WO1997015658A1 - Wortmannin phosphoinositide 3-kinase interaction site - Google Patents

Abstract

The site of interaction of wortmannin on the catalytic subunit of PI3-kinase, p110α is identified. At physiological pH (6.5-8) wortmannin reacted specifically with p110α. PtdIns (4.5)P2, ATP and ATP-analogs (adenine and 5'(4-fluorosulfonylbenzoyl)-adenine, FSBA) completed effectively with wortmannin, while substances containing nucleophilic amino acid side chain functions had no effect at the same concentrations. Proteolytic fragments of wortmannin-treated, recombinant p110α are mapped using anti-wortmannin and anti-p110α peptide antibodies, thus limiting the target site within a 10 kD fragment, co-localizing with the ATP binding site. Site-directed mutagenesis of all candidate residues within this region shows that only the conservative Lys802 to Arg mutation abolished wortmannin binding. Inhibition of PI3-kinase occurs therefore by the formation of an enamine following the attack of Lys802 on the furan ring (at C20) of wortmannin. The Lys802Arg mutant was also unable to bind FSBA, and was catalytically inactive in lipid and protein kinase assays, indicating a crucial role for Lys802 in the phosphotransfer reaction. In contrast, an Arg916Pro mutation abolished the catalytic activity, while covalent wortmannin binding remained intact.

Description

WORTMANNIN PHOSPHOINOSITIDE 3-KINASE INTERACTION SITE

The present invention relates to phosphoinositide (PI) 3-kinases, and also homologs and analogues thereof, and the modulation of the catalytic activity of these molecules by modulator ligands. Particularly, the invention relates to the wort armin interaction with PI3-kinase and its inhibitory effects on PI3 kinase activity. More particularly the invention concerns the molecular structure of the PI3 kinase interaction site, its determination and manipulation and also ligands identified or designed having regard to the molecular structure of said interaction site, and yet further still manipulation of inhibitors which can interact with said site. In addition, the invention concerns modelling techniques for imaging or viewing the interaction between a PI3 kinase interaction site and ligands adapted to bind thereto.

The expanding family of Ptdlns 3-kinases consists of enzymes composed of various catalytic subunits of the pl lOα (26,75), pllOβ

(27), pl lOγ (65) and Vps34p type (56,74). The pllOα and β subtypes form tight heterodimers with a p85 regulatory subunit, which embodies a SH3, two SH2 and a BCR domain (17,46,60). Interaction between pl lOα or β and p85 is mediated through the N-terminus of the catalytic subunit and the inter-SH2 region of p85 (14). This heterodimeric phosphoinositide 3-kinase (PI 3-kinase) is activated during the translocation to autophosphorylated growth factor receptors

(30) or their substrates (e.g. the insulin receptor substrate- 1), where the

SH2 domains of p85 interact with phosphorylated YXXM motives ((2,19,32,48), reviews in (11,80)). Other PI 3-kinases seem to propagate signals from seven transmembrane helix receptors as they are activated by G-protein βγ-subunits (62,65). The first of this kind has been cloned recently and named pl lOγ (65). It has been shown that PI 3-kinases are the teπriinal enzymes in the synthesis of D-3 phosphorylated phosphoinositides (24,63) and that these Upids are poor substrates for phosphatidylinositol specific phospholipases C (58). PI 3-kinases produce therefore novel lipid second messengers that are believed to be a key step in receptor signalling by growth factors, cytokines and hormones (30,71).

The yeast Vps34 gene product represents yet another class of Ptdlns 3- kinases: associated with and activated by a Vpsl5p serme/threonine kinase, Vps34p phosphorylates solely Ptdlns to Ptdlns 3-P and, in contrast to the pl lOα-γ, does not accept Ptdlns 4-P and PtdTns(4,5) ₂ as a substrate (56). Point mutations which the conserved ATP-binding site of Vps34p result in loss of Ptdlns 3-kinase activity and lead to the missorting of proteins otherwise targeted to the yeast vacuole, which demonstrates that a functional lipid kinase is essential for cellular sorting processes (25,61).

In higher eukaryotes, manipulation of PI 3-kinase activity was achieved by mutation of the docking sites of the p85 regulatory subunits on growth factor receptors (36,78) or overexpression of dominant negative p85 (Δp85) (77). These labour intensive approaches, however, are limited to growth factor-mediated activation of the lipid kinase. The use of chemical, low molecular weight mhibitors of PI 3-kinase, on the other hand, provides a fast and easy way to explore the importance of this enzyme in any cell surface receptor-mediated signalling pathway. Wortmannin has recentiy been described as a potent and specific inhibitor of PI 3-kinase (1,45,82,83) and has since been widely used because of its advantageous properties.

Wortmannin is cell permeable, commercially available and has, at concentrations where its fully inhibits PI 3-kinase, very little effects on other signalling molecules. At submicromolar concentrations, wortmannin does not interfere with the activity of protein kinase C (PKC), calmodulin-dependent, cAMP-dependent and cGMP-dependent protein kinases (42), mitogen activated protein kinase (MAPK, (44)), _p70S6kin^ase ^Q-J ^ _{Λe PDG}p _receptor tyrosine kinase (82). Under these conditions, wortmannin has no influence on the levels of Ptdlns 4-P and PtdIns(4,5)_P₂ in resting cells (1), and was reported not to inhibit Ptdlns 4- kinase in vitro (45), but to inhibit a novel, weekly membrane-associated Ptdlns 4-kinase at elevated concentrations (41). The release of calcium from intracellular stores in response to serpentine receptor ligand-binding remains unaffected by wortmannin, illustrating that activation of phospholipase C, inositol(l,4,5) ₃ and diacylglycerol production remain intact in the presence of the inhibitor. With an IC₅₀ of about 200 nM, wortmaruiin has recently been shown to inhibit DNA-dependent protein kinase (DNA-PK_CS), which is involved in the control of DNA-repair mechanisms (22).

Wortmannin at nM concentrations, on the other hand, has been shown to inhibit the activation of neutrophil NADPH oxidase by N-formyl-Met-Leu- Phe, complement factor 5a, leukotriene B₄, platelet activation factor and interleukin 8, but has no effect on the phorbol ester-induced response or the enzyme's activity in vitro (1,3,45). Later, platelet-derived growth factor receptor-mediated fibroblast motility and membrane mffling (77,82), histairiine release in mast cells (83), insulin-stimulated glucose uptake in various tissues (29,44,85), nerve growth factor (NGF)-dependent survival and differentiation of PC 12 cells (33,84), platelet aggregation (35) and vesicular sorting of cathepsin D to lysosomes (7,13) were added to the list of wortmarinin-inhibited - and possibly PI 3-kinase controlled cellular processes.

Although it is not yet clear how inactivation of Ptdlns 3-kinase by wortmannin affects these cell responses, a number of candidate downstream enzymatic activities have been found to be affected by the presence of the inhibitor or to be modulated by PtdIns(3,4,5) ^>3. Serum activation of die MAPK pathway, phosphorylations of glycogen synthase kinase-3 (12,76) and p70^S6 kinase (10), phospholipase D activity (5,20,49), the activation of the GTP-binding protein rac (9,23) and the PKC 6,e,τ (69) and ζ (40) isoforms and recently PKB/c-Ark protein kinase (8,18) have been proposed to depend on functional PI 3-kinase.

The family of Ptdlns 3-kinases is constantly expanding, and related proteins include members of the ATM-related genes (for a review see (87)) and the targets of the immunosuppressant rapamycin (TOR, FRAP, RAFT1, (6,37,51,52)) with so far unknown activities. It has been suggested that TORs are downstream of PI 3-kinase due to the differential actions of rapamycin and wortmannin on p70 phosphorylation (8,10). A speculative interaction of wortmannin with TORs was also considered (72). The importance that wortmamrin has gained as a tool to explore Ptdlns 3-kinase signalling demands a better understanding of its inhibitory mechanism. In this work, we identified the reaction site of wortmannin on PI 3-kinase. Additionally, we show that the lysine residue involved in wortmannin binding is likely to play a central role in the phosphotransfer reaction. Together, the results elucidate the inactivation mechanism of Ptdlns 3-kinases by wortmannin and provided a fix point for the alignment of the lipid kinase's catalytic centre with PKA.

It is therefore an object of the invention to exploit information relating to a knowledge of the active site of PI-3 kinase and its homologs or analogues.

Accordingly, a first aspect of the present invention provides an interaction site on PI-3 kinase, or a homolog or analogue thereof, which site when exposed to a modulator modulates the activity of PI3- kinase, the interaction site comprising a molecular shape which is adapted to interact with at least a part of the modulator so as to modulate the PI3- kinase activity.

In a preferred embodiment of the invention said interaction site comprises at least one element capable of forming a covalent linkage with said modulator, ideally the element is a lysine and ideally still the lysine is positioned at position 802 of the PI3 kinase sequence, or an equivalent position in a homolog or analogue thereof. In yet a further preferred embodiment of the invention said interaction site comprises at least one negatively charged element suitably positioned so as to enhance provision of the aforedescribed covalent linkage. Preferably, said negative element is glutamine and it is ideally provided at position 821 of the PI3 kinase sequence, or an equivalent position in an homolog or analogue thereof.

In yet a further preferred embodiment of the invention said interaction site comprises at least one amino acid selected from the group comprising proline, asparagine or isoleucine. Preferably, at least one proline is provided and it is located at position 786 of the sequence structure of PI3 kinase, or an equivalent position in an homolog or analogue thereof. Preferably further still at least one asparagine is provided and it is ideally located at position 787 of the sequence of PI3 kinase, or an equivalent position in an homolog or analogue thereof. Preferably further still at least one isoleucine is provided and it is located at position 788 of the sequence structure of PI3 kinase, or an equivalent position in a homolog or analogue thereof.

A further aspect of the invention indues a method for modulating the activity of PI3 kinase, or a homolog or analogue thereof, comprising deleting or altering any of the afore described features of the interaction site.

According to a yet further aspect of the invention there is provided an anti-body raised against the whole or a part of the said interaction site which antibody is most suitably monoclonal. Alternatively, there is provided a ligand adapted to bind with said interaction site which ligand either comprises, or is suitably provided with, interactive elements designed, or likely, to interact with the aforementioned features of the said interaction site.

The invention also includes a ligand reactive with a PI3-kinase activity modulator, said ligand comprising at least a portion of a PI3-kinase.

In a second aspect the invention provides a modulator of PI3-kinase activity other than wortmannin wherein the modulator is capable of interacting with the wortmannin-inhibition site of PI3-kinase.

The term "Ptdlns 3-kinase" was recently used for the human homolog of Vps34p utilizing solely Ptdlns as a substrate. We therefore use phosphoinositide 3-kinase (PI3-kinase) for enzymes phosphorylating Ptdlns. Ptdlns 4- and Ptdlns (4.5) ₂ in vitro. Ptdlns 3-kinase is also used where a statement applies to the family of these enzymes.

PI3-kinases include enzymes comprising various catalytic sub-units of the pl lOα, pl lOβ, pl lOγ and Vps34p type. The PI3-kinase may therefore comprise the pl lOα or β sub-type combined with the p85 regulatory sub-unit to form heterodimers. P13-kinase activity is the ability to phosphorylate phosphatydylinositol (Ptdlns).

The interaction site or hgand will have a molecular configuration which is adapted to receive at least a portion of the modulator. Preferably, the interaction site or ligand comprises at least one amino acid residue selected from lysine, histidine, asparagine, isoleucine, proline, serine, asparagine or glutamine.

The modulation of the PI3-kinase activity may be manifest in a variety of ways, preferably inhibition (competitive or non-competitive) or stimulation but other manifestations such as changes in kinetic parameters eg V_mJ may arise. The modulation may be manifest in terms of the potential range of phosphorylatable substrates. The range of Ptdlns homologs which are phosphorylatable may be extended by the interaction of the modulator with the modulation site on the PI3-kinase.

The modulator may be an inhibitor, an agonist or an antagonist. The types of modulator are not restricted and may include low molecular weight molecules eg wortmannin or higher molecular weight molecules such as polypeptides and proteins. The polypeptides or proteins may be synthesized chemically or may be produced by recombinant means by expression of an appropriate nucleic acid construct in an appropriate prokaryotic or eukaryotic host cell. The modulator may be an antibody or an antibody fragment which is monoclonal or polyclonal in origin.

The ligand of the second aspect of the invention may comprise a protein or a polypeptide which can be synthesized chemically or produced by recombinant means through expression of an appropriate nucleic acid construct in an appropriate prokaryotic or eukaryotic host cell.

The interaction site or ligand includes functionally equivalent variants due to one or more amino changes eg substitutions. Other variants within the scope of the invention may arise from amino acid changes and substitutions which do not give rise to functional changes. Interaction sites or ligands may vary in structure due to amino acid changes to the extent that no apparent change in PI3-kinase activity takes place but that there is a modulation or inhibition of other aspects such as the inability of the interaction site or hgand to bind covalently to the modulator.

Monoclonal or polyclonal antibodies may be raised against the interaction site or ligand. Alternatively, anti-idiotypic antibodies may be raised via monoclonal or polyclonal routes to monoclonal or polyclonal antibodies reaction against wortmarinin.

In a third aspect the invention provides a method of stabilising the interaction of wortmannin with PI3-kinase comprising contacting wortmannin and PI3-kinase in the presence of a reducing agent. A preferred reducing agent is NaCNBH₃.

In a further aspect of the invention there is provided a method of identifying or designing hgands capable of modulating PI3-kinase activity comprising:

(a) generating a molecular model of the wortmannin-inhibition site of PI3-kinase,

(b) identifying existing ligands or designing new hgands which interact at least in part with the inhibition site model, and optionally,

(c) contacting the putative hgand with PI3-kinase under conditions suitable for monitoring PI3-kinase activity and then monitoring that activity and deterrnining the degree and extent to which the activity is modulated.

In a yet further aspect of the invention there is provided a product identified or designed according to the previous method.

In a further aspect of the invention there is provided a method of generating a molecular model of the wortmannin inhibition site of PI3- kinase comprising:

(a) identifying conserved amino acid residues within me amino acid sequence for PI3-kinase,

(b) predicting the tertiary structure for such conserved residues,

(c) aligning the conserved structural sequences of PI3-kinase with equivalent portions of cAMP-dependent protein kinase (PKA), and

(d) ahgning the lysine residues of the active sites of PI3 -kinase and PKA whilst disregarding the common P-loop structures, (e) determ-Lning the most common conformation for remaήiing amino acid residues and then deteπrjining the thermodynamically most stable conformation for remaining amino acid residues, optionally

(f) inserting the wortmannin structure into the ATP binding site of the model for PI3-kinase, and

(g) adjusting the structure of the model to the most thermodynamically favourable state.

The invention will now be described in detail with reference to the drawings.

Fig. 1. Conditions for covalent modification of PI 3-kinase by wortmannin. Purified recombinant PI 3-kinase pl l0α/p85α complex was incubated with wortmannin as described in materials and methods. Wortmannin-labelled proteins were subsequently applied to SDS-PAGE and probed with anti-wortmarinin antisera on immunoblots. a) Binding of 100 nM wortmannin was achieved at the indicated pH in 0.2 M phosphate buffers, b) PI 3-kinase complex was incubated with increasing concentrations of wortmannin.

Fig. 2. Substrates competing with wortmannm-binding. PI 3-kinase was subjected to wortmannm-labeUing (100 nM) in the presence of the enzyme's substrates Ptdlns (PI), PtdIns(4,5)_P₂, ATP, ATP-analogues and nucleophilic substances. The reaction of PI 3-kinase with wortmannin was verified as in Fig. 1. a) Recombinant pl l0α/p85α complex was added to lipid suspensions made up from the indicated lipids (phosphoinositide mixtures contained additionally one quarter of phosphatidylserine; Triton-X 100 (TX-100) was 0.1%). Wortmannin or DMSO only (-wortmannin) were subsequently added. b) Immobilized GST-pl l0α/p85α complex was preincubated with 1 mM of ATP, 5'-ρ-fluorosulfonylbenzoyladenine (FSBA), adenine, β- mercaptoethanol iβ-ME), ethanolamine (EtOHamine) or β- guanidinopropionic acid (GuPrAc) for 30 min. at 37jC. The samples were cooled to OjC before incubation with wortmannin.

Fig. 3. Partial digests of GST-pll0α/p85 complex by Factor Xa and Glu-C. Wortmannin-labelled PI 3-kinase immobilized on glutathione- sepharose beads was digested as described in material and methods. Beads (B) and supernatants (S) following protease treatment and undigested PI 3-kinase (C), were denatured separately and examined on immunoblots for the presence of wortmannin. The membranes were then stripped and reprobed with anti-pllO C-terminal antibodies.

Fig. 4. Localization of wortmannin-labelled peptides. Immobilized GST-pll0α/p85α complex was incubated with 200 nM wortmannin, digested with indicated proteases (Lys-C 2 μg when not otherwise indicated, trypsin 0.04 μg, Arg-C 2 μg), applied to TRICINE gel electrophoresis and immunoblotting. a) Detection of wortmannin- labelled peptides with anti-wortmannin antisera (W) or peptides including the C-terminus with anti-pllO C-terminal antiserum (CT). b) Lys-C concentration dependent formation of wortmannin-labelled peptides (α-wortmannin) or peptides containing the internal sequence 734-748 of pllOα (α-pllO 734-48).

Triangles mark the position of peptide molecular weight standards (to the left: SIGMA 2.5-17 kD, to the right: Pharmacia 14-94 kD).

Fig. 5. The wortmannin target site deduced from digest patterns as displayed in Figs. 3 and 4. Selected peptides were aligned with full length pl lOα using the C-terminus or the internal sequence 734-748 as markers. Peptides detected by anti-wortmannin antibodies are displayed in gray. Numbers indicate the apparent molecular weight in kD (*expected molecular weights) as obtained by the indicated proteolytic treatment.

Fig. 6. Wortmannin- and FSBA-binding on overexpressed pl lOα Lys/ Arg mutants, a-d) Human embryonic kidney 293 cells were co- transfected with p85 and pllOα DNA as indicated at the bottom of d. PI 3-kinase was immunoprecipitated from cell lysates with monoclonal anti-p85 antibodies (a, b) or anti-pllOα rabbit antisera (c,d; see methods for details). Expression of p85α or pl lOα was verified by staining for total protein (a,c). a) Coomassie blue-stained SDS-PAGE of immunoprecipitates from 293 cells. b) anti-p85α immunoprecipitates were incubated with 100 nM wortmannin and probed subsequently on immunoblots for covalent binding of wortmannin using anti-wortmannin antisera. c) Colloidal gold total protein staining of the PVDF-membrane after wortmannin detection in d. d) anti-pllOα immunoprecipitates treated as the samples in b). e) Recombinant wild type and mutant pi 10α/85 complexes were isolated from insect cells on glutathione (wt) or phosphotyrosine beads (KR). Samples were labelled with 200 μM FSBA and subsequent immunoblots were developed with anti-FSBA antibodies (left), before the membranes were stained for total protein (right) as in c. wt: wild type pl lOα, KRl: pll0 ^K733H, KR2: _P110α^C769s/K776R, KR3: pll0α^K802R, KR4: _pπθα^{C862S K863/K867R}.

Fig. 7. Lipid and protein kinase activity of pllOα lysine mutants. PI 3-kinase was immunoprecipitated with anti-p85α antisera from 293 cells transfected with p85α and pllOα DNA as in Fig. 6. a) Immunoprecipitates were subjected to PI 3-kinase activity assays. The radioactivity incorporated in [³²P]-Ptdlns-3P is displayed in relation to the activity obtained from cells transfected with p85α DNA only (n = 3, ^±SEM). b) Immunoprecipitates were assayed for p85α phosphorylation.

Fig. 8. Covalent reaction of wortmannin with an inactive PI 3-kinase complex, a) Purified recombinant GST-pl l0α/p85 or Arg→Pro mutant GST-pl lO /p85 complex (twice the amount of protein) were exposed to the indicated concentrations of wortmannin in the presence of 0.1 % Triton X-100 before SDS-PAGE and immunoblotting with anti-wortmannin antibodies, b) [³²P]-Ptdlns-3 formed by GST-pl l0α/p85α (wt) or the GST-pl l0α^R9,6P/p85α (RP) complex in a PI 3-kinase assay.

Fig. 9. Modelling of the wortmannin-pllOα complex, a) Three dimensional representation of the back-bone of pi 10a with wortmannin docked into the putative catalytic site. Lys802 is represented by balls- and-sticks and is coloured blue, while the stabilizing Glu821 is shown in red. Wortmannin is represented as van der Waals spheres in white. b) A schematic representation of the wortmannin molecule and protein amino acids that surround it within the binding site. Dashed lines indicate possible hydrogen bonds and/or electrostatic interactions between residues and wortmannin. Half circles indicate hydrophobic interactions.

Material and Methods:

Anti-wortmannin, anti-pllO antisera and immunoblots: Polyclonal antibodies were raised in rabbits. Wortmannin was dissolved in 50 μ\ dimethylsulfoxide at a concentration of 40 mg/ml and subsequently mixed with 950 μl of 2 mg bovine serum albumin (BSA) in phosphate buffered saline (PBS, 8mM Na₂HP0₄.2H₂0,1.4 mM KH₂P0₄, 2.6 mM KC1, 136 mM NaCl, pH 7.4) for 24" at room temperature. Serum collected after 5 injections of wortmannin-BSA conjugate was passed over a BSA-sepharose column to remove anti-BSA antibodies. Anti-FSB A antibodies were produced in a similar way. The peptides F L V E Q M R R P D F M D A L - C y s a n d C y s - GGWTTKMDWIFHTIKQHALN corresponding to amino acids 734- 748 and 1049-1068 (C-terminus) of pllOα, respectively, were coupled to keyhole limpet hemocyanine using the bifunctional linker Sulfosuccinimidy 1 4- (N-maleimidomethy 1) cy clohexane- 1 -carboxylate (SMCC, Pierce). IgG fractions from antipeptide antisera were purified on y-bind sepharose (Pharmacia). Anti-p85α rabbit antisera were obtained by immunisation with a (His)₆-tag fusion protein containing amino acids 6-112 of bovine p85α.

Wortmannin-labelling of PI 3-kinase, proteolytic digests:

Immobilized GST-PI 3-kinase fusion protein or soluble p85/pll0α complex was taken up in PBS and usually incubated with 100 nM wortmannin on ice for 15 min. PI 3-kinase immobilized on glutathione or phosphotyrosine beads was washed with PBS/Triton X-100 0.5% to remove excess inhibitor before denaturation, while soluble protein was either denatured directly or precipitated according to (79) . See legends to figures for modified conditions and wortmannin concentrations. Where not indicated otherwise, samples were denatured in the presence of /3-mercaptoethanol and applied to standard SDS-PAGE (38) and transferred to PVDF membranes (Millipore) according to (70). Primary antibodies were detected by goat anti-rabbit horseradish peroxidase-conjugate (SIGMA) and enhanced chemiluminescence (ECL, Amersham). Reactions of 200 μM FSBA with PI 3-kinase complex, immobilized on glutathione or phosphotyrosine beads, were carried out in PBS with 0.1% Triton X-100 for 15 min. at 37; C. The beads were washed 4x with PBS/0.5% Triton X-100 and denatured in sample buffer containing additionally 5mM DTT.

PI 3-kinase samples for digestion with various proteases were labelled with 200 nM wortmannin as indicated above. Subsequently NaCNBH₃ was added (6.5 mM) for lh, to reduce and stabilize the Schiff-base formed between the kinase and wortmannin. Whenever soluble PI 3- kinase was used, labelling was performed in the respective digestion buffers. GST-PI 3-kinase was labelled and reduced in PBS, the beads were subsequently washed twice with digestion buffer. Large fragments were obtained with factor Xa (from Promega, 1.5μg protease in 0.1 M NaCl, 20 mM TRIS HC1 pH 8.0) and Glu-C {S. aureus V8 protease, Boehringer, 0.4 μg in 50 mM NH HC0₃ , pH 7.8) and were separated on 10% SDS-PAGE. Peptides smaller than 30 kD were generated with Lys-C (Promega, 2 μg in 25 mM TRIS, lmM EDTA, pH 7.7) , trypsin (SIGMA, 0.04 μg in 50 mM TRIS, 0.1 mM CaCl₂, pH 7.8) and Arg-C (clostripain, Promega, 0.2 μg; the kinase was denatured in 1 mM DTT, 1 M Urea, 20 mM TRIS, pH 7.8 for 5 min. at 90 jC and subsequently digested) . Digestions were carried out at 37 jC for 3 h in a total volume of 80 μl, before the products were denatured and separated on TRICINE gels prepared as described by (57).

Introduction of point mutations into pl lOα: The Transformer Site Directed Mutagenesis Kit from Clontech with Afl III/ Bgl II trans oligo primers and the respective mutagenesis primers were used to produce the desired Lys- > His, Lys- > Arg and Cys- > Ser codon exchanges in the pllOα cDNA, cloned as a BamHI-Hindlll fragment (26) in the multicloning site of pUC19, and to introduce at the same time new restriction sites for plasmid selection (the mutants were named KRl : K733H, KR2: C769S/K776R, KR3: K802R). The triple mutant KR4 (C862S/K863R/K867R) was produced by the ligation of a synthetic adaptor into the Hinf I/Pst I region of pllOα DNA (corresponding to amino acids 860-871).

All nucleotide exchanges were confirmed by DNA sequenci? (Sequenase 2.0, US Biochemicals). The respective mutations were subsequently reintroduced as Xbal-PstI or BbsI-PstI fragments into wild type pllOα harboured in the pSCTl and p36C expression vectors (see below) to generate the desired mutant proteins. General procedures were carried out as described in (53).

Production of recombinant proteins: For transient expression of the p85 /pll0α Ptdlns complex, human embryo kidney 293 cells (ATCC) were transfected with 1 μg of adenovirus late promoter expression vector pMT2 (31) containing the coding sequence of p85 (26,46) and 10 μg of the CMV promoter-based expression vector pSCTl (50) carrying the wild type and KR mutants of pllOα. Expression vectors and 5 μg calf thymus carrier DNA were preprecipitated for 10 min. in a calcium phosphate solution and this was subsequently added to 10 cm petri dishes of 293 cells cultured in DMEM (Gibco/BRL) with 5% FCS (the detailed procedure will be published elsewhere). The medium was exchanged after 12 h and cells were lysed 48 h later in 20 mM Tris pH 8.0, 138 mM NaCl, 2.7 mM KC1, supplemented with 5% glycerol, 1 mM MgCl₂, 1 mM CaCl₂, 5 mM EDTA, 20 mM NaF, 1 mM sodium-o-vanadate, 20 μM leupeptin, 18 μM pepstatin and 1% NP-40. Cleared cell lysates (12'000g, 15 min. 4jC) were subjected to immunoprecipitation, wortmannln-binding and kinase assays. The protein applied to SDS-PAGE or kinase assays corresponded typically to V₄ of a 10 cm petri dish. Endogenous p85 could not be detected in 293 cells, while p85/3 was present (T4 anti-p85/3 antibodies were kindly given by I. Gout). Recombinant pllOα bacculo virus was essentially produced as in (26). Briefly, Sf9 cells cultured in EPL41 (Gibco/BRL) with 10% FCS were cotransfected with 5 μg of each p36C transfer vector (47) containing mutated pl lOα cDNA and 0.25 μg linear BacculoGold DNA (Pharmingen) essentially as described by the manufacturer. Recombinant virus was plaque purified and amplified (43) . For protein production, cells were harvested 60 h post transfection with recombinant pllOα and p85α virus, centrif uged and lysed by mechanical disruption in a Dounce homogenizer in 20 mM HEPES, 2 mM DTT, 10 mM NaF, 100 μM Na₃V0₄, 5 mM EDTA, pH 7.5 supplemented with the protease inhibitors PMSF (ImM), aprotinin (lOμg/ml), leupeptin (20μM) and pepstatin (18μM) . The lysate was cleared by ultracentrifugation and the supernatant was incubated with phosphotyrosine-actigel beads (Sterogene). The beads with the associated p85 /pl l0α PI 3-kinase complex were washed (20 mM phosphate buffer pH 7.5 with 0.2 M NaCl) and used directly for experiments. The production of the p85 /GST-pll0α^R916P is described in (15).

Immunoprecipitations: Recombinant PI 3-kinase subunits p85α, pl lOα or complexes were immunoprecipitated from 293 cell ly sates with anti-C terminal antibodies and anti-p85 antisera using Protein A- Sepharose (Pharmacia), or with monoclonal anti-p85α antibodies (U13, a kind gift from I. Gout) using anti-mouse IgG agarose beads (Sigma) to immobilize the immunecomplexes. Precipitates were subsequently washed 3 times with 0.1 M Tris-HCl, pH 7.4, 0.5 M LiCl, followed by three washing steps with the respective reaction buffers. PI 3-kinase and protein kinase assays: The assay for PI 3-kinase activity in immunoprecipitates and protein from insect cells was essentially carried out as described in (30) . PI 3-kinase samples were incubated with an [y-³²P]ATP (Amersham, 3000 Ci/mmol), Ptdlns, PS, Mg^{2 +} mix for 10 min. at 30°C, before lipids were extracted and Ptdlns 3-P was quantified after thin layer chromatography (1 ,30).

The protein serine kinase activity of PI 3-kinase was essayed as described before (15), except for minor modifications. Immunoprecipitated or PI 3-kinase immobilized on phosphotyrosine beads was washed twice with kinase buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 5 mM EDTA, 5 mM DTT, 10 mM MnCl₂, 0.01% Triton X-100), before an equal volume of kinase buffer with double concentrated ATP (to give a final concentration of 20 μM and 10 μCi [³²P]yATP /experiment) was added to start the reaction. After a 20 min. incubation at 30 ;C, samples were denatured, separated by SDS- PAGE and exposed to Kodak X-Omat films for autoradiography.

Molecular modelling: An initial model of the pl lOα catalytic domain was build based on the crystallographic structure of PKA with bound PKA inhibitor-peptide and ATP (39). To improve the alignment of these two low homology molecules special attention was given to residues conserved within each family and to functional residues (for a review see (64)). Each family of structures was aligned using the MULT ALIGN method of Barton (4). Predicted (89) and X-ray secondary structure information was used as further constraints for the alignment. The final alignment (45% homology and 11% identity) was then used to construct the model using the suite of programs within Quanta™. The target (pl lOα) protein and template (PKA) were aligned. Where the target sequence matched the template molecule, the residue coordinates from the template were transformed directly to the target. Where no equivalent atoms were found in the template molecule for the target protein, reference was made to a side-chain rotamer library. This defines the most common conformation found for each side-chain type (66). Gaps in the target sequence were subjected to local energy minimization to bring the core ends together and to alleviate local conformational strain. Insertions in the target sequence were modelled by searching a fragment database of high resolution structures to find an appropriate template. The final structure was subjected to 100 steps of steepest gradient minimization by the CHARMM program to make minor shifts in the coordinate positions, thereby alleviating steric clashes between atoms and obtaining a reasonable peptide geometry. Wortmannin was fitted into the ATP binding site using ATP as a template. The docking of wortmannin was refined manually and the complex was minimised. This cycle of adjustment was repeated several times in order to obtain the model presented here.

Results

Specific interaction of wortmannin with pllOα:

It has been demonstrated that wortmannin inhibits the catalytic activity of PI 3-kinase directly and does not interfere with the formation of the P85/pll0 complex or its interaction with growth factor receptors (82,83). Using 17-[³H], 17-hydroxy wortmannin we could recently show that wortmannin binds covalently to a 110 kD protein in neutrophil cytosol, which copurified with PI 3-kinase activity (68).

While conditions for Ptdlns 3-kinase assays are usually optimized to produce a maximal effect of wortmannin, the parameters for the present study were adapted to increase specificity rather than sensitivity so as to avoid the detection of non-specific interactions of nucleophilic amino acid residues with wortmannin.

When wortmannin at nM concentrations was incubated with recombinant bovine p85α/pll0α complex, the inhibitor-protein conjugates could be detected after separation on reducing SDS-PAGE using anti-wortmannin antibodies. At physiological pH, wortmannin reacted covalently and specifically with pl lOα, but not with p85α. At pHs below 6, no wortmannin binding could be detected. On the other hand, when the pH was increased beyond 8.5, wortmannin attached to p85α as well (Fig. la). Non-specific wortmannin-binding was also observed when the concentration of the inhibitor was increased to 1 μM or more (Fig. lb), while at lower concentrations pllOα alone was stained. Non-specific labelling of p85 could be somewhat reduced by the inclusion of detergent (0.1% Triton X-100, compare Fig. 1 and 8). Detectable labelling under the stringent conditions used (0°C for 15 min.) could be achieved with as little as 5 nM wortmannin (Fig. lb). Identical results were obtained with immobilized GST-pll0α/p85 complex (data not shown) . At high pH or excessive inhibitor concentrations, wortmannin reacts with virtually any protein containing lysine residues, a feature that was exploited when producing anti-wortmannin antibodies (see methods).

Competition of PI 3-kinase substrates with wortmannin

Wortmannin could either act as a competitive or non-competitive inhibitor of PI 3-kinase. To test the two possibilities, PI 3-kinase was incubated with sonicated mixtures of phosphoinositides and PS, before wortmannin was added. While lipid suspensions containing Ptdlns (4, 5)P₂ protected PI 3-kinase from modification by wortmannin, Ptdlns, PS or Triton-XlOO alone had no effect up to 1 mg/ml (Fig. 2a). These results suggest that wortmannin might interact with the head group binding site for Ptdlns (4, 5)_P₂.

It has been previously reported that high concentrations of ATP reduce the inhibitory action of wortmannin on PI 3-kinase (82). The same was true for the covalent reaction of wortmannin with PI 3-kinase: ATP, and the ATP analogues FSBA and adenine at 1 mM all interfered with the alkylation of PI 3-kinase when added before the inhibitor (Fig. 2b). FSBA reacts covalently with reactive nucleophilic amino acid residues and was previously used to map the nucleotide-binding site of various ATP and GTP utilizing enzymes (see (34) for refs.). Although FSBA reacts to a certain extent non-specifically (see (34) and results below), the present results suggest that wortmannin alkylates a nucleophilic residue within the ATP-binding site of PI 3-kinase. To mimic the nucleophilic attack of amino acid side chains on wortmannin, substances similar to the respective amino acids but lacking the α-amino group were used. An excess of β- mercaptoethanol, ethanolamine, imidazole and guanidinopropionic acid did not compete with PI 3-kinase and did not destroy the reactivity of wortmannin under the conditions used (Fig. 2b). This demonstrates that the nucleophilicity of the wortmannin-reactive amino acid residue is rather determined by its environment within the enzyme's catalytic site than by its characteristics in aqueous solution. The lack of any effect of ethanolamine also shows that the primary amino group in ATP and its analogues does not have the potential to inactivate wortmannin under the given conditions.

Proteolytic mapping of the wortmannin-binding site

When wortmannin-labelled GST-pl l0α/p85α complex was partially digested, Factor Xa produced relatively small amounts of the expected pllOα full length protein cleaved adjacent to the IleGluGlyArg- intersection of GST-pllO. The major wortmannin-labelled product of

Factor Xa digestion was of 52 kD, while Glu-C {S. aureus V8 protease) generated a wortmannin-containing 55 kD peptide (Fig. 3). After the removal of anti-wortmannin antibodies, immunoblots were reprobed with anti-pl lO C-terminal antibodies. The superposition of the obtained signals from anti-wortmannin and anti-pllO C-terminal immunoblots showed that GST-pl lOα, pl lOα, as well as the 52 and the 52 kD peptides contained both the wortmannin-binding site and the C-terminus of pllOα (Fig. 3). The 52 and the 55 kD peptides remained firmly attached to the glutathione beads, as none of the 52 and very little of the 55 kD peptide could be detected in the digest's supernatant. Separate experiments in the absence of wortmannin showed that PI 3-kinase activity of the immobilized GST-pll0 /p85 complex was increased 2.3 times by Factor Xa and 3.8 times by Glu-C cleavage, indicating that the enzyme remained fully functional after the above protease treatment.

Once the N-terminal half of pi 10 could be excluded from binding wortmannin covalently, the interaction site was further mapped with proteases generating smaller peptides. In the course of these studies, we realized that binding of wortmannin to PI 3-kinase was covalent, but reversible. Attempts to isolate wortmannin-labelled peptides failed because of the acid lability of the wortmannin-enzyme interaction D and the fact that labelled peptides could not be recovered from various HPLC columns at neutral pH. Assuming that the acid labile function might be a Schiff-base, model reactions with lysine- and morpholino- adducts of wortmannin (kindly obtained from T.G. Payne, Basel) were carried out, to find ways to stabilize the putative amine- wortmannin interaction (for reactions of wortmannin with amines see (21)). Thus, NaCNBH₃ was finally chosen to reduce wortmannin-pl 10. Reduction with NaCNBH₃, as described in methods, considerably decreased the release of wortmannin from pllOα when exposed to low pH (pH 3-4, data not shown) . Due to its superior half -life, reduced, wortmannin- labelled pllOα was used subsequently for peptide analysis by TRICINE gel electrophoresis and immunodetection. Proteases were always used at various concentrations to follow the formation of peptides, which were then analyzed by superimposing anti-wortmannin and anti-peptide immunoblots. Because stripping and reprobing of membranes increased background chemiluminescence, parallel and not serial immunoblots are shown in Fig. 4.

Digestion with Lys-C protease produced two groups of wortmannin- labelled peptides with apparent molecular weights around 15 and 7 kD, which were not cross-detected with anti-C terminal antibodies (Fig 4a). Trypsin and Arg-C, on the other hand, produced various C-terminal peptides that remained undetected with anti-wortmannin antibodies (in the range of 7-15 kD). Peptides staining for wortmannin and the C- terminus were found upward from 40 kD in Arg-C partial digests.

As C-terminal peptides above 40 kD were found to be labelled by wortmannin and those below 15 kD were not (as illustrated in Fig. 5), a region of approximately 25 kD (amino acids 720 to 935) was estimated for the wortmannin-binding site. This localization corresponds to the ATP-binding site.

A complete digestion of pllOα with Lys-C generates theoretically 72 peptides, of which 12 are larger than 3 kD D and are therefore putative candidates for the obtained wortmannin-labelled peptide family with an apparent molecular weight of 7 kD (Fig. 4). Only 4 of these peptides, however, fall into the region of amino acids 720-935. In case the e- amino group of a lysine reacted with wortmannin, one would expect that the lysine-specific protease would not cleave C-terminal to the alkylated residue. A Lys-C digest would then result in 32 peptides larger than 3 kD, with 6 candidates located within the 25 kD wortmannin-binding region (see Table I). Of these peptides, however, only two (F734-K802 and R777-K863) would give rise to a 15 kD wortmannin-labelled peptide with one additional uncleaved amide bound distal to one of the neighbouring lysines. The 15 kD wortmannin-signal being the F734-K863 peptide would also explain the appearance of multiple bands, since it can be further extended to the lysines 729, 724 or 867.

To check this possibility, wortmannin-labelled pl lOα was digested with Lys-C, and immunoblots were probed with anti-wortmannin and anti- peptide antibodies (pl lOα 34-748). These experiments showed clearly, that the 15 kD peptides contain the wortmannin-attachment site as well as the peptide sequence 734-748. However, the wortmannin-labelled 7 kD peptides that were formed as the 15 kD group was digested further remained undetected by anti-pllOα 34-748 antisera (Fig 4b). Because of its proximity to the gel front, the wortmannin-unlabelled 734-776 fragment produced only a weak signal with an apparent molecular weight of about 3.2 kD in anti-peptide immunoblots (calculated M_r: 4.9 kD, data not shown). These results clearly exclude Lys776 from being the covalent binding site of wortmannin and point to an interaction at Lys802.

Wortmannin and FSBA-binding to Lys/ Arg mutants of pl lOα

Lysines within the putative wortmannin-binding site K733-K867 were replaced by arginines (histidine for Lys733) using site-directed mutagenesis. Selected cysteins within the same region were mutated to serine at the same time. Mutated pllOα was subsequently co- expressed with p85α in 293 cells and immunoprecipitated with monoclonal anti-p85α antibodies or anti-pllOα C-terminal antibodies from cell lysates. Expression levels of wild type pi 10α/p85 complex and mutant PI 3-kinase protein were all equal as demonstrated by Coomassie blue- and colloidal gold-staining in Fig. 6a, b. The amount of endogenous PI 3-kinase from 293 cells was too low to be detected. When anti-PI 3-kinase immunoprecipitates were subjected to wortmannin-labelling, wild type pl lOα (wt), pll0α^K733H (KRl), _{pl l0α}C769S/K776R _{(KR2) and pl l0α}C862S/K863/K867R _{(KR4) a}„ _bound

wortmannin covalently, while the pl l0α^κ802R (KR3) mutant was undetected by anti-wortmannin antisera (Fig. 6b, d). Identical results were obtained with recombinant PI 3-kinase complexes (wt, KR2-4) isolated from the bacculovirus/Sf9 cell expression system (data not shown). A covalent interaction of wortmannin with Lys802, as confirmed by the KR3 mutation, is in agreement with the results obtained with substrate competition and proteolyic digests.

Wild type and mutant PI 3-kinase was also incubated with FSBA and subsequently examined with anti-FSBA antisera for the presence of sulfonated protein. Although FSBA is considered as a specific probe for ATP-binding sites (34), we observed non-specific modification of p85 (Fig. 6e) and IgG heavy and light chains in immunoprecipitates (data not shown) . Further experiments were therefore carried out with recombinant pl l0/p85 complex isolated from insect cells. With this purified, immobilized protein FSBA-staining of wild type, KR2 and KR4-pll0α was prominent, while the labelling of KR3-pl l0α was virtually absent (Fig. 6e) . Although samples were extensively washed and denatured in the presence of /3ME and DTT, FSBA-binding to p85 persisted.

Lipid and protein kinase activity of Lys/ Arg mutants of pllOα

The transiently overexpressed and immunoprecipitated proteins were assayed for lipid and protein kinase activity. While 293 cells transfected with the p85α and one of the wt, KRl , KR2 or KR4-pll0α vectors produced 35-40 times the amount of Ptdlns 3-P compared to cells that received p85 only, the PI 3-kinase activity from p85/KR3 transfected cells achieved only levels of mock-transfected cells (Fig. 7a). That the remaining activity in KR3 transfected cells is due to coprecipitation of endogenous PI 3-kinase is illustrated by the fact that the p85α/KR3-pll0 complex purified from Sf9 insect cells yielded less than 0.1% of the activity of the wt, KR2 and KR4 protein from the same source (data not shown) .

As in the lipid kinase assay, the KR3 mutant pl lOα was impaired in its ability to phosphorylate co-expressed p85α, while the wt and the rest of the KR mutants all heavily phosphorylated the p85α band (Fig 7b) . The same results were obtained with protein from Sf9 cells, where the KR3 mutant showed no phosphorylation of p85α at all (data not shown) .

It has been shown previously that mutations within the conserved DXHXXN kinase motif completely destroy the lipid kinase activity of Vps34p (56) as well as the lipid and protein kinase activity of pllOα (DRHNSN in pllOα, R = Arg916, (15)). When we used a p85α/GST- pl lO R916P mutant (RP) PI 3-kinase to assay wortmannin binding, the inactive pi 10 subunit still bound wortmannin covalently. The reaction with the inhibitor was observed at similar concentrations as with the wild type lipid kinase and occurred in a specific manner (Fig 8.). The labelling efficiency of the RP mutant, however, was reduced and more protein was needed to yield comparable signals in the anti- wortmannin immunoblots. This indicates, that the interaction of wortmannin with the catalytic subunit of PI 3-kinase is optimal with an intact ATP-binding site, but is not strictly coupled to kinase activity.

Structure-prediction of the wortmannin-pllOα complex

The availability of the crystal structure of PKA prompted us to construct a model of the catalytic domain of pllOα so that, using our assignment of the wortmannin-binding residue, a possible structure of the pllO - wortmannin interaction site could be put forward. Modelling was performed as described in Materials and Methods. The overall root mean square (RMS) deviation of the modelled structure to the original template structure was 2.5 A after minimization. Two significant insertions had to be added to the pllOα model with respect to PKA: the first is 8 residues long and partially covers a grove in pllOα which is equivalent to the binding site of the PKA inhibitor peptide. However, a smaller substrate such as PI would still fit into the active site of pl lOα together with ATP. The second insertion (10 residues) is located in such a way as to replace the N-terminal helix which is found in PKA but seems to be absent in the Pl-kinase catalytic domains. The environment of the ATP binding site contains the conserved Lys802 (equivalent to Lys72 in PKA) and the stabilizing Glu821 (Glu91 in PKA). Lys802 could also interact with the a- and ^-phosphates of ATP (data not shown) as proposed for Lys72 in PKA (39). When wortmannin was docked into this site using the ATP molecule as template the C20 became positioned so that it was near Lys802. Residues that may stabilize wortmannin in this site are Pro786, Asρ787, Ile788, Ser919, Asρ933, His936, and Lys802. Wortmannin fits well into this pocket (see Fig. 9). Lys802 could attack C20 of wortmannin, while still being stabilized by Glu821. Although the model of the pllOα catalytic centre as shown in Fig. 9. is preliminary, and further studies are in progress to improve it, it is in good agreement with the biochemical data described above.

Discussion

PI 3-kinase has been proposed to play an important role in many biological processes activated by diverse extracellular signals. Because the molecular targets of 3-phosphorylated phosphoinositides have yet to be identified, the involvement of PI 3-kinase in a particular response has to be demonstrated using several independent experimental systems, such as mutation of specific recruitment sites on receptors, expression of "dominant-negative" forms of PI 3-kinase and the use of specific inhibitors - the most widely used being wortmannin.

Wortmannin received its first broad attention as an inhibitor of the agonist- induced superoxide anion production in neutrophils. Although the cellular target of wortmannin was unknown at that time, Baggiolini and coworkers (3) demonstrated that the furan ring structure of the substance was important for its action, as derivatives with an opened or protected furan ring were ineffective in biological assays. After the identification of wortmannin as a specific inhibitor of PI 3-kinase (1,45,82,83), it was found that 17-[³H],17-hydroxy wortmannin labelled a 110 kD PI 3-kinase from neutrophil cytosol. The resistance of this association to denaturation under reducing conditions suggested a covalent interaction of wortmannin and the catalytic pi 10 subunit of PI 3-kinase (68) .

To determine the exact site of this covalent interaction, we optimised the reaction conditions for wortmannin with PI 3-kinase, aiming at a most specific reaction of the inhibitor with pl lOα. When added in excess, the physiological substrates of PI 3-kinase, Ptdlns (4, 5)P₂ and ATP, both prevented the reaction of wortmannin with pl lOα. The kinase motifs DXHXXN and DFG are conserved in pl lOα (26). It has been shown that point mutations within these motifs destroy the lipid kinase activities of pllOα (15) and Vps34p (56). Because of the preceding glycine rich region (Gly 837, 842 and 846 in pllOα, see Table B) Lys863 was originally aligned by Hiles at al. (26) with Lys 72 of cAMP dependent kinase (67). This roughly localized the ATP-binding site within the 842G-DFG935 region of pllOα.

The charged 941KKKKFGYKRER951 stretch of pllOα resembles the

K(X)_nKXKK (n = 3-7) motif that was found to bind PtdIns(4,5)P₂ in gelsolin (86) and might therefore constitute the binding site for the 4,5- phosphates of the lipid. The fact that Ptdlns (4, 5)P₂, but not Ptdlns, PS or detergent diminished the wortmannin-pllO interaction may indicate that the concentrated positive charges participate somehow in the non-covalent binding of wortmannin. The competitors in this experiment, however, were present in mixed micelles and vesicles. As it was observed before that the physical properties of lipids can influence their effects on lipid binding proteins (e.g. on gelsolin (28)), these results must be viewed with caution. Together, these results define a minimal putative 13 kD region for non-covalent wortmannin-binding overlapping with the ATP- and Ptdlns (4, 5)P₂-binding sites from approximately G824 to R951.

The localisation of the covalent reaction site was carried out by proteolysis of wortmannin-labelled pll0α/p85α complex and lead to the identification of a wortmannin-labelled 15 kD peptide containing the F734-K863 core, while the peptides with apparent M_r of 7 kD contained the R777-K863 sequence. As summarised in Fig. 5, the latter peptide overlaps with the putative ATP-binding site and has a calculated mass of 10 kD. As this sequence contains an internal lysine residue that is not recognised by Lys-C, one must assume that this lysine - Ly802 - is alkylated by wortmannin.

To evaluate the putative reaction sites within the target region defined by proteolytic digests (Lys802 within the wortmannin-labelled R777-K863 peptide, see Results and Fig. 5), we used site directed mutagenesis to exchange putative wortmannin-reactive lysines. The aim of these manipulations was to leave the positive charges in place, while reducing the nucleophilicity of the respective side chains. Lysines were therefore exchanged by arginines (except for the K733H mutation) . Because arginines are always protonated under physiologic conditions and histidines are secondary amines, both are weak nucleophiles. The observation that the exchange of Lys733, 776, 863 and 867 had no consequences and only the KR3 mutant (K802R) was no longer able to bind wortmannin covalently, confirmed the interpretation of the peptide and substrate competition data. This is also in agreement with the fact that Lys-C digestion of the wortmannin-preincubated KR2 PI 3-kinase accumulated the 15 kD wortmannin-labelled peptides, while the 7 kD peptides were not present anymore (data not shown) .

In analogy to chemical reactions performed to produce wortmannin derivatives (21), the e-amino group of Lys802 attacks C20 of wortmannin leading to the opening of the furan ring and the formation of an enamine. The enamine is in equilibrium with a Schiff-base, which is relatively stable at physiological pH but is easily hydrolysed under acidic conditions and can be reduced by NaCNBH₃ in its imine form. A similar reaction mechanism is likely to be found for demethoxyviridin, another inhibitor of PI 3-kinase effective at nM concentrations (5,81), because the substance contains an identical furan ring system as wortmannin.

The covalent character of the interaction also explains the low inhibitory concentrations of wortmannin, when compared to quercetin derivatives. Quercetin and the more specific LY294002 derivative (59,73) have no site for a nucleophilic attack and were reported to inhibit Ptdlns kinases at μM concentrations. This is comparable with the effect of wortmannin derivatives with hydrolysed furane ring structures (3,83) . LY294002 was shown to inhibit PI 3-kinase competitively in respect to ATP (73), while wortmannin was classified as a non-competitive inhibitor (45). However, as we show here, wortmannin does react covalently within the ATP- binding site and thus produces the characteristics of an apparent non- covalent inhibitor by reducing the amount of functional enzyme and the apparent V_maχ, while the K^ of unreacted PI 3-kinase remains unchanged.

FSBA was previously used to map the ATP-binding site of various proteins, among them PKA and MLCK. The isolation of sulfonated peptides led to the identification of the FSBA-reactive residues, which are Lys72 in PKA (88) and Lys548 in MLCK (34). From the three KR mutants produced in insect cells, only the KR3-pll0α did not react with FSBA, while the K776R, K763R and K767R replacements had no effect. Although we obtained non-specific FSBA-staining of p85, the differential staining of the pi 10 subunits indicates that K802 is the residue modified by FSBA.

The sulfonated residues in PKA and MLCK were shown to play crucial roles in the phosphotransfer reaction (39,67) and both enzymes were irreversibly inhibited by FSBA. When wild type and KR mutant PI 3- kinase were tested for lipid and kinase activity, wild-type and KRl , 2 and 4 phosphorylated Ptdlns, as well as the co-expressed p85 regulatory subunit. KR3 on the other hand, was found to be inactive in both assays. Altogether, our results demonstrate that K802 has a similar function as K72 in PKA, which was proposed to interact with the - and /3-phosphate groups of ATP (39) . An alignment of Ptdlns 3-kinases, TORs, DNA-PK_CS, Ptdlns 4-kinases and two protein kinases around the Lys802-pl l0 conserved region illustrates that the identified lipid kinases are devoid of a typical P-loop (Table II, for a review see on protein kinase motives see (54)). The TOR genes D without an attributed activity so far D contain as the only members of this family two conserved glycine residues, that might constitute a degenerated P-loop.

Based on the alignments in Table II it is likely that K548 of MLCK will turn out to be the wortmannin-binding residue, as it has been verified that MLCK is irreversibly inhibited by μM concentrations of wortmannin (42) . Ptdlns 4-kinase is inactivated by FSBA, and it has been proposed that the unknown sulfonylated residue is a lysine (55). Lys602 seems here to be the most likely candidate at present.

Reflections about the reaction sequence of wortmannin and FSBA with the catalytic centre of kinases may yield some general predictions: both substances associate in a first step non-covalently with the substrate- binding site of the respective kinase. Once the inhibitors are held in place by non-covalent interactions, the high local concentrations of the reaction partners (e.g. Lys802 and wortmannin) drive the reaction immediately towards the covalent conjugate. While the first step depends on the dissociation constant of the inhibitor, the on-rate of the covalent reaction is dictated by the nucleophilicity of the attacking lysine residue. As this residue is present in all lipid kinases and most protein kinases, differences and specificities in wortmannin reactivity must be sought at the level of non-covalent interactions. This is supported by the differential reactivity of yeast and human Vps34p with wortmannin, as both enzymes catalyze identical reactions and contain the K802 - equivalent residue.

PI 3-kinase, non-functional by the R916P mutation within the DXHXXN motive, still could bind wortmannin covalently, although less efficiently than wild type kinase. Wortmannin also reacted with PI 3-kinase in the presence of inactivating concentrations of detergent or with PI 3-kinase inactivated by the pre-phosphorylation of p85 by the intrinsic protein kinase activity of pi 10. Together, this indicates that kinase activity and intact substrate binding are not obligatory for a covalent reaction of wortmannin, although the latter is optimal if the ATP-binding site is intact. As wortmannin and FSBA both bind within the ATP-binding site of pre-phoshorylated p85/pll0 complex, one might assume that the phosphorylation of Ser608 on p85α (15) D deactivating the PI 3-kinase complex D blocks rather lipid binding than the access of ATP to the catalytic site. A contribution of p85 in the binding of Ptdlns (4, 5)P was suggested before (16).

The identified covalent interaction site and the proposed reaction mechanism presented here explains the highly efficient inhibition of Ptdlns 3-kinases by wortmannin, as the inhibitor not only blocks substrate binding, but also alkylates a Lys residue, whose nucleophilicity is crucial for the catalytic process. Wortmannin can be predicted to react at high concentrations with known or yet to be identified protein and lipid kinases if they show homologies within the PI 3-kinase Lys802 region and contain classical kinase domains (like DXHXXN, DFG, see e.g. MLCK) . Lipid- binding sites and an acid residue corresponding to Glu821 could increase the sensitivity to wortmannin, as the former expose hydrophobic surfaces for wortmannin interaction and the latter might increase the nucleophilicity of the Lys802 homolog residue. The present data contribute to the correct interpretation of data obtained with the inhibitor in respect to Ptdlns 3-kinase involvement in various cell responses. When using wortmannin, the covalent nature of the reaction has to be recalled and the reaction temperature, buffer composition and the incubation time have to be considered. As shown above, non-specific reactions are to be expected at pHs above 8, where the nucleophilicity of lysines is dramatically increased due to deprotonation. As for all covalent reactions, IC_5Q values are misleading if they are not strictly used comparatively and were not obtained under identical conditions. Although wortmannin inhibits PI 3-kinases at very low concentrations, careful controls are essential to rule out cross-interferences with other signalling pathways.

The model of the pllOα-wortmannin complex that we have derived is certainly speculative and would need to be confirmed by crystallography. However, it is in very good agreement with the experimental data presented here and serves as a template for further site-directed mutagenesis studies in order to test and improve its validity. The localization and modelling of the wortmannin-binding site provides the basis for ongoing experimental approaches exploring the non-covalent wortmannin interaction within the catalytic cleft of PI 3-kinase. This information might lead to the design of subunit specific inhibitors for PI 3-kinases and related enzymes. Unlike wortmannin, such inhibitors might uncover the possibility to inhibit one specific member of the PI 3-kinase family while others function normally - be it for research or therapeutic uses.

References

1. Arcaro, A. and M. P. Wymann. 1993. Wortmannin is a potent phosphatidylinositol 3-kinase inhibitor; the role of phosphatidylinositol 3,4,5- trisphosphate in neutrophil responses. Biochem. J. 296:297-301.

2. Backer, J. M., M. G. Myers,Jr., S. E. Shoelson, D. J. Chin, X. J. Sun, M.

Miralpeix, P. Hu, B. Margolis, E. Y. Skolnik, and J. Schlessinger. 1992. Phosphatidylinositol 3 '-kinase is activated by association with IRS- 1 during insulin stimulation. EMBO J 11:3469-3479. 3. Baggiolini, ML, B. De ald. J. Schnyder, W. Ruch, P. H. Cooper, and T. G.

Payne. 1987. Inhibition of the phagocytosis-induced respiratory burst by the fungal metabolite wortmannin and some analogues. Exp. Cell Res. 169:408- 418.

4. Barton, G. J. and M. J. Sternberg. 1987. A Strategy for the Rapid Multiple Alignment of Protein Sequences: Confidence Levels from Tertiary Structure

Comparisons. J. Mol. Biol. 198:327-337.

5. Bonser, R. W., N. T. Thompson, R. W. Randall, J. E. Tateson, G. D.

Spacey, H. F. Hodson, and L. G. Garland. 1991. Demethoxyviridin and wortmannin block phospholipase C and D activation in the human neutrophil. Br. J. Pharmacol. 103: 1237-1241.

6. Brown, E. J., M. W. Albers, T. B. Shin, K. Ichikawa, C. T. Keith, W. S.

Lane, and S. L. Schreiber. 1994. A mammalian protein targeted by Gl- arresting rapamycin-receptor complex. Nature 369:756-758.

7. Brown, W. J., D. B. DeWald, S. D. Emr, H. Plutner, and W. E. Balch. 1995. Role for Phosphatidylinositol 3-Kinase in the Sorting and Transport of

Newly Synthesized Lysosomal Enzymes in Mammalian Cells. J. Cell Biol. 130:781-796.

8. Burgering, B. M. T. and P. J. Coffer. 1995. Protein kinase B (c-Akt) in phosphatidylinositol-3-OH kinase signal transduction. Nature 376:599-603. 9. Chuang, T. -H., B. P. Bohl, and G. M. Bokoch. 1993. Biologically Active Lipids Are Regulators of RacGDI Complexation. J. Biol. Chem. 268:26206- 26211.

10. Chung, J., T. C. Grammer, K. P. Lemon, A. Kazlauskas, and J. Blenis.

1994. PDGF- and insulin-dependent pp70^S6k activation mediated by phosphatidylinositol-3-OH kinase. Nature 370:71-75.

11. Claesson- Welsh, L. 1994. Platelet-derived Growth Factor Receptor Signals. J.

Biol. Chem. 269:32023-32026.

12. Cross, D. A., D. R. Alessi, J. R. Vandenheede, H. E. McDowell, H. S.

Hundal, and P. Cohen. 1994. The inhibition of glycogen synthase kinase-3 by insulin or insulin- like growth factor 1 in the rat skeletal muscle cell line

L6 is blocked by wortmannin, but not by rapamycin: evidence that wortmannin blocks activation of the mitogen-activated protein kinase pathway in L6 cells between Ras and Raf. Biochem. J 303:21-26.

13. Davidson, H. W. 1995. Wortmannin Causes Mistargeting of Procathepsin D. Evidence for the Involvement of Phosphatidylinositol 3-Kinase in Vesicular

Transport to Lysosomes. J. Cell Biol. 130:797-805.

14. Dhand, R., K. Hara, I. Hues, B. Bax, I. Gout, G. Panayotou, M. J. Fry, K.

Yonezawa, M. Kasuga, and M. D. Waterfield. 1994. PI 3-kinase: structural and functional analysis of intersubunit interactions. EMBO J. 13:511-521. 15. Dhand, R., I. Hues, G. Panayotou, S. Roche, M. J. Fry, I. Gout, N. F.

Totty, O. Truong, P. Vicendo, K. Yonezawa, M. Kasuga, S. A. Courtneidge, and M. D. Waterfield. 1994. PI 3-kinase is a dual specific enzyme: autoregulation by an intrinsic protein serine-kinase activity. EMBO J. 13:522-533. 16. End, P., I. Gout, M. J. Fry, G. Panayotou, R. Dhand, K. Yonezawa, M.

Kasuga, and M. Waterfield. 1993. A Biosensor Approach to Probe the

Structure and Function of the p85 Subunit of the Phosphatidylinositiol 3- kinase Complex. J. Biol. Chem. 268: 10066-10075.

17. Escobedo, J. A., S. Navankasattusas, W. M. Kavanaugh, D. Milfay, V. A. Fried, and L. T. Williams. 1991. cDNA Cloning of a Novel 85 kD Protein That Has SH2 Domains and Regulates Binding of PI3-Kinase to the PDGF β- Receptor. Cell 65:75-82.

18. Franke, T. F., S. -II Yang, T. O. Chan, K. Datta, A. Kazlauskas, D. K.

Morrison, D. R. Kaplan, and P. N. Tsichlis. 1995. The Protein Kinase Encoded by the Akt Proto-Oncogene Is a Target of the PDGF-Activated

Phosphatidylinositol 3-kinase. Cell 81:727-736.

19. Fry, M. J., G. Panayotou, R. Dhand, F. Ruiz Larrea, I. Gout, O. Nguyen,

S. A. Courtneidge, and M. D. Waterfield. 1992. Purification and characterization of a phosphatidylinositol 3-kinase complex from bovine brain by using phosphopeptide affinity columns. Biochem. J. 288:383-393.

20. Gelas, P., V. von Tscharner, M. Record, M. Baggiolini, and H. Chap. 1992.

Human neutrophil phospholipase D activation by N-formylmethionyl- leucylphenylalanine reveals a two-step process for the control of phosphatidylcholine breakdown and oxidative burst. Biochem. J. 287:67-72. 21. Haefliger, W. and D. Hauser. 1975. EinfYhrung der Corticoid-Seitenkette bei

Wortmannin. Helv. Chim. Acta 58:1629-1633.

22. Hartley, K. O., D. Gell, G. C. M. Smith, H. Zhang, N. Divecha, M. A.

Connelly, A. Admon, S. P. Lees-Miller, C. W. Anderson, and S. P. Jackson. 1995. DNA-Dependent Protein Kinase Catalytic Subunit: A Relative of Phosphatidylinositol 3-Kinase and the Ataxia Telangiectasia Gene Product.

Cell 82:849-856.

23. Hawkins, P. T., A. Eguinoa, R. -G. Qiu, D. Stokoe, F. T. Cooke, R.

Walters, S. Weπnstrsm, L. Claesson- Welsh, T. Evans, M. Symons, and L. R. Stephens. 1995. PDGF stimulates an increase in GTP-Rac via activation of phosphoinositide 3-kinase. Curr. Biol. 5:393-403.

24. Hawkins, P. T., T. R. Jackson, and L. R. Stephens. 1992. Platelet-derived growth factor stimulates synthesis of Ptdlns (3, 4, 5) P3 by activating a PtdIns(4,5)P₂ 3-OH kinase. Nature 358: 157-159.

25. Herman, P. K. and S. D. Emr. 1990. Characterization of VPS34, a gene required for vacuolar protein sorting and vacuole segregation in Saccharomyces cerevisiae. Mol. Cell Biol. 10:6742-6754.

26. Hues, I. D., M. Otsu, S. Volinia, M. J. Fry, I. Gout, R. Dhand, G.

Panayotou, F. Ruiz Larrea, A. Thompson, and N. F. Totty. 1992. Phosphatidylinositol 3-Kinase: Structure and Expression of the 110 kd Catalytic Subunit. Cell 70:419-429.

27. Hu, P., A. Mondino, E. Y. Skolnik, and J. Schlessinger. 1993. Cloning of a novel, ubiquitously expressed human phosphatidylinositol 3-kinase and identification of its binding site on p85. Mol. Cell Biol. 13:7677-7688.

28. Janmey, P. A., K. Iida, H. L. Yin, and T. P. Stossel. 1987. Polyphosphoinositide Micelles and Polyphosphoinositide-containing Vesicles

Dissociate Endogenous Gelsolin-Actin Complexes and Promote Actin Assembly from the Fast-growing End of Actin Filaments Blocked by Gelsolin. J Biol. Chem. 262: 12228-12236.

29. Kanai, F., K. Ito, M. Todaka, H. Hayashi, S. Kamohara, K. Ishii, T. Okada, O. Hazeki, M. Ui, and Y. Ebina. 1993. Insulin-stimulated GLUT4 translocation is relevant to the phosphorylation of IRS-1 and the activity of PI3-kinase. Biochem. Biophys. Res. Commun. 195:762-768.

30. Kaplan, D. R., M. Whitman, B. Schaffhausen, D. C. Pallas, M. White, L.

C. Cantley, and T. M. Roberts. 1987. Common elements in growth factor stimulation and oncogenic transformation: 85 kd phosphoprotein and phosphatidylinositol kinase activity. Cell 50: 1021-1029.

31. Kaufmann, R. J., P. Murtha, D. E. Ingolia, C. -Y. Yeung, and R. E.

Kellems. 1989. The phosphorylation state of eucaryotic initiation factor 2 alters translational efficiency of specific mRNAs. Mol. Cell Biol. 9:946. 32. Kazlauskas, A. and J. A. Cooper. 1990. Phosphorylation of the PDGF receptor β subunit creates a tight binding site for phosphatidylinositol 3 kinase. EMBO

J. 9:3279-3286. 33. Kimura, K., S. Hattori, Y. Kabuyama, Y. Shizawa, J. Takayanagi, S.

Nakamura, S. Toki, Y. Matsuda, K. Onodera, and Y. Fukui. 1994. Neurite outgrowth of PC 12 cells is suppressed by wortmannin, a specific inhibitor of phosphatidylinositol 3-kinase. J Biol. Chem. 269:18961-18967. 34. Komatsu, H. and M. Ikebe. 1993. Affinity labelling of smooth-muscle myosin light-chain kinase with 5'-(p-(fluorosulphonyl)benzoyl)adenosine. Biochem. J 296:53-58. 35. Kovacsovics, T. J., C. Bachelot, A. Toker, C. J. Vlahos, B. Duckworth, L.

C. Cantley, and J. H. Hartwig. 1995. Phosphoinositide 3-kinase inhibition spares actin assembly in activating platelets but reverses platelet aggregation. J Biol. Chem. 270:11358-11366.

36. Kundra, V., J. A. Escobedo, A. Kaslauskas, H. K. Kim, S. G. Rhee, L. T. Williams, and B. R. Zetter. 1994. Regulation of chemotaxis by the platelet- derived growth factor receptor- ? . Nature 367:474-476.

37. Kunz, J., R. Henriquez, U. Schneider, M. Deuter Reinhard, N. R. Mowa, and M. N. Hall. 1993. Target of rapamycin in yeast, TOR2, is an essential phosphatidylinositol kinase homolog required for GI progression. Cell 73:585- 596.

38. Laemmli, U. R. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680.

39. Madhusudan, , E. A. Trafny, N. H. Xuong, J. A. Adams, L. F. Ten Eyck,

S. S. Taylor, and J. M. Sowadski. 1994. cAMP -dependent protein kinase: crystallographic insights into substrate recognition and phosphotransfer. Protein

Sci. 3:176-187.

40. Nakanishi, H., K. A. Brewer, and J. H. Exton. 1993. Activation of the ζ

Isozyme of Protein Kinase C by Phosphatidylinositol 3,4,5-Trisphosphate. J.

Biol. Chem. 268:13-16. 41. Nakanishi, S., K. J. Cart, and T. Balla. 1995. A wortmannin-sensitive phosphatidylinositol 4-kinase that regulates hormone-sensitive pools of inositolphospholipids. Proc. Natl. Acad. Sci. U. S. A. 92:5317-5321. 42. Nakanishi, S., S. Kakita, I. Takahashi, K. Kawahara, E. Tsukuda, T. Sano,

K. Yamada, M. Yoshida, H. Kase, Y. Matsuda, Y. Hashimoto, and Y. Nonomura. 1992. Wortmannin, a microbial product inhibitor of myosin light chain kinase. J. Biol. Chem. 267:2157-2163. 43. O'Reilly, D. R., L. K. MUler, and V. A. Luckow. 1992. Bacculovirus Expression Vectors - A Laboratory Manual. W.H. Freeman & Co., New York. 44. Okada, T., Y. Kawano, T. Sakakibara, O. Hazeki, and M. Ui. 1994.

Essential Role of Phosphatidylinositol 3-Kinase in Insulin-induced Glucose Transport and Antilipolysis in Rat Adipocytes - Studies with a Selective Inhibitor Wortmannin. J. Biol. Chem. 269:3568-3573.

45. Okada, T., L. Sakuma, Y. Fukui, O. Hazeki, and M. Ui. 1994. Blockage of Chemotactic Peptide-induced Stimulation of Neutrophils by Wortmannin as a

Result of Selective Inhibition of Phosphatidylinositol 3-Kinase. J. Biol. Chem. 269:3563-3567.

46. Otsu, M., I. D. Hiles, I. Gout, M. J. Fry, F. Ruiz-Larrea, G. Panayotou, A.

Thompson, R. Dhand, J. Hsuan, N. Totty, A. D. Smith, S. J. Morgan, S. A. Courtneidge, P. J. Parker, and M. D. Waterfield. 1991.

Characterisation of Two 85 kd Proteins That Associate with Receptor Tyrosine Kinases, Middle-T/ppόO^0"^ Complexes, and PI3-Kinase. Cell 65:91-104.

47. Page, M. J. 1989. p36C: an improved bacculovirus expression vector for producing high levels of mature recombinant proteins. Nucleic Acids Res. 17:454.

48. Panayotou, G., B. Bax, I. Gout, M. Federwisch, B. Wroblowski, R. Dhand,

M. J. Fry, T. L. Blundell, A. Wollmer, and M. D. Waterfield. 1992. Interaction of the p85-Subunit of PI 3-Kinase and Its N- Terminal SH2-Domain with a PDGF Receptor Phosphorylation Site - Structural Features and Analysis of Conformational Changes. EMBO J. 11:4261-4272.

49. Reinhold, S. L., S. M. Prescott, G. A. Zimmerman, and T. M. Mclntyre.

1990. Activation of human neutrophil phospholipase D by three separable mechanisms. FASEB J. 4:208-214.

50. Rusconi, S., Y. Severne, O. Georgiev, I. Galli, and S. A. Wieland. 1990. A novel expression assay to study transcriptional activators. Gene 89:211-221.

51. Sabatini, D. M., H. Erdjument-Bromage, M. Lui, P. Tempst, and S. H. Snyder. 1994. RAFT: A Mammalian Protein That Binds to FKBP12 in a Rapamycin-Dependent Fashion and It Homologous to Yeast TORs. Cell 78:35- 43.

52. Sabers, C. J., M. M. Martin, G. J. Brunn, J. M. Williams, F. J. Dumont, G. Wiederrecht, and R. T. Abraham. 1995. Isolation of a protein target of the FKBP12-rapamycin complex in mammalian cells. J. Biol. Chem. 270:815- 822.

53. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning - A

Laboratory Manual. Cold Spring Harbor Laboratory Press, New York. 54. Saraste, M., P. R. Sibbald, and A. Wittinghofer. 1990. The P-loop - a common motiv in ATP- and GTP-binding proteins. Trends Biochem. Sci. 15:430-434.

55. Scholz, G., G. J. Barritt, and F. Kwok. 1992. Affinity labelling of the active site of brain phosphatidylinositol 4-kinase with 5'-fluorosulphonylbenzoyl- adenosine. Eur. J. Biochem. 210:461-466.

56. Schu, P. V., K. Takegawa, M. J. Fry, J. H. Stack, M. D. Waterfield, and

S. D. Emr. 1993. Phosphatidylinositol 3-Kinase Encoded by Yeast VPS34 Gene Essential for Protein Sorting. Science 260:88-91.

57. SchSgger, H. and G. Von Jagow. 1987. Tricine-Sodium Dodecyl Sulfate- Polyacrylamide Gel Electrophoresis for the Separation of Proteins in the Range of 1 to 100 kDa. Anal. Biochem. 166:368-379.

58. Serunian, L. A., M. T. Haber, T. Fukui, J. W. Kim, S. G. Rhee, J. M.

Lowenstein, and L. C. Cantley. 1989. Polyphosphoinositides produced by phosphatidylinositol 3-kinase are poor substrates for phospholipases C from rat liver and bovine brain. J. Biol. Chem. 269:31552-31562.

59. Shibasaki, F., Y. Fukui, and T. Takenawa. 1993. Different properties of monomer and heterodimer forms of phosphatidylinositol 3-kinases. Biochem. J. 289:227-231.

60. Skolnik, E. Y., B. Margolis, M. Mohammadi, E. Lowenstein, R. Fischer, A. Drepps, A. Ullrich, and J. Schlessinger. 1991. Cloning of PI3 Kinase- Associated p85 Utilizing a Novel Method for Expression/Cloning of Target Proteins for Receptor Tyrosine Kinases. Cell 65:83-90.

61. Stack, J. H., D. B. DeWald, K. Takegawa, and S. D. Emr. 1995. Vesicle- mediated Protein Transport: Regulation of Interactions between the Vpsl5 Protein Kinase and the Vps34 Ptdlns 3-Kinase Essential for Protein Sorting to the Vacuole in Yeast. J. Cell Biol. 129:321-334.

62. Stephens, L., A. Smrcka, F. T. Cooke, T. R. Jackson, P. C. Sternweis, and

P. T. Hawkins. 1994. A Novel Phosphoinositide 3-kinase Activity in Myeloid-

Derived Cells Is Activated by G Protein βγ Subunits. Cell 77:83-93. 63. Stephens, L. R., K. T. Hughes, and R. F. Irvine. 1991. Pathway of phosphatidylinositol(3,4,5)-trisphosphate synthesis in activated neutrophils.

Nature 351:33-39. 64. Sternberg, M. J. and M. J. Zvelebil. 1990. Prediction of protein structure from sequences. Eur. J. Cancer 26: 1163-1166. 65. Stoyanov, B., S. Volinia, T. Hanck, I. Rubio, M. Loubtchenkov, D. Malek,

S. Stoyanova, B. Vanhaesebroeck, R. Dhand, B. NΫrnberg, P. Gierschik,

K. Seedorf, J. J. Hsuan, M. D. Waterfield, and R. Wetzker. 1995.

Cloning and Characterization of a G Protein-Activated Human

Phosphatidylinositide-3 Kinase. Science 269:690-693. 66. Summers, N. L. and M. Karplus. 1988. . J. Mol. Biol. 210:268-272.

67. Taylor, S. S., D. R. Knighton, J. Zheng, L. F. Ten Eyck, and J. M.

Sowadski. 1992. Structural Framework for the Protein Kinase Family. Annu. Rev. Cell Biol. 8:429-462.

68. Thelen, M., M. P. Wymann, and H. Langen. 1994. Wortmannin binds specifically to phosphatidylinositol 3-kinase while inhibiting G protein coupled receptor signalling in neutrophil leukocytes. Proc. Natl. Acad. Sci. U. S. A. 91:4960-4964.

69. Toker, A., M. Meyer, K. K. Reddy, J. R. Falck, R. Aneja, S. Aneja, A.

Parra, D. J. Burns, L. M. Ballas, and L. C. Cantley. 1994. Activation of protein kinase C family members by the novel polyphosphoinositides Ptdlns- 3.4-P2 and PtdIns-3,4,5-P3. J Biol. Chem. 269:32358-32367.

70. Towbin, H., T. Staehelin. and J. Gordon. 1979. Electrophoretic Transfer of proteins from SDS and acid/urea polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U. S. A. 76:4350- 4353.

71. Tra nor-Kaplan, A. E., A. L. Harris, B. Thompson, P. Taylor, and L. A.

Sklar. 1988. An inositol tetrakisphosphate-containing phospholipid in activated neutrophils. Nature 334:353-356.

72. Ui, M., T. Okada, K. Hazeki, and O. Hazeki. 1995. Wortmannin as a unique probe for an intracellular signalling protein, phosphoinositide 3-kinase. Trends

Biochem. Sci. 20:303-307.

73. Vlahos, C. J., W. F. Matter, K. Y. Hui, and R. F. Brown. 1994. A Specific

Inhibitor of Phosphatidylinositol 3-kinase, 2-(4-Morpholinyl)-8-phenyl-4H-l- benzopyran-4-one (LY294002). J. Biol. Chem. 269:5241-5248. 74. Volina, S., R. Dhand, B. Vanhaesebroeck, L. K. MacDougall, R. Stein, M.

J. Zvelebil. J. Do min, Ch. Panaretou, and M. D. Waterfield. 1995. A human phosphatidylinositol 3-kinase complex related to the yeast Vps34p- Vρsl5ρ protein sorting system. EMBO J. 14:3339-3348.

75. Volina, S., I. D. Hues, E. Ormondroyd, D. Nizetic, R. Antonacci, M. Rocchi, and M. D. Waterfield. 1994. Molecular cloning, cDNA sequence, and chromosomal localization of the human phosphatidylinositol 3-kinase pllOα (PIK3CA) gene. Genomics 24:472-477.

76. Welsh, G. I., E. J. Foulstone, S. W. Young, J. M. Tavare, and C. G. Proud.

1994. Wortmannin inhibits the effects of insulin and serum on the activities of glycogen synthase kinase-3 and mitogen-activated protein kinase. Biochem. J

303: 15-20.

77. Wennstrsm, S., P. T. Hawkins, F. Cooke, K. Hara, K. Yonezawa, M.

Kasuga, T. Jackson, L. Claesson- Welsh, and L. Stephens. 1994. Activation of phosphoinositide 3-kinase is required for PDGF-stimulated membrane ruffling. Curr. Biol. 4:385-393. 78. Wennstrsm, S., A. Siegbahn, K. Yokote, A. -K. Arvidsson, C. -H. Heldin, S. Mori, and L. Claesson- Welsh. 1994. Membrane ruffling and chemotaxis transduced by the PDGF /3-receptor require the binding site for phosphatidylinositol 3' kinase. Oncogene 9:651-660. 79. Wessel, D. and U. I. FlΫgge. 1984. A Method for the Quantitative Recovery of

Protein in Dilute Solution in the Presence of Detergents and Lipids. Anal. Biochem. 138:141-143.

80. White, M. F. and C. R. Kahn. 1994. The Insulin Signaling System. J. Biol. Chem. 269:1-4. 81. Woscholski, R., T. Kodaki, M. McKinnon, M. D. Waterfield, and P. J.

Parker. 1994. A comparison of demethoxyviridin and wortmannin as inhibitors of phosphatidylinositol 3-kinase. FEBS Lett. 342: 109-114.

82. Wymann, M. P. and A. Arcaro. 1994. PDGF-induced phosphatidylinositol 3- kinase activation mediates actin rearrangements in fibroblasts. Biochem. J. 298:517-520.

83. Yano, H., S. Nakanishi, K. Kimura, N. Hanai, Y. Saitoh, Y. Fukui, Y.

Nonomura, and Y. Matsuda. 1993. Inhibition of Histamine Secretion by Wortmannin through the Blockade of Phosphatidylinositol 3-Kinase in RBL- 2H3 Cells. J. Biol. Chem. 268:25846-25856. 84. Yao, R. and G. M. Cooper. 1995. Requirement for phosphatidylinositol-3 kinase in the prevention of apoptosis by nerve growth factor. Science 267:2003-2006.

85. Yeh, J. I., E. A. Gulve, L. Rameh, and M. J. Birnbaum. 1995. The effects of wortmannin on rat skeletal muscle. Dissociation of signaling pathways for insulin- and contraction-activated hexose transport. J Biol. Chem. 270:2107- 2111.

86. Yu, F. -X., H. -Q. Sun, P. A. Janmey, and H. L. Yin. 1992. Identification of a Phosphoinositide-binding Sequence in an Actin Monomer-binding Domain of Gelsolin. J. Biol. Chem. 267: 14616-14621.

87. Zakian, V. A. 1995. ^T -Related Genes: What Do They Tell Us about Functions of the Human Gene? . Cell 82:685-687. 88. Zoller, M. J., N. C. Nelson, and S. S. Taylor. 1989. Affinity Labeling of cAMP-dependent Protein Kinase with p-Fluorosulfonylbenzoyl Adenosine. J. Biol. Chem. 256:10837-10842.

89. Zvelebil, M. J., G. J. Barton, W. R. Taylor, and M. J. Sternberg. 1987. Prediction of protein secondary structure and active sites using the alignment of homologous sequences. J. Mol. Biol. 195:957-961.

Abbreviations:

AT, ataxia telangiectasia

/3-ME, /3-mercaptoethanol

DNA-PK_CS, DNA-dependent protein kinase DTT, Dithiothreithol

EtOHamine, Ethanolamine

FSBA, 5'-p-Fluorosulfonylbenzoyladenine

GST, Glutathione S-transferase

GuPrAc, /3-Guanidinopropionic Acid MAPK, mitogen activated protein kinase

MLCK, smooth muscle myosin light chain kinase

PI 3-kinase, phosphoinositide 3-kinase

Ptdlns, phosphatidylinositol

PKA, cAMP-dependent protein kinase PKC, protein kinase C

SDS-PAGE, sodiumdodecylsulfate polyacrylamide gel electrophoresis

SH, Src homology domain

TOR, target of rapamycin TX-100, Triton X-100

Table I.

Predicted Lys-C peptide pattern for wortmannin-labelled pllOα

Peptides > 3 kD starting at amino acids > 600 and are listed by increasing molecular weights. The molecular weight was calculated under the assumption that a wortmannin-lysine adduct at the indicated position masks the following cleavage site for Lys-C. Table II

Alignment of Wortmannin-reactive lysine in pllOα with related proteins

FSBA/WT BT pllOα SAKRPLWLN ENPDIMSELLFQNNEIIFKNGDDLRQD LTLQIIRIM 820 FSBA-/WT=K802,nM HS p110α SAKRPLWLNWENPDIMSELLFQNNEIIFKNGDDLRQDMLTLQIIRIM 820 nM HS p110/3 KYMDSKMKPLWLVYNNKVFGEDSVGVIFKNGDDLRQDMLTLQMLRLM 823 nM HS p110r SKKKPLWLEFKCADPTALSN-ETIGIIFKHGDDLRQDMLILQILRIM 850 nM/B*

HS Vps34 TATLFKSALMPAQLFFKTEDGGKYPVIFKHGDDLRQDQLILQIISLM 654 nM/B* SC Vps34 TSKVFKSSLSPLKITFKTTLNQPYALMFKVGDDLRQDQLWQIISLM 642 >μM AT Vps34 SSLFKSALHPLRLTFRTPEEGRSCKLIFKKGDDLRQDQLWQMVWLM 582 ?

SC TOR2 VFSVISSKQRPRKFCIKGSDGKDYKYVLKGHEDIRQDSLVMQLFG V 2146 ? HS FRAP Ξ QVITSKQRPRKLTLMGSNGHEFVFLLKGHEDLRQDERVMQLFG V 2205 ? RAFT1 SLQVITSKQRPRKLTLMGSNGHEFVFLLKGHEDLRQDERVMQLFGLV 2205 ?

HS P14K LRCRSDSEDECSTQEADGQKISWQAAIFKVGDDCRQDMLALQIIDLF 620 >μM SC P14K DQATKKERIRKTSEYGHFEN D CSVIAKTGDDLRQEAFAYQMIQAM 814 μM

MLCK RLGSGKFGQVFRLVEKK TGKVWAGKFFKAYSAKEKENIRDEIS K=548 FSBA-K548/μM/B

PKA TLGTGSFGRVMLVKHKE SGNHYAMKILDKQKWKLKQIEHTLN K=72 FSBA-K72

BT p11Oα pygclsigdcvglie vrnshtimqiqckggl galqf k.=863 Miss¬ alignment

Lipid and protein kinase sequences were aligned around Lys802 (K) of the pl lOα catalytic submit of PI3-kinase. The concentration range at which wortmannin inhibits the respective activities is indicated (nM, μM). B, covalent wortmannin-binding was observed (for refs. see text, *in preparation). FSBA-X and WT = X denote identified reaction sites of kinases with either FSBA or wortmannin. AT, Arabidopsis thaliana; BT, bovine; HS, human, SC, yeast.

Claims

1. An interaction site on PI-3 kinase, or a homolog or analogue thereof, which site when exposed to a modulator modulates the activity of PI-3 kinase, the interaction site comprising a molecular shape which is adapted to interact with at least a part of the modulator so as to modulate the PI-3 kinase activity.

2. An interaction site according to claim 1 wherein the site includes an element capable of forming a covalent linkage with said modulator.

3. An interaction site according to claims 1 or 2 wherein a covalent link is formed with said modulator by virtue of at least one lysine amino acid.

4. An interaction site according to any preceding claim wherein the site is characterised by the provision of lysine at position 802 of the PI-3 kinase, or an equivalent position in a homolog or analogue thereof.

5. An interaction site according to any preceding claim which interaction site includes at least one negatively charged element positioned so as to enhance the binding activity of the covalent linkage or lysine.

6. An interaction site according to any preceding claim which site includes at least one glutamine.

7. An interaction site according to any preceding claim wherein the site is further characterised by a glutamine at position 821 of the PI-3 kinase, or equivalent position in an homolog or analogue thereof.

8. An interaction site according to any preceding claim comprising at least one amino acid selected from proline, asparagine or isoleucine.

9. An interaction site according to any preceding claim further characterised in that a proline is provided at position 786 of the

PI-3 kinase, or an equivalent position in a homolog or analogue thereof.

10. An interaction site according to any preceding claim further characterised in that an asparagine is provided at position 787 of the PI-3 kinase, or an equivalent position in a homolog or analogue thereof.

11. An interaction site according to any preceding claim further characterised in that an isoleucine is provided at position 788 of the PI-3 kinase, or an equivalent position in a homolog or analogue thereof.

12. An antibody raised against the interaction site according to any one of claims 1 - 11.

13. An antibody according to claim 12 wherein said antibody is monoclonal.

14. A method for modulating the activity of PI-3 kinase, or a homolog or analogue thereof, comprising deleting or altering any of the features of the interaction site according to claims 1 to 11.

15. A method according to claim 14 wherein said alteration comprises the substitution of said specified amino acid for an alternative amino acid or any other selected element.

16. A method of identifying or designing ligands capable of modulating PI-3 kinase activity comprising:

a) generating a molecular model of the wortmarmin-inhibition site of PI-3 kinase;

b) identifying existing hgands or designing new hgands which interact at least in part with the inhibition site model, and optionally;

c) contacting the putative hgand with PI-3 kinase under conditions suitable for monitoring PI-3 kinase activity and then monitoring that activity and deteπnining the degree and extent to which the activity is modulated.

17. Ligands identified or designed by the method of claim 16.

18. Ligands which are capable of interacting with the PI- 3 kinase interaction site according to claims 1 to 11 and which have been further modified, having regarding to said interaction sites with a view to modulating the ability of said hgand to interact with said site.

19. Ligands according to claim 19 which comprise suitable interactive elements for interacting with the features of said interaction sites as described in claims 1 to 11.

20. A method of generating a molecular model of the wortmannin inhibition site of PI-3 kinase comprising:

a) identifying conserved amino acid residues within the amino acid sequence for PI-3 kinase;

b) predicting the tertiary structure for such conserved residues;

c) aligning the conserved structural sequences of PI-3 kinase with equivalent portions of a functionally similar kinase such as cAMP-dependent protein kinase (PKA), and

d) ahgning the lysine residues of the active sites of PI- 3 kinase and said PKA whilst disregarding the common P- loop structures;

e) determining the most common conformation for remaining amino acid residues and then determining the thermodynamically most stable conformation for remaining amino acid residues, optionally;

f) inserting the wortmannin structure into the ATP binding site of the model for PI-3 kinase; and

g) adjusting the structure of the model to the most thermodynamically favourable state.