heme environment protected from the actions of the bulk solvent. Chapter 6 provides the kinetic and mechanistic background of IDO1, as an introduction to analysis of the IDO1 under different experimental conditions. The kinetics of human and walrus IDO1, and of a loop deletion mutant of human IDO1, were analysed in the presence of different redox cycling systems. Wild-type human IDO1, the human deletion mutant, and wild-type walrus IDO1 generally behaved identically under the conditions examined. A surprising exception was that wild-type walrus IDO1 showed hyperbolic kinetics with respect to L-tryptophan when IDO1 was reduced by cytochrome b5, whilst substrate inhibition was observed for all other circumstances. The possible implications of kinetic results are further discussed in Chapter 6. Chapter 7 describes work on collaborative projects relevant to the kynurenine pathway. A carborane cage is a molecular structure made of carbons and boron in a spherical arrangement. Their use in medicinal chemistry is relatively unstudied compared to more common hydrocarbon structures such as adamantane or benzene. Collaborators synthesised a series of known IDO1inhibitors with carborane cages in place of benzyl rings, and tested them against recombinant human IDO1. Molecular mechanics simulations were then run as part of this PhD project to visualise the differences between the benzyl rings and the carborane cages. The cages were generally well tolerated despite the apparent loss of a favourable pi-cation interaction between the benzyl analogues and an arginine at the active site entrance. Simulations indicated that this arginine performs a gating function with these types of inhibitors, as has previously been the case indicated for the substrate L-tryptophan. Also looked at was the possible interaction between quinolinic acid, a downstream metabolite of the kynurenine pathway, and the excitory amino acid transporter 3 (EAAT3). Comparison of a structure with glutamic acid with those created through docking simulations with quinolinic acid led to the hypothesis that quinolinic acid could competitively bind in the same location as glutamic acid. This would indicate a new mechanism of quinolinic acid transport across neuronal cells, as well as a new mechanism of action for the control of neurological function via competition between two excitory amino acids. The thesis ends with Chapter 8, which summarises briefly the results obtained herein, and provides suggestions for future work of interest based on these observations.