Classifying protein kinase structures guides use of ligand-selectivity profiles to predict inactive conformations: structure of lck/imatinib complex

We report a clustering of public human protein kinase structures based on the conformations of two structural elements, the activation segment and the C-helix, revealing three discrete clusters. One cluster includes kinases in catalytically active conformations. Each of the other clusters contains a distinct inactive conformation. Typically, kinases adopt at most one of the inactive conformations in available X-ray structures, implying that one of the conformations is preferred for many kinases. The classification is consistent with selectivity profiles of several well-characterized kinase inhibitors. We show further that inhibitor selectivity profiles guide kinase classification. For example, selective inhibition of lck among src-family kinases by imatinib (Gleevec) suggests that the relative stabilities of inactive conformations of lck are different from other src-family kinases. We report the X-ray structure of the lck/imatinib complex, confirming that the conformation adopted by lck is distinct from other structurally-characterized src-family kinases and instead resembles kinases abl1 and kit in complex with imatinib. Our classification creates new paths for designing small-molecule inhibitors.     

Structural insights into the autoactivation mechanism of p21-activated protein kinase

p21-activated kinases (PAKs) play an important role in diverse cellular processes. Full activation of PAKs requires autophosphorylation of a critical threonine/serine located in the activation loop of the kinase domain. Here we report crystal structures of the phosphorylated and unphosphorylated PAK1 kinase domain. The phosphorylated PAK1 kinase domain has a conformation typical of all active protein kinases. Interestingly, the structure of the unphosphorylated PAK1 kinase domain reveals an unusual dimeric arrangement expected in an authentic enzyme-substrate complex, in which the activation loop of the putative "substrate" is projected into the active site of the "enzyme." The enzyme is bound to AMP-PNP and has an active conformation, whereas the substrate is empty and adopts an inactive conformation. Thus, the structure of the asymmetric homodimer mimics a trans-autophosphorylation complex, and suggests that unphosphorylated PAK1 could dynamically adopt both the active and inactive conformations in solution.     

The immune inhibitory receptor osteoactivin is upregulated in monocyte-derived dendritic cells by BCR–ABL tyrosine kinase inhibitors

Multiple approaches presently aim to combine targeted therapies using tyrosine kinase inhibitors with immunotherapy. Ex vivo-generated dendritic cells are frequently used in such strategies due to their unique ability to initiate primary T-cell immune responses. Besides governing tumor cell growth, many kinases targeted by tyrosine kinase inhibitors are involved in the development and function of dendritic cells and thus tyrosine kinase inhibitor therapy may cause immunoinhibitory side effects. We here report that exposure of developing human monocyte-derived dendritic cells to the BCR–ABL inhibitors imatinib, dasatinib, and nilotinib results in profound upregulation of the transmembrane glycoprotein osteoactivin that has recently been characterized as a negative regulator of T-cell activation. Thus, in line with osteoactivin upregulation, exposure to tyrosine kinase inhibitors resulted in significantly reduced stimulatory capacity of dendritic cells in mixed lymphocyte reactions that could be restored by the addition of blocking anti-osteoactivin antibody. Our data demonstrate that tyrosine kinase inhibitor-mediated inhibition of dendritic cell function is, at least in great part, mediated by upregulation of the immune inhibitory molecule osteoactivin.     

A Src-like inactive conformation in the abl tyrosine kinase domain

The improper activation of the Abl tyrosine kinase results in chronic myeloid leukemia (CML). The recognition of an inactive conformation of Abl, in which a catalytically important Asp-Phe-Gly (DFG) motif is flipped by approximately 180 degrees with respect to the active conformation, underlies the specificity of the cancer drug imatinib, which is used to treat CML. The DFG motif is not flipped in crystal structures of inactive forms of the closely related Src kinases, and imatinib does not inhibit c-Src. We present a structure of the kinase domain of Abl, determined in complex with an ATP-peptide conjugate, in which the protein adopts an inactive conformation that resembles closely that of the Src kinases. An interesting aspect of the Src-like inactive structure, suggested by molecular dynamics simulations and additional crystal structures, is the presence of features that might facilitate the flip of the DFG motif by providing room for the phenylalanine to move and by coordinating the aspartate side chain as it leaves the active site. One class of mutations in BCR-Abl that confers resistance to imatinib appears more likely to destabilize the inactive Src-like conformation than the active or imatinib-bound conformations. Our results suggest that interconversion between distinctly different inactive conformations is a characteristic feature of the Abl kinase domain.     

Isolation of specific and biologically active peptides that bind cells from patients with acute myeloid leukemia (AML)

Imatinib was the first BCR-ABL-targeted agent approved for the treatment of patients with chronic myeloid leukemia (CML) and confers significant benefit for most patients; however, a substantial number of patients are either initially refractory or develop resistance. Point mutations within the ABL kinase domain of the BCR-ABL fusion protein are a major underlying cause of resistance. Of the known imatinib-resistant mutations, the most frequently occurring involve the ATP-binding loop (P-loop). In vitro evidence has suggested that these mutations are more oncogenic with respect to other mutations and wild type BCR-ABL. Dasatinib and nilotinib have been approved for second-line treatment of patients with CML who demonstrate resistance (or intolerance) to imatinib. Both agents have marked activity in patients resistant to imatinib; however, they have differential activity against certain mutations, including those of the P-loop. Data from clinical trials suggest that dasatinib may be more effective vs. nilotinib for treating patients harboring P-loop mutations. Other mutations that are differentially sensitive to the second-line tyrosine kinase inhibitors (TKIs) include F317L and F359I/V, which are more sensitive to nilotinib and dasatinib, respectively. P-loop status in patients with CML and the potency of TKIs against P-loop mutations are key determinants for prognosis and response to treatment. This communication reviews the clinical importance of P-loop mutations and the efficacy of the currently available TKIs against them.     

Dasatinib treatment can overcome imatinib and nilotinib resistance in CML patient carrying F359I mutation of BCR-ABL oncogene

Point mutations of bcr-abl tyrosine kinase are the most frequent causes of imatinib resistance in chronic myeloid leukaemia (CML) patients. In most CML cases with BCR-ABL mutations leading to imatinib resistance the second generation of tyrosine kinase inhibitors (TKI- e.g. nilotinib or dasatinib) may be effective. Here, we report a case of a CML patient who during imatinib treatment did not obtain clinical and cytogenetic response within 12 months of therapy. The sequencing of BCR-ABL kinase domains was performed and revealed the presence of a F359I point mutation (TTC-to-ATC nucleotide change leading to Phe-to-Ile amino acid substitution). After 1 month of nilotinib therapy a rapid progression of clinical symptoms was observed. In the presence of the F359I point mutation only dasatinib treatment overcame imatinib and nilotinib resistance.     

Secondary c‐Kit mutations confer acquired resistance to RTK inhibitors in c‐Kit mutant melanoma cells

Activation of the c‐Kit receptor tyrosine kinase is rare in melanoma, but occurs in 20‐40% of melanoma arising on mucosal membranes, acral skin and skin with chronic sun‐induced damage. Many activating c‐Kit mutations have been shown to be highly sensitive to imatinib mesylate, although the majority of patients with c‐Kit mutant melanoma eventually progress on this inhibitor. We examined acquired resistance to imatinib and the newer generation inhibitor nilotinib in resistant c‐kit mutant melanoma sublines. Four imatinib‐resistant and six nilotinib‐resistant sublines had acquired additional, secondary c‐Kit mutations. The secondary A829P c‐Kit mutation rendered cells resistant to imatinib, but did not suppress the activity of the tyrosine kinase inhibitors nilotinib and dasatinib. Sublines with an additional T670I c‐Kit mutation showed resistance to imatinib, nilotinib and dasatinib, but responded to sunitinib. The concurrent inhibition of the MAPK and PI3K pathways was also effective at promoting apoptosis in the parent and derived resistant sublines. Our data provide a rationale for treating patients with melanoma progressing on imatinib or nilotinib with alternative RTK inhibitors or inhibitors targeting the MAPK and PI3K signalling cascades.     

Real-Time Analysis of Imatinib- and Dasatinib-Induced Effects on Chronic Myelogenous Leukemia Cell Interaction with Fibronectin

Attachment of stem leukemic cells to the bone marrow extracellular matrix increases their resistance to chemotherapy and contributes to the disease persistence. In chronic myelogenous leukemia (CML), the activity of the fusion BCR-ABL kinase affects adhesion signaling. Using real-time monitoring of microimpedance, we studied in detail the kinetics of interaction of human CML cells (JURL-MK1, MOLM-7) and of control BCR-ABL-negative leukemia cells (HEL, JURKAT) with fibronectin-coated surface. The effect of two clinically used kinase inhibitors, imatinib (a relatively specific c-ABL inhibitor) and dasatinib (dual ABL/SRC family kinase inhibitor), on cell binding to fibronectin is described. Both imatinib and low-dose (several nM) dasatinib reinforced CML cell interaction with fibronectin while no significant change was induced in BCR-ABL-negative cells. On the other hand, clinically relevant doses of dasatinib (100 nM) had almost no effect in CML cells. The efficiency of the inhibitors in blocking the activity of BCR-ABL and SRC-family kinases was assessed from the extent of phosphorylation at autophosphorylation sites. In both CML cell lines, SRC kinases were found to be transactivated by BCR-ABL. In the intracellular context, EC50 for BCR-ABL inhibition was in subnanomolar range for dasatinib and in submicromolar one for imatinib. EC50 for direct inhibition of LYN kinase was found to be about 20 nM for dasatinib and more than 10 µM for imatinib. Cells pretreated with 100 nM dasatinib were still able to bind to fibronectin and SRC kinases are thus not necessary for the formation of cell-matrix contacts. However, a minimal activity of SRC kinases might be required to mediate the increase in cell adhesivity induced by BCR-ABL inhibition. Indeed, active (autophosphorylated) LYN was found to localize in cell adhesive structures which were visualized using interference reflection microscopy.     

Structural Mechanisms Determining Inhibition of the Collagen Receptor DDR1 by Selective and Multi-Targeted Type II Kinase Inhibitors

The discoidin domain receptors (DDRs), DDR1 and DDR2, form a unique subfamily of receptor tyrosine kinases that are activated by the binding of triple-helical collagen. Excessive signaling by DDR1 and DDR2 has been linked to the progression of various human diseases, including fibrosis, atherosclerosis and cancer. We report the inhibition of these unusual receptor tyrosine kinases by the multi-targeted cancer drugs imatinib and ponatinib, as well as the selective type II inhibitor DDR1-IN-1. Ponatinib is identified as the more potent molecule, which inhibits DDR1 and DDR2 with an IC₅₀ of 9 nM. Co-crystal structures of human DDR1 reveal a DFG-out conformation (DFG, Asp-Phe-Gly) of the kinase domain that is stabilized by an unusual salt bridge between the activation loop and αD helix. Differences to Abelson kinase (ABL) are observed in the DDR1 P-loop, where a β-hairpin replaces the cage-like structure of ABL. P-loop residues in DDR1 that confer drug resistance in ABL are therefore accommodated outside the ATP pocket. Whereas imatinib and ponatinib bind potently to both the DDR and ABL kinases, the hydrophobic interactions of the ABL P-loop appear poorly satisfied by DDR1-IN-1 suggesting a structural basis for its DDR1 selectivity. Such inhibitors may have applications in clinical indications of DDR1 and DDR2 overexpression or mutation, including lung cancer.     

Dasatinib, imatinib and staurosporine capture compounds - Complementary tools for the profiling of kinases by Capture Compound Mass Spectrometry (CCMS)

Capture Compound Mass Spectrometry (CCMS) is a platform technology for the functional isolation of subproteomes. Here we report the synthesis of two new kinase Capture Compounds (CCs) based on the tyrosine-kinase specific inhibitors dasatinib and imatinib and compare their interaction profiles to that of our previously reported staurosporine-CCs. CCs are tri-functional molecules: they comprise a sorting function (e.g. the small molecule or drug of interest) which interacts with target proteins, a photo-activatable reactivity function to covalently trap the interacting proteins, and a sorting function to isolate the CC-protein conjugates from complex biological samples for protein identification by liquid chromatography/mass spectrometry (LC-MS/MS). We present data of CCMS experiments from human HepG2 cells and compare the profiles of the kinases isolated with dasatinib, imatinib and staurosporine CC, respectively. Dasatinib and imatinib have a more selective kinase binding profile than staurosporine. Moreover, the new CCs allow isolation and identification of additional kinases, complementing the staurosporine CC. The family of kinase CCs will be a valuable tool for the proteomic profiling of this important protein class. Besides sets of expected kinases we identified additional specific interactors; these off-targets may be of relevance in the view of the pharmacological profile of dasatinib and imatinib.     

c-Src binds to the cancer drug imatinib with an inactive Abl/c-Kit conformation and a distributed thermodynamic penalty

The cancer drug imatinib inhibits the tyrosine kinases c-Abl, c-Kit, and the PDGF receptor. Imatinib is less effective against c-Src, which is difficult to understand because residues interacting with imatinib in crystal structures of Abl and c-Kit are conserved in c-Src. The crystal structure of the c-Src kinase domain in complex with imatinib closely resembles that of Abl*imatinib and c-Kit*imatinib, and differs significantly from the inactive "Src/CDK" conformation of the Src family kinases. Attempts to increase the affinity of c-Src for imatinib by swapping residues with the corresponding residues in Abl have not been successful, suggesting that the thermodynamic penalty for adoption of the imatinib-binding conformation by c-Src is distributed over a broad region of the structure. Two mutations that are expected to destabilize the inactive Src/CDK conformation increase drug sensitivity 15-fold, suggesting that the free-energy balance between different inactive states is a key to imatinib binding.     

Impact of BCR-ABL mutations on patients with chronic myeloid leukemia

Therapies that target BCR-ABL in chronic myeloid leukemia, including imatinib, dasatinib and nilotinib, have dramatically improved patient outcome. BCR-ABL mutations, however, contribute to treatment resistance by disrupting drug contact sites or causing conformational changes thus making contact sites inaccessible. Clinical data indicate that developing BCR-ABL mutations during imatinib treatment is predictive for shorter progression-free survival, and that outcomes may depend on mutation type or location. In vitro, dasatinib and nilotinib inhibit most imatinib-resistant BCR-ABL mutations, except for T315I. In clinical studies, other mutations associated with treatment resistance include V299L, T315A, and F317I/L for dasatinib and Y253F/H, E255K/V, and F359C/V for nilotinib. Evaluating patients with clinical signs of resistance for BCR-ABL mutations is an important component of disease monitoring, potentially facilitating selection of subsequent therapy. First-line treatment with dasatinib or nilotinib instead of imatinib may reduce emergence of resistance but novel agents are needed to overcome the problematic T315I mutation.     

Quantitative chemical proteomics reveals mechanisms of action of clinical ABL kinase inhibitors

We describe a chemical proteomics approach to profile the interaction of small molecules with hundreds of endogenously expressed protein kinases and purine-binding proteins. This subproteome is captured by immobilized nonselective kinase inhibitors (kinobeads), and the bound proteins are quantified in parallel by mass spectrometry using isobaric tags for relative and absolute quantification (iTRAQ). By measuring the competition with the affinity matrix, we assess the binding of drugs to their targets in cell lysates and in cells. By mapping drug-induced changes in the phosphorylation state of the captured proteome, we also analyze signaling pathways downstream of target kinases. Quantitative profiling of the drugs imatinib (Gleevec), dasatinib (Sprycel) and bosutinib in K562 cells confirms known targets including ABL and SRC family kinases and identifies the receptor tyrosine kinase DDR1 and the oxidoreductase NQO2 as novel targets of imatinib. The data suggest that our approach is a valuable tool for drug discovery.     

Inhibitors of the Abl kinase directed at either the ATP- or myristate-binding site

The ATP-competitive inhibitors dasatinib and nilotinib, which bind to catalytically different conformations of the Abl kinase domain, have recently been approved for the treatment of imatinib-resistant CML. These two new drugs, albeit very efficient against most of the imatinib-resistant mutants of Bcr–Abl, fail to effectively suppress the Bcr–Abl activity of the T315I (or gatekeeper) mutation. Generating new ATP site-binding drugs that target the T315I in Abl has been hampered, amongst others, by target selectivity, which is frequently an issue when developing ATP-competitive inhibitors. Recently, using an unbiased cellular screening approach, GNF-2, a non-ATP-competitive inhibitor, has been identified that demonstrates cellular activity against Bcr–Abl transformed cells. The exquisite selectivity of GNF-2 is due to the finding that it targets the myristate binding site located near the C-terminus of the Abl kinase domain, as demonstrated by genetic approaches, solution NMR and X-ray crystallography. GNF-2, like myristate, is able to induce and/or stabilize the clamped inactive conformation of Abl analogous to the SH2-Y527 interaction of Src. The molecular mechanism for allosteric inhibition by the GNF-2 inhibitor class, and the combined effects with ATP-competitive inhibitors such as nilotinib and imatinib on wild-type Abl and imatinib-resistant mutants, in particular the T315I gatekeeper mutant, are reviewed.     

Rational design of inhibitors that bind to inactive kinase conformations

The majority of kinase inhibitors that have been developed so far--known as type I inhibitors--target the ATP binding site of the kinase in its active conformation, in which the activation loop is phosphorylated. Recently, crystal structures of inhibitors such as imatinib (STI571), BIRB796 and sorafenib (BAY43-9006)--known as type II inhibitors--have revealed a new binding mode that exploits an additional binding site immediately adjacent to the region occupied by ATP. This pocket is made accessible by an activation-loop rearrangement that is characteristic of kinases in an inactive conformation. Here, we present a structural analysis of binding modes of known human type II inhibitors and demonstrate that they conform to a pharmacophore model that is currently being used to design a new generation of kinase inhibitors.     

Dasatinib preferentially induces apoptosis by inhibiting Lyn kinase in nilotinib-resistant chronic myeloid leukemia cell line

Nilotinib is approved for treatment of newly diagnosed chronic myeloid leukemia (CML) and it is shown superiority over imatinib in first-line treatment for patients of CML. In this study, we established a nilotinib-resistant cell line, K562NR, and evaluated the resistance to nilotinib and efficacy of dasatinib. We found activation of Lyn plays a dominant role in survival of the nilotinib-resistant cell line. We found dasatinib induces the apoptosis of nilotinib-resistant cells and inhibits Lyn kinase activity. This novel nilotinib-resistant CML cell line may help to explore novel therapy for CML.     

Crystal structures of the N-terminal kinase domain of human RSK1 bound to three different ligands: Implications for the design of RSK1 specific inhibitors

The p90 ribosomal S6 kinases (RSKs) also known as MAPKAP-Ks are serine/threonine protein kinases that are activated by ERK or PDK1 and act as downstream effectors of mitogen-activated protein kinase (MAPK). RSK1, a member of the RSK family, contains two distinct kinase domains in a single polypeptide chain, the regulatory C-terminal kinase domain (CTKD) and the catalytic N-terminal kinase domain (NTKD). Autophosphorylation of the CTKD leads to activation of the NTKD that subsequently phosphorylates downstream substrates. Here we report the crystal structures of the unactivated RSK1 NTKD bound to different ligands at 2.0 A resolution. The activation loop and helix alphaC, key regulatory elements of kinase function, are disordered. The DFG motif of the inactive RSK1 adopts an "active-like" conformation. The beta-PO(4) group in the AMP-PCP complex adopts a unique conformation that may contribute to inactivity of the enzyme. Structures of RSK1 ligand complexes offer insights into the design of novel anticancer agents and into the regulation of the catalytic activity of RSKs.     

Enhancement of ABL kinase catalytic efficiency by a direct binding regulator is independent of other regulatory mechanisms

ABL family tyrosine kinases are tightly regulated by autoinhibition and phosphorylation mechanisms. These kinases maintain an inactive conformation through intramolecular interactions involving SH3 and SH2 domains. RIN1, a downstream effector of RAS, binds to the ABL SH3 and SH2 domains and stimulates ABL tyrosine kinase activity. RIN1 binding to the ABL2 kinase resulted in a large decrease in Km and a small increase in Vmax toward an ABL consensus substrate peptide. The enzyme efficiency (k(cat)/Km) was increased more than 5-fold by RIN1. In addition, RIN1 strongly enhanced ABL-mediated phosphorylation of CRK, PSTPIP1, and DOK1, all established ABL substrates but with unique protein structures and distinct target sequences. Importantly RIN1-mediated stimulation of ABL kinase activity was independent of activation by SRC-mediated phosphorylation. RIN1 increased the kinase activity of both ABL1 and ABL2, and this occurred in the presence or absence of ABL regulatory domains outside the SH3-SH2-tyrosine kinase domain core. We further demonstrate that a catalytic site mutation associated with broad drug resistance, ABL1T315I, remains responsive to stimulation by RIN1. These findings are consistent with an allosteric kinase activation mechanism by which RIN1 binding promotes a more accessible ABL catalytic site through relief of autoinhibition. Direct disruption of RIN1 binding may therefore be a useful strategy to suppress the activity of normal and oncogenic ABL, including inhibitor-resistant mutants that confound current therapeutic strategies. Stimulation through derepression may be applicable to many other tyrosine kinases autoinhibited by coupled SH3 and SH2 domains.     

Structural Consequences of Multisite Phosphorylation in the BAK1 Kinase Domain

Multisite phosphorylation is an important mechanism of post-translational control of protein kinases. The effects of combinations of possible phosphorylation states on protein kinase activity are difficult to study experimentally because of challenges in isolating a particular phosphorylation state; surprising little effort on this topic has been expended in computational studies. To understand the effects of multisite phosphorylation on the plant protein kinase brassinosteroid insensitive 1-associated kinase 1 (BAK1) conformational ensemble, we performed Gaussian accelerated molecular dynamics simulations on eight BAK1 mod-forms involving phosphorylation of the four activation-loop threonine residues and binding of ATP-Mg²⁺. We find that unphosphorylated BAK1 transitions into an inactive conformation with a “cracked” activation loop and with the αC helix swung away from the active site. T450 phosphorylation can prevent the activation loop from cracking and keep the αC helix in an active-like conformation, whereas phosphorylation of T455 only slightly stabilizes the activation loop. There is a general trend of reduced flexibility in interlobe motion with increased phosphorylation. Interestingly, the αC helix is destabilized when the activation loop is fully phosphorylated but is again stabilized with ATP-Mg²⁺ bound. Our results provide insight into the mechanism of phosphorylation-controlled BAK1 activation while at the same time represent the first, to our knowledge, comprehensive, comparative study of the effects of combinatorial phosphorylation states on protein kinase conformational dynamics.