From structure--based to knowledge--based drug design through x-ray protein crystallography: sketching glycogen phosphorylase binding sites

The knowledge derived from the three-dimensional structure of a macromolecular receptor either in the native form or in complex with different ligands has given new insights to the development of improved drug candidates contributing to the drug development pipeline. The structure-based drug design approach has been tested on a number of macromolecular targets implicated in various diseases such as hypertension, glaucoma, HIV and influenza. This approach has also been employed for the development of new antidiabetic agents targeting glycogen phosphorylase (GP), an enzyme that modulates glucose levels in blood circulation. The key role of x-ray protein crystallography in the structure-based inhibitor design process is presented by the case of rabbit muscle GP (RMGPb) that shares increased homology with the liver isoenzyme. The properties of the allosteric binding sites of RMGPb are revealed by filing the interactions formed upon binding of characteristic functional groups and documenting the changes induced in the residues lining the site of interest.     

Myocardial glycogen dynamics: New perspectives on disease mechanisms

Cardiac glycogen regulation involves a complex interplay between multiple signalling pathways, allosteric activation of enzymes, and sequestration for autophagic degradation. Signalling pathways appear to converge on glycogen regulatory enzymes via insulin (glycogen synthase kinase 3β, protein phosphatase 1, allosteric action of glucose‐6‐phosphate), β–adrenergic (phosphorylase kinase protein phosphatase 1 inhibitor), and 5′ adenosine monophosphate‐activated protein kinase (allosteric action of glucose‐6‐phosphate, direct glycogen binding, insulin receptor). While cytosolic glycogen synthesis and breakdown are relatively well understood, recent findings relating to phagic glycogen degradation highlight a new area of investigation in the heart. It has been recently demonstrated that a specific glycophagy pathway is operational in the myocardium. Proteins involved in recruiting glycogen to the forming phagosome have been identified. Starch–binding domain‐containing protein 1 is involved in binding glycogen and mediating membrane anchorage via interaction with a homologue of the phagosomal protein light‐chain 3. Specifically, it has been shown that starch–binding domain‐containing protein 1 and light‐chain 3 have discrete phagosomal immunolocalization patterns in cardiomyocytes, indicating that autophagic trafficking of glycogen and protein cargo in cardiomyocytes can occur via distinct pathways. There is strong evidence from glycogen storage diseases that phagic/lysosomal glycogen breakdown is important for maintaining normal cardiac glycogen levels and does not simply constitute a redundant ‘alternative’ breakdown route for glycogen. Advancing understanding of glycogen handling in the heart is an important priority with relevance not only to genetic glycogen storage diseases but also to cardiac metabolic stress disorders such as diabetes and ischaemia.     

Identification, synthesis, and characterization of new glycogen phosphorylase inhibitors binding to the allosteric AMP site

Inhibition of glycogen phosphorylase (GP) has attracted considerable attention during the last five to 10 years as a means of treating the elevated hepatic glucose production seen in patients with type 2 diabetes. Several different GP inhibitors binding to various binding sites of the GP enzyme have been reported in the literature. In this paper we report on a novel class of compounds that have been identified as potent GP inhibitors. Their synthesis, mode of binding to the allosteric AMP site as well as in vitro data on GP inhibition are shown. The most potent inhibitor was found to be 4-[2,4-bis-(3-nitrobenzoylamino)phenoxy]phthalic acid (4j) with an IC(50) value of 74 nM. This compound together with a closely related analogue was further characterized by enzyme kinetics and in primary rat hepatocytes.     

The SeeDs approach: integrating fragments into drug discovery

Finding novel compounds as starting points for optimization is a major challenge in drug discovery research. Fragment-based methods have emerged in the past ten years as an effective way to sample chemical diversity with a limited number of low molecular weight compounds. The structures of the fragments(s) binding to the protein can then be used to design new compounds with increased affinity, specificity and novelty. This article describes the Vernalis approach to fragment based drug discovery, called SeeDs (Structural exploitation of experimental Drug startpoints). The approach includes the design of a fragment library, identification of fragments that bind competitively to a target by ligand-based NMR techniques and protein crystal structures to characterize binding. Fragments that bind are then evolved to hits, either by growing the fragment or by combining structural features from a number of compounds. The process is illustrated with examples from recent medicinal chemistry programmes to discover compounds against the oncology targets Hsp90 and PDK1. In addition, we summarise our experience with using molecular docking calculations to predict fragment binding and anecdotes on the selectivity and binding modes for fragments seen against a range of targets.     

The effect of isoproterenol and beta-sympathomimetic drugs on human placental glycogen metabolism

Human placental glycogen metabolism is studied, in vitro, by means of three parameters: glycogen level, glycogen phosphorylase and glycogen synthase activities. No significant decrease of glycogen content is observed after incubation of placental fragments with 1 × 10^?4M isoproterenol or salbutamol. In placental homogenates, isoproterenol induces an increase of glycogen phosphorylase a activity and a simultaneous decrease of glycogen synthase a activity. This effect is time and dose-dependent, β sympathomimetic drugs (isoxsuprine, salbutamol) have only a slight action on the activity of both these placental enzymes.     

The role of cyclic AMP-mediated regulation of glycogen metabolism in levamisole-perfused Ascaris suum muscle

The effects of levamisole on muscle contraction and glycogen metabolism have been examined in isolated muscle-cuticle sections of the roundworm Ascaris suum. Muscle contraction occurred when various levels of levamisole were perfused through the preparation. At a levamisole concentration of 0.42 mM, the period of contraction lasted only about 6 min and was followed by a period of relaxation. During this relaxation period, there was an activation of glycogen synthase (EC 2.4.1.11), as evidenced by a decrease in the Ka values of glucose 6-phosphate for glycogen synthase to 0.26 mM from control values of 0.50 mM. The glycogen phosphorylase (EC 2.4.1.1) activity ratio decreased from 0.85 to 0.65, which indicated an inactivation of this enzyme. Concomitant with this activation of glycogen synthase and inactivation of phosphorylase there was an increased synthesis of glycogen. In addition, the presence of levamisole prevented both the serotonin-induced cyclic AMP accumulation and the activation of the cyclic AMP-dependent protein kinase (EC 2.7.1.37). However, levamisole did not significantly affect the changes in glycogen synthase and phosphorylase brought about by perfusion with the neurostimulator acetylcholine. Collectively, the data indicated that levamisole caused a transient muscle contraction followed by muscle relaxation, and the muscle relaxation effect appeared to be the result of a levamisole-inhibited cyclic AMP-mediated pathway of glycogen utilization.     

Synthesis and Evaluation of Novel Oleanolic Acid Derivatives as Potential Antidiabetic Agents

Antidiabetic agents simultaneously inhibiting hepatic glucose production and stimulating hepatic glucose consumption could apply a better control over hyperglycemia. A series of oleanolic acid derivatives with bulky substituents at C‐3 position were designed and synthesized in order to search for this kind of agents. All of the compounds were evaluated biologically in vitro using glycogen phosphorylase and HepG2 cells. The results indicated that several derivatives exhibited moderate‐to‐good inhibitory activities against glycogen phosphorylase. Compound 8g showed the best inhibition with an IC₅₀ value of 5.4 μm . Moreover, most of the derivatives were found to increase the glucose consumption in HepG2 cells in a dose‐dependent manner. The possible binding mode of compound 8g with glycogen phosphorylase was also explored by docking study. 8g was found to have hydrogen bonding interactions with Arg193, Arg310, and Arg60 of the allosteric site.     

High-throughput X-ray crystallography for drug discovery

Knowledge of the three-dimensional structures of protein targets has the potential to greatly accelerate drug discovery, but technical challenges and time constraints have traditionally limited its use to lead optimization. Its application is now being extended beyond structure determination into new approaches for lead discovery. Structure-activity relationships by nuclear magnetic resonance have been widely used to detect ligand binding and to give some indication of the location of the binding site. X-ray crystallography has the advantage of defining ligand-binding sites with greater certainty. High-throughput approaches make this method applicable to screening to identify molecular fragments that bind protein targets, and to defining precisely their binding sites. X-ray crystallography can then be used as a rapid technique to guide the elaboration of the fragments into larger molecular weight compounds that might be useful leads for drug discovery.     

Paracetamol-induced stimulation of glycogenolysis in isolated mouse hepatocytes is not directly associated with cell death

Paracetamol intoxication in vivo is known to be accompanied by depletion of hepatic glycogen stores. We have demonstrated a dose-dependent stimulation of glycogenolysis by paracetamol in glycogen-rich hepatocytes isolated from the mouse. Concentrations of paracetamol that produced plasma membrane damage were also found to activate glycogen phosphorylase a and deplete cellular glycogen contents. However, paracetamol-mediated stimulation of glycogenolysis could be dissociated from the events associated with paracetamol-induced cell killing. Both N-acetylcysteine and 2,4-dichloro-6-phenylphenoxyethylamine markedly reduced the extent of hepatocellular plasma membrane damage induced by paracetamol, yet neither agent prevented the activation of phosphorylase a nor the depletion of glycogen. These findings suggest that the hepatic glycogen depletion that accompanies paracetamol intoxication in vivo is due, at least in part, to a direct effect of the drug on the liver.     

Characterization of the effects of adenosine 5'-[beta-thio]-diphosphate in rat liver.

1. In rat liver cells micromolar concentrations of adenosine 5'-[beta-thio]diphosphate (ADP beta S), activate glycogen phosphorylase by an adenosine 3':5'-cyclic monophosphate (cyclic AMP)- independent mechanism. 2. As with adenosine 5'-triphosphate (ATP), ADP beta S also inhibits the rise in cyclic AMP after glucagon. 3. Cytosolic Ca2+ measured in single cells is rapidly increased with a pattern similar for ADP beta S and for ATP. 4. At variance with ATP, ADP beta S hardly increases inositol 1,4,5-trisphosphate (IP3) levels. 5. Phorbol myristic acetate, which inhibits only slightly the glycogenolytic effect of ATP, almost completely abolishes this effect of ADP beta S. 6. With adenosine 5'-[beta-[35S]thio]diphosphate (ADP beta[35S]) as radioligand, we detected specific purinoceptors on rat liver plasma membranes. Binding consists of a major binding component with KD = 0.7 microM and Bmax = 51 pmol mg-1 of protein, probably mediating the activation of glycogen phosphorylase, and a minor high affinity, low capacity binding component with no obvious function. 7. It is concluded that the differences in biological effects between ATP and ADP beta S may involve different receptors and/or different transduction mechanisms and that ADP beta[35S] can be used to detect the specific binding sites for ADP beta S.     

Glycogen phosphorylase activation by two different alpha 1-adrenergic receptor subtypes: methoxamine selectively stimulates a putative alpha 1-adrenergic receptor subtype (alpha 1a) that couples with Ca2+ influx

We compared the effects of methoxamine on alpha 1-adrenergic receptor-mediated phosphorylase activation in rat hepatocytes and rabbit aorta. Although methoxamine is a potent agonist in activating phosphorylase of rabbit aorta, it had little effect in rat hepatocytes. Using the phenoxybenzamine inactivation method, we found that the quantitative relationship between 125I-BE2254 (125I-BE) binding capacity and maximal norepinephrine-stimulated phosphorylase activation was nonlinear in rabbit aorta, whereas it was linear in rat hepatocytes. The potency of methoxamine in inhibiting specific 125I-BE binding is significantly (p less than 0.05) higher in rabbit aorta (Kd, 96.4 +/- 7.7 microM), compared with rat hepatocytes (Kd, 283 +/- 16 microM). However, these quantitative differences could not fully explain the blunted [Ca2+]c and phosphorylase responses to methoxamine in rat hepatocytes. Treatment with chlorethylclonidine dose dependently suppressed 125I-BE binding sites and norepinephrine-induced phosphorylase activation in rat hepatocytes, whereas in rabbit aorta it resulted in only a 31% decrease in 125I-BE binding sites, with little effect on phosphorylase activation. Furthermore, alpha 1-adrenergic receptor-mediated cellular events of phosphatidylinositol (PI) hydrolysis and phosphorylase activation were unaffected by the removal of extracellular Ca2+ in rat hepatocytes, whereas both responses were markedly attenuated in rabbit aorta. The results indicate that two different alpha 1-adrenergic receptor subtypes activate glycogen phosphorylase, through different mechanisms for increasing [Ca2+]c in the two systems. In rat hepatocytes, alpha 1 receptors are closely linked to PI hydrolysis and Ca2+ release from intracellular stores and cause phosphorylase activation. In rabbit aorta, on the other hand, activation of alpha 1 receptors increases [Ca2+]c by Ca2+ influx from the extracellular fluid as well as by Ca2+ release, and both PI hydrolysis and phosphorylase activation are caused mainly by the Ca2+ entry. Methoxamine interacts with both chlorethylclonidine-sensitive and -insensitive alpha 1 receptor subtypes but selectively stimulates the alpha 1 receptor subtype that closely couples with the Ca2+ influx.     

Advances in glycogen phosphorylase inhibitor design

The regulation of glycogen metabolism is a major therapeutic strategy for blood glucose control in type 2 diabetes. Because glycogen phosphorylase catalyzes the first step in the phosphorolysis of glycogen, it has become a potential key target for controlling hyperglycemia in this disease. This review focuses on advances in new, mostly synthetic, molecules that inhibit glycogen phosphorylase, and describes progress in our understanding of the mechanism of action of these inhibitors gained through X-ray crystallographic studies.     

Theoretical analysis of the binding of salicylate by human serum albumin: The relationship between free and bound drug and therapeutic levels

Summary The binding of salicylate by human serum albumin was analyzed by use of a computer program using previously published association constants and binding capacities for the two sets of binding sites on the protein. The analysis consisted of computing free and bound salicylate for a range of therapeutic and toxic concentrations from 181 to 7246 µmole/L (25 to 1000 mg/L). At low and therapeutic levels the total amount of bound drug would exceed the amount of free drug. At higher levels, which included therapeutic and toxic ranges, the amount of free drug would equal or exceed the amount of bound salicylate. At low levels of drug in the plasma, up to 2000 µmole/L the high affinity sites (Site 1), would bind most of the drug, but as the concentration of drug increased this site would approach saturation and the low affinity Site 2 would bind increasing amounts of salicylate. At high salicylate levels the amount of drug bound by the low affinity sites would exceed the amount bound by the high affinity sites. Computation also showed that when the total amount of protein in the analysis was reduced, from 5,4,3 to 2 gm%, as in hypoalbuminemia, the total amount of drug bound by the protein would decrease and the quantity of free drug would increase. The amount of drug bound by each of the two sets of sites also fell as the concentration of protein decreased. Some of the possible clinical implications of these findings are discussed.     

Plasma/serum protein binding determinations

The binding of a drug to serum or plasma proteins enables the transport of drugs via the blood to sites of action throughout the body. For expediency we will use serum proteins throughout this discussion with the understanding that one can substitute the term plasma proteins in all experimental instances. Only the fraction of drug unbound from serum proteins is available to diffuse from the vascular system and accumulate in tissues thereby enabling interaction with therapeutic targets and accessibility to xenobiotic clearance pathways. Therefore, the extent of drug binding to serum proteins can have a significant impact on pharmacokinetic (PK) parameters such as clearance rates and volume of distribution. In addition, because only the unbound drug is available to interact with therapeutic targets, the extent of serum binding can have significant effects on the pharmacodynamic properties of a compound as well [1, 2] Determining the fraction of drug bound to serum proteins is a standard parameter evaluated in the process of drug discovery. Although the clinical importance of changes in serum protein binding has been questioned [3-8] the need for serum protein binding studies in the discovery and preclinical development stages is essential for the pharmacokinetic modeling of drugs [1, 3, 9]. The extent of serum protein binding is an important parameter used in many in vivo modeling calculations to estimate the volume of distribution, organ clearance, and for scale-up of pharmacokinetic and pharmacodynamic parameters from animal models to humans [3, 10, 11]. The convergence of several trends in the pharmaceutical industry including high speed chemical synthesis technologies, the increasing use of in silico ADME modeling together with early in vivo evaluations of lead compounds has increased the demand for serum protein binding determinations [12].     

Rationally engineered proteins or antibodies with absent or reduced immunogenicity

One challenge associated with the clinical use of protein therapeutics destined for chronic administration is the potential for the development of unwanted anti-drug immune reactions. The molecular basis for this reactivity is the binding of peptide fragments (epitopes) derived from the breakdown of the protein drug to the HLA receptors expressed by the patient's immune cells. If these epitopes are recognized as "foreign" by the immune system, specific helper T lymphocytes (HTL), are activated, which initiate and direct the formation of antibodies against the protein drug. These antibodies can bind and neutralize the protein drug, resulting in either decreased efficacy or total ineffectiveness of the drug. Moreover, various safety concerns, such as allergic reactions and other adverse events, are also frequently associated with the formation of anti-drug antibodies. Herein, we describe the development of "ImmunoStealth", an integrated bioinformatics, biochemical and cellular immunology approach that specifically addresses the issue of unwanted immune responses against protein therapeutics. Unwanted HTL epitopes are identified using in silico sequence analysis methods and high throughput in vitro biochemical evaluations and thereafter confirmed using cellular immunogenicity assays. The "offending" epitopes within the drug are then rationally modified to alter their HLA binding capacity, and thus render them non-recognizable by the immune system. This technology will ultimately facilitate the design of safer, more potent and more economical drugs.     

Metformin and its liver targets in the treatment of type 2 diabetes

Although a number of assessments disagree, the preponderance of the evidence indicates that the major therapeutic action of metformin in type 2 diabetes (DM2) is on the liver, and glucose production (EGP) in particular. At the level of this organ, the actions of metformin can be characterized as pleiotropic. The major questions addressed here are therefore: (i) the methodological aspects of the determination of glucose fluxes: when glucose production is not found to be elevated in type 2 diabetes, it is not surprising that little action of metformin on this flux is found. The issues of populations examined, experimental protocols, and quantitative methods of flux determination are important in answering this question. Early morning EGP is increased and constitutes a valid target for metformin. (ii) the multiple targets of metformin: metformin acts at a number of sites and interacts with metabolites and hormones. Some of these actions may be expressed at different doses. Although their net effect is therapeutic, not all are oriented towards lowering hyperglycemia, perhaps explaining the more modest effect of this drug than could be anticipated from individual actions. Sites of metformin action can therefore be considered as a compilation of valid therapeutic targets in DM2. Gluconeogenesis, glycogenolysis and glycogen synthesis can be altered by metformin, although in vivo, this also depends on the methodology. Component processes from substrate supply and liver uptake, through a number of glucogenic enzymes, as well as glycogen synthase and phosphorylase have all been shown to be affected. (iii) unifying concepts: reported actions of metformin on the mitochondrial respiratory chain, free fatty acid metabolism, AMP-activated protein kinase, and on membrane proteins directly may all explain subsets of actions that are seen, providing more integrated targets for consideration in the therapy of DM2.     

Effect of cannabis extract on uterine glycogen metabolism in prepubertal rats under normal and estradiol-treated conditions.

Repeated administration of cannabis extract on two uterine glycogen metabolising enzymes, glycogen phosphorylase and glycogen synthetase in prepubertal rats, treated with or without estradiol benzoate reduces the glycogen content of uterus by increasing phosphorylase activity (both total and form a) and by decreasing glycogen synthetase activity. In estradiol treated rats, however, cannabis extract has been found to inhibit the estradiol-induced rise in type a phosphorylase activity, glycogen synthetase activity being inhibited as in the previous case. Hence reduction of glycogen content in uterus by cannabis extract in estradiol-treated rats appears to be primarily due to decreased synthesis. The results indicate the antiestrogenic effect of this drug at the level of uterine glycogen metabolizing enzymes.     

In vitro characterization of binding and stability of single-chain Fv Ni-NTA-liposomes

Recently, we presented a new method for the generation of single-chain Fv (scFv) immunoliposomes, which circumvents the necessity to introduce additional reactive groups in the protein. This method is based on immobilizing scFv fragments via their C-terminal hexahistidyl-tag on liposomes containing nickel-complexed dioleoyl-glycero-succinyl-nitrilotriacetic acid (Ni-NTA-DOGS) as an anchor lipid within the lipid bilayer. Here, we have extended this approach to various other scFv fragments and further demonstrate strong and selective binding of these liposomes to target cells in vitro. In order to evaluate suitability for in vivo applications, we investigated the influence of human plasma on stability and binding behaviour of scFv Ni-NTA-liposomes in vitro using scFv A5 directed against human endoglin (CD105) as a model antibody. We could show that the binding activity to target cells is rapidly lost in the presence of human plasma. Incorporation of polyethylene glycol (PEG) chains into the lipid bilayer did not protect against loss of binding capability. Further studies showed that loss of binding is mainly due to displacement of Ni-NTA-bound scFv fragments caused by plasma proteins. In conclusion, the system allows for a rapid and flexible generation of target cell specific immunoliposomes for in vitro applications but lacks stability for in vivo applications.     

Sensitivity of glycogen phosphorylase isoforms to indole site inhibitors is markedly dependent on the activation state of the enzyme

BACKGROUND AND PURPOSE: Inhibition of hepatic glycogen phosphorylase is a potential treatment for glycaemic control in type 2 diabetes. Selective inhibition of the liver phosphorylase isoform could minimize adverse effects in other tissues. We investigated the potential selectivity of two indole site phosphorylase inhibitors, GPi688 and GPi819. EXPERIMENTAL APPROACH: The activity of glycogen phosphorylase was modulated using the allosteric effectors glucose or caffeine to promote the less active T state, and AMP to promote the more active R state. In vitro potency of indole site inhibitors against liver and muscle glycogen phosphorylase a was examined at different effector concentrations using purified recombinant enzymes. The potency of GPi819 was compared with its in vivo efficacy at raising glycogen concentrations in liver and muscle of Zucker (fa/fa) rats. KEY RESULTS: In vitro potency of indole site inhibitors depended upon the activity state of phosphorylase a. Both inhibitors showed selectivity for liver phosphorylase a when the isoform specific activities were equal. After 5 days dosing of GPi819 (37.5 micromol kg(-1)), where free compound levels in plasma and tissue were at steady state, glycogen elevation was 1.5-fold greater in soleus muscle than in liver (P < 0.05). CONCLUSIONS AND IMPLICATIONS: The in vivo selectivity of GPi819 did not match that seen in vitro when the specific activities of phosphorylase a isoforms are equal. This suggests T state promoters may be important physiological regulators in skeletal muscle. The greater efficacy of indole site inhibitors in skeletal muscle has implications for the overall safety profile of such drugs.     

Pharmacological regulation of hepatic glucose production

The discovery of antidiabetic agents that inhibit hepatic glucose production is a popular and potentially fruitful research area for the pharmaceutical research community. Metformin, a marketed agent with this mechanism of action, is widely used for the treatment of type 2 diabetes, however, more efficacious agents are sought. A number of promising proteins are being targeted for modulation by new compounds, including the glucagon receptor, glycogen phosphorylase, glucocorticoid receptor, 11 beta-hydroxysteroid dehydrogenase-1, fructose-1,6-bisphosphatase, carnitine palmitoyltransferase-1, glycogen synthase kinase-3, glucose-6-phosphate T1 translocase and the A2B receptor. Compounds designed to work against these targets are at the early clinical or preclinical phase of study. Glucagon receptor antagonists, glycogen phosphorylase inhibitors, 11 beta-hydroxysteroid dehydrogenase-1 inhibitors, carnitine palmitoyltransferase-1 inhibitors and fructose-1,6-bisphosphatase inhibitors are, or have been, clinically evaluated. Preclinical studies against the other targets have yielded compounds that demonstrate efficacy in diabetic animal models and clinical activity will continue.