Research

Chemical Biology and Chemical Genetics with Organometallic Compounds

Introduction

Altering cellular functions with small synthetic molecules is a general approach for the design of drugs (medicinal chemistry) and molecular probes (chemical biology, chemical genetics). Medicinal chemistry and chemical biology are focused predominately on the design and utilization of organic molecules. These are usually designed to bind to a biomolecular target in a reversible fashion by a combination of weak interactions, such as hydrogen bonding, electrostatic attraction, polar interactions, and hydrophobic contacts (Figure 1a).

In contrast, inorganic compounds find applications mainly for their reactivity (e.g. cisplatin as DNA-reactive therapeutic) or imaging properties (e.g. gadolinium complexes as MRI diagnostics). In all current metallopharmaceuticals, the metal-ion bears the key feature of the mechanism, which includes ligand exchange processes. For example, the highly efficient anticancer drug cisplatin reacts with DNA by crosslinking guanine bases leading eventually to apoptosis (Figure 1b). Gadolinium complexes such as Gd-DOTA, on the other hand, coordinate reversibly to water molecules and are the most widely used magnetic resonance imaging (MRI) contrast agents due to the very high magnetic moment and symmetric electronic ground state of the Gd3+ ion (Figure 1c).


Figure 1. Organic and inorganic compounds in biological and medicinal chemistry: a) A typical organic small molecule enzyme inhibitor. b) Cisplatin, (H3N)2PtCl2, crosslinks DNA and represents the class of reactive metal compounds. c) Gadolinium complexes are used for MRI contrast agents and represent the class of metal complexes used for imaging purposes.

Our Approach

Central to our program is the development of new generations of molecules that combine organic and inorganic elements and thus bridge organic and inorganic biological and medicinal chemistry. In particular, our group envisions the utilization of metal-containing compounds as functional scaffolds. In these molecules, the organic ligands are mainly responsible for the recognition process, whereas the metal center, in combination with the entire coordination sphere, organizes the orientation of the organic ligand sphere, can participate in coordination chemistry, catalysis, and imaging processes (Figure 2). Therefore, we believe that the addition of a metal atom to the otherwise organic scaffold opens new opportunities for the design of biological active molecules and tools for the investigation of biological processes.


Figure 2. Opportunities for metallo-organic compounds in chemical biology.

Metal complexes as protein kinase inhibitors: Design strategy and proof-of-principle

Our group has made progress with investigating the scope of using organometallic compounds as structural scaffolds for the design of bioactive molecules. In contrast to conventional metal-containing bioactive compounds or imaging agents, in these molecules the metal center plays a solely structural role by organizing the organic ligands in the three-dimensional receptor space.

We started our program by using the class of indolocarbazole alkaloids as a lead structure for ATP-competitive protein kinase inhibitors. This family of inhibitors, with staurosporine (1) (Figures 3 and 4) as one of its most potent but nonspefic members, shares the indolo[2,3-a]carbazole moiety 2a (lactam form) or 2b (imide form, arcyriaflavin A) which binds to the adenine binding site by undergoing two hydrogen bonds to the backbone of the hinge between the N-terminal and C-terminal kinase domain. The lactam or imide NH-group acts as a hydrogen-bond donor to a backbone carbonyl, and the lactam or imide carbonyl accepts a hydrogen from a backbone amide. For this class of inhibitors, specificity for a particular protein kinase can be achieved by the moiety which is attached to the indole nitrogen atoms.


Figure 3. Schematic representation of the ATP binding pocket of the cyclin dependent protein kinase 2 (CDK2) in complex with a) ATP, b) staurosporine, and c) an indolocarbazole-derived ruthenium complex.


Figure 4. Designing ruthenium complexes as protein kinase inhibitors by mimicking ATP-competitive indolocarbazole alkaloids.


Figure 5. IC50 curves of ruthenium complex 5, its methylated derivative, the ligand 3, and staurosporine.

We envisioned that by replacing the indolocarbazole alkaloid scaffold with metal complexes in which the structural features of the indolocarbazole heterocycle 2a/b is retained in one of the ligands, metal complexes could be targeted to the ATP-binding site of protein kinases (Figure 3a-c). Potent and specific inhibitors for a particular kinase could then be obtained by assembling elaborate structures around the metal center. The key component of our design is the novel pyridocarbazole ligand 3 (Figure 4), derived from arcyriaflavin A 2b by just replacing one indole moiety with a pyridine. Ligand 3 can serve as a bidentate ligand for metal complexes 4. Additional ligands in the coordination sphere of the metal can now substitute for the carbohydrate moiety of staurosporine, with the metal center serving as a "glue" to unite all parts. We recently performed a proof-of-principle of this concept by developing a ruthenium compound 5 as a low nanomolar and ATP-competitive inhibitor of the Glycogen Synthase Kinase 3 (GSK-3). This compound 5 is air-stable, stable in water, and can even withstand millimolar concentrations of thiols (no reaction in presence of 5 mM β-mercaptoethanol overnight). This stability is due to the kinetical inertness of coordinative bonds to ruthenium.

Protein kinase inhibition with an organometallic compound inside of mammalian cells

We were next asking the question whether organometallic compounds of type 5 can inhibit GSK-3 inside of a living mammalian cell. We used compound 6, which bears an additional hydroxyl group at the indole moiety (Figure 5, top right). Organoruthenium compound 6 is not only more water soluble compared to 5 (it can be dissolved at 1 mM in 3% DMSO/H2O) but displays also an even higher potency with an IC50 for the β-isoform GSK-3β of 0.5 nM (Ki = 150 pM).


Figure 5. Cells transfected with a β-catenin-responsive luciferase reporter were treated with different concentrations of 6, the methylated ruthenium complex 7, and the ligand 8 for a time period of 24 hours. LiCl serves as a positive control and DMSO as a negative control. Left: Relative luciferase activity determined after cell lysis, addition of the substrate luciferin, and luminescence measurement. Bottom right: Western blot of the cell lysate qualitatively detecting the amount of β-catenin.

GSK-3 is a negative regulator of the wnt signal transduction pathway that phosphorylates β-catenin. Phosphorylated β-catenin is unstable and is degraded rapidly by the proteasome. Wnt signaling inhibits GSK-3, leading to stabilization of β-catenin protein. β-catenin then accumulates and serves as a transcriptional coactivator through its interaction with the T-cell factor (Tcf) family of transcription factors. Thus inhibition of GSK-3 by pharmacological inhibitors or by wnt signaling leads to increased β-catenin levels and activation of wnt dependent transcription.

In order to test cellular accumulation of β-catenin protein as a response to inhibition of GSK-3 by the ruthenium complex 6, we used human embryonic kidney cells (HEK293T) that have stably incorporated a Tcf-luciferase transcription reporter (OT-Luc cells). This transciption reporter generates luciferase in response to increased concentrations of β-catenin. Exposure of OT-Luc cells to varying concentrations of 6 over a period of 24 hours yields a strong upregulation of luciferase in the concentration window of 3 µM down to 100 nM (left side of Figure 5). For example, at a concentration of 1 µM 6, luciferase activity is enhanced by a factor of around 1500. Intriguingly, at the same time, the methylated control 7 as well as the free ligand 8 do not yield any significant increase in luciferase activity at a concentration of 1 µM (1.6 and 2.4 fold, respectively). As a positive control, the established selective GSK-3 inhibitor LiCl leads to an increase of the luminescence signal by a factor of 1190 at a concentration of 30 mM.

To verify the accumulation of β-catenin directly, we analyzed the cellular β-catenin concentration by western-blotting after incubation with 6. Figure 5 (bottom right) demonstrates qualitatively an increase in β-catenin protein in the presence of 6, but not 7 or 8, at 1 µM. These experiments demonstrate that the ruthenium complex 6 crosses the cell membrane and activates the wnt pathway at low micromolar and even nanomolar concentrations.

It is also worth noting that 6 does not show any signs of cytotoxicity at concentrations of 3 µM or less, supporting our assumption that the kinetically inert ruthenium center is only an innocent non-reactive bystander. The integrity of the complex 6 inside the cell is also reinforced by the fact that the entire ligand assembly is necessary for the activity inside the cell and that the free ligand 8 does not activate the wnt pathway itself at similar concentrations. This is a remarkable observation considering the millimolar concentrations of highly "metallophilic" thiols such of glutathione, cysteine, and cysteine containing proteins within the cell.

Finally, in collaboration with the group of Dr. Peter Klein (Medical School, University of Pennsylvania), we investigated the in vivo effects of 6. Injection of 6 into a ventral blastomere at the 32-64 cell stage caused formation of a complete secondary dorsal axis (Figure 6), similar to effects observed with injection of LiCl. Thus, it can be concluded that ruthenium complex 6 inhibits GSK-3 and activates wnt signaling in vivo in Xenopus embryos.

Crystal structures of ruthenium compounds bound to protein kinase Pim-1


Figure 6. Left: Superimposed co-crystal structures of ruthenium compound (S)-6 and staurosporine with Pim-1. Right: (S)-6 as a spacefilling model in the binding pocket of Pim-1 demonstrates the shape match.