Structure-Based Design of Staurosporine (STS)-Based Fluorescent Probes for Broad-Ranging Kinase Profiling
Keywords: Protein kinase, Staurosporine, Fluorescent probe, TR-FRET based binding assay
ABSTRACT
Selectivity profiling of compounds is important for kinase drug discovery. To this end, we aimed to develop a broad-range protein kinase assay by synthesizing a novel staurosporine-derived fluorescent probe based on staurosporine and kinase-binding related structural information. Upon structural analysis of staurosporine with kinases, a 4′-methylamine moiety of Staurosporine (STS) was found to be located on the solvent side of the kinases, to which several linker units can be conjugated by either alkylation or acylation. However, such conjugation was suggested to reduce the binding affinities of the modified compound for several kinases, owing to the elimination of hydrogen bond donor moiety of NH-group from 4′-methylamine and/or steric hindrance by acyl moiety. Based on this structural information, we designed and synthesized a novel staurosporine-based probe without methyl group in order to retain the hydrogen bond donor, similar to unmodified staurosporine. The broad range of the kinase binding assay demonstrated that our novel fluorescent probe is an excellent tool for developing broad-ranging kinase binding assay.
INTRODUCTION
More than 500 species of human protein kinases have been discovered thus far, most of which are potential therapeutic targets. Protein kinases play integral roles in signal transduction, cellular processes, and maintenance of cellular homeostasis and communication. Structurally, protein kinases possess a highly conserved ATP-binding pocket that can be exploited for binding small compounds. Owing to high similarity of the ATP binding pocket across kinase families, development of selective kinase inhibitors is considerably challenging. To overcome this challenge, development of a high-throughput and wide-range protein kinase panel assay is necessary. Fluorescent probes interacting with a wide range of protein kinases are an essential tool for the development of a wide-range protein kinase assay. Staurosporine is a natural product originally isolated from Streptomyces staurosporeus, and is well known as a wide-range human protein kinase inhibitor. Staurosporine binds to the ATP-binding site of various protein kinases with similar orientation. Its lactam and indolocarbazole rings interact with hinge region and its 4′-methylamine moiety interacts with the hinge region and/or catalytic loop by a hydrogen bond and ionic interaction in the solvent accessible region. Because of these biological and structural characteristics, several chemical probes have been synthesized based on staurosporine, conjugating it with a fluorescent core through a linker extended from the 4′-methylamine position.
Figure 1. Overall Strategy for the Development of Novel Kinase Binding Assay.(a) Structure of staurosporine, (b) compound structure of staurosporine probes, (c) schematic view of TR-FRET based protein-probe binding assay and expected key hydrogen bond of the 4′-methylamine of staurosporine and kinase.
Generally, conjugation of a fluorescent substance and a linker unit reduces the binding affinity of the resulting probe to the target protein compared with that of the parent compound, because of steric hindrance and change in the polarity of the compound. Staurosporine-based chemical probes have been conjugated with a linker unit by alkylation or acylation utilizing the nucleophilicity of its 4′-methylamine position. However, these chemical probes had reduced binding affinity for several kinds of protein kinases presumably due to the elimination of hydrogen bond donor moiety from 4′-methylamine moiety and/or steric hindrance by acyl moiety.
On the basis of these structural information, we designed a novel fluorescent probe (compound 8a) that would retain hydrogen bond and ionic interaction of staurosporine with kinases by eliminating methyl group of other known staurosporine-based chemical probes. The compound 8a is predicted to retain its binding affinity for many types of protein kinases compared to the known staurosporine-based chemical probes. To evaluate the effect of the removal of methyl group, we also synthesized 8b, which is a methyl capped derivative of 8a.
Several types of probe compound based comprehensive protein kinase assay have been reported till date, such as active site-directed biochemical competition binding assay, biochemical kinase activity assay, chemical proteomics assay, and in cell bioluminescent assay. Among these assays, we focused on the time-resolved fluorescence resonance energy transfer (TR-FRET) based protein kinase assay since it has strong advantages including high-throughput performance and low false-positive signals from intrinsic fluorescence. We evaluated the binding affinity of probes based on 8a and 8b for more than 250 kinases by TR-FRET-based saturation binding assay.
MATERIALS AND METHODS
Synthesis
Synthesis of BODIPY-FL conjugated staurosporine derivatives via PEG4 linker (8a-b) is described in Scheme 1. Demethylated staurosporine 2a was synthesized by hydrogenation of oxime derivative using 5% Pt/C. Staurosporine derivatives 2a-b were alkylated with N-Boc protected PEG4 bromide 4, which was readily prepared from N-Boc protected PEG4 alcohol 3, to afford PEG4 linker conjugated staurosporine derivatives 5a-b. Removal of Boc group, followed by amidation using BODIPY-FL succinimidyl ester 7 provided 8a-b. Photochemical properties such as quantum efficiency of fluorescence, absorption spectra and emission spectra of 8a and 8b were calculated under free state and protein binding state.
TR-FRET Based Saturation Binding Assay
The experimental procedure to prepare the recombinant kinase protein is shown in Supporting information. All tested protein kinases are shown in Supplemental Table 1. The experimental procedure of TR-FRET based saturation binding assay is shown in Supporting information. Criteria for detecting specific binding were set as S/B ratio > 1.3 and Kd < 1500 nM for either probe.
Structural Analysis
To analysis of the effect of N-Me group on kinase affinity and selectivity, distances between the hydrogen of 4′-methylamine and amino acid residues were calculated by LIGPLOT (v.4.5.3, EMBl-EBL) using protein structures from PDB. The 4′-methylamine of staurosporine interacts with the main chain of catalytic loop region or the side chain of hinge region, or both for most kinases.
RESULTS AND DISCUSSION
Kinase Binding Affinity Analysis
Kd values and S/B ratio of 8a and 8b for 288 protein kinases tested are summarized in Supplemental Table 2. Out of all the 288 kinases tested, specific saturation binding curve for 203 kinases was generated for at least one of the probes for the detection of saturation-specific binding. Compound 8a showed high specificity for tyrosine kinase (TK) and tyrosine kinase-like (TKL) family kinases, and 8b showed specificity for protein kinase A, G, and C (AGC) family kinases.
Kd values of 8a and 8b for 25 protein kinases, whose cocrystal structures with staurosporine have been published, are summarized in Table 1. While 8a retained the secondary amine at 4′-methylamine position with an unshared electron pair and a hydrogen donor, similar to staurosporine, 8b did not. Additionally, steric hindrance due to the extra methyl moiety of 8b might inhibit its interaction with some kinases.
A reported staurosporine-based chemical probe, which was conjugated with a linker unit by acylation of the staurosporine 4′-methylamine position, showed a Kd of 8,000 nM for MAP3K5 (ASK1), whereas 8a showed Kd of 136 nM and 8b of 248 nM for the kinase. Previous reported Kd value of native staurosporine for MAP3K5 was 110 nM. The affinity of 8a for MAP3K5 was almost the same as that of native staurosporine. These results suggested that 8a almost unaffected by steric hindrance from probe formation for the binding of MAP3K5.
Hydrogen Bond Interaction Analysis
Furthermore, the affinity of 8a for CDK2, CHEK1, and RPS6KA1 having two hydrogen bonds on 4′-methylamine of staurosporine was stronger than that of 8b. These results indicated that 8a strictly retained its interaction at 4′-methylamine by ionic forces with the side chain of hinge-region amino-acid residue and hydrogen-bond with the main chain carbonyl of catalytic loop amino-acid residue, as indicated by the protein crystal structures in PDB.
On the contrary, some kinases that had no hydrogen bond on the 4′-methylamine of staurosporine, such as STK16, ALK, and GSK3B, showed stronger affinity for 8b than for 8a. The 4′-methylamine moiety does not interact with any residue of these kinases, and the extra methyl of 8b might interact with the amino acid residues by hydrophobic interaction.
Zhao and colleagues reported that the affinities for CHEK1 and CDK2 of the staurosporine analog SB-218078, devoid of 4′-methylamine moiety, were lower than that of staurosporine. R549, which is a well-known CDK family selective inhibitor (including CDK2), interacts with amino acid residues at the entrance of ATP pocket by ionic forces, similar to staurosporine. When the 1-methanesulfonyl-4-piperidine moiety of R549 was changed to methyl, its affinity for CDK2 reduced more than 100 times. These results indicated that the kinases that had higher affinity for 8a than 8b required hydrophilic interaction in the ATP-binding pocket to ensure tight affinity.
STK16-IN-1, the highly selective STK16 inhibitor, has a hydrophobic structure in the solvent-accessible region of STK16, which is the main reason for its high selectivity across human protein kinases. CH5424802, the highly selective ALK inhibitor, does not show hydrophilic interaction in the ATP entrance region. Taken together, these results indicated that the kinases that showed higher affinity for 8b than for 8a do not require hydrophilic interaction and rather tend to have hydrophobic interaction at the ATP entrance region.
Kinase Family Specificity
Regarding the binding affinity preferences, 8a showed high specificity for tyrosine kinase (TK) and tyrosine kinase-like (TKL) family kinases, and 8b showed specificity for protein kinase A, G, and C (AGC) family kinases. For example, 8b had a greater affinity for PKC family kinases, such as PRKCA, PRKCB, PRKCD, PRKCE, PRKCH, PRKCQ, and PRKCZ compared to 8a, thus suggesting that the extra methyl group of 8b might acquire the binding energy through the interaction with amino acid residues of AGC family kinases, although this group might inhibit effective binding to other types of kinases.
As the comprehensive kinase profile data show, Dasatinib and Bosutinib were more selective for TK family kinases than for AGC family kinases. Regarding the study of Dasatinib, the methyl moiety of solvent region had decreased affinity and amino moiety had increased affinity for the primary target of Dasatinib. Furthermore, the study of Bosutinib indicated that the carbon chain length located at ATP entrance region influenced the affinity for SRC inhibition. On the other hand, GSK690693 is selective to AGC family kinases as seen from comprehensive kinase profile data, and affinity of GSK690693 for AKT1, AKT2, and AKT3 was unaffected by the structure of the solvent-exposed region.
Application to Kinase Panel Assays
A wide range of kinase-binding BODIPY-FL-conjugates of staurosporine would enable development of the comprehensive kinase binding assay that estimates kinase selectivity. In our experience, S/B ratio > 1.5 of fluorescent probe appears to be the optimal criteria for the development of TR-FRET-based competitive binding assay. A total of 172 kinases of all the kinases tested for 8a, and 138 kinases of all the kinases tested for 8b were found to meet these criteria. Furthermore, 39 kinases were fulfilled the above criteria for only 8a while 5 kinases were for only 8b. These results together suggest that kinase-binding profile of 8a almost encompasses that of 8b.
CONCLUSION
We rationally designed novel staurosporine-based fluorescent probes based on the structural information, to retain their binding affinity for a wide range of protein kinases. The novel staurosporine-based probe has certain advantages over previously reported probes in the development of TR-FRET-based kinase profiling assay. As a result of comparison with the binding affinity of 8a for MAP3K5 and the reported staurosporine-based probe of that, 8a almost unaffected by steric hindrance from probe formation for interaction with the kinase.
Furthermore, comparison with interact kinases of 8a and that of Kinase Tracer 236, which is already commercialized fluorescent probe for kinase binding assay by Thermo Fisher Scientific but compound structure is not opened, 8a have a potential to develop binding assay for the kinases which are not recommended to use Kinase Tracer 236 such as PIM3, CDC42BPG, MAP2K7, TXK and SIK3. This novel probe can be a strong tool to analyze kinase selectivity in kinase drug discovery.
