<p>Macrolides are a large group of clinically relevant antibiotics that inhibit protein synthesis by binding to the large ribosomal subunit in the peptide exit tunnel, close to the peptidyl transferase center (PTC). We have shown that the peptide length of the resulting peptidyl-tRNA drop-off products is proportional to the distance between the PTC and the respective macrolide in the tunnel. This indicates that macrolides act by sterically blocking the nascent peptide exit path.</p><p>A substantial amount of read-through into full-length product was observed for some macrolides and depends on the relation between the dissociation rate constants for peptidyl-tRNA and the macrolide, respectively. The dissociation rate constant for josamycin is 60 times lower than the dissociation rate constant for erythromycin, which explains why no read-through is seen for josamycin in contrast to erythromycin.</p><p>Macrolides do not compete with binding of ternary complexes, hence they are non-competitive inhibitors. However, the text-book description is not valid for macrolide antibiotics, and we show that this is due to the equilibrium assumption generally used to describe non-competitive inhibitors. Our results suggest that a more thorough mechanistic investigation is required to classify inhibitors than what has been proposed previously.</p><p>Further, we have examined the phenomenon of peptide mediated resistance to macrolides. Our results show that expression of a resistance peptide increases the dissociation rate constant for erythromycin.</p><p>In addition, we have examined the accuracy of protein synthesis on three different levels: (<i>i</i>) How do the three initiation factors accomplish fast and accurate initiation of protein synthesis, (<i>ii</i>) how does proof-reading work on the isoleucyl-tRNA synthetase, and (<i>iii</i>) what is the accuracy in the tRNA selection and how is it accomplished? Our data propose a change of the view on all these mechanisms.</p><p>In conclusion this thesis presents new results on protein synthesis, macrolide antibiotics and macrolide resistance.</p>

Macrolides are a large group of clinically relevant antibiotics that inhibit protein synthesis by binding to the large ribosomal subunit in the peptide exit tunnel, close to the peptidyl transferase center (PTC). We have shown that the peptide length of the resulting peptidyl-tRNA drop-off products is proportional to the distance between the PTC and the respective macrolide in the tunnel. This indicates that macrolides act by sterically blocking the nascent peptide exit path. A substantial amount of read-through into full-length product was observed for some macrolides and depends on the relation between the dissociation rate constants for peptidyl-tRNA and the macrolide, respectively. The dissociation rate constant for josamycin is 60 times lower than the dissociation rate constant for erythromycin, which explains why no read-through is seen for josamycin in contrast to erythromycin. Macrolides do not compete with binding of ternary complexes, hence they are non-competitive inhibitors. However, the text-book description is not valid for macrolide antibiotics, and we show that this is due to the equilibrium assumption generally used to describe non-competitive inhibitors. Our results suggest that a more thorough mechanistic investigation is required to classify inhibitors than what has been proposed previously. Further, we have examined the phenomenon of peptide mediated resistance to macrolides. Our results show that expression of a resistance peptide increases the dissociation rate constant for erythromycin. In addition, we have examined the accuracy of protein synthesis on three different levels: (i) How do the three initiation factors accomplish fast and accurate initiation of protein synthesis, (ii) how does proof-reading work on the isoleucyl-tRNA synthetase, and (iii) what is the accuracy in the tRNA selection and how is it accomplished? Our data propose a change of the view on all these mechanisms. In conclusion this thesis presents new results on protein synthesis, macrolide antibiotics and macrolide resistance.

<p>CYP2C9 is one of our major drug metabolising enzymes and belongs to the cytochrome P450 (CYP) super family. The aim of this thesis was to gain an understanding of the quantitative structure–activity relationships (QSAR) of CYP2C9 substrates and inhibitors. This information will be useful in predicting drug metabolism and the potential for drug–drug interactions. To achieve this, a well characterised data set of structurally diverse, competitive CYP2C9 inhibitors was identified in our laboratory. Several computational methodologies, many based on GRID molecular interaction fields, were applied or developed in order to handle issues such as compound alignment and bioactive conformer selection. First, a traditional 3D QSAR was carried out in GOLPE, generating a predictive model. In this model the selection of a bioactive conformer and alignment was based on docking in a homology model of CYP2C9. Secondly, we introduced the concept of alignment independent descriptors from ALMOND. These descriptors were used to generate quantitatively and qualitatively predictive models. We subsequently derived conformation independent descriptors from molecular interaction fields calculated in FlexGRID. This enabled the derivation of 3D QSAR models without taking into account the selection of an alignment or a bioactive conformer. A subsequent programming effort enabled the conversion of this model back to 3D aligned pharmacophores. Similar alignment independent descriptors were also used in the development of the software MetaSite® that predicts the site of metabolism for CYP2C9 ligands. Finally, as crystal information on this isoform emerged, the performance of molecular dynamics simulations and homology models and the flexibility of the protein were evaluated using statistical analyses.</p><p>These modelling efforts have resulted in detailed knowledge of the structural characteristics in ligand interactions with the cytochrome P450 2C9 isoform.</p>

CYP2C9 is one of our major drug metabolising enzymes and belongs to the cytochrome P450 (CYP) super family. The aim of this thesis was to gain an understanding of the quantitative structure–activity relationships (QSAR) of CYP2C9 substrates and inhibitors. This information will be useful in predicting drug metabolism and the potential for drug–drug interactions. To achieve this, a well characterised data set of structurally diverse, competitive CYP2C9 inhibitors was identified in our laboratory. Several computational methodologies, many based on GRID molecular interaction fields, were applied or developed in order to handle issues such as compound alignment and bioactive conformer selection. First, a traditional 3D QSAR was carried out in GOLPE, generating a predictive model. In this model the selection of a bioactive conformer and alignment was based on docking in a homology model of CYP2C9. Secondly, we introduced the concept of alignment independent descriptors from ALMOND. These descriptors were used to generate quantitatively and qualitatively predictive models. We subsequently derived conformation independent descriptors from molecular interaction fields calculated in FlexGRID. This enabled the derivation of 3D QSAR models without taking into account the selection of an alignment or a bioactive conformer. A subsequent programming effort enabled the conversion of this model back to 3D aligned pharmacophores. Similar alignment independent descriptors were also used in the development of the software MetaSite® that predicts the site of metabolism for CYP2C9 ligands. Finally, as crystal information on this isoform emerged, the performance of molecular dynamics simulations and homology models and the flexibility of the protein were evaluated using statistical analyses. These modelling efforts have resulted in detailed knowledge of the structural characteristics in ligand interactions with the cytochrome P450 2C9 isoform.

Alzheimer’s disease (AD) is a chronic, progressive neurodegenerative disorder, characterized clinically by a progressive loss of memory, cognitive function, ability to care for oneself and psychiatric symptoms. First-line agents for the treatment of AD are ChE inhibitors (DON, GAL), whose modest clinical efficacy and the high incidence of dose-limiting toxicities limit their clinical utility. In addition to AD, ChE inhibitors (PYR) are used for other medical conditions, such as myasthenia gravis (MG). Furthermore, ChE inhibitors (PYR) are used by military personnel prophylactically if impending exposure to chemical warfare agents, e.g., soman, is suspected. The purpose of this research project was to understand the effect of age on the in-vitro sensitivity of ChE inhibitors in human RBCs and plasma. Understanding possible covariates, such as age and gender, may assist in optimizing dosing regimens of ChE inhibitors and/or developing newer ChE inhibitors with better adverse effect profiles. Plasma PHYS concentrations were measured by a validated HPLC-FD method. RBC AChE activity and plasma BuChE activity were measured by a modified Ellman’s colorimetric method using the model substrates, acetylthiocholine and butyrylthiocholine, respectively. The kinetics of RBC and plasma ChE activity followed Michaelis-Menten kinetics. Acetylthiocholine was found to be a nonselective substrate (RBC AChE Km = 73 μM; plasma BuChE Km = 117 μM); while butyrylthiocholine was a selective substrate for plasma BuChE (RBC AChE Km = 130,000 μM; plasma BuChE Km = 72 μM). For the following studies, RBC AChE activity was measured using acetylthiocholine as the substrate and plasma BuChE activity was measured using butyrylthiocholine as the substrate. This research project was performed in two parts: First, mechanistic studies of PHYS, PYR, DON and GAL, explored and determined the mechanism of in-vitro inhibition of RBC AChE and plasma BuChE inhibition, as well as the in-vitro degradation of PHYS in human whole blood, plasma and RBC. PHYS was rapidly degraded in human whole blood, RBC and plasma and followed Michaelis-Menten kinetics but its degradation clearance - scaled to whole blood clearance - was only predicted to account for 4-6% (i.e., 195-261 ml/min) of the reported total body clearance for PHYS (4500 ml/min). RBCs were responsible for 60% of the whole blood clearance while plasma accounted for 40% of the whole blood clearance. Inhibition results indicated that both PHYS and PYR were nonselective and rapid suicide ChE inactivators. PYR inactivated RBC AChE more rapidly at low concentrations and inactivated plasma BuChE more rapidly at high concentrations, but inactivated both more rapidly than PHYS. PHYS was a more potent inactivator than PYR with a Ki for RBC AChE of 0.011 μM and 0.063 μM, respectively, and 0.023 μM and 0.036 μM, respectively for plasma BuChE. DON was found to be a noncompetitive inhibitor for RBC AChE (Ki,noncomp = 114 μM), but a competitive inhibitor for plasma BuChE (Ki,comp = 213 μM). GAL was found to be a competitive inhibitor for both RBC AChE (Ki,comp = 66 μM) and plasma BuChE (Ki,comp = 358 μM). The second part involved a clinical study with ten young and nine elderly healthy subjects, balanced for gender, who donated blood for an in-vitro study in order to assess any age- and gender-related differences in in-vitro sensitivity to RBC AChE and plasma BuChE inhibition to all four ChE inhibitors. Elderly adults were found to be 2-3-fold less sensitive compared to the young adults for PHYS (BuChE Ki,pss; 0.010 and 0.015 μM, young and elderly, respectively) and PYR (AChE Ki,pss; 0.12 and 0.25 μM, young and elderly, respectively) only, while neither DON nor GAL showed any age-related differences in sensitivity. The observed differences for PHYS and PYR may be due to kinetic differences in ChE inactivation between young and aged adults, rather then a difference in binding affinities/potencies. These carbamate ChE inhibitors, presumably, have a slower decarbamoylation rate in younger adults than elderly adults, which leads to the observed difference in in-vitro sensitivity. The above in-vitro results were consistent with results of a meta-analysis: In a study by Knapp et al. (1991), young males (n=6), receiving 18 mg, 24 mg and 30 mg PHYS tablets, showed similar ex-vivo plasma BuChE sensitivity to (28 %/(ng/ml)) as the in-vitro sensitivity for young males in the current study (33 %/(ng/ml)). On the other hand, in the study by Men (2004), elderly males (n=8) and females (n=8), receiving 6.7 μg/kg PHYS as 30-minute infusion, showed similar ex-vivo RBC AChE sensitivity (12 %/(ng/ml)) as the in-vitro sensitivity for elderly subjects in the current study (9.7 %/(ng/ml)). This suggests that in-vitro measurement of ChE sensitivity is predictive of ex-vivo sensitivity in clinical studies. The study results suggest that elderly adults may require a 2-3-fold higher blood concentration than young adults to achieve the same ChE inhibition. This may explain why for epistigmine, an investigational carbamate ChE inhibitor for the treatment of AD, the maximum tolerated dose observed in young adults (40 mg single dose) was lower than for older adults (90 mg/day). Higher sensitivity in young adults prevented further dose escalation, while all elderly subjects tolerated higher doses. This research may have implications for other diseases and conditions, most notably MG and as a prophylaxis of nerve gases poisoning. As patients with MG age, they may become less sensitive to PYR, the most common symptomatic treatment for MG, and an increase in dose may be required. Further, older military personnel assigned to receive PYR, may require increased doses to achieve the targeted 10% RBC AChE inhibition, necessary to protect against nerve gas poisoning.