Using isoform-specific inhibitors to target lipid kinases

In this issue of Cell, Knight et al. (2006) present a detailed pharmacological analysis of the PI3 kinase (PI3-K) family using small molecule inhibitors. They conclude that p110alpha, a PI3-K isoform, has a critical role in insulin signaling. In a related Cancer Cell paper, Fan et al. (2006) show that blocking activation of both p110alpha and the kinase mTOR with a small molecule inhibitor limits the growth of gliomas.     

Insights into the oncogenic effects of /PIK3CA/ mutations from the structure of p110α/p85α

Phosphatidylinositide-3-kinases (PI3K) initiate a number of signaling pathways by recruiting other kinases, such as Akt, to the plasma membrane. One of the isoforms, PI3Kα, is an oncogene frequently mutated in several cancer types. These mutations increase PI3K kinase activity, leading to increased cell survival, cell motility, cell metabolism, and cell cycle progression. The structure of the complex between the catalytic subunit of PI3Kα, p110α, and a portion of its regulatory subunit, p85α reveals that the majority of the oncogenic mutations occur at the interfaces between p110 domains and between p110 and p85 domains. At these positions, mutations disrupt interactions resulting in changes in the kinase domain that may increase enzymatic activity. The structure also suggests that interaction with the membrane is mediated by one of the p85 domains (iSH2). These findings may provide novel structural loci for the design of new anti-cancer drugs.     

Ras couples phosphoinositide 3-OH kinase to the epithelial Na+ channel

Aldosterone induces the expression of the small G protein K-Ras. Both K-Ras and its 1st effector phosphoinositide 3-OH kinase (PI3-K) are necessary and sufficient for the activation of ENaC increasing channel open probability. The cell signaling mechanism by which K-Ras enhances ENaC activity, however, is uncertain. We demonstrate here that K-Ras significantly activates human ENaC reconstituted in Chinese hamster ovary cells approximately 3-fold. Activation in response to K-Ras was sensitive to the irreversible PI3-K inhibitor wortmannin but not the competitive LY294002 inhibitor of this phospholipid kinase. Similarly, a PI3-K 1st effector-specific Ras mutant (G12:C40) enhanced ENaC activity in a wortmannin but not LY294002 sensitive manner. Constitutively active PI3-K also enhanced ENaC activity but in a wortmannin and LY294002 sensitive manner with the effects of PI3-K and K-Ras not being additive. The activation of ENaC by PI3-K was also sensitive to intracellular GDPbetaS. Constitutively active PI3-K that is incapable of interacting with K-Ras (K227E p110alpha) acted as dominant negative with respect to the regulation of ENaC even in the presence of K-Ras. K-Ras is known to directly interact with PI3-K with aldosterone promoting this interaction. Here we demonstrate that K-Ras also interacts with ENaC through an, as yet, undetermined mechanism. We conclude that K-Ras enhances ENaC activity by localizing PI3-K near the channel and stimulating of PI3-K activity.     

Cancer-specific mutations in phosphatidylinositol 3-kinase

Cancer-specific mutations in the catalytic subunit of phosphatidylinositol 3-kinase (PI3K) p110 alpha occur in diverse tumors in frequencies that can exceed 30%. The majority of these mutations map to one of three hot spots in the gene, and the rest are distributed over much of the PI3K coding sequence. Most of the cancer-specific mutations induce a gain of function that results in oncogenicity, elevated lipid kinase activity and constitutive signaling through the kinases Akt and TOR. The location of the mutations on a model structure of p110 alpha indicates several distinct mechanisms for the gain of function. The mutated p110 alpha proteins are promising cancer targets. Although identification of mutant-specific small-molecule inhibitors seems technically challenging, the therapeutic benefits from such inhibitors could be extremely important.     

The C-terminal conserved domain of DNA-PKcs, missing in the SCID mouse, is required for kinase activity

DNA-PKcs, the catalytic subunit of DNA-dependent protein kinase (DNA-PK), has a phosphoinositol 3-kinase (PI 3-K) domain close to its C-terminus. Cell lines derived from the SCID mouse have been utilised as a model DNA-PKcs-defective system. The SCID mutation results in truncation of DNA-Pkcs at the extreme C-terminus leaving the PI 3-K domain intact. The mutated protein is expressed at low levels in most SCID cell lines, leaving open the question of whether the mutation abolishes kinase activity. Here, we show that a SCID cell line that expresses the mutant protein normally has dramatically impaired kinase activity. We estimate that the residual kinase activity typically present in SCID fibroblast cell lines is at least two orders of magnitude less than that found in control cells. Our results substantiate evidence that DNA-PKcs kinase activity is required for DSB rejoining and V(D)J recombination and show that the extreme C-terminal region of DNA-PKcs, present in PI 3-K-related protein kinases but absent in bona fide PI 3 lipid kinases, is required for DNA-PKcs to function as a protein kinase. We also show that expression of mutant DNA-PKcs protein confers a growth disadvantage, providing an explanation for the lack of DNA-PKcs expression in most SCID cell lines.     

Membrane localization of phosphatidylinositol 3-kinase is sufficient to activate multiple signal-transducing kinase pathways.

Phosphatidylinositol (PI) 3-kinase is a cytoplasmic signaling molecule recruited to the membrane by activated growth factor receptors. The p85 subunit of PI 3-kinase links the catalytic p110 subunit to activated growth factor receptors and is required for enzymatic activity of p110. In this report, we describe the effects of expressing novel forms of p110 that are targeted to the membrane by either N-terminal myristoylation or C-terminal farnesylation. The expression of membrane-localized p110 is sufficient to trigger downstream responses characteristic of growth factor action, including the stimulation of pp70 S6 kinase, Akt/Rac, and Jun N-terminal kinase (JNK). These responses can also be triggered by expression of a form of p110 (p110*) that is cytosolic but exhibits a high specific activity. Finally, targeting of pl10* to the membrane results in maximal activation of downstream responses. Our data demonstrate that either membrane-targeted forms of p110 or a form of p110 with high specific activity can act as constitutively active PI 3-kinases and induce PI 3-kinase-dependent responses in the absence of growth factor stimulation. The results also show that PI 3-kinase activation is sufficient to stimulate several kinases that appear to function in different signaling pathways.     

Activation of phosphoinositide 3-kinase by interaction with Ras and by point mutation.

We have reported previously that Ras interacts with the catalytic subunit of phosphoinositide 3-kinase (PI 3-kinase) in a GTP-dependent manner. The affinity of the interaction of Ras-GTP with p85alpha/p110alpha is shown here to be approximately 150 nM. The site of interaction on the p110alpha and beta isoforms of PI 3-kinase lies between amino acid residues 133 and 314. A point mutation in this region, K227E, blocks the GTP-dependent interaction of PI 3-kinase p110alpha with Ras in vitro and the ability of Ras to activate PI 3-kinase in intact cells. In addition, this mutation elevates the basal activity of PI 3-kinase in intact cells, suggesting a direct influence of the Ras binding site on the catalytic activity of PI 3-kinase. Using an in vitro reconstitution assay, it is shown that the interaction of Ras-GTP, but not Ras-GDP, with PI 3-kinase leads to an increase in its enzymatic activity. This stimulation is synergistic with the effect of tyrosine phosphopeptide binding to p85, particularly at suboptimal peptide concentrations. These data show that PI 3-kinase is regulated by a number of mechanisms, and that Ras contributes to the activation of this lipid kinase synergistically with tyrosine kinases.     

Autophosphorylation of p110delta phosphoinositide 3-kinase: a new paradigm for the regulation of lipid kinases in vitro and in vivo

Phosphoinositide 3-kinases (PI3Ks) are lipid kinases which also possess an in vitro protein kinase activity towards themselves or their adaptor proteins. The physiological relevance of these phosphorylations is unclear at present. Here, the protein kinase activity of the tyrosine kinase-linked PI3K, p110delta, is characterized and its functional impact assessed. In vitro autophosphorylation of p110delta completely down-regulates its lipid kinase activity. The single site of autophosphorylation was mapped to Ser1039 at the C-terminus of p110delta. Antisera specific for phospho-Ser1039 revealed a very low level of phosphorylation of this residue in cell lines. However, p110delta that is recruited to activated receptors (such as CD28 in T cells) shows a time-dependent increase in Ser1039 phosphorylation and a concomitant decrease in associated lipid kinase activity. Treatment of cells with okadaic acid, an inhibitor of Ser/Thr phosphatases, also dramatically increases the level of Ser1039-phosphorylated p110delta. LY294002 and wortmannin blocked these in vivo increases in Ser1039 phosphorylation, consistent with the notion that PI3Ks, and possibly p110delta itself, are involved in the in vivo phosphorylation of p110delta. In summary, we show that PI3Ks are subject to regulatory phosphorylations in vivo similar to those identified under in vitro conditions, identifying a new level of control of these signalling molecules.     

PKCalpha reduces the lipid kinase activity of the p110alpha/p85alpha PI3K through the phosphorylation of the catalytic subunit

The modulation of phosphoinositide 3-kinase (PI3K) activity influences the quality of cellular responses triggered by various receptor tyrosine kinases. Protein kinase C (PKC) has been reported to phosphorylate signalling molecules upstream of PI3K and thereby it may affect the activation of PI3K. Here, we provide the first evidence for a direct effect of a PKC isoenzyme on the activity of PI3K. PKCalpha but not PKCepsilon phosphorylated the catalytic subunit of the p110alpha/p85alpha PI3K in vitro in a manner inhibited by the PKC inhibitor bisindolylmaleimide I (BIM I). The incubation of PI3K with active PKCalpha resulted in a significant decrease in its lipid kinase activity and this effect was also attenuated by BIM I. We conclude that PKCalpha is able to modulate negatively the lipid kinase activity of the p110alpha/p85alpha PI3K through the phosphorylation of the catalytic subunit.     

A pharmacological map of the PI3-K family defines a role for p110alpha in insulin signaling

Phosphoinositide 3-kinases (PI3-Ks) are an important emerging class of drug targets, but the unique roles of PI3-K isoforms remain poorly defined. We describe here an approach to pharmacologically interrogate the PI3-K family. A chemically diverse panel of PI3-K inhibitors was synthesized, and their target selectivity was biochemically enumerated, revealing cryptic homologies across targets and chemotypes. Crystal structures of three inhibitors bound to p110gamma identify a conformationally mobile region that is uniquely exploited by selective compounds. This chemical array was then used to define the PI3-K isoforms required for insulin signaling. We find that p110alpha is the primary insulin-responsive PI3-K in cultured cells, whereas p110beta is dispensable but sets a phenotypic threshold for p110alpha activity. Compounds targeting p110alpha block the acute effects of insulin treatment in vivo, whereas a p110beta inhibitor has no effect. These results illustrate systematic target validation using a matrix of inhibitors that span a protein family.     

Synthesis and biological evaluation of 4-morpholino-2-phenylquinazolines and related derivatives as novel PI3 kinase p110alpha inhibitors

A series of 4-morpholino-2-phenylquinazolines and related derivatives were prepared and evaluated as inhibitors of PI3 kinase p110alpha. In this series, the thieno[3,2-d]pyrimidine derivative 15e showed the strongest inhibitory activity against p110alpha, with an IC(50) value of 2.0 nM, and inhibited proliferation of A375 melanoma cells with an IC(50) value of 0.58 microM. Moreover, 15e was found to be selective for p110alpha over other PI3K isoforms and protein kinases, making it the first example of a selective PI3K p110alpha inhibitor.     

eIF4E binding protein 1 and H-Ras are novel substrates for the protein kinase activity of class-I phosphoinositide 3-kinase

Class-I phosphoinositide 3-kinases (PI 3-kinases) are dual specificity enzymes that possess both lipid and protein kinase activity. While the best characterized property of this protein kinase is as an autokinase activity, there have also been reports it can phosphorylate exogenous substrates including peptides, IRS-1 and PDE-3B. The identification of two novel potential protein substrates of PI 3-kinase is described here. By employing in vitro kinase assays using recombinant proteins as the substrates, it is shown that the translational regulator 4EBP1 becomes phosphorylated by the p110alpha and p110gamma isoforms of class-I PI 3-kinases. The lipid kinase activity of both these isoforms is increased by allosteric binding of H-Ras or betagamma subunits of heterotrimeric G proteins, but we find this is not the case for the protein kinase activity. Surprisingly though, a site on H-Ras is phosphorylated by p110alpha and p110gamma. This raises the possibility that these proteins could serve as physiological substrates for the protein kinase activity of PI 3-kinase and suggests this activity operates in a physiological context by phosphorylating substrates other than the PI 3-kinase itself. This may be particularly important in regulating the interaction of Ras with PI 3-kinase.     

Isoform Specific Inhibitors of PI3 Kinase in Glioma

The PI3 kinase pathway is among the most frequently activated signaling pathways in human cancer and represents an attractive target for small molecule inhibitor based therapies. The PI3Ks show considerable diversity however, and it remains unclear which kinases in this family should be targeted in cancer. We recently screened a panel of potent and structurally diverse drug-like molecules that target this enzyme family in glioma, a malignancy that shows frequent activation of PI3K signaling. Although PI3K alpha was the major isoform driving malignant progression in glioma, blockade of PI3K alpha was not sufficient to maximally inhibit glioma cells. A single agent that inhibited both PI3K alpha and mTOR targeted two points in a pathway with multiple levels of feedback, and was essential for shutting down the proliferation of glioma cells. This result suggests a potentially effective strategy for cancer therapy based on dual inhibition of these two PI3K family members.     

A switch in the kinase domain of rat testis 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase.

The bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase plays an essential role in the regulation of glucose metabolism by both producing and degrading Fru-2,6-P(2) via its distinct catalytic activities. The 6-PF-2-K and Fru-2,6-P(2)ase active sites are located in separate domains of the enzyme. The kinase domain is structurally related to the superfamily of mononucleotide binding proteins that includes adenylate kinase and the G-proteins. We have determined three new structures of the enzymatic monomer, each with a different ligand in the ATP binding site of the 6-PF-2-K domain (AMP-PNP, PO(4), and water). A comparison of these three new structures with the ATPgammaS-bound 6-PF-2-K domain reveals a rearrangement of a helix that is dependent on the ligand bound in the ATP binding site of the enzyme. This helix motion dramatically alters the position of a catalytic residue (Lys172). This catalytic cation is analogous to the Arg residue donated by the rasGAP protein, and the Arg residue at the core of the GTP or GDP sensing switch motion seen in the heterotrimeric G-proteins. In addition, a succinate molecule is observed in the Fru-6-P binding site. Kinetic analysis of succinate inhibition of the 6-PF-2-K reaction is consistent with the structural findings, and suggests a mechanism for feedback inhibition of glycolysis by citric acid cycle intermediates. Alterations in the 6-PF-2-K kinetics of several proteins mutated near both the switch and the succinate binding site suggest a mode of communication between the ATP- and F6P binding sites. Together with these kinetic data, these new structures provide insights into the mechanism of the 6-PF-2-K activity of this important bifunctional enzyme.     

Effects of oncogenic p110alpha subunit mutations on the lipid kinase activity of phosphoinositide 3-kinase

The PIK3CA gene, encoding the p110alpha catalytic subunit of Class IA PI3Ks (phosphoinositide 3-kinases), is frequently mutated in many human tumours. The three most common tumour-derived alleles of p110alpha, H1047R, E542K and E545K, were shown to potently activate PI3K signalling in human epithelial cells. In the present study, we examine the biochemical activity of the recombinantly purified PI3K oncogenic mutants. The kinetic characterizations of the wt (wild-type) and the three 'hot spot' PI3K mutants show that the mutants all have approx. 2-fold increase in lipid kinase activities. Interestingly, the phosphorylated IRS-1 (insulin receptor substrate-1) protein shows activation of the lipid kinase activity for the wt and H1047R but not E542K and E545K PI3Kalpha, suggesting that these mutations represent different mechanisms of lipid kinase activation and hence transforming activity in cancer cells.     

Enzyme activity effects of N-terminal His-tag attached to catalytic sub-unit of phosphoinositide-3-kinase

NTT (N-terminal tags) on the catalytic (p110) sub-unit of PI 3-K (phosphoinositol 3-kinase) have previously been shown to increase cell signalling and oncogenic transformation. Here we test the impact of an NT (N-terminal) His-tag on in vitro lipid and protein kinase activity of all class-1 PI 3-K isoforms and two representative oncogenic mutant forms (E545K and H1047R), in order to elucidate the mechanisms behind this elevated signalling and transformation observed in vivo. Our results show that an NT His-tag has no impact on lipid kinase activity as measured by enzyme titration, kinetics and inhibitor susceptibility. Conversely, the NT His-tag did result in a differential effect on protein kinase activity, further potentiating the elevated protein kinase activity of both the helical domain and catalytic domain oncogenic mutants with relation to p110 phosphorylation. All other isoforms also showed elevated p110 phosphorylation (although not statistically significant). We conclude that the previously reported increase in cell signalling and oncogenic-like transformation in response to p110 NTT is not mediated via an increase in the lipid kinase activity of PI 3-K, but may be mediated by increased p110 autophosphorylation and/or other, as yet unidentified, intracellular protein/protein interactions. We further observe that tagged recombinant protein is suitable for use in in vitro lipid kinase screens to identify PI 3-K inhibitors; however, we recommend that in vivo (including intracellular) experiments and investigations into the protein kinase activity of PI 3-K should be conducted with untagged constructs.     

Phosphatidylinositol 3-Kinase: Increased Activity and Protein Level in Amyotrophic Lateral Sclerosis

Abstract: Enzyme activities and protein levels of several protein and lipid kinases were measured in postmortem tissue from patients who died with amyotrophic lateral sclerosis (ALS) as well as from control subjects. Patients who died with ALS had increased activities and protein levels of phosphatidylinositol 3-kinase (PI 3-K) in particulate fractions of spinal cord tissue compared with control subjects. The PI 3-K activity increased with PI 3-K protein level, indicating no change in specific PI 3-K activity in ALS. No differences in PI 3-K activities were found in cytosolic fractions of spinal cord, or in motor and visual cortices, from ALS patients compared with those from controls. PI 3-K activities and protein levels were unchanged in brain tissue from patients who died with Alzheimer's disease compared with those from controls. PI 3-K is a lipid kinase that is important for cell survival and is activated in response to many growth factors. Increased PI 3-K activities in particulate fractions of spinal cord from ALS patients may be related to the increase of PI 3-K protein levels found in this tissue. The protein kinases Erk2, protein kinase B (PKB), and p70 ribosomal S6 kinase (S6K) showed no differences in activities in spinal cord tissue between ALS patients and controls. However, the amounts of PKB and S6K protein were significantly higher in ALS patients, whereas Erk2 protein amount was unchanged compared with controls. Protein kinase C activity was increased in spinal cord tissue from ALS patients, which is consistent with our previous report. The increased activity of PI 3-K in spinal cord tissue from patients with ALS implicates the involvement or activation of PI 3-K in ALS, as either a cause or a consequence of the neuron loss. The lack of up-regulation in the activities of PKB and S6K in ALS tissue supports an impairment in signal transduction cascades mediated by PI 3-K in this neurodegenerative disease.     

Ras couples phosphoinositide 3-OH kinase to the epithelial Na channel

Aldosterone induces the expression of the small G protein K-Ras. Both K-Ras and its 1st effector phosphoinositide 3-OH kinase (PI3-K) are necessary and sufficient for the activation of ENaC increasing channel open probability. The cell signaling mechanism by which K-Ras enhances ENaC activity, however, is uncertain. We demonstrate here that K-Ras significantly activates human ENaC reconstituted in Chinese hamster ovary cells ∼3-fold. Activation in response to K-Ras was sensitive to the irreversible PI3-K inhibitor wortmannin but not the competitive LY294002 inhibitor of this phospholipid kinase. Similarly, a PI3-K 1st effector-specific Ras mutant (G12:C40) enhanced ENaC activity in a wortmannin but not LY294002 sensitive manner. Constitutively active PI3-K also enhanced ENaC activity but in a wortmannin and LY294002 sensitive manner with the effects of PI3-K and K-Ras not being additive. The activation of ENaC by PI3-K was also sensitive to intracellular GDPβS. Constitutively active PI3-K that is incapable of interacting with K-Ras (K227E p110α) acted as dominant negative with respect to the regulation of ENaC even in the presence of K-Ras. K-Ras is known to directly interact with PI3-K with aldosterone promoting this interaction. Here we demonstrate that K-Ras also interacts with ENaC through an, as yet, undetermined mechanism. We conclude that K-Ras enhances ENaC activity by localizing PI3-K near the channel and stimulating of PI3-K activity.     

Activation loop sequences confer substrate specificity to phosphoinositide 3-kinase alpha (PI3Kalpha ). Functions of lipid kinase-deficient PI3Kalpha in signaling

Phosphoinositide 3-kinases (PI3Ks) are dual specificity lipid and protein kinases. While the lipid-dependent PI3K downstream signaling is well characterized, little is known about PI3K protein kinase signaling and structural determinants of lipid substrate specificity across the various PI3K classes. Here we show that sequences C-terminal to the PI3K ATP-binding site determine the lipid substrate specificity of the class IA PI3Kalpha (p85/p110alpha). Transfer of such activation loop sequences from class II PI3Ks, class III PI3Ks, and a related mammalian target of rapamycin (FRAP) into p110alpha turns the lipid substrate specificity of the resulting hybrid protein into that of the donor protein, while leaving the protein kinase activity unaffected. All resulting hybrids lacked the ability to produce phosphatidylinositol 3,4,5-trisphosphate in intact cells. Amino acid substitutions and structure modeling showed that two conserved positively charged (Lys and Arg) residues in the activation loop are crucial for the functionality of class I PI3Ks as phosphatidylinositol 4,5-bisphosphate kinases. By transient transfecion of 293 cells, we show that p110alpha hybrids, although unable to support lipid-dependent PI3K signaling, such as activation of protein kinase B/Akt and p70(S6k), retain the capability to associate with and phosphorylate insulin receptor substrate-1, with the same specificity and higher efficacy than wild type PI3Kalpha. Our data lay the basis for the understanding of the class I PI3K substrate selectivity and for the use of PI3Kalpha hybrids to dissect PI3Kalpha function as lipid and protein kinase.