Drugging the Undruggable: How Isoquinolines and PKA Initiated the Era of Designed Protein Kinase Inhibitor Therapeutics.

In 1984, Japanese researchers led by the biochemist Hiroyoshi Hidaka described the first synthetic protein kinase inhibitors based on an isoquinoline sulfonamide structure (Hidaka et al. Biochemistry, 1984 Oct 9; 23(21): 5036-41. doi: 10.1021/bi00316a032). These led to the first protein kinase inhibitor approved for medical use (fasudil), an inhibitor of the AGC subfamily Rho kinase. With potencies strong enough to compete against endogenous ATP, the isoquinoline compounds established the druggability of the ATP binding site. Crystal structures of their protein kinase complexes, including with cAMP-dependent protein kinase (PKA), showed interactions that, on the one hand, could mimic ATP but, on the other hand, could be optimized for high potency binding, kinase selectivity, and diversification away from adenosine. They also showed the flexibility of the glycine-rich loop, and PKA became a major prototype for crystallographic and nuclear magnetic resonance (NMR) studies of protein kinase mechanism and dynamic activity control. Since fasudil, more than 70 kinase inhibitors have been approved for clinical use, involving efforts that progressively have introduced new paradigms of data-driven drug discovery. Publicly available data alone comprise over 5000 protein kinase crystal structures and hundreds of thousands of binding data. Now, new methods, including artificial intelligence techniques and expansion of protein kinase targeting approaches, together with the expiration of patent protection for optimized inhibitor scaffolds, promise even greater advances in drug discovery. Looking back to the time of the first isoquinoline hinge binders brings the current state-of-the-art into stark contrast. Appropriately for this Perspective article, many of the milestone papers during this time were published in Biochemistry (now ACS Biochemistry).

Inhibitor binding to mutants of protein kinase A with GGGxxG and GxGxxA glycine-rich loop motifs.

The conserved GxGxxG motif of protein kinases forms a beta turn at the tip of the flexible glycine-rich loop and creates much of the ATP pocket binding surface. Notable exceptions to this sequence include GGGxxG in ABL kinase and GxGxxA in protein kinase C isoforms. We constructed the corresponding mutants of PKA, T51G, and G55A, and tested quinazoline inhibitors that were designed to bind via glycine-rich loop interactions, testing also staurosporine for comparison. The quinazoline inhibitors have significantly reduced binding strengths in both mutants. In striking contrast to these results, the binding of the "pan-kinome" inhibitor staurosporine is strengthened in the mutants. Surface plasmon resonance (SPR) shows that the tightened binding of staurosporine arises from increased kon rates, changes not offset by more moderately increased koff rates. The SPR results fit best to a two step binding process for staurosporine in wild type PKA, but not the mutants.

Two SnRK2-Interacting Calcium Sensor Isoforms Negatively Regulate SnRK2 Activity by Different Mechanisms.

SNF1-related protein kinases 2 (SnRK2s) are key signaling elements regulating abscisic acid-dependent plant development and responses to environmental stresses. Our previous data showed that the SnRK2-interacting Calcium Sensor (SCS) inhibits SnRK2 activity. Use of alternative transcription start sites located within the Arabidopsis (Arabidopsis thaliana) AtSCS gene results in two in-frame transcripts and subsequently two proteins, that differ only by the sequence position of the N terminus. We previously described the longer AtSCS-A, and now describe the shorter AtSCS-B and compare the two isoforms. The two isoforms differ substantially in their expression profiles in plant organs and in response to environmental stresses, in their calcium binding properties, and in their conformational dynamics in the presence and absence of Ca2+ Only AtSCS-A has the features of a calcium sensor. Both forms inhibit SnRK2 activity, but while AtSCS-A requires calcium for inhibition, AtSCS-B does not. Analysis of Arabidopsis plants stably expressing 35S::AtSCS-A-c-myc or 35S::AtSCS-B-c-myc in the scs-1 knockout mutant background revealed that, in planta, both forms are negative regulators of abscisic acid-induced SnRK2 activity and regulate plant resistance against water deficit. Moreover, the data highlight biochemical, biophysical, and functional properties of EF-hand-like motifs in plant proteins.

Density Functional Studies on Secondary Amides: Role of Steric Factors in Cis/Trans Isomerization.

Cis/trans isomerization of amide bonds is a key step in a wide range of biological and synthetic processes. Occurring through C-N amide bond rotation, it also coincides with the activation of amides in enzymatic hydrolysis. In recently described QM studies of cis/trans isomerization in secondary amides using density functional methods, we highlighted that a peptidic prototype, such as glycylglycine methyl ester, can suitably represent the isomerization and complexities arising out of a larger molecular backbone, and can serve as the primary scaffold for model structures with different substitution patterns in order to assess and compare the steric effect of the substitution patterns. Here, we describe our theoretical assessment of such steric effects using tert-butyl as a representative bulky substitution. We analyze the geometries and relative stabilities of both trans and cis isomers, and effects on the cis/trans isomerization barrier. We also use the additivity principle to calculate absolute steric effects with a gradual increase in bulk. The study establishes that bulky substitutions significantly destabilize cis isomers and also increases the isomerization barrier, thereby synergistically hindering the cis/trans isomerization of secondary amides. These results provide a basis for the rationalization of kinetic and thermodynamic properties of peptides with potential applications in synthetic and medicinal chemistry.

Cis/Trans Isomerization in Secondary Amides: Reaction Paths, Nitrogen Inversion, and Relevance to Peptidic Systems.

Cis/trans isomerization of 2°-amide bonds is a key step in a wide range of important processes. Here we present a theoretical assessment of cis/trans isomerization of 2°-amide bonds using B3LYP density functional methods, describing two reaction paths and corresponding geometry changes during isomerization of N-methylacetamide (NMA) and glycylglycine methyl ester (GGMe). The isomerization begins via a common path, as the extended π-bonding of the amide bond maintains approximate planarity of the O-C-N-H dihedral angle, with only gradually increasing pyramidalization of the nitrogen atom, until a bifurcation point is reached. Both subsequent paths comprise two phases, an "ω phase" (characterized by a major change in C-C-N-C dihedral) and a "θ phase" (characterized by major change in O-C-N-H dihedral), with two distinct transition states. The θ phase involves inversion of the pyramidal amide-nitrogen geometry. Both reaction paths converge at another bifurcation point near the opposite geometry. Studies on the larger GGMe show in addition that the multiple additional rotamers do not change the qualitative properties of the isomerization, but do affect the energies of the differing transition states. These detailed results provide significant new insights into cis/trans isomerization paths in 2°-amides, and serve as a basis for theoretical studies on larger peptidic systems.

Perspective on computational and structural aspects of kinase discovery from IPK2014.

Recent advances in understanding the activity and selectivity of kinase inhibitors and their relationships to protein structure are presented. Conformational selection in kinases is studied from empirical, data-driven and simulation approaches. Ligand binding and its affinity are, in many cases, determined by the predetermined active and inactive conformation of kinases. Binding affinity and selectivity predictions highlight the current state of the art and advances in computational chemistry as it applies to kinase inhibitor discovery. Kinome wide inhibitor profiling and cell panel profiling lead to a better understanding of selectivity and allow for target validation and patient tailoring hypotheses. This article is part of a Special Issue entitled: Inhibitors of Protein Kinases.

Assessing protein kinase target similarity: Comparing sequence, structure, and cheminformatics approaches.

In just over two decades, structure based protein kinase inhibitor discovery has grown from trial and error approaches, using individual target structures, to structure and data driven approaches that may aim to optimize inhibition properties across several targets. This is increasingly enabled by the growing availability of potent compounds and kinome-wide binding data. Assessing the prospects for adapting known compounds to new therapeutic uses is thus a key priority for current drug discovery efforts. Tools that can successfully link the diverse information regarding target sequence, structure, and ligand binding properties now accompany a transformation of protein kinase inhibitor research, away from single, block-buster drug models, and toward "personalized medicine" with niche applications and highly specialized research groups. Major hurdles for the transformation to data driven drug discovery include mismatches in data types, and disparities of methods and molecules used; at the core remains the problem that ligand binding energies cannot be predicted precisely from individual structures. However, there is a growing body of experimental data for increasingly successful focussing of efforts: focussed chemical libraries, drug repurposing, polypharmacological design, to name a few. Protein kinase target similarity is easily quantified by sequence, and its relevance to ligand design includes broad classification by key binding sites, evaluation of resistance mutations, and the use of surrogate proteins. Although structural evaluation offers more information, the flexibility of protein kinases, and differences between the crystal and physiological environments may make the use of crystal structures misleading when structures are considered individually. Cheminformatics may enable the "calibration" of sequence and crystal structure information, with statistical methods able to identify key correlates to activity but also here, "the devil is in the details." Examples from specific repurposing and polypharmacology applications illustrate these points. This article is part of a Special Issue entitled: Inhibitors of Protein Kinases.

Addressing the Glycine-Rich Loop of Protein Kinases by a Multi-Facetted Interaction Network: Inhibition of PKA and a PKB Mimic.

Protein kinases continue to be hot targets in drug discovery research, as they are involved in many essential cellular processes and their deregulation can lead to a variety of diseases. A series of 32 enantiomerically pure inhibitors was synthesized and tested towards protein kinase A (PKA) and protein kinase B mimic PKAB3 (PKA triple mutant). The ligands bind to the hinge region, ribose pocket, and glycine-rich loop at the ATP site. Biological assays showed high potency against PKA, with Ki values in the low nanomolar range. The investigation demonstrates the significance of targeting the often neglected glycine-rich loop for gaining high binding potency. X-ray co-crystal structures revealed a multi-facetted network of ligand-loop interactions for the tightest binders, involving orthogonal dipolar contacts, sulfur and other dispersive contacts, amide-π stacking, and H-bonding to organofluorine, besides efficient water replacement. The network was analyzed in a computational approach.

The structure of a dual-specificity tyrosine phosphorylation-regulated kinase 1A-PKC412 complex reveals disulfide-bridge formation with the anomalous catalytic loop HRD(HCD) cysteine.

Dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYRK1A) is a protein kinase associated with neuronal development and brain physiology. The DYRK kinases are very unusual with respect to the sequence of the catalytic loop, in which the otherwise highly conserved arginine of the HRD motif is replaced by a cysteine. This replacement, along with the proximity of a potential disulfide-bridge partner from the activation segment, implies a potential for redox control of DYRK family activities. Here, the crystal structure of DYRK1A bound to PKC412 is reported, showing the formation of the disulfide bridge and associated conformational changes of the activation loop. The DYRK kinases represent emerging drug targets for several neurological diseases as well as cancer. The observation of distinct activation states may impact strategies for drug targeting. In addition, the characterization of PKC412 binding offers new insights for DYRK inhibitor discovery.

Label-free measurement of antimicrobial peptide interactions with lipid vesicles and nanodiscs using microscale thermophoresis.

One strategy to combat antimicrobial resistance is the discovery of new classes of antibiotics. Most antibiotics will at some point interact with the bacterial membrane to either interfere with its integrity or to cross it. Reliable and efficient tools for determining the dissociation constant for membrane binding (KD) and the partitioning coefficient between the aqueous- and membrane phases (KP) are therefore important tools for discovering and optimizing antimicrobial hits. Here we demonstrate that microscale thermophoresis (MST) can be used for label-free measurement of KD by utilising the intrinsic fluorescence of tryptophan and thereby removing the need for chromophore labelling. As proof of principle, we have used the method to measure the binding of a set of small cyclic AMPs to large unilamellar vesicles (LUVs) and two types of lipid nanodiscs assembled by styrene maleic acid (SMA) and quaternary ammonium SMA (SMA-QA). The measured KD values correlate well with the corresponding measurements using surface plasmon resonance (SPR), also broadly reflecting the tested AMPs' minimal inhibition concentration (MIC) towards S. aureus and E. coli. We conclude that MST is a promising method for fast and cost-efficient detection of peptide-lipid interactions or mapping of sample conditions in preparation for more advanced studies that rely on expensive sample preparation, labelling and/or instrument time.

Structural origins of AGC protein kinase inhibitor selectivities: PKA as a drug discovery tool.

The era of structure-based protein kinase inhibitor design began in the early 1990s with the determination of crystal structures of protein kinase A (PKA, or cyclic AMP-dependent kinase). Although many other protein kinases have since been extensively characterized, PKA remains a prototype for studies of protein kinase active conformations. It serves well as a model for the structural properties of AGC subfamily protein kinases, clarifying inhibitor selectivity profiles. Its reliable expression, constitutive activity, simple domain structure, and reproducible crystallizability have also made it a useful surrogate for the discovery of inhibitors of both established and emerging AGC kinase targets.

Anomalous dispersion analysis of inhibitor flexibility: a case study of the kinase inhibitor H-89.

With its ability to show the interactions between drug-target proteins and small-molecule ligands, X-ray crystallography is an essential tool in drug-discovery programmes. However, its usefulness can be limited by crystallization artifacts or by the data resolution, and in particular when assumptions of unimodal binding (and isotropic motion) do not apply. Discrepancies between the modelled crystal structure and the physiological range of structures generally prevent quantitative estimation of binding energies. Improved crystal structure resolution will often not aid energy estimation because the conditions which provide the highest rigidity and resolution are not likely to reflect physiological conditions. Instead, strategies must be employed to measure and model flexibility and multiple binding modes to supplement crystallographic information. One useful tool is the use of anomalous dispersion for small molecules that contain suitable atoms. Here, an analysis of the binding of the kinase inhibitor H-89 to protein kinase A (PKA) is presented. H-89 contains a bromobenzene moiety that apparently binds with multiple conformations in the kinase ATP pocket. Using anomalous dispersion methods, it was possible to resolve these conformations into two distinct binding geometries.

Mutants of protein kinase A that mimic the ATP-binding site of Aurora kinase.

We describe in the present paper mutations of the catalytic subunit α of PKA (protein kinase A) that introduce amino acid side chains into the ATP-binding site and progressively transform the pocket to mimic that of Aurora protein kinases. The resultant PKA variants are enzymatically active and exhibit high affinity for ATP site inhibitors that are specific for Aurora kinases. These features make the Aurora-chimaeric PKA a valuable tool for structure-based drug discovery tasks. Analysis of crystal structures of the chimaera reveal the roles for individual amino acid residues in the binding of a variety of inhibitors, offering key insights into selectivity mechanisms. Furthermore, the high affinity for Aurora kinase-specific inhibitors, combined with the favourable crystallizability properties of PKA, allow rapid determination of inhibitor complex structures at an atomic resolution. We demonstrate the utility of the Aurora-chimaeric PKA by measuring binding kinetics for three Aurora kinase-specific inhibitors, and present the X-ray structures of the chimaeric enzyme in complex with VX-680 (MK-0457) and JNJ-7706621 [Aurora kinase/CDK (cyclin-dependent kinase) inhibitor].

VX680 binding in Aurora A: π-π interactions involving the conserved aromatic amino acid of the flexible glycine-rich loop.

The regulation of protein kinases requires flexibility, especially near the ATP binding site. The cancer drug target Aurora A is inhibited by the ATP site inhibitor VX680, and published crystal structures show two distinct conformations. In one, a refolded glycine-rich loop creates a stacked π-π interaction between the conserved aromatic residue of the glycine-rich loop hairpin turn (F144) and the inhibitor. This refolding, associated with binding to a peptide derived from the cofactor TPX2, is absent in the other structure. We use surface plasmon resonance to measure VX680 binding to native and mutant F144A Aurora A kinase domains, with and without the TPX2 peptide. Results show that the F144 aromatic side chain contributes 2 kcal/mol to the VX680 binding energy, independent of the TPX2 peptide. This indicates that distinct VX680 bound conformations of Aurora A cannot be simply correlated with TPX2 binding and that Aurora A retains flexibility when inhibitor-bound. Molecular dynamics simulations show that alternate geometries for the π-π interactions are feasible in the absence of the rigidifying packing interactions seen in the crystal lattice.

Protein kinase inhibition of clinically important staurosporine analogues.

The isolation in 1977 of the microbial alkaloid staurosporine inaugurated research into several distinct series of related natural and synthetic compounds. This has especially included research into applications as anticancer drugs, beginning with the observation of low nanomolar inhibition of protein kinases. At present, several staurosporine cognates are in advanced clinical trials as anticancer agents, with the potential to join the 10 other protein kinase inhibitors now approved for clinical use. Staurosporine is a broadly selective and potent protein kinase inhibitor, with submicromolar binding to the vast majority of the protein kinases tested, and binding most of them more tightly than 100 nM. Crystal structures have shown the extended buried surface area interactions between the protein kinase adenine binding site and the extended aromatic plane of the inhibitor, together with protein-saccharide interactions in the ribose binding site. Together with structures of closely related analogues, there are now some 70 X-ray crystal structures in the Protein Data Bank that enable analysis of target binding properties of the clinical compounds. In this manuscript we review the discovery of these compounds, revisit crystal structures and review the observed interactions. These support the interpretation of kinase selectivity profiles of staurosporine and its analogues, including midostaurin (PKC412), for which a co-crystal structure is not yet available. Further, the mix of purely natural, biosynthetically and chemically modified compounds described here offer insights into prospects and strategies for drug discovery via bioprospecting.

Novel Scaffolds for Dual Specificity Tyrosine-Phosphorylation-Regulated Kinase (DYRK1A) Inhibitors.

DYRK1A is one of five members of the dual-specificity tyrosine (Y) phosphorylation-regulated kinase (DYRK) family. The DYRK1A gene is located in the Down syndrome critical region and regulates cellular processes related to proliferation and differentiation of neuronal progenitor cells during early development. This has focused research on its role in neuronal degenerative diseases, including Alzheimer's and Down syndrome. Recent studies have also shown a possible role of DYRK1A in diabetes. Here we report a variety of scaffolds not generally known for DYRK1A inhibition, demonstrating their effects in in vitro assays and also in cell cultures. These inhibitors effectively block the tau phosphorylation that is a hallmark of Alzheimer's disease. The crystal structures of these inhibitors support the design of optimized and novel therapeutics.

Structural basis for the regulation of insulin-like growth factors by IGF binding proteins.

Insulin-like growth factor binding proteins (IGFBPs) control the extracellular distribution, function, and activity of IGFs. Here, we report an X-ray structure of the binary complex of IGF-I and the N-terminal domain of IGFBP-4 (NBP-4, residues 3-82) and a model of the ternary complex of IGF-I, NBP-4, and the C-terminal domain (CBP-4, residues 151-232) derived from diffraction data with weak definition of the C-terminal domain. These structures show how the IGFBPs regulate IGF signaling. Key features of the structures include (1) a disulphide bond ladder that binds to IGF and partially masks the IGF residues responsible for type 1 IGF receptor (IGF-IR) binding, (2) the high-affinity IGF-I interaction site formed by residues 39-82 in a globular fold, and (3) CBP-4 interactions. Although CBP-4 does not bind individually to either IGF-I or NBP-4, in the ternary complex, CBP-4 contacts both and also blocks the IGF-IR binding region of IGF-I.

Dynamical models of mutated chronic myelogenous leukemia cells for a post-imatinib treatment scenario: Response to dasatinib or nilotinib therapy.

Targeted inhibition of the oncogenic BCR-ABL1 fusion protein using the ABL1 tyrosine kinase inhibitor imatinib has become standard therapy for chronic myelogenous leukemia (CML), with most patients reaching total and durable remission. However, a significant fraction of patients develop resistance, commonly due to mutated ABL1 kinase domains. This motivated development of second-generation drugs with broadened or altered protein kinase selectivity profiles, including dasatinib and nilotinib. Imatinib-resistant patients undergoing treatment with second-line drugs typically develop resistance to them, but dynamic and clonal properties of this response differ. Shared, however, is the observation of clonal competition, reflected in patterns of successive dominance of individual clones. We present three deterministic mathematical models to study the origins of clinically observed dynamics. Each model is a system of coupled first-order differential equations, considering populations of three mutated active stem cell strains and three associated pools of differentiated cells; two models allow for activation of quiescent stem cells. Each approach is distinguished by the way proliferation rates of the primary stem cell reservoir are modulated. Previous studies have concentrated on simulating the response of wild-type leukemic cells to imatinib administration; our focus is on modelling the time dependence of imatinib-resistant clones upon subsequent exposure to dasatinib or nilotinib. Performance of the three computational schemes to reproduce selected CML patient profiles is assessed. While some simple cases can be approximated by a basic design that does not invoke quiescence, others are more complex and require involvement of non-cycling stem cells for reproduction. We implement a new feedback mechanism for regulation of coupling between cycling and non-cycling stem cell reservoirs that depends on total cell populations. A bifurcation landscape analysis is also performed for solutions to the basic ansatz. Computational models reproducing patient data illustrate potential dynamic mechanisms that may guide optimization of therapy of drug resistant CML.

Data driven polypharmacological drug design for lung cancer: analyses for targeting ALK, MET, and EGFR.

Drug design of protein kinase inhibitors is now greatly enabled by thousands of publicly available X-ray structures, extensive ligand binding data, and optimized scaffolds coming off patent. The extensive data begin to enable design against a spectrum of targets (polypharmacology); however, the data also reveal heterogeneities of structure, subtleties of chemical interactions, and apparent inconsistencies between diverse data types. As a result, incorporation of all relevant data requires expert choices to combine computational and informatics methods, along with human insight. Here we consider polypharmacological targeting of protein kinases ALK, MET, and EGFR (and its drug resistant mutant T790M) in non small cell lung cancer as an example. Both EGFR and ALK represent sources of primary oncogenic lesions, while drug resistance arises from MET amplification and EGFR mutation. A drug which inhibits these targets will expand relevant patient populations and forestall drug resistance. Crizotinib co-targets ALK and MET. Analysis of the crystal structures reveals few shared interaction types, highlighting proton-arene and key CH-O hydrogen bonding interactions. These are not typically encoded into molecular mechanics force fields. Cheminformatics analyses of binding data show EGFR to be dissimilar to ALK and MET, but its structure shows how it may be co-targeted with the addition of a covalent trap. This suggests a strategy for the design of a focussed chemical library based on a pan-kinome scaffold. Tests of model compounds show these to be compatible with the goal of ALK, MET, and EGFR polypharmacology.