EP1370689A1 - Molecular target validation method - Google Patents

Abstract

A method of investigating the target for a molecule suspected to have inhibitory activity on a biochemical pathway, the method comprising the steps of: Providing further yeast strains which differ from a first yeast strain, which is deficient for one or more genes encoding proteins involved in a yeast biochemical pathway and in.which the deficiency is complemented by the expression of one or more genes encoding proteins of a mammalian biochemical pathway and in which an endpoint indicative of activation of the yeast biochemical pathway can be inhibited by an inhibitory molecule when applied to the first yeast strain under conditions which would normally lead to the activation of yeast biochemical pathway, in that they include mutations in the mammalian gene or one of the mammalian genes; Exposing the further yeast strains to the inhibitory molecule under conditions which would normally lead to the activation of the yeast biochemical pathway; and for each strain, looking for inhibition of the endpoint, Whereby non-inhibition of the endpoint in the presence of the inhibitory molecule indicates that the mutated mammalian gene of that strain is likely to encode the target of the inhibitory molecule.

Description

Molecular Target Validation Method

The present invention relates to methods and materials for identifying and/or validating a molecular species, e.g. a protein, as the target for the action of another molecular species, e.g. a drug, and for identifying mutations in such a target species that may correspond to naturally occurring mutations that contribute to a disease state.

The development and use of selective cell-permeable inhibitor compounds is central to the elucidation of many biochemical processes. The results of these studies j provide valuable information towards rational drug design' and ultimately the cause of human disease. If an inhibitor compound is specific for a particular protein then the effects of that compound are used to implicate that protein target in the regulation of a physiological process. Inhibitor specificity can be verified in vitro against a panel of potential inhibitor substrates.

However, often it may not prove feasible to test every known potential substrate. It is also not possible to test potential cellular substrates that have not been characterised or include substrates for which sensitivity to the inhibitor could not have been predicted. It is therefore important to have methods of target validation which establish whether the cellular effects observed upon inhibitor treatment are due to inhibition of the target, or result from non-specific effects.

A recent approach has been the co-crystallisation of the inhibitor and target. This has led to the identification of a residue in SAPK2a/p38 (stress activated protein kinase 2a/p38, a protein of the mitogen activated protein kinase, or MAPK, family) that is critical for inhibition by the compound SB 203580 (Tong, L. et al . , 1997 Nature Struct. Biol. 4, 311-316) . This type of analysis is very powerful, providing information on the binding site for the drug and an explanation of the compound selectivity. Further, it provides a rational basis for altering residues that are key for the sensitivity of a particular protein to a compound.

Indeed a resistant mutant of SAPK2a/p38 has been designed on the basis of such co-crystallisation studies to have a mutation of a residue critical for SB 203580 inhibition. The inducible expression of the mutant SAPK2a/p38 has been used to assess whether the cellular effects of SB

203580 are mediated by inhibition of SAPK2a/p38 or another enzyme, i.e. to validate SAPK2a/p38 as the target for the various cellular activities of SB 203580 (Eyers,

P. A., 1999 FEBS Lett. 451, 191-196).

However, several factors determine the success of the method which underlies this target validation approach, namely that of co-crystallisation. For example the protein target may or may not crystallise easily; solved protein structures from homologous proteins are often required as they form a useful basis for crystal structure determination of the new target; the inhibitors may be required at many fold molar excess in the co- crystallisation conditions, which can lead to problems of solubility, expense and identification of "false" inhibitor binding sites on a molecule.

The present inventors have thus identified a need for alternative methods of target validation, which may employ methods that are more amenable to a wider scientific skill base, that use lower concentrations of inhibitor compounds, and/or that are suitable for automation. In particular, the present inventors have identified a need for methods of target validation which are suitable for use when a suspected target for an inhibitory molecule has been established, e.g. by in vitro criteria, but which do not require detailed knowledge of the site of action of the inhibitory molecule on the suspected target.

WO94/23039 teaches that in yeast strains which are deficient in MAPKK (i.e. MAPK kinase) and/or MAPKKK (i.e. MAPK kinase kinase) gene activity, the deficiency may be complemented by the coexpression of mammalian MAPKKK and MAPKK genes. Further, inhibition of an endpoint indicative of yeast MAPK pathway activation, under conditions in which the pathway would normally be activated, can be used to test candidate inhibitors for the ability to inhibit the mammalian MAPK pathway gene products.

However, from the results obtainable using the methods and materials of WO94/23039, it is not necessarily deducible which protein in the MAPK pathway is the target of the MAPK pathway inhibitor.

The present invention is a development of the method disclosed in O94/23039 and provides in particular a method of validating a molecular species as the target of an inhibitory molecule.

Accordingly, in a first aspect, the present invention provides a method of investigating the target for a molecule suspected to have inhibitory activity on a biochemical pathway, the method comprising the steps of: providing further yeast strains which differ from a first yeast strain, which is deficient for one or more genes encoding proteins involved in a yeast biochemical pathway and in which the deficiency is complemented by the expression of one or more genes encoding proteins of a mammalian biochemical pathway and in which an endpoint indicative of activation of the yeast biochemical pathway can be inhibited by an inhibitory molecule when applied to the first yeast strain under conditions which would normally lead to the activation of the yeast biochemical pathway, in that they include mutations in the mammalian gene or one of the mammalian genes; exposing the further yeast strains to the inhibitory molecule under conditions which would normally lead to the activation of the yeast biochemical pathway; and for each strain, looking for inhibition of the endpoint, whereby non-inhibition of the endpoint in the presence of the inhibitory molecule indicates that the mutated mammalian gene of that strain is likely to encode the target of the inhibitory molecule.

The first yeast strain may also be provided and exposed to the inhibitor and the endpoint may be looked for in the first strain also, e.g. as a control, although this is_^not an essential part of the method of the invention.

Preferably a plurality of different further yeast strains are provided and screened, preferably simultaneously.

The identification of a mutated strain in which the endpoint is not inhibited on exposure to the inhibitory molecule suggests that the mutation has led to the production of an active, but inhibitor-insensitive, form of a protein of the biochemical pathway. Usually the method will also include, as a control, looking for the endpoint in the absence of inhibitor and such strains will be expected also to show non-inhibition of the endpoint in the absence of inhibitor.

Usually, the method will be used to validate a previously suspected target of the inhibitor. In such a case, the further yeast strains will usually all possess mutations only of the gene encoding the suspected target of the inhibitor. Identification of a suspected target of the inhibitory molecule may be accomplished, prior to carrying out the method of the invention, by any appropriate means, for example by known in vi tro techniques. Typically, purified components of a biochemical pathway are assayed for their ability to induce an endpoint indicative of activation of the pathway in the presence and absence of the inhibitor. If more than one protein is present then the target protein is typically determined by a process of elimination and/or assay of individual steps in the multi-step pathway.

However, it is also contemplated that the steps indicated above may be used when no suspected target has been identified, for example by providing different further yeast strains which respectively include mutations in different genes. In such a case, different strains respectively having mutations in different genes may be provided simultaneously. Preferably, for each further strain, it is known which gene is mutated relative to the first strain.

Following the identification of a mutated strain having resistance to the inhibitor compound, the protein encoded by the mutated gene is preferably overexpressed in a model of the in vivo system in which the inhibitor compound has been shown to exert an effect (e.g. inhibition of a pathway whose activity contributes to a disease state) . Loss of effect of the inhibitor compound under such conditions further validates that the inhibitor compound acts in vivo via the target protein.

If an inhibitor binds preferentially to the target protein in an inactive state, then it is possible that an inhibitor-insensitive yeast strain is expressing (i.e. the mutant gene encodes) a constitutively active form of a protein of the pathway, whereby the protein (and the endpoint) is uncoupled from its upstream activators and non-inhibition of the biochemical pathway is observed. This may be verified by for example transforming a second yeast strain with the mutant gene, the second yeast strain being devoid of the upstream activators of the mutated gene. If the pathway is active in this second yeast strain then an interpretation is that the mutation is causing constitutive activation of the protein. Indeed, the identification of such mutations may be relevant to disease states where sustained increased activity of the mammalian biochemical pathway contributes to the progression of the disease e.g. sustained mammalian MAPK pathway activity contributing to uncontrolled cell proliferation in cancer progression.

For example, patient samples may be screened for genes having corresponding mutations (e.g. mutations encoding proteins which differ from the wild type protein in the same residue (s) as the constitutively active mutant protein encoded by the mutated gene in the yeast strain, preferably having the same mutation (s) ) .

Accordingly it is also contemplated that mutant genes can be transformed into a yeast strain devoid of an upstream regulator and assessed for their ability to induce an endpoint indicative of activation of the pathway (preferably assessed in the presence and absence of the inhibitor) . This relies on the wild type gene not inducing the endpoint indicative of activation of the pathway. If the pathway is active in this yeast strain then an interpretation is that the mutation is causing constitutive activation of the protein. These mutant proteins may be sensitive or insensitive to the inhibitor compound. As outlined previously, identification of such mutations may be relevant to disease states where sustained increased activity of the mammalian biochemical pathway contributes to the progression of the disease.

The yeast strain may be deficient in an upstream regulator of the gene being mutated by being deficient for at least two genes encoding proteins involved in a yeast biochemical pathway, one corresponding to the mutated mammalian gene, the other encoding an upstream activator of the gene. Complementation of the deficiency by expression of only the mutated mammalian gene, in the absence of the mammalian gene encoding the upstream activator, is therefore indicative of the mutated gene encoding a constitutively active form of the protein.

Alternatively, where the yeast gene corresponding to the upstream activator is not capable of activating the mammalian gene (as is the case for yeast MAPKKK and mammalian MAPKK) , then the yeast strain need only be deficient in the yeast gene corresponding to the mammalian gene (e.g. MAPKK) being mutated.

Usually the biochemical pathway will be an enzymatic pathway, i.e. enzymatic activity will be involved in at least one step of the pathway. It is also recognised, however, that steps of a biochemical pathway may for example involve a conformational change in a protein of the pathway, rather than chemical modification of the protein or substrate.

A preferred biochemical pathway is a MAPK pathway, since it has been shown that various mammalian (and other, e.g. Xenopus) MAPK pathway genes can complement deficiencies in yeast MAPK pathways. See O94/23039 and other references cited therein.

While the examples use a rabbit MKK1 gene, it will be understood that genes of other mammals, particularly humans, will be of interest.

As is described in WO94/23039, there are a number of MAPK pathways in mammals and in different yeasts. The names and sources of different MAPK genes, different MAPK kinase (MAPKK) genes and different MAPKK kinase (MAPKKK) genes, and different activators therefor are disclosed in WO94/23039 for yeast and mammalian MAPK pathways and are preferred for use in accordance with the present invention.

In particularly preferred embodiments, the first yeast strain is deficient in at least a yeast MAPKKK and/or an MAPKK gene, which deficiency is complemented by the expression of mammalian MAPKKK and MAPKK genes. The first yeast strain may also be deficient in a yeast MAPK gene and express a complementing mammalian MAPK gene. Especially preferred first yeast strains are those disclosed in WO94/23039, in particular a -byrl (i.e. MAPKK) and/or byr2 (i.e. MAPKKK) mutant strain of Schizosaccharomyces pombe in which the byrl and/or byr2 deficiency is complemented by coexpression of mammalian c-Raf or B-Raf with mammalian MKK1 (i.e. a MAPKK, also referred to as MAPKK1 or MEK1) .

Yeast strains expressing mammalian MAPK may also express one or more mammalian MAPK phosphatases, which may act as the inhibitor for the pathway, as defined above. The MAPK phosphatase (s) may be inducibly expressible, to control exposure of the strain to it/them.

It is also thought that other mammalian intracellular biochemical pathways have closely homologous pathways in yeast, whereby deficiencies in proteins involved in the yeast pathway can be complemented by expression of mammalian genes which encode proteins involved in the homologous mammalian pathways. Any such pathway is appropriate for, and may be used in, the present invention, which can make use of any mammalian protein that is functional in yeast and for which there is a detectable endpoint, representative of activation of that protein and/or a biochemical pathway including that protein (see Figure 1) .

Other pathways suitable for use in accordance with the present invention are those involving stress-activated protein kinases and cyclin-dependent protein kinases, since the proteins involved in these pathways are highly conserved between species, including between yeast and mammal species. For example, human Cdc2 complements a fission yeast cdc2 mutant strain (Lee, M. and Nurse, P. (1987) Nature 327, 31-35), so the present invention can apply also to the regulation of cyclin-dependent kinases (CDKs) by CDK activating kinase species (CAKs) .

Preferably the endpoint of the pathway is the activation of a reporter gene, the reporter gene being under the control of a promoter of a gene which is expressed on activation of the biochemical pathway and which is downstream of the gene which is mutated. The reporter gene may be integrated by homologous recombination into the genome of the host yeast strain in which the mutant protein will be tested. In a preferred embodiment, a reporter gene (e.g. E. coli lacZ gene encoding β- galactosidase) is under the control of a promoter sequence of the pherornone-induced gene ma tl -aFm that is induced by the action of the MAPK pathway as described in WO94/23039 (matl -Pm is referred to as matPm in WO94/23039) . Alternatively, the endpoint may be a cellular event that normally occurs on activation of the pathway, e.g. mating and sporulation for yeast MAPK pathways or progression through the cell division cycle for CDK pathways. There are various established qualitative and quantitative assays for the yeast pheromone response (MAPK pathway) that are well known in the art .

For screening single yeast colonies, a β-galactosidase plate assay may be employed, preferably the highly sensitive assay described in Duttweiler, H. M. (1996) Trends Genet. 12, 340-341. Other reporter genes and assays therefor more suited to automation (e.g. using mul iwell assay plates) are well known in the art and may be used instead of β-galactosidase. Preferably the screening of the invention is automated.

Conditions which would normally lead to the activation of the yeast biochemical pathway may be provided by for example causing the yeast strain to express constitutively active forms of proteins which are upstream in the biochemical pathway of the protein for which the gene is being mutated. Referring to MAPK pathways, the first yeast strain may be transformed with constitutively active forms of Ras. Other methods of constitutively activating MAPK pathways are known from WO94/23039.

Any suitable yeast strain and/or species may be employed in the practice of the invention. Schizosaccharomyces pombe is the yeast used in the examples, though other yeasts are also expected to be suitable, for example Saccharomyces cerevisiae, as used in WO94/23039.

Mutation of the gene may be accomplished by random mutagenesis, particularly (but not exclusively) if there is no prior knowledge of the residues in the target protein that interact with the inhibitor compound. This approach may identify novel residues that are important when the target protein binds inhibitor in its in vivo context rather than in a crystallised state. Mutations affecting such residues may be missed if mutagenesis is based solely on knowledge of the residues involved in binding in the crystallised state.

Random mutagenesis may be performed using available methods e.g. chemical mutagenesis, alanine-scanning mutagenesis, error-prone PCR or by propagating the gene (e.-g. the gene encoding the suspected target protein) in an appropriate plasmid in a mutator strain, e.g. the XLl- Red strain of E. coli (Stratagene, La Jolla, CA, USA). The protocol for this procedure is described in Greener and Callahan (1993) Strategies 7, 32-34.

If, however, some prior knowledge is available of residues that may be important for inhibitor interaction, then a site-directed mutagenesis protocol can be employed using standard methods. Indeed this approach can be applied to particular regions of the protein that have been identified as important in inhibitor interaction by a prior random mutagenesis approach.

Preferably a library of mutants of the mammalian gene of interest is transformed into a suitable yeast strain, resulting in a panel of yeast transformants that express differently (and preferably randomly) mutated variants of the protein of interest. These yeast transformants can be used (e.g. using standard replica plating techniques) to screen a selection of different inhibitor compounds suspected to inhibit the protein encoded by the gene of interest. This is in contrast to the mutants that are produced using co-crystallisation studies to identify the key inhibitor binding residues on the target protein, as mutation of these residues is predicted to be necessarily highly inhibitor-specific.

Panels of different transformants that respectively express mutant versions of different proteins of the biochemical pathway may be used to screen inhibitors even if the target is not known or suspected, though this is not generally preferred.

In -a second aspect, the present invention provides a panel of yeast transformants as defined above. Such a panel facilitates the rapid screening of mutants against a range of inhibitors.

In a third aspect, the present invention provides a system for investigating the target for a molecule having inhibitory activity on a biochemical pathway, wherein a plurality of mutant yeast strains as provided in the first aspect are simultaneously or sequentially (preferably simultaneously) exposed to the inhibitor and monitored for occurrence or non-occurrence of the endpoint .

In a fourth aspect, the invention provides the use of a plurality of mutant yeast strains as provided in the first aspect in a system for investigating the target for a molecule having inhibitory activity on a biochemical pathway.

As with the method of the first aspect, the system may employ a multiwell (e.g. 96 well) assay plate. Preferably the system is automated.

Aspects corresponding to the second, third and fourth aspects are also provided in relation to the above- described method for screening for constitutively active forms of the protein.

When a yeast transformant is identified that exhibits pathway activation in the presence of an inhibitor compound, the mutant gene may be sequenced. The nucleic acid mutations and corresponding amino acid alterations can be identified (typically by first rescuing a plasmid containing that gene) to provide information on residues that may be critical for activation of the target protein, and/or for inhibitor-compound interaction and mechanism of action. Typically, this will be combined with purifying and testing the mutant form of the protein for inhibitor insensitivity and/or constitutive activity in an in vi tro reconstitution assay with purified mammalian pathway components (e.g. purified mammalian MAPK cascade enzymes) . Moreover, such mutations may be present in individuals, and may have implications for the treatment of disorders in which the pathway is implicated (such as the use of pathway inhibitors to treat disorders, particularly proliferative disorders such as cancer; this is especially relevant for the treatment of cancer with inhibitors of MAPK pathways) . Patient samples may therefore be screened for mutations corresponding to those identified in yeast strains which are insensitive to a particular inhibitor. Patients having such corresponding mutations may therefore have reduced susceptibility to treatment with that inhibitor, and treatment with another inhibitor may be more appropriate.

Similarly, mutations identified in constitutively active forms of the protein may exist in naturally "active" forms of the protein that may lead to or predispose to a disease state. Screening diseased tissues for the specific activating mutations identified in this invention may lead to important insights about the contribution of the target protein and its activated pathway to the diseased state.

Furthermore, the inhibitor-insensitive mutant (s) may be included in an in vivo system where the inhibitor has been shown to exert a cellular effect. The cellular effect of the inhibitor would be predicted to be abolished or reduced in cells overexpressing the mutant protein (compared with the in vivo system in the absence of the mutant protein) , given the low relative expression of the wild type (inhibitor-sensitive) protein. This would then validate the wild-type protein as the in vivo target for the inhibitor. Sequencing may be combined with modelling studies of the mutant protein, e.g. in silico modelling based on sequence information (e.g. using commercially available modelling packages) and/or modelling using data from structural studies of the purified mutant protein (e.g. X-ray crystallography and/or NMR) .

Information from such sequencing and modelling may be used in rational drug design.

It is also contemplated that the invention, in all its aspects, could be applied to enhancers or potentiators of a biochemical pathway. Consequently, in further and broader aspects, the invention relates to a method corresponding to that of the first aspect for investigating the target of a pathway modulator and to a system, panel and use corresponding to the second to fourth aspects. However, there is a tendency for such pathways to be overactive in disease states, so inhibitors will usually be of more interest.

References herein to "investigating" the target of a pathway modulator or inhibitor include both determining which protein of the pathway is likely to be the target when no prior knowledge of the suspected target is available, and providing further evidence or even substantial proof that a suspected target is indeed the target. The latter alternative, which is the preferred application of the present invention, is also referred to herein as "target validation". The requirement for such target validation has been emphasised recently with respect to HIV inhibitors (DeClerq, E (2000), Identification of the real molecular target for HIV inhibitors. TIPS 21, p.167). Embodiments of the invention, in its various aspects, will now be described in detail, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 shows in outline the method and system of the present invention; and

Figure 2 shows an embodiment in which mutant forms of mammalian MKK1 are co-expressed with mammalian c-Raf in the presence and absence of an inhibitor of mammalian

MKK1. Filled circles represent yeast colonies expressing β-galactosidase activity (representative of MAPK activation) . Open circles represent yeast colonies with no β-galactosidase activity. Checked circles represent yeast colonies expressing a higher level of β- galactosidase activity than the filled circles and the hatched circles represent a median level of β- galactosidase activity.

Figure 3 shows a screen of S. pombe transformants to identify mutants of mammalian MKK1 that are resistant to PD 184352 using a β-galactosidase reporter gene plate assay. Plate A represents colonies of S . pombe strain CB270 transformed with pREP42 and pARTl vectors alone in quadrant (a) , co-transformed with c-Raf-mycpREP42 and wild type (WT) rMKKlpARTl in quadrant (b) and co- transformed with c-Raf-mycpREP42 and a library of randomly mutated rMKKlpARTl plasmids in (c) . The vehicle for the inhibitor compounds, dimethyl sulphoxide (DMSO) , was added to the agar. Plate B represents a replica of plate A. PD 184352 (lμM final) was added to plate B. The β-galactosidase overlay plate assay was performed as described in Duttweiler, H. M. (1996) TIG 12 340-341) . Arrow 1 denotes an example of a colony of yeast transformants that activate the reporter gene to an increased level relative to that of the transformants in quadrant (b) . This reporter gene activity is not inhibited by PD 184352 (compare plates A and B) . Arrow 2 denotes a colony of yeast transformants in which the reporter gene activity is at a comparable level to that of the transformants in quadrant (b) but this activity is not inhibited significantly in the presence of PD 184352.

Figure 4 shows β-galactosidase reporter gene activity of constitutive and inhibitor-resistant mutant forms of mammalian MKK1 in S. pombe strain CB270.

Pools of 15 transformants per condition were grown in liquid culture in the presence of thiamine (lOμM final) and then 3μl of this growing culture was spotted onto an agar plate containing thiamine and grown overnight at 30°C. These "dot colonies" were then replica plated to plates containing dimethyl sulphoxide in panel A and PD 184352 (2μM final) in panel B. These plates did not contain thiamine which allows expression of the c-Rafmyc protein.

The "dot colonies" in the upper row of Panel A represent colonies of S. pombe strain CB270 transformed with pREP42 and pARTl vectors (v) , transformed with c-Raf-mycpREP42 alone (R) , transformed with wild type rMKKlpARTl alone (WT MKK1) or co-transformed with c-Raf-mycpREP42 and wild type rMKKlpARTl (R + WT MKK1) . The lower row of panel A represents colonies of S. pombe strain CB270 transformed with mutated versions of rMKKl, designated P2 and P3-MKK1 in the presence or absence of c-RafmycpREP42 as indicated. Panel B represents a replica plate of panel A with the exception that PD 184352 (2μM final) was added to the plate in panel B.

Figure 5 shows the constitutive activation of p42 MAPK in HeLa cells overexpressing mutant P2 MKKl . Wild type rabbit MKKl or mutant P2 rabbit MKKl were subcloned into the pCMV-Tag3C mammalian expression vector (commercially available from Stratagene, La Jolla, CA, USA) which results in addition of a c-myc epitope tag at the amino terminus of the MKKl protein. HeLa cell culture and stimulation were carried out as described in Mody, N et al . , (2001) FEBS Letters, 502, 21-24. HeLa cells (at approximately 40% confluence) were transfected with these plasmids using the Fugene™6 transfection reagent (commercially available from Boehringer Mannheim) as per the manufacturer's instructions. A ratio of 0.5μg of MKKlpCMVTag3C DNA:2μl of Fugene™6 reagent was used routinely per well of a 6 well plate. Wild type MKKl (WT MKKl) was expressed for a period of 48 hours whereas mutant P2 MKKl was expressed for 24 hours. Duplicate transfections of P2 MKKl are represented. Prior to stimulation with recombinant human Epidermal Growth Factor (EGF 25ng/ml final) available commercially from In Vitrogen Life Technologies, cells were cultured overnight in the absence of serum. The serum starved cells were pre-treated with PD 184352 (lOOnM final) or the equivalent concentration of DMSO as a control for 1 hour prior to EGF stimulation (5 mins) . Cell lysates were prepared as described by Mody et al . , (2001) FEBS Letters, 502, 21-24 and the phosphorylation status of endogenous p42 and p44 MAPK examined using standard protein separation and immunoblotting procedures (Fig.5A) . The primary antibody used was a rabbit polyclonal antiphospho-p44/42 MAPK antibody available commercially from Cell Signaling Technology, Ine (Beverly, MA, USA). This antibody was used at a 1:1000 dilution and detects human p42 and p44 MAPK only when catalytically activated by phosphorylation at Thr202 and Tyr204. Affinity purified anti-rabbit antibodies coupled to horseradish peroxidase available commercially from Pierce (Rockford, IL, USA) were used at a 1:5000 dilution. Immunoreactive proteins were detected by enhanced chemiluminescence (Amersham Pharmacia Biotech UK, Buckinghamshire, UK) and the light emitted detected on X-ray film. Lanes 1-3 represent untransfected HeLa cells (UT) , lanes 4-6 HeLa cells overexpressing wild type MKKl (WT MKKl) and Lanes 7-12 represent independent duplicates of HeLa cells overexpressing mutant P2 MKKl. Levels of phospho-p44/p42 MAPK were determined under basal conditions, (lanes 1,4,7 and 10), following EGF stimulation (lanes 2, 5, 8 and 11) and following EGF stimulation in the presence of PD 184352 (lanes 3, 6, 9 and 12) .

In 5B the levels of phosphorylated p42MAPK were quantified using a Fujifilm LAS-1000 CCD camera and the data analysed using Aida 200 (version 2, March 05 1998) software.

Figure 6 shows a shift in the dose dependancy for inhibition of p44/42 MAPK activation by PD 184352 in HeLa cells overexpressing mutant P3 MKKl. Wild type MKKl or mutant P3 MKKl were subcloned into the pCMV-Tag3C mammalian expression vector and the experiment performed as described in Figure 5 with slight modifications. Wild type MKKl (WT MKKl) and mutant P3 MKKl were expressed for a period of 48 hours. Prior to stimulation with recombinant human Epidermal Growth Factor (EGF 25ng/ml final, 5 mins) cells were cultured overnight in the absence of serum. The serum-starved cells were pre- treated with a range of PD 184352 concentrations (0.1- lOOOnM final as indicated) or the equivalent concentration of DMSO (ED) as a control for 1 hour prior to EGF stimulation. Basal levels of phospho-p44/p42 MAPK (B) were measured in cells that were not treated with EGF or PD inhibitor compound.

In 6A phospho-p44/42 MAPK protein was detected by enhanced chemiluminescence and the light emitted detected on X-ray film.

In 6B the level of phospho-p42 MAPK was quantified as in Figure 5 and the data analysed using Kaleidagraph-cl (version 3.5.1, December, 2000) synergy software.

PD 098059 [2-(2'-amino-3'-methoxyphenyl) -oxanaphthalen-4- one] is a synthetic flavone compound developed by Parke- Davis Pharmaceutical Research Division, Warner-Lambert Co., Ann Arbor, MI 48105, USA. It is available commercially from Calbiochem-Novabiochem Biosciences UK, Nottingham NG9 2JR, UK. Amongst many cellular effects it has been demonstrated to reverse the phenotype of several ras-transformed cell lines (Dudley D. T. et al . , (1995.) Proc. Na tl . Acad. Sci . USA 92, 7686-7689). It has been demonstrated that this compound binds selectively to the inactive form of MKKl thereby preventing its activation by c-Raf and other upstream activators (Alessi D. R. et al . , (1995) J. Biol. Chem., 270, 27489-27494). Given these observations and the role of the MAPK pathway in cell proliferation there is considerable interest in the discovery of more potent and soluble MAPK pathway inhibitors as anti-cancer drugs. Two further compounds, U0126 and PD 184352, have been described recently that similarly to PD 098059 inhibit MKKl in a non-competitive manner with respect to both MKKl substrates, ATP and MAPK (Favata M. F. et al . ,

(1998) J. Biol. Chem., 273, 18623-18632; Sebolt-Leopold, J. S. et al . , (1999) Nature Med., 5, 810-816). The exact details of the individual modes of action of all of these MKKl inhibitor compounds in a cellular context remain to be elucidated. The methodology of the invention is designed to investigate pathway inhibitors, in this case MKKl inhibitors, that may exhibit different mechanisms of action e.g. both by blocking kinase activation or indeed blocking kinase activity.

The present inventors have used PD 184352 in a yeast strain in which a deficiency in byrl (a MAPKK gene of S. pombe) is complemented by co-expression of mammalian c- Raf (a mammalian MAPKKK) with MKKl (a mammalian MAPKK) . Such yeast strains are generally known from WO94/23039.

PD 184352 [2- (2-chloro-4-iodo-phenylamino) -N- cyclopropylmethoxy-3, 4-difluoro-benzamide] , has been developed by Parke-Davis Pharmaceutical Research Division and is commercially available from Calbiochem-Novabiochem Biosciences, UK.

The inventors have demonstrated the inhibition of an observable endpoint with compounds, particularly PD 184352 (but also PD 098059, for which see below), that have been reported to inhibit MKKl.

According to the present invention, they have now mutated the mammalian MKKl in an attempt to find kinase active, inhibitor-insensitive forms of the protein, as a proof- of-principle study to demonstrate that the methods of the invention can be used to validate that the action of a inhibitor is mediated via the candidate protein target it is directed towards.

In particular, the goals of the study were:

1. To use the reconstitution of the mammalian c-Raf and MKKl kinase activity in S. pombe to screen random mutants of MKKl and identify kinase-active but inhibitor-insensitive (e.g. PD 098059- or PD 184352- insensitive) mutant (s).

2. To identify the mutated residue (s) in the inhibitor- insensitive MKKl mutant (s) and hence gain information on the potential inhibitor binding site, thereby providing a basis for rational drug design.

3. To identify whether the kinase-active but inhibitor- insensitive MKKl mutant (s) could be constitutively active MKKl mutants. Such mutants are identified by their loss of upstream activator dependence in the S. pombe MAPK reconstitution assay. Definition of the altered residues provides information about both MKKl regulation and identifies candidate residues that may be mutated in vivo in disease states e.g. cancer progression.

Experimental rationale:

Mammalian c-Raf is expressed in S. pombe under the control of the nmtl promoter (as disclosed in WO94/23039) . This promoter is thiamine repressible and therefore induced in the absence of thiamine in the growth media. Wild type mammalian MKKl or mutated mammalian MKKl is expressed under the control of the adh promoter and is expressed constitutively. The rationale behind the use of the differently regulated promoters is to accommodate the mechanism of action of the inhibitor compound PD 098059 which is believed to bind selectively to the inactive form of MKKl thereby preventing its activation by upstream activators. The constitutive expression of the MKKl protein should ensure exposure of the non-activated form of MKKl to the inhibitor compound PD 098059 prior to the expression of its upstream activator c-Raf. This experimental design will not preclude study of inhibitors that act to inhibit the kinase activity of MKKl.

A highly sensitive β-galactosidase plate assay is used which allows detection of increases in reporter gene activity at the single colony level (Duttweiler, H. M. (1996) TIG 12 340-341) .

The following description will refer primarily to PD 184352, which is a preferred MKKl inhibitor. However, preliminary experiments showed similar results (reported in the priority application, GB 0106641.4) for PD 098059 also. Differences between the experimental details used in -the earlier PD 098059 experiments and the PD 184352 experiments reported here are indicated in parentheses.

PD 184352, however, is the preferred inhibitor, because it is more potent than PD 098059 and currently the subject of clinical trials for the treatment of cancer.

PD 184352 (1 μM final; lOμM final for PD 098059) is added to the plate medium and is active and maximally inhibitory under these conditions. The ability to assay single colonies for inhibitor sensitivity provides a powerful method of screening for colonies expressing mutant MKKl proteins that show reduced or total insensitivity to inhibition by PD 184352.

Replica plating is a standard procedure that in this application enables identical colonies to be treated with different inhibitor compounds and at varying concentrations of each inhibitor. Further, it ensures that viable colonies that display altered sensitivity to PD 184352 can be recovered from an identical plate that has not been exposed to inhibitor compound.

Subsequent to rescue of a plasmid from a yeast transformant that displays altered sensitivity to the inhibitor compound the gene encoding the target protein (e.g. MKKl) is sequenced completely and mutations identified. A recombinant version of this novel mutated target protein (MKKl) is purified and is examined using both in vi tro and in vivo approaches.

The in vitro analysis is undertaken to determine whether the mutated target protein retains all of the biochemical characteristics of the wild type protein with the exception of sensitivity to the inhibitor compound. Detailed kinetic and pharmacological studies are performed using purified, recombinant proteins of the mammalian pathway of interest. This is particularly useful as in the initial screening procedure the mutated mammalian proteins may have interacted with proteins of the yeast pathway. This in vitro approach will confirm that the mutated target proteins can still interact with proteins of the mammalian pathway. Finally, one can demonstrate that the mutated, inhibitor- insensitive target protein when expressed in an in vivo model system of choice (that is normally sensitive to the effects of the inhibitor molecule) confers inhibitor compound insensitivity to this model system. This provides final confirmation (validation) that the wild type version of the target protein is the actual in vivo target of the inhibitor compound.

The experimental protocol to screen for mutated, inhibitor-insensitive target proteins:

Mammalian c-Raf and MKKl kinases are coexpressed from multicopy plasmids in the byrl mutant S. pombe strain that has the Spkl (MAPK) -responsive reporter gene integrated stably into its genome.

A reporter construct consisting of the promoter sequence of the pheromone-induced gene matl -Pm (which is induced by the MAPK pathway) upstream of the E. coli lacZ gene encoding β-galactosidase was integrated by homologous recombination at the his5 locus in a S . pombe strain that has the jbyrl locus disrupted by ura4⁺ . The ura4⁺ gene in this disrupted jbyrl locus is then made non-functional by homologous recombination of an internal fragment of the ura^'X gene, so as to allow transformation of the strain with plasmids carrying the uraX gene for selection. Transformants resistant to 5-fluoroorotic acid (FOA) , which selects against cells containing the normal ura X gene product, are selected. The full genotype of the strain is CB270: h⁹⁰ jbyrl :: ura4^~ΔRS adeβ-216 leul-32 ura4D-18 his5: ^•. ma tl -PmAΥ -lacZ . Two additional "TR" motifs were inserted into the promoter of the matl -Pm gene using standard molecular biology techniques. These are "T-rich" sequences that are important for the binding of the transcription factor Stell that is activated by S . pombe MAPK (Spkl) .

Referring firstly to Figure 3, the S. pombe strain CB270 was cotransformed with multicopy plasmids encoding wild type human c-Raf and either wild type or mutated versions of rabbit MKKl using a standard lithium acetate procedure (Fission Yeast Handbook at http : //www . bio . uva . nl/pombe/handbook) . c-Raf was myc epitope tagged at the c-terminus and under the control of a mutated nmtl promoter that does not provide full induction strength but is still repressed by thiamine (Basi, G., Schmid, E and Maundrell, K. , (1993) Gene, 123, 131-136). This plasmid is designated c-Raf-mycpREP42. Rabbit MKKl is under the control of a promoter for the adh gene, which is constitutively expressed, and either wild type forms or randomly mutated versions of this plasmid are introduced into the S. pombe CB270 strain. This plasmid is designated rMKKlpARTl.

The rMKKlpARTl plasmid was mutated randomly using error- prone PCR to generate a mutant library as described by Leung and co-workers (Leung D. W et al . , (1989) Technique, 1, 11-15) .

(In the earlier PD 098059 experiments, the rMKKlpARTl plasmid was mutated randomly by transformation and amplification in the XLl-Red E. coli mutator strain as outlined in the experimental protocol provided with this strain (Greener, A and Callahan, M (1993) Strategies 7, 32-34) . The XLl-Red mutator strain is available commercially from Stratagene, 11011 North Torrey Pines Road, La Jolla, California 92037.) Single colonies from CB270 transformed with vectors (pREP42 and pARTl) , c-Raf-mycpREP42 or rMKKlpARTl alone do not produce detectable β-galactosidase activity in a sensitive plate assay (Duttweiler, H. M (1996) TIG 12 340-341) . Indeed, only when c-Raf-mycpREP42 and rMKKlpARTl are co-transformed into CB270 is activity detectable as blue staining single colonies (as shown in quadrant (b) of Figure 3) . The β-galactosidase plate assay is performed exactly as described by the aforementioned authors with the exception that the PD 184352 inhibitor compound or the equivalent concentration of the vehicle, DMSO, are included (or not) in the liquid agar prior to pouring the plates. The assay is performed on replica plates as outlined in Figure 2 and shown in

Figure 3. The inventors have established conditions where PD 184352 (1 μM final concentration; lOμM for PD 098059) abolishes the c-Raf-stimulated activation of wild type MKKl as assessed by complete inhibition of the β- galactosidase activity in the single colony plate assay (compare quadrant (b) in plates A and B of Figure 3) .

Transformants co-expressing either wild type or mutant MKKl proteins and wild type c-Raf are selected by their ability to overcome an auxotrophic deficiency. These transformants are selected in the presence of thiamine (lOμM final concentration) to repress the expression of c-Raf. Replicas of these plates are then grown in the absence of thiamine (to allow induction of the expression of c-Raf as the colonies grow) and absence or presence of the PD 184352 inhibitor compound (1 μM final concentration; lOμM for PD 098059) . The MKKl protein has been expressed throughout the selection and replica procedure and should bind the inhibitor (or not in the case of an insensitive mutant protein) prior to being activated by c-Raf that will be expressed maximally when the nmtl promoter is derepressed fully (16-20 hours) . Indeed, the procedure has been designed to accommodate a mechanism proposed for the PD 098059 inhibitor in preventing the activation of MKKl by upstream activators rather than inhibiting c-Raf-activated MKKl (Alessi et al . , 1995, J. Biol., Chem., 27489-27494). However, the methodology is also suited to inhibitors that act to inhibit the actual kinase activity of MKKl.

A PD 184352 concentration of 1 μM final (lOμM final for PD 098059) in the plates is sufficient compound to inhibit completely the c-Raf-stimulated β-galactosidase activity in yeast colonies expressing both c-Raf and wild type MKKl (compare plates A and B, quadrants (b) in Figure 3) . Using this initial screen concentration of PD 184352 it is proposed that a mutant MKKl protein could be identified in a yeast colony that remained blue to an equal intensity in the absence and presence of the compound (see colonies 1 and 2 in Figure 3) . There are a number of possible scenarios with mutated MKKl proteins displaying altered kinase activity relative to the wild- type protein even in the absence of compound i.e. inactive, equally active, less active or more active (see Fig. 2) . Indeed in the presence of the compound these mutants could all retain the ability to bind the compound or may be totally inhibitor insensitive. Equally these MKKl mutants could display varying degrees of inhibitor sensitivity (a possibility omitted from Fig. 2 for simplicity) . Yeast colonies that display any β- galactosidase activity when assayed in the presence of 1 μM PD 184352 (lOμM PD 098059) are then picked and re- streaked to single colony and single colonies are re- tested for PD 184352-insensitivity.

In the case of PD 098059 it is well established that agonists that strongly activate the MAPK pathway require more PD 098059 to inhibit this activation than those that stimulate this pathway to a lesser extent (Alessi D. R. et al . , 1995, J. Biol. Chem., 27489-27494). Hence, the mutated MKKl proteins that are most likely to aid in elucidation of the inhibitor binding site on MKKl are those that display a similar level of kinase activity to the wild type protein in the absence of inhibitor (see colony 2, plate A, Figure 3) , as opposed to those that exhibit enhanced kinase activity and merely require higher concentrations of inhibitor to block their activation (see colony 1, plate A, Figure 3) . The same principle is generally applicable to other inhibitors and pathways.

MKKl plasmids are then rescued from cultures derived from the single yeast colonies that remain insensitive to PD 184352 using the method described by Topal and co-workers (Topal, A., et al . , (1997) Elsevier Trends Journals Technical Tips Online, http://tto.biomednet.com). These cultures are grown in media containing uracil that, in this case, will encourage the loss of the c-Raf multicopy plasmid reducing the chances of co-rescue of this plasmid.

The rescued rMKKlpARTl plasmids are then re-transformed into strain CB270 in the presence and also the absence of c-Raf-mycpREP42 plasmid and re-screened as described previously and outlined in Figure 2. This re-screening procedure should confirm the PD 184352-insensitivity of the MKKl proteins expressed from the rescued rMKKlpARTl plasmids. Mutant MKKl proteins that are selected in the first screen on the basis of PD 184352 insensitivity may be constitutively active forms and this will be confirmed by activation of β-galactosidase activity in the absence of the upstream regulator, c-Raf.

Figure 4 demonstrates that re-transformation of strain CB 270 with rescued mutant MKKl plasmids (that were generated by error prone PCR and expressed in the pARTl plasmid) designated P2-MKK1 and P3-MKK1 does confer resistance to PD 184352 (2μM final) . Co-transformation of wild type MKKl with c-Raf activates the β- galactosidase reporter gene readout. This activity is completely inhibited in the presence of the PD 184352 compound (compare dot "R+ wtMKKl" in the top row of panels A and B, Figure 4) . However a significant amount of β-galactosidase activity remains in the presence of PD 184352 when CB270 is co-transformed with mutant P2-MKK1 or mutant P3-MKK1 and c-Raf (compare the dots in the lower row of panels A and B) . Transformation of mutant P2-MKKlpARTl alone results in the activation of the β- galactosidase reporter gene readout indicating the loss of dependance on the upstream activator, c-Raf and the isolation of a constitutively active mutant of MKKl. This constitutive activity was however sensitive to inhibition by PD 184352.

The rescued plasmids encoding kinase active, c-Raf- dependent, PD 184352-insensitive MKKl mutants and constitutively active, PD 184352-insensitive MKKl mutants are then sequenced using standard procedures and the mutations identified. This sequence information can be used to corroborate the results obtained with the rescued PD 184352-insensitive plasmids by generating the specific mutant rMKKlpARTl plasmids using site-directed mutagenesis (SDM) . This serves as an additional test that the mutatio (s) do confer increased resistance to PD 184352. SDM of the wild type MKKl protein in the yeast expression vector (pARTl) is performed using standard techniques (e.g. using the Quikchange Site-Directed Mutagenesis Kit supplied by Stratagene, 11011 North Torrey Pines Road, La Jolla, California 92037). When using the SDM approach it is important to ensure only the desired mutations have been introduced into the expression plasmid and this is done using standard, automated DNA sequencing procedures. The mutant MKKl proteins can then be re-tested in the yeast assay prior to purification and in vitro analysis. The experimental protocol for the purification of wild- type and mutated target proteins:

Mutant MKKl cDNA can then be subcloned from the pARTl plasmid into a suitable vector for protein expression and purification e.g. the pGEX series of vectors for protein expression and purification in E. coli (available commercially from Amersham Pharmacia Biotech UK, Buckinghamshire, UK) . The pGEX series of vectors enable subcloning of the gene of interest as a fusion with Schistosoma japonicum glutathione S-transferase (GST) . The resulting GST fusion protein can be purified by affinity chromatography on glutathione-Sepharose beads using standard procedures. Alternatively, if the gene for the wild type version of the target protein of interest is already cloned into a vector suitable for expression and purification then site-directed mutagenesis of this plasmid is performed as outlined previously.

In this embodiment the mutated MKKl proteins were prepared by SDM of wild type rabbit MKKl that had been cloned into pGEX3X as described in Alessi D. R. , et al . , (1995) Methods in Enzymology, 255, 279-290. The MKKl proteins are expressed as GST fusion proteins that have been modified at the carboxyl terminus to encode six histidine residues. The mutated and wild type MKKl proteins can thus be affinity purified on glutathione- Sepharose followed by nickel nitrilotriacetate agarose purification using standard procedures.

The experimental protocol for in vi tro kinase assays:

Purified versions of mutated and wild type MKKl (or the target proteins of interest) are then tested in an in vitro reconstituted assay with purified components of the pathway of interest. The yeast screen has identified these mutated proteins in a cellular background of yeast proteins (in this embodiment yeast MAPK pathway homologues) and it is a useful further step to verify that these mutated proteins can indeed interact with purified mammalian components e.g. upstream kinase regulators and the downstream target kinase, p42MAPK.

In this embodiment the mutated MKKl proteins are assayed in a "coupled kinase assay" for their ability to activate p42MAPK , which is then assayed by phosphorylation of myelin basic protein (MBP) as outlined in detail in

Alessi D. R. et al . , (1995) Methods in Enzymology 255, 279-290. The assay can be performed in the absence and presence of inhibitor compounds e.g. PD 184352 or PD 098059 and the mutants tested as to whether they retain inhibitor insensitivity in this in vitro assay. Purified MAPKKK e.g. c-Raf acts as the upstream activating kinase in this in vitro assay and mutants not giving further activation of the pathway in the presence of this enzyme suggest identification of constitutively active forms of MKKl (or the relevant downstream protein) .

The experimental protocol for in vivo assays:

The cDNAs for the inhibitor-insensitive mutant proteins are subcloned into a vector suitable for over-expression in established mammalian cell lines in which the inhibitor compound of interest has been demonstrated to exert an effect. Loss of effect of the compound in the mutant over-expressing cells provides strong evidence of target validation. Equally the cDNAs for constitutively active mutant proteins are subcloned into such vectors and tested for their ability to activate the pathway of interest (e.g. the MAPK pathway) in the absence of activation by upstream regulators.

The PD 184352 compound has been shown to inhibit the phosphorylation and activity of p44/p42 MAPK in a human epithelial cell line, HeLa (Mody, N et al . , (2001) FEBS Letters, 502, 21-24) . In Figures 5 and 6 both wild type MKKl (WT MKKl) and the mutant MKKl plasmids designated P2-MKK1 and P3-MKK1 were subcloned from the pARTl vector into the pCMVTag3c expression vector and transfected into HeLa cells as outlined in detail previously. The pCMVTag3C vector adds an amino terminal c-myc epitope to the MKKl proteins and overexpression of the transfected MKKl proteins can be detected with a commercially available Myc-Tag 9B11 monoclonal antibody (Cell Signaling Technology, Inc., Beverly, MA, USA). Overexpression of equivalent levels of the transfected

MKKl proteins was confirmed for the experiments described in Figures 5 and 6 (data not shown) .

The P2-MKK1 mutant was identified in the yeast reporter gene assay as a constitutively active mutant that when expressed in the absence of c-Raf increased β- galactosidase activity (panel A, Figure 4) . However, this constitutive activity remained sensitive to inhibition by PD 184352. The c-Raf-stimulated activity of the P2-MKK1 mutant did show resistance to complete inhibition by PD 184352 (compare panels A and B, Figure 4) .

Figure 5A represents a Western immunoblot of HeLa cell lysates probed with an antiserum that recognises the endogenous p42 and p44 forms of MAPK only when they are activated catalytically by phosphorylation on threonine and tyrosine residues (Thr202/Tyr204) in human MAPK. Interestingly, overexpression of mutant P2-MKK1 in the HeLa cells increased the level of phosphorylation of p42MAPK in the basal, non-epidermal growth factor (EGF) - stimulated condition (Lanes 7 and 10 in Figure 5A) . This constitutive activity was not observed in the untransfected or wild type MKKl transfected HeLa cells (lanes 1 and 4, Figure 5A) . Quantification of this data demonstrated an increase in phopsho-p42MAPK levels in the absence of EGF-stimulation in HeLa cells overexpressing mutant P2-MKK1 (Figure 5B) . Using a commercially available antiserum, that detects total (phosphorylation- state independent) p44/42 MAPK, the expression of the total endogenous p44/p42 MAPK remained unchanged for each experimental condition tested in Figure 5. This observation confirmed that the increased levels of phospho-p42 MAPK could not be explained by mutant P2-MKK1 altering the expression of endogenous MAPK (data not shown) . The EGF-stimulated increase in phospho-p44/42 MAPK was inhibited totally by pre-incubation of the HeLa cells with PD 184352 (lOOnM) prior to EGF-stimulation in all the conditions tested (lanes 3,6,9 and 12). Thus, in a HeLa cell based assay MKKl mutant P2 can activate p42MAPK in a constitutive manner demonstrating that such mutants can be isolated from the yeast based screen. In this example mutant P2-MKK1 is however not insensitive to the PD 184352 inhibitor compound. Further analysis of the mutated amino acid residues will provide novel insights into the mechanism of activation of MKKl and may identify mutations in MKKl that exist in cancerous disease states.

Figure 6A represents a Western immunoblot of HeLa cell lysates probed with the anti-phospho p44/42 MAPK antiserum used in Figure 5 (described in detail previously) . This figure demonstrates that HeLa cells overexpressing mutant P3-MKK1 cause a shift in the dose dependency of PD 184352-mediated inhibition of p44/42 MAPK phosphorylation. Significant levels of phosphorylation of p42/p44 MAPK are observed in the cells overexpressing mutant P3-MKK1 at concentrations of PD 184352 that are completely inhibitory in cells overexpressing wild type MKKl (Figure 6A) . Quantification of the levels of phospho-p42MAPK and analysis of the dose-dependent inhibition curves using Kaleidagraph software resulted in estimates for the IC₅o values (a 50% inhibitory concentration) for the transfected HeLa cells (Figure 6B) . Approximate values of IC₅₀ =0.9nM for the wild type transfected HeLa cells and IC₅₀ =33nM for the mutant P3-MKK1 transfected HeLa cells demonstrate that a greater concentration of the PD inhibitor is required to inhibit MAPK phosphorylation in cells overexpressing mutant P3 MKKl. This mutant P3- MKKl-mediated effect is likely to be an underestimate due to the contribution of endogenous wild type MKKl, in the transfected HeLa cells, to both the activation of p42 MAPK and its inhibition by PD 184352.

Thus mutant P3-MKK1 represents a mutant MKKl protein isolated in the yeast screen for PD 184352-insensitive MKKl mutants that when overexpressed in a HeLa cell confers increased resistance to PD 184352. This indicates that MKKl is indeed the in vivo target for the PD 184352 compound and that the mutated residues in mutant P3 MKKl play a role in conferring the sensitivity of the MKKl target to this inhibitor compound.

Human tumour cell lines, particularly those derived from colon tumours (e.g. colon 26, HT-29 and colo205), provide an excellent system to test the effects of MKKl inhibitors, e.g. PD 184352, as demonstrated by the work of Sebolt-Leopold and her co-workers (Sebolt-Leopold J. S., et al (1999) Nature Med. , 5, 810-816). Over- expression of mutant MKKl proteins, e.g. mutant P3-MKKl, in these cell lines and observation of the loss of the inhibitory effect of PD 184352 on a number of the biological assays described by these authors will provide strong evidence that MKKl is the in vivo target of PD 184352.

PD 098059 has been shown to exert many in vivo effects (Cohen, P (1997) Trends Cell Biol . , 7, 353-361). Suitable rodent models are the ras-transformed cell lines e.g. K-Balb and KNRK cells (which are ras-transformed BALB 3T3 mouse and normal rat kidney cells, respectively) . Two phenotypes typical of ras-transformed cells are altered by continuous PD 098059 treatment of these cell lines. The ability to grow in soft agar is reduced and cells change their cell morphology from the typical rounded and loosely attached transformed phenotype to flatter and more spread out ^'forms (Dudley D. T., et al . , (1995) PNAS, 92, 7686-7689). The PD 098059- insensitive mutant forms of MKKl, as isolated in yeast screens using the PD 098059 inhibitor, (and wild type MKKl as a control) are over-expressed in these cell lines. Abolishing or significantly reducing the ability of this compound to reduce soft agar growth and alter cell morphology would provide strong evidence that MKKl was the in vivo target for the inhibitor compound. Hence the suspected target would be validated.

Identification of a panel of mutants should allow validation of MKKl as the in vivo target for the PD inhibitor compounds PD 184352 and PD 098059. These mutants will give important information on the binding site(s) for these inhibitors. Interestingly, recent studies suggest that the U0126 inhibitor compound may share a common or overlapping binding site on MKKl (Favata et al. (1998) J. Biol. Chem. 273, 18623-18632) . With this information alternative inhibitor compounds can be designed on a rational basis, providing potentially important advances in therapeutic intervention.

The identification of a panel of constitutively active MKKl mutant proteins, and the mutated residues therein, will provide information about how the wild-type protein is _^regulated.

Screening and identifying cancer cell lines harbouring these "activating mutations" will provide important information about cancer progression. References

Alessi, D. R. , Cohen, P., Ashworth, A., Cowley, S.,

Leevers S. J. , and Marshall, C. J. (1995) Assay and expression of mitogen-activated protein kinase, MAP kinase kinase, and Raf. Methods Enzymol . , 255, 279-290.

Alessi D. R. , Cuenda, A., Cohen, P., Dudley D. T. and Saltiel, A. R. (1995) PD 098059 is a specific inhibitor of the activation of mitogen-activated protein kinase kinase in vitro and in vivo. J. Biol . Chem . , 270, 27489- 27494

Basi, G., Schmid, E and Maundrell, K. (1993) TATA box mutations in the Schizosaccharomyces pombe nmtl promoter affect transcription efficiency but not the transcription start point or thiamine repressibility . Gene, 123, 131- 136 Cohen, P. (1997) The search for physiological substrates of MAP and SAP kinases in mammalian cells. Trends Cell Biol . , 7, 353-361.

DeClerq, E., (2000) Identification of the real molecular target for HIV inhibitors. TIPS 21, p.167

Dudley, D. T., Pang, L., Decker, S. J. , Bridges, A. J. , and Saltiel, A. R. (1995; A synthetic inhibitor of the mitogen-activated protein kinase cascade. Proc . Natl . Acad. Sci . USA 92, 7686-7689

Duttweiler, H. M. (1996) A highly sensitive and non- lethal β-galactosidase plate assay for yeast. Trends Genet . 12, 340-341

Eyers, P. A., Ijssel, P., Quinlan, R. A., Goedert, M. and Cohen, P. (1999) Use of a drug-resistant mutant of stress-activated protein kinase 2a/p38 to validate the in vivo specificity of SB 203580. FEBS lett . 451, 191-196.

Favata M. F., Horiuchi, K. Y., Manos, E. J. , Daulerio, A. J. , Stradley, D. A., Feeser, W. S., Van Dyk, D. E., Pitts, W. J., Earl, R. A., Hobbs, F. , Copeland, R. A., Magolda, R. L., Scherle, P. A., and Trzaskos, J. M. (1998) Identification of a novel inhibitor of mitogen- activated protein kinase kinase J. Biol . Chem . , 273, 18623-18632.

Greener A. and Callahan M. (1993) XLl-Red: A Highly efficient random mutagenesis strain. Strategies 1 , 32-34 Lee, M and Nurse P. (1987) Complementation used to clone a human homologue of the fission yeast cell cycle control gene cdc2. Nature 327, 31-35.

Leung, D. W. , Chen, E and Goeddel D. V. (1989) A method for random mutagenesis of a defined DNA segment using a modified polymerase chain reaction. Technique, 1 , 11-15 Mody, H., Leitch, J. , Armstrong, C, Dixon, J and Cohen, P (2001) Effects of MAP kinase cascade inhibitors on the MKK5/ERK5 pathway. FEBS Lett . 502, 21-24

Sebolt-Leopold, J. S., Dudley, D. T., Herrera, R. , Becelaere, K, V., Wiland, A., Gowan, R. C, Tecle, H., Barrett, S. D. Bridges, A., Przybranowski, S., Leopold, W. R. and Saltiel, A. R. (1999) Blockade of the MAP kinase pathway suppresses growth of colon tumours in vivo . Nature Med. , 5, 810-816

Tong, L., Pav, S., White, D. M. , Rogers S., Crane K, M. , Cywin C.L., Brown M. L. and Pargellis C. A. (1997) A highly specific inhibitor of human p38 MAP kinase binds in the ATP pocket. Na ture Struct . Biol . 4, 311-316.

Topal, A., Karaer, S. and Temizkan G. (1997) A simple method for rescuing autonomous plasmids from fission yeast. Elsevier Trends Journals Technical Tips Online, http : //tto . biomednet . com

Claims

1. A method of investigating the target for a molecule suspected to have inhibitory activity on a biochemical pathway, the method comprising the steps of: providing further yeast strains which differ from a first yeast strain, which is deficient for one or more genes encoding proteins involved in a yeast biochemical pathway and in which the deficiency is complemented by the expression of one or more genes encoding proteins of a mammalian biochemical pathway and in which an endpoint indicative of activation of the yeast biochemical pathway can be inhibited by an inhibitory molecule when applied to the first yeast strain under conditions which would normally lead to the activation of the yeast biochemical pathway, in that they include mutations in the mammalian gene or one of the mammalian genes; exposing the further yeast strains to the inhibitory molecule under conditions which would normally lead to the activation of the yeast biochemical pathway; and for each strain, looking for inhibition of the endpoint, whereby non-inhibition of the endpoint in the presence of the inhibitory molecule indicates that the mutated mammalian gene of that strain is likely to encode the target of the inhibitory molecule.

2. The method of claim 1, further comprising providing, as a control, the first yeast strain; exposing the first yeast strain to the inhibitor; and, for the first yeast strain, looking for the endpoint.

3. The method of claim 1 or claim 2, wherein a plurality of different further yeast strains are provided and screened.

4. The method of any preceding claim including, as a control, looking for the endpoint in the absence of inhibitor.

5. The method of any preceding claim wherein the further yeast strains possess mutations only of a gene encoding a suspected target of the inhibitor.

6. The method of any preceding claim further comprising the step, following the identification of a mutated strain having resistance to the inhibitor compound, of overexpressing the protein encoded by the mutated gene in a model of the in vivo system in which the inhibitor compound has been shown to exert an effect.

7. The method of any preceding claim further comprising the step, following identification of an inhibitor- insensitive mutated strain, of transforming a second yeast strain with the mutated gene of the inhibitor- insensitive mutated strain, wherein the second yeast differs from said first yeast strain in that prior to said transformation it is deficient in both a gene corresponding to the mutant gene and an upstream activator for the wild type mammalian gene corresponding to ...the mutated gene, wherein observation of the endpoint is indicative of the mutated gene encoding a constitutively active form of the protein.

8. The method of claim 7 further comprising the step, following identification of a mutated gene encoding a constitutively active form of the protein, of screening patient samples for the presence of a gene having a corresponding mutation.

9. The method of any preceding claim further comprising the step, following identification of a mutated gene encoding an inhibitor-insensitive form of the protein, of screening patient samples for the presence of a gene having a corresponding mutation.

10. The method of any preceding claim, wherein the biochemical pathway is an enzymatic pathway.

11. The method of claim 10, wherein the pathway is a MAPK pathway.

12. The method of claim 11, wherein the first yeast strain is deficient in a yeast MAPKKK gene and/or a yeast MAPKK genes, the deficiency being complemented by the expression of mammalian MAPKKK and MAPKK genes.

13. The method of claim 12 wherein the first yeast strain is a jbyrl (i.e. MAPKK) and/or jbyr2 (i.e. MAPKKK) mutant strain of Schizosaccharomyces pombe in which the jbyrl and/or Jbyr2 deficiency is complemented by coexpression of mammalian c-Raf or B-Raf with mammalian MKKl.

14. The method of any preceding claim wherein the endpoint of the pathway is the activation of a reporter gene.

15. The method of claim 14 wherein the reporter gene is under the control of a promoter sequence of the pheromone-induced gene matl -Pm .

16. The method of any preceding claim wherein the mutations are generated by random mutagenesis.

17. The method of claim 16 wherein the mutations are generated by error-prone PCR.

18. A system for investigating the target for a molecule • having inhibitory activity on a biochemical pathway, wherein a plurality of said further yeast strains as defined in any preceding claim are simultaneously or sequentially exposed to the inhibitor and monitored for occurrence or non-occurrence of the endpoint.

19. The use of a plurality of said further yeast strains as defined in any one of claims 1 to 15 in a system for investigating the target for a molecule having inhibitory activity on a biochemical pathway.

20. A method of screening mutated forms of a gene encoding a protein involved in a biochemical pathway for a mutant encoding a constitutively active form of the protein, the method comprising the steps of: providing further yeast strains that differ from a first yeast strain, which is deficient for a first yeast gene encoding a protein involved in a yeast biochemical pathway and in which the deficiency is complemented by the expression of a first mammalian gene which encodes a corresponding protein of a mammalian biochemical pathway and a second mammalian gene which encodes an upstream activator of said corresponding protein, in that they lack an upstream activator of said corresponding protein and in that they include mutations in the first mammalian gene; and for each strain, looking for an endpoint indicative of activation of the biochemical pathway, whereby observation of the endpoint indicates that the mutated first mammalian gene of that strain is likely to encode a constitutively active form of the protein.

21. A method according to claim 20, wherein said further strains are also deficient for a second yeast gene which encodes an upstream activator of the protein encoded by said first yeast gene.

22. A method according to claim 20, wherein said corresponding mammalian protein is not activated by the endogenous upstream activator of the protein encoded by said first yeast gene.

23. The method of any one of claims 20 to 22, further comprising the step, following identification of a mutated gene encoding a constitutively active form of the protein, of determining the effect on the constitutively active form of the protein of an inhibitor of the wild type form of the protein.

24. The method of any one of claims 20 to 23 further comprising the step, following identification of a mutated gene encoding a constitutively active form of the protein, of screening patient samples for the presence of a gene having a corresponding mutation.

25. A system for screening for a mutant encoding a constitutively active form of a protein involved in a biochemical pathway, wherein a plurality of said further yeast strains as defined in any one of claims 20 to 22 are simultaneously or sequentially monitored for occurrence or non-occurrence of the endpoint.

26. The use of a plurality of said further yeast strains as defined in any one of claims 20 to 22 in a system for screening for a mutant encoding a constitutively active form of a protein involved in a biochemical pathway.

27. A panel comprising a plurality of said further yeast strains as defined in any preceding claim.