Crystallized P38 complexes

Abstract

This invention provides certain crystallized, protein kinase-ligand complexes, in particular P38-ligand complexes, and their structure coordinates. The structure coordinates are based on the structure of a phosphorylated P38&ggr; complex which has now been solved and which reveals new structural information useful for understanding the activated states of other, related kinase proteins as described herein. The key structural features of the proteins, particularly the shape of the substrate binding site, are useful in methods for designing or identifying selective inhibitors of the protein kinases, particularly P38&ggr; and in solving the structures of other proteins with similar features. The structure coordinates may be encoded in a data storage medium for use with a computer for graphical three-dimensional representation of the structure and for computer-aided molecular design of new inhibitors.

Description

TECHNICAL FIELD OF INVENTION

[0001] This application claims priority from U.S. Provisional Applications Serial No. 60/112,354 filed Dec. 16, 1998, and U.S. Provisional Application Serial No. 60/163,373 filed Nov. 3, 1999.

[0002] This invention relates to certain crystallized kinase protein-ligand complexes, particularly complexes of crystallized P38 protein, and more particularly complexes of P38&ggr; protein. This invention also relates to crystallizable compositions from which the protein-ligand complexes may be obtained. This invention also relates to computational methods of using structure coordinates of the protein complex to screen for and design compounds that interact with the protein, particularly P38 protein or homologues thereof.

BACKGROUND OF THE INVENTION

[0003] Mammalian cells respond to extracellular stimuli by activating signaling cascades that are mediated by members of the mitogen-activated protein (MAP) kinase family. Mammalian mitogen-activated protein (MAP) kinases are proline-directed serine/threonine kinases that facilitate signal translocation in cells [Davis, Mol. Reprod. Dev. 42, 459-467 (1995); Cobb et al., J. Biol. Chem. 270, 14843-14846 (1995); Marshall, Cell 80, 179-185 (1995)]. MAP kinases include the extracellular-signal regulated kinases (ERKs), the c-Jun NH2-terminal kinases (JNKs) and the P38 kinases, which have similar sequences and three-dimensional structures [Taylor & lipopolysaccharides (LPS), UV, anisomycin, or osmotic shock, and by cytokines, such as interleukin-1 (IL-1) and tissue necrosis factor (TNF). Inhibition of P38&agr; kinase leads to a blockade on the production of both IL-1 and TNF. IL-1 and TNF stimulate the production of other proinflammatory cytokines such as IL-6 and IL-8 and have been implicated in acute and chronic inflammatory diseases and in post-menopausal osteoporosis [Kimble et al., Endocrinol., 136, 3054-61 (1995)].

[0004] Based upon this finding it is believed that P38&agr;, along with other MAPKs, has a role in mediating cellular response to inflammatory stimuli, such as leukocyte accumulation, macrophage/monocyte activation, tissue resorption, fever, acute phase responses and neutrophilia. In addition, the MAPKs, such as P38&agr;, have been implicated in cancer, thrombin-induced platelet aggregation, immunodeficiency disorders, autoimmune diseases, cell death, allergies, osteoporosis and neurodegenerative disorders. Inhibitors of P38&agr; also appear to be involved in pain management through inhibition of prostaglandin endoperoxide synthase-2 induction. Other diseases associated with Il -1, IL-6, IL-8 or TNF overproduction are set forth in WO 96/21654. P38&ggr; MAP kinase (also known as ERK6 and stress activated protein kinase-3 or SAPK3) is a newly discovered member of the MAP kinase family. However, unlike the other P38 family members which are expressed in many tissues, P38&ggr; is expressed at highest levels in skeletal muscle [Li et al., Biochem Biophys Res Commun 228, 334-340 (1996); Enslen et al., J Biol Chem 273, 1741-1748 (1998); Raingeaud et al., J. Biol. Chem. 270, 7420-7426 (1995)]. Thus P38&ggr; may have a unique function related to muscle morphogenesis, and it may be a potential target for treating degenerative diseases occurring in muscle tissue.

[0005] Compounds that selectively inhibit P38&ggr; and not P38&agr; would be highly desirable. It would be useful to have new treatments for muscle degenerative diseases using compounds that do not suppress the inflammatory response or other functions of P38&agr;. However, the design of inhibitors that are selective for any particular MAP kinase, such as P38&ggr;, is challenging due to the structural similarity of the MAP kinases. Therefore, it would be advantageous to have a detailed understanding of the structures of the various MAP kinases in order to exploit any subtle differences that may exist among them.

[0006] A general approach to designing inhibitors that are selective for an enzyme target is to determine how a putative inhibitor interacts with the three dimensional structure of the enzyme. For this reason it is useful to obtain the enzyme protein in crystal form and perform X-ray diffraction techniques to determine its three dimensional structure coordinates. If the enzyme is crystallized as a complex with a ligand, one can determine both the shape of the enzyme binding pocket when bound to the ligand, as well as the amino acid residues that are capable of close contact with the ligand. By knowing the shape and amino acid residues in the binding pocket, one may design new ligands that will interact favorably with the enzyme. With such structural information, available computational methods may be used to predict how strong the ligand binding interaction will be. Such methods thus enable the design of inhibitors that bind strongly, as well as selectively to the target enzyme.

[0007] Crystal structures are known for some of the MAP kinases; for example, unphosphorylated JNK3, unphosphorylated P38&agr;, and ERK2 in both phosphorylated and unphosphorylated forms. Phosphorylated ERK2 is reported to exist as a dimer in both solution and as a crystal. The unphosphorylated forms of JNK3, ERK2 and P38&agr;, on the other hand, are reported to be monomeric. [Tong et al., Nat Struct Biol 4, 311-316 (1997); Wilson and Su, Chem Biol 4, 423-431 (1997); Xie et al., Structure 6, 983-991 (1998); Zhang et al., Nature 367, 704-711 (1994); Canagarajah et al., Cell 90, 859-869 (1997); Wilson and Su, J Biol Chem 271, 27696-27700 (1996)]

[0008] The crystal structure reported for P38&agr; is based on unphosphorylated protein. However, it is the phosphorylated or activated form of the enzyme that is able to phosphorylate its substrate enzyme. In order to disrupt the phosphorylation of the substrate, and produce the desired clinical effect, a small molecule inhibitor would likely act by blocking a phosphorylated form of P38. Thus, the most suitable target for drug design is the active or phosphorylated form. While the structure of the unphosphorylated enzyme is often used for drug design purposes, there is an inherent uncertainty as to whether the phosphorylated and unphosphorylated forms would bind a designed inhibitor with equal affinity.

[0009] A class of pyridinylimidazole compounds are known to inhibit P38&agr; MAP kinase [Lee et al., Nature 372, 739-746 (1994)]. These inhibitors have been shown to bind in the ATP binding site of P38&agr; [Young et al., J Biol Chem 272, 12116-12121 (1997); Tong et al., Nat Struct Biol 4, 311-316 (1997); Wilson et al., Chem Biol 4, 423-431 (1997)]. However, the pyridinylimidazoles reportedly do not inhibit the activity of ERK2, JNK3, or P38&ggr;. This observed selectivity is interesting because the amino acid sequence in the ATP binding site of the various kinases are known to be highly conserved [Fox et al., Protein Science 7, 2249-2255 (1998); Xie et al., supra; Wilson and Su, supra; Enslen et al., J Biol Chem 273, 1741-1748 (1998)].

[0010] As there is a need for compounds that selectively inhibit a particular MAP kinase, it would be desirable to have improved methods that facilitate the design of such compounds. For this purpose, knowledge of the three dimensional structure coordinates of an activated P38 protein would be useful. Such information would aid in identifying and designing potential inhibitors of particular P38 proteins which, in turn, are expected to have therapeutic utility.

SUMMARY OF THE INVENTION

[0011] This invention provides certain crystallized, protein kinase-ligand complexes, in particular P38-ligand complexes, and their structure coordinates. The structure coordinates are based on the structure of a phosphorylated P38&ggr;-ligand complex that has now been solved and which reveals new structural information useful for understanding the activated states of other, related kinase proteins as described herein. The key structural features of the proteins, particularly the shape of the substrate binding site, are useful in methods for designing or identifying selective inhibitors of the protein kinases, particularly P38, and in solving the structures of other proteins with similar features.

[0012] The invention also provides a computer which which is programmed with the structure coordinates of the activated P38 binding site. Such a computer, appropriately programmed and attached to the necessary viewing device, is capable of displaying a three-dimensional graphical representation of a molecule or molecular complex comprising such binding sites or similarly shaped homologous binding pockets.

[0013] The invention also provides a method for determining at least a portion of the three-dimensional structure of other molecules or molecular complexes which contain at least some features that are structurally similar to P38&ggr;, particularly P38&agr;, P38&bgr;, P38&dgr; and other P38 isoforms. This is achieved by using at least some of the structural coordinates obtained for a phosphorylated P38 complex.

BRIEF DESCRIPTION OF THE FIGURES

[0014] FIG. 1 lists the atomic structure coordinates for phosphorylated P38&ggr; in complex with MgAMP-PNP as derived by X-ray diffraction from a crystal of that complex. The following abbreviations are used in FIG. 1:

[0015] “Atom type” refers to the element whose coordinates are measured. The first letter in the column defines the element.

[0016] “X, Y, Z” crystallographically define the atomic position of the element measured.

[0017] “B” is a thermal factor that measures movement of the atom around its atomic center.

[0018] “Occ” is an occupancy factor that refers to the fraction of the molecules in which each atom occupies the position specified by the coordinates. A value of “1” indicates that each atom has the same conformation, i.e., the same position, in all molecules of the crystal.

[0019] FIG. 1a is an overview of the phosphorylated P38&ggr;.

[0020] FIG. 2 is a superimposition of unphosphorylated P38&ggr; and phosphorylated P38&ggr;.

[0021] FIG. 3 is a detailed stereo view of the activation loop.

[0022] FIG. 4 is a stereo view of the AMP-PNP bound in the active site.

[0023] FIG. 5 is a comparison of the active sites of activated P38&ggr; with P38&agr; (a) and cAPK or cyclic AMP dependent protein kinase (b).

[0024] FIG. 6 is a comparison of activated phosphorylation loops from P38&ggr; (dark orange), ERK2 (dark blue), and cAPK (red).

[0025] FIG. 7 shows a diagram of a system used to carry out the instructions encoded by the storage medium of FIGS. 8 and 9.

[0026] FIG. 8 shows a cross section of a magnetic storage medium.

[0027] FIG. 9 shows a cross section of a optically-readable data storage medium.

DETAILED DESCRIPTION OF THE INVENTION

[0028] This invention provides certain crystallized, protein kinase-ligand complexes, in particular P38-ligand complexes, and their structure coordinates. The structure coordinates are based on the structure of a phosphorylated P38&ggr; complex that has now been solved and which reveals new structural information regarding the activated states of other, related kinase proteins as described herein. The key structural features of the protein, particularly the shape of the substrate binding site, are useful in methods for designing inhibitors of the P38 and in solving the structures of other proteins with similar features.

[0029] In describing protein structure and function, reference is made to amino acids comprising the protein. The amino acids may also be referred to by their conventional abbreviations, as shown in the table below. 1 A = Ala = Alanine T = Thr = Threonine V = Val = Valine C = Cys = Cysteine L = Leu = Leucine Y = Tyr = Tyrosine I = Ile = Isoleucine N = Asn = Asparagine P = Pro = Proline Q = Gln = Glutamine F = Phe = Phenylalanine D = Asp = Aspartic Acid W = Trp = Tryptophan E = Glu = Glutamic Acid M = Met = Methionine K = Lys = Lysine G = Gly = Glycine R = Arg = Arginine S = Ser = Serine H = His = Histidine

[0030] This invention also provides a crystallizable composition from which the crystallized protein is obtained. The crystallizable composition preferably comprises a phosphorylated P38 protein complexed with a substrate or ligand. The ligand may be any ligand capable of binding to the P38 protein, and is preferably a ligand that binds to the ATP binding site of the protein. Examples of such ligands are small molecule inhibitors of the particular P38 as well as non-hydrolyzable ATP analogs and suicide substrates. Non-hydrolyzable ATP analogs useful in the crystallizable compositions of this invention include AMP-PCH2P, AMP-PSP and AMP-PNP where the oxygen linking the second and third phosphates of the ATP analogs is replaced by CH2, S and NH, respectively. An example of a suicidal substrate is 5′-(p-fluorosulfonyl benzoyl)adenosine (FSBA). Preferably, the crystallizable compositions of this invention comprise AMP-PNP as the substrate. It is preferred that the composition further comprise divalent cations, especially magnesium or magnanese cations, which may be introduced in any suitable manner. For example, the cations may be introduced by incubating the desired ligand with a suitable metal salt such as MgCl2 prior to incubation with the phosphorylated P38 protein.

[0031] It has been found that the crystallization of the phosphorylated P38 protein is sensitive to buffer conditions. Thus, in a preferred embodiment, the crystallizable compositions of this invention further comprise a suitable glycol such as ethylene glycol, polyethylene glycol (PEG), PEG-monomethyl ether or mixtures thereof, preferably PEG 4000, as an aqueous solution containing between about 10 to 35% of the glycol by volume of solution, a salt, such as sodium acetate at about 50 to 200 mM, a reducing agent, such as dithiothreitol (DTT) at between about 1 to 10 mM, a detergent such as C12E9 at about 0.01 to 0.05%, and a buffer that maintains pH at between about 8.0 and 9.0. An example of a suitable buffer is 100 mM Tris at pH 8.5.

[0032] By applying standard crystallization protocols to the above described crystallizable compositions, crystals of the phosphorylated P38 protein complex may be obtained. Thus, one aspect of this invention relates to a method of preparing phosphorylated P38-containing crystals. The method comprises the steps of

[0033] (a) obtaining a crystallizable composition comprising a phosphorylated P38 protein, divalent cations, and a ligand capable of binding to the protein, and

[0034] (b) subjecting the composition of step (a) to conditions which promote crystallization.

[0035] FIG. 1 shows the structure coordinates of a phosphorylated P38&ggr; protein complexed with MgAMP-PNP. The manner of obtaining these structure coordinates, interpretation of the coordinates and their utility in understanding the protein structure, as described herein, will be understood by those of skill in the art and by reference to standard texts such as Crystal Structure Analysis, Jenny Pickworth Glusker and Kenneth N. Trueblood, 2nd Ed. Oxford University Press, 1985, New York; and Principles of Protein Structure, G. E. Schulz and R. H. Schirmer, Springer-Verlag, 1985, New York.

[0036] Those of skill in the art understand that a set of structure coordinates for an enzyme or an enzyme-complex or a portion thereof, is a relative set of points that define a shape in three dimensions. Thus, it is possible that an entirely different set of coordinates could define a similar or identical shape. Moreover, slight variations in the individual coordinates will have little effect on overall shape. In terms of binding pockets, these variations would not be expected to significantly alter the nature of ligands that could associate with those pockets.

[0037] These variations in coordinates may be generated because of mathematical manipulations of the P38&ggr;/MgAMP-PNP structure coordinates. For example, the structure coordinates set forth in FIG. 1 could be manipulated by crystallographic permutations of the structure coordinates, fractionalization of the structure coordinates, integer additions or subtractions to sets of the structure coordinates, inversion of the structure coordinates or any combination of the above.

[0038] Alternatively, modifications in the crystal structure due to mutations, additions, substitutions, and/or deletions of amino acids, or other changes in any of the components that make up the crystal could also account for variations in structure coordinates. If such variations are within an acceptable standard error as compared to the original coordinates, the resulting three-dimensional shape is considered to be the same. Thus, for example, a ligand that bound to the active site binding pocket of P38&ggr; would also be expected to bind to another binding pocket whose structure coordinates defined a shape that fell within the acceptable error.

[0039] The term “binding pocket” refers to a region of the protein that, as a result of its shape, favorably associates with a ligand or substrate. The term “P38&ggr;-like binding pocket” refers to a portion of a molecule or molecular complex whose shape is sufficiently similar to the P38&ggr; binding pockets as to bind common ligands. This commonality of shape may be quantitatively defined by a root mean square deviation (rmsd) from the structure coordinates of the backbone atoms of the amino acids that make up the binding pockets in P38&ggr; (as set forth in FIG. 1). The method of performing this rmsd calculation is described below.

[0040] The “active site binding pockets” or “active site” of P38&ggr; refers to the area on the P38&ggr; enzyme surface where the substrate binds. In resolving the crystal structure of phosphorylated P38&ggr; in complex with MgAMP-PNP, applicants have determined that P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 are within 5 Å of and therefore close enough to interact with MgAMP-PNP. These amino acids are hereinafter referred to as the “SET 5A amino acids.” Thus, a binding pocket defined by the structural coordinates of those amino acids, as set forth in FIG. 1; or a binding pocket whose root mean square deviation from the structure coordinates of the backbone atoms of those amino acids of not more than about 1.15 angstroms (Å) is considered a P38&ggr;-like binding pocket of this invention.

[0041] Applicants have also determined that in addition to the P38&ggr; amino acids set forth above, Pro32, Cys42, Ser43, Val53, Ile55, Lys57, Leu58, Thr59, Arg70, Glu74, Gly88, Leu107, Val108, Leu116, Gly117, Pro156, Leu159, Val161, Lys168, Phe172, Ala175, and Thr188 are within 8 Å of bound MgAMP-PNP and therefore are also close enough to interact with that substrate. These amino acids, in addition to the SET 5A amino acids, are hereinafter referred to as the “SET 8A amino acids.” Thus, in a preferred embodiment, a binding pocket defined by the structural coordinates of the amino acids within 8 Å of bound MgAMP-PNP, as set forth in FIG. 1; or a binding pocket whose root mean square deviation from the structure coordinates of the backbone atoms of those amino acids of not more than about 1.15 Å is considered a preferred P38&ggr;-like binding pocket of this invention.

[0042] It will be readily apparent to those of skill in the art that the numbering of amino acids in other isoforms of P38 may be different than that set forth for P38&ggr;. Corresponding amino acids in other isoforms of P38 are easily identified by visual inspection of the amino acid sequences or by using commercially available homology software programs, as further described below.

[0043] Various computational analyses may be used to determine whether a protein or the binding pocket portion thereof is sufficiently similar to the P38&ggr; binding pockets described above. Such analyses may be carried out in well known software applications, such as the Molecular Similarity application of QUANTA (Molecular Simulations Inc., San Diego, Calif.) version 4.1, and as described in the accompanying User's Guide.

[0044] For the purpose of this invention, a rigid fitting method was conveniently used to compare protein structures. Any molecule or molecular complex or binding pocket thereof having a root mean square deviation of conserved residue backbone atoms (N, C&agr;, C, O) of less than about 1.15 Å when superimposed on the relevant backbone atoms described by structure coordinates listed in FIG. 1 are considered identical. More preferably, the root mean square deviation is less than about 1.0 Å.

[0045] The P38 X-ray coordinate data, when used in conjunction with a computer programmed with software to translate those coordinates into the 3-dimensional structure of p38&ggr; may be used for a variety of purposes, especially for purposes relating to drug discovery. Such software for generating three-dimensional graphical representations are known and commercially available. The ready use of the coordinate data requires that it be stored in a computer-readable format. Thus, in accordance with the present invention, data capable of being displayed as the three dimensional structure of P38&ggr; and portions thereof and their structurally similar homologues is stored in a machine-readable storage medium, which is capable of displaying a graphical three-dimensional representation of the structure.

[0046] Therefore, another embodiment of this invention provides a machine-readable data storage medium, comprising a data storage material encoded with machine readable data which, when used by a machine programmed with instructions for using said data, displays a graphical three-dimensional representation of a molecule or molecular complex comprising a binding pocket defined by structure coordinates of the P38&ggr; SET 5A amino acids, or preferably the P38&ggr; SET 8A amino acids, or a homologue of said molecule or molecular complex, wherein said homologue comprises a binding pocket that has a root mean square deviation from the backbone atoms of said amino acids of not more than about 1.15 Å.

[0047] Even more preferred is a machine-readable data storage medium that is capable of displaying a graphical three-dimensional representation of a molecule or molecular complex that is defined by the structure coordinates of all of the amino acids in FIG. 1 or a homologue of said molecule or molecular complex, wherein said homologue has a root mean square deviation from the backbone atoms of all of the amino acids in FIG. 1 of not more than about 1.15 Å.

[0048] According to an alternate embodiment, the machine-readable data storage medium comprises a data storage material encoded with a first set of machine readable data which comprises the Fourier transform of the structure coordinates set forth in FIG. 1, and which, when using a machine programmed with instructions for using said data, can be combined with a second set of machine readable data comprising the X-ray diffraction pattern of another molecule or molecular complex to determine at least a portion of the structure coordinates corresponding to the second set of machine readable data.

[0049] For example, the Fourier transform of the structure coordinates set forth in FIG. 1 may be used to determine at least a portion of the structure coordinates of other P38s, such as P38&bgr;, and P38&dgr; and isoforms of P38&bgr;, P38&dgr; or P38&ggr;. The structure coordinates in FIG. 1 and the Fourier transform of the coordinates are especially useful for determining the coordinates of other P38s in phosphorylated form.

[0050] According to an alternate embodiment, this invention provides a computer for producing a three-dimensional representation of a molecule or molecular complex, wherein said molecule or molecular complex comprises a binding pocket defined by the P38&ggr; SET 5A amino acids, or preferably the P38&ggr; SET 8A amino acids, or a homologue of said molecule or molecular complex, wherein said homologue comprises a binding pocket that has a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, wherein said computer comprises:

[0051] (a) a machine readable data storage medium comprising a data storage material encoded with machine-readable data, wherein said machine readable data comprises the structure coordinates of P38&ggr; or portions thereof;

[0052] (b) a working memory for storing instructions for processing said machine-readable data;

[0053] (c) a central-processing unit coupled to said working memory and to said machine-readable data storage medium, for processing said machine-readable data into said three-dimensional representation; and

[0054] (d) an output hardware coupled to said central processing unit, for receiving said three Dimensional representation.

[0055] FIG. 7 demonstrates one version of these embodiments. System 10 includes a computer 11 comprising a central processing unit (“CPU”) 20, a working memory 22 which may be, eg., RAM (random-access memory) or “core” memory, mass storage memory 24 (such as one or more disk drives or CD-ROM drives), one or more cathode-ray tube (“CRT”) display terminals 26, one or more keyboards 28, one or more input lines 30, and one or more output lines 40, all of which are interconnected by a conventional bi-directional system bus 50.

[0056] Input hardware 36, coupled to computer 11 by input lines 30, may be implemented in a variety of ways. Machine-readable data of this invention may be inputted via the use of a modem or modems 32 connected by a telephone line or dedicated data line 34. Alternatively or additionally, the-input hardware 36 may comprise CD-ROM drives or disk drives 24. In conjunction with display terminal 26, keyboard 28 may also be used as an input device.

[0057] Output hardware 46, coupled to computer 11 by output lines 40, may similarly be implemented by conventional devices. By way of example, output hardware 46 may include CRT display terminal 26 for displaying a graphical representation of a binding pocket of this invention using a program such as QUANTA as described herein. Output hardware might also include a printer 42, so that hard copy output may be produced, or a disk drive 24, to store system output for later use.

[0058] In operation, CPU 20 coordinates the use of the various input and output devices 36, 46 coordinates data accesses from mass storage 24 and accesses to and from working memory 22, and determines the sequence of data processing steps. A number of programs may be used to process the machine-readable data of this invention. Such programs are discussed in reference to the computational methods of drug discovery as described herein. Specific references to components of the hardware system 10 are included as appropriate throughout the following description of the data storage medium.

[0059] FIG. 8 shows a cross section of a magnetic data storage medium 100 which can be encoded with a machine-readable data that can be carried out by a system such as system 10 of FIG. 7. Medium 100 can be a conventional floppy diskette or hard disk, having a suitable substrate 101, which may be conventional, and a suitable coating 102, which may be conventional, on one or both sides, containing magnetic domains (not visible) whose polarity or orientation can be altered magnetically. Medium 100 may also have an opening (not shown) for receiving the spindle of a disk drive or other data storage device 24. The magnetic domains of coating 102 of medium 100 are polarized or oriented so as to encode in manner which may be conventional, machine readable data such as that described herein, for execution by a system such as system 10 of FIG. 7.

[0060] FIG. 9 shows a cross section of an optically-readable data storage medium 110 which also can be encoded with such a machine-readable data, or set of instructions, which can be carried out by a system such as system 10 of FIG. 7. Medium 110 can be a conventional compact disk read only memory (CD-ROM) or a rewritable medium such as a magneto-optical disk which is optically readable and magneto-optically writable. Medium 100 preferably has a suitable substrate 111, which may be conventional, and a suitable coating 112, which may be conventional, usually of one side of substrate 111.

[0061] In the case of CD-ROM, as is well known, coating 112 is reflective and is impressed with a plurality of pits 113 to encode the machine-readable data. The arrangement of pits is read by reflecting laser light off the surface of coating 112. A protective coating 114, which preferably is substantially transparent, is provided on top of coating 112.

[0062] In the case of a magneto-optical disk, as is well known, coating 112 has no pits 113, but has a plurality of magnetic domains whose polarity or orientation can be changed magnetically when heated above a certain temperature, as by a laser (not shown). The orientation of the domains can be read by measuring the polarization of laser light reflected from coating 112. The arrangement of the domains encodes the data as described above.

[0063] As mentioned above, the P38&ggr; X-ray coordinate data is useful for screening and identifying drugs that inhibit P38, especially phosphorylated P38. For example, the structure encoded by the data may be computationally evaluated for its ability to associate with putative substrates or ligands. Such compounds that associate with p38&ggr; may inhibit p38&ggr;, and are potential drug candidates. Additionally or alternatively, the structure encoded by the data may be displayed in a graphical three-dimensional representation on a computer screen. This allows visual inspection of the structure, as well as visual inspection of the structure's association with the compounds.

[0064] Thus, according to another embodiment, this invention relates to a method for evaluating the potential of a compound to associate with a molecule or molecular complex comprising a binding pocket defined by the structure coordinates of the P38&ggr; SET 5A amino acids, or preferably the P38&ggr; SET 8A amino acids, or a homologues of said molecule or molecular complex, wherein said homologue comprises a binding pocket that has a root mean square deviation from the backbone atoms of said amino acids of not more than about 1.15 Å.

[0065] This method comprises the steps of:

[0066] a) creating a computer model of the binding pocket using structure coordinates wherein the root mean square deviation between said structure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1 is not more than about 1.15 Å;

[0067] b) employing computational means to perform a fitting operation between the chemical entity and said computer model of the binding pocket; and

[0068] c) analyzing the results of said fitting operation to quantify the association between the chemical entity and the binding pocket model.

[0069] The term “chemical entity”, as used herein, refers to chemical compounds or ligands, complexes of at least two chemical compounds, and fragments of such compounds or complexes.

[0070] Even more preferably, the method evaluates the potential of a chemical entity to associate with a molecule or molecular complex defined by the structure coordinates of all of the P38&ggr; amino acids, as set forth in FIG. 1, or a homologue of said molecule or molecular complex having a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å.

[0071] Alternatively, the structural coordinates of the P38&ggr; binding pocket can be utilized in a method for identifying a potential agonist or antagonist of a molecule comprising a P38&ggr;-like binding pocket. This method comprises the steps of:

[0072] (a) using atomic coordinates of the P38&ggr; SET 5A amino acids±a root mean square deviation from the backbone atoms of said amino acids of not more than about 1.15 Å, to generate a three-dimensional structure of molecule comprising a P38&ggr;-like binding pocket;

[0073] (b) employing said three-dimensional structure to design or select said potential agonist or antagonist;

[0074] (c) synthesizing said agonist or antagonist; and

[0075] (d) contacting said agonist or antagonist with said molecule to determine the ability of said potential agonist or antagonist to interact with said molecule.

[0076] More preferred is the use of the atomic coordinates of the P38&ggr; SET 8A amino acids, ±a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, to generate a three-dimensional structure of molecule comprising a p38&ggr;-like binding pocket. Most preferred is when the atomic coordinates of all the amino acids of P38&ggr; according to FIG. 1±a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, are used to generate a three-dimensional structure of molecule comprising a P38&ggr;-like binding pocket.

[0077] For the first time, the present invention permits the use of molecular design techniques to identify, select or design potential inhibitors of p38, based on the structure of a phosphorylated p38&ggr;-like binding pocket. Such a predictive model is valuable in light of the high costs associated with the preparation and testing of the many diverse compounds that may possibly bind to the p38 protein.

[0078] According to this invention, a potential p38 inhibitor may now be evaluated for its ability to bind a P38&ggr;-like binding pocket prior to its actual synthesis and testing. If a proposed compound is predicted to have insufficient interaction or association with the binding pocket, preparation and testing of the compound is obviated. However, if the computer modeling indicates a strong interaction, the compound may then be obtained and tested for its ability to bind. Testing to confirm binding may be performed using assays such as described in Example 6.

[0079] A potential inhibitor of a P38&ggr;-like binding pocket may be computationally evaluated by means of a series of steps in which chemical entities or fragments are screened and selected for their ability to associate with the P38&ggr;-like binding pockets.

[0080] One skilled in the art may use one of several methods to screen chemical entities or fragments for their ability to associate with a P38&ggr;-like binding pocket. This process may begin by visual inspection of, for example, a P38&ggr;-like binding pocket on the computer screen based on the P38&ggr; structure coordinates in FIG. 1 or other coordinates which define a similar shape generated from the machine-readable storage medium. Selected fragments or chemical entities may then be positioned in a variety of orientations, or docked, within that binding pocket as defined above. Docking may be accomplished using software such as Quanta and Sybyl, followed by energy minimization and molecular dynamics with standard molecular mechanics force fields, such as CHARMM and AMBER.

[0081] Specialized computer programs may also assist in the process of selecting fragments or chemical entities. These include:

[0082] 1. GRID (P. J. Goodford, “A Computational Procedure for Determining Energetically Favorable Binding Sites on Biologically Important Macromolecules”, J. Med. Chem., 28, pp. 849-857 (1985)). GRID is available from Oxford University, Oxford, UK.

[0083] 2. MCSS (A. Miranker et al., “Functionality Maps of Binding Sites: A Multiple Copy Simultaneous Search Method.” Proteins,: Structure, Function and Genetics, 11, pp. 29-34 (1991)). MCSS is available from Molecular Simulations, San Diego, Calif.

[0084] 3. AUTODOCK (D. S. Goodsell et al., “Automated Docking of Substrates to Proteins by Simulated Annealing”, Proteins: Structure, Function, and Genetics, 8, pp. 195-202 (1990)). AUTODOCK is available from Scripps Research Institute, La Jolla, Calif.

[0085] 4. DOCK (I. D. Kuntz et al., “A Geometric Approach to Macromolecule-Ligand Interactions”, J. Mol. Biol., 161, pp. 269-288 (1982)). DOCK is available from University of California, San Francisco, Calif.

[0086] Once suitable chemical entities or fragments have been selected, they can be designed or assembled into a single compound or complex. Assembly may be preceded by visual inspection of the relationship of the fragments to each other on the three-dimensional image displayed on a computer screen in relation to the structure coordinates of P38&ggr;. This would be followed by manual model building using software such as Quanta or Sybyl [Tripos Associates, St. Louis, Mo.].

[0087] Useful programs to aid one of skill in the art in connecting the individual chemical entities or fragments include:

[0088] 1. CAVEAT (P. A. Bartlett et al., “CAVEAT: A Program to Facilitate the Structure-Derived Design of Biologically Active Molecules”, in Molecular Recognition in Chemical and Biological Problems”, Special Pub., Royal Chem. Soc., 78, pp. 182-196 (1989); G. Lauri and P. A. Bartlett, “CAVEAT: a Program to Facilitate the Design of Organic Molecules”, J. Comput. Aided Mol. Des. , 8, pp. 51-66 (1994)). CAVEAT is available from the University of California, Berkeley, Calif.

[0089] 2. 3D Database systems such as ISIS (MDL Information Systems, San Leandro, Calif.). This area is reviewed in Y. C. Martin, “3D Database Searching in Drug Design”, J. Med. Chem., 35, pp. 2145-2154 (1992).

[0090] 3. HOOK (M. B. Eisen et al, “HOOK: A Program for Finding Novel Molecular Architectures that Satisfy the Chemical and Steric Requirements of a Macromolecule Binding Site”, Proteins: Struct., Funct., Genet., 19, pp. 199-221 (1994). HOOK is available from Molecular Simulations, San Diego, Calif.

[0091] Instead of proceeding to build an inhibitor of a P38&ggr;-like binding pocket in a step-wise fashion one fragment or chemical entity at a time as described above, inhibitory or other P38&ggr; binding compounds may be designed as a whole or “de novo” using either an empty binding site or optionally including some portion(s) of a known inhibitor(s). There are many de novo ligand design methods including:

[0092] 1. LUDI (H.-J. Bohm, “The Computer Program LUDI: A New Method for the De Novo Design of Enzyme Inhibitors”, J. Comp. Aid. Molec. Design, 6, pp. 61-78 (1992)). LUDI is available from Molecular Simulations Incorporated, San Diego, Calif.

[0093] 2. LEGEND (Y. Nishibata et al., Tetrahedron, 47, p. 8985 (1991)). LEGEND is available from Molecular Simulations Incorporated, San Diego, Calif.

[0094] 3. LeapFrog (available from Tripos Associates, St. Louis, Mo.).

[0095] 4. SPROUT (V. Gillet et al, “SPROUT: A Program for Structure Generation)”, J. Comput. Aided Mol. Design, 7, pp. 127-153 (1993)). SPROUT is available from the University of Leeds, UK.

[0096] Other molecular modeling techniques may also be employed in accordance with this invention [see, e.g., Cohen et al., “Molecular Modeling Software and Methods for Medicinal Chemistry, J. Med. Chem., 33, pp. 883-894 (1990); see also, M. A. Navia and M. A. Murcko, “The Use of Structural Information in Drug Design”, Current Opinions in Structural Biology, 2, pp. 202-210 (1992); L. M. Balbes et al., “A Perspective of Modern Methods in Computer-Aided Drug Design”, in Reviews in Computational Chemistry, Vol. 5, K. B. Lipkowitz and D. B. Boyd, Eds., VCH, New York, pp. 337-380 (1994); see also, W. C. Guida, “Software For Structure-Based Drug Design”, Curr. Opin. Struct. Biology, 4, pp. 777-781 (1994)].

[0097] Once a compound has been designed or selected by the above methods, the efficiency with which-that entity may bind to a P38&ggr; binding pocket may be tested and optimized by computational evaluation. For example, an effective P38&ggr; binding pocket inhibitor must preferably demonstrate a relatively small difference in energy between its bound and free states (i.e., a small deformation energy of binding). Thus, the most efficient P38&ggr; binding pocket inhibitors should preferably be designed with a deformation energy of binding of not greater than about 10 kcal/mole, more preferably, not greater than 7 kcal/mole. P38&ggr; binding pocket inhibitors may interact with the binding pocket in more than one of multiple conformations that are similar in overall binding energy. In those cases, the deformation energy of binding is taken to be the difference between the energy of the free entity and the average energy of the conformations observed when the inhibitor binds to the protein.

[0098] An entity designed or selected as binding to a P38&ggr; binding pocket may be further computationally optimized so that in its bound state it would preferably lack repulsive electrostatic interaction with the target enzyme and with the surrounding water molecules. Such non-complementary electrostatic interactions include repulsive charge-charge, dipole-dipole and charge-dipole interactions.

[0099] Specific computer software is available in the art to evaluate compound deformation energy and electrostatic interactions. Examples of programs designed for such uses include: Gaussian 94, revision C (M. J. Frisch, Gaussian, Inc., Pittsburgh, Pa. ©1995); AMBER, version 4.1 (P. A. Kollman, University of California at San Francisco, ©1995); QUANTA/CHARMM (Molecular Simulations, Inc., San Diego, Calif. ©1995); Insight II/Discover (Molecular Simulations, Inc., San Diego, Calif. ©1995); DelPhi (Molecular Simulations, Inc., San Diego, Calif. (©1995); and AMSOL (Quantum Chemistry Program Exchange, Indiana University). These programs may be implemented, for instance, using a Silicon Graphics workstation such as an Indigo2 with “IMPACT” graphics. Other hardware systems and software packages will be known to those skilled in the art.

[0100] Another approach enabled by this invention, is the computational screening of small molecule databases for chemical entities or compounds that can bind in whole, or in part, to a P38&ggr; binding pocket. In this screening, the quality of fit of such entities to the binding site may be judged either by shape complementarity or by estimated interaction energy [E. C. Meng et al., J. Comp. Chem., 13, 505-524 (1992)].

[0101] According to another embodiment, the invention provides compounds which associate with a P38&ggr;-like binding pocket produced or identified by the method set forth above.

[0102] The structure coordinates set forth in FIG. 1 can also be used to aid in obtaining structural information about another crystallized molecule or molecular complex. This may be achieved by any of a number of well-known techniques, including molecular replacement.

[0103] Therefore, in another embodiment this invention provides a method of utilizing molecular replacement to obtain structural information about a molecule or molecular complex whose structure is unknown comprising the steps of:

[0104] a) crystallizing said molecule or molecular complex of unknown structure;

[0105] b) generating an X-ray diffraction pattern from said crystallized molecule or molecular complex; and

[0106] c) applying at least a portion of the structure coordinates set forth in FIG. 1 to the X-ray diffraction pattern to generate a three-dimensional electron density map of the molecule or molecular complex whose structure is unknown.

[0107] By using molecular replacement, all or part of the structure coordinates of the P38&ggr;/MgAMP-PNP complex as provided by this invention (and set forth in FIG. 1) can be used to determine the structure of another crystallized molecule or molecular complex more quickly and efficiently than attempting an ab initio structure determination.

[0108] Molecular replacement provides an accurate estimation of the phases for an unknown structures Phases are a factor in equations used to solve crystal structures that can not be determined directly. Obtaining accurate values for the phases, by methods other than molecular replacement, is a time-consuming process that involves iterative cycles of approximations and refinements and greatly hinders the solution of crystal structures. However, when the crystal structure of a protein containing at least a homologous portion has been solved, the phases from the known structure provide a satisfactory estimate of the phases for the unknown structure.

[0109] Thus, this method involves generating a preliminary model of a molecule or molecular complex whose structure coordinates are unknown. bv orienting and positioning the relevant portion of the P38&ggr;/MgAMP-PNP complex according to FIG. 1 within the unit cell of the crystal of the unknown molecule or molecular complex so as best to account for the observed X-ray diffraction pattern of the crystal of the molecule or molecular complex whose structure is unknown. Phases can then be calculated from this model and combined with the observed X-ray diffraction pattern amplitudes to generate an electron density map of the structure whose coordinates are unknown. This, in turn, can be subjected to any well-known model building and structure refinement techniques to provide a final, accurate structure of the unknown crystallized molecule or molecular complex [E. Lattman, “Use of the Rotation and Translation Functions”, in Meth. Enzymol., 115, pp. 55-77 (1985); M. G. Rossmann, ed., “The Molecular Replacement Method”, Int. Sci. Rev. Ser., No. 13, Gordon & Breach, New York (1972)].

[0110] The structure of any portion of any crystallized molecule or molecular complex that is sufficiently homologous to any portion of the P38&ggr;/MgAMP-PNP complex can be resolved by this method.

[0111] In a preferred embodiment, the method of molecular replacement is utilized to obtain structural information about another P38, such as P38&agr;, P38&bgr;, P38&dgr;, or isoforms of P38&bgr;, P38&dgr; or P38&ggr;. The structure coordinates of P38&ggr; as provided by this invention are particularly useful in solving the structure of other isoforms of P38&ggr; or P38&ggr; complexes.

[0112] Furthermore, the structure coordinates of P38&ggr; as provided by this invention are useful in solving the structure of P38&ggr; proteins that have amino acid substitutions, additions and/or deletions (referred to collectively as “P38&ggr; mutants”, as compared to naturally occurring P38&ggr; isoforms). These P38&ggr; mutants may optionally be crystallized in co-complex with a chemical entity, such as a non-hydrolyzable ATP analogue or a suicide substrate. The crystal structures of a series of such complexes may then be solved by molecular replacement and compared with that of wild-type p38&ggr;. Potential sites for modification within the various binding sites of the enzyme may thus be identified. This information provides an additional tool for determining the most efficient binding interactions such as, for example, increased hydrophobic interactions, between P38&ggr; and a chemical entity or compound.

[0113] All of the complexes referred to above may be studied using well-known X-ray diffraction techniques and may be refined versus 1.5-3A resolution X-ray data to an R value of about 0.22 or less using computer software, such as X-PLOR [Yale University, ©1992, distributed by Molecular Simulations, Inc.; see, e.g., Blundell & Johnson, supra; Meth. Enzymol., vol. 114 & 115, H. W. Wyckoff et al., eds., Academic Press (1985)]. This information may thus be used to optimize known P38&ggr; inhibitors, and more importantly, to design new P38&ggr; inhibitors.

[0114] The structure coordinates described above may also be used to derive the dihedral angles, &phgr; and &psgr;, that define the conformation of the amino acids in the protein backbone. As will be understood by those skilled in the art, the &phgr;n angle refers to the rotation around the bond between the alpha carbon and the nitrogen, and the &psgr;n angle refers to the rotation around the bond between the carbonyl carbon and the alpha carbon. The subscript “n” identifies the amino acid whose conformation is being described [for a general reference, see Blundell and Johnson, Protein Crystallography, Academic Press, London, 1976].

[0115] Surprisingly, it has now been found that for the crystalline P38&ggr;-ligand complex, the conformation of Gly113 is very different from the conformations reported for corresponding amino acids in other protein kinases. In order to compare the conformations of P38&ggr; and other protein kinases at a particular amino acid site, such as Gly113, along the polypeptide backbone well-known procedures may be used for doing sequence alignments of the amino acids. Such sequence alignments allow for the equivalent or corresponding sites to be compared. One such method for doing a sequence alignment is the “bestfit” program available from Genetics Computer Group which uses the local homology algorithm described by Smith and Waterman in Advances in Applied Mathematics 2; 482 (1981).

[0116] A suitable amino acid sequence alignment will require that the proteins being aligned share a minimum percentage of identical amino acids. Generally, a first protein being aligned with a second protein should share in excess of about 35% identical amino acids with the second protein. Hanks et al., Science, 241, 42 (1988); Hanks and Quinn, Methods in Enzymology, 200, 38 (1991).

[0117] Equivalents of the Gly113 residue of p38&ggr; may also be identified by its functional position. Gly113 is the amino acid residue that immediately follows sequentially the amino acid residue that donates, or is capable of donating, a hydrogen bond to the N1 nitrogen of the adenosine ring of ATP or an ATP analog, if such ATP or ATP analog were to be in the binding pocket comprising the Gly113 residue. The ability of the amino acid to donate such a hydrogen bond occurs as the result of the spatial position of the amino acid in the binding pocket or the protein. As used herein, the term “corresponding amino acid” or “equivalent amino acid” refers to a particular amino acid in a protein kinase that corresponds to another, particular amino acid in a different protein kinase as determined by sequence alignment and/or its functional position.

[0118] Table 1 shows the sequence alignments for selected protein kinases where corresponding amino acids are shown in the same column. The amino acid numbering is based on the assignments given in the Swiss-Prot database which is an international protein sequence database distributed by the European Bioinformatics Institute (EBI) in Geneva, Switzerland. The database can be found at www.ebi.ac.uk/swissprot. Erk6_HUMAN is the database protein name for P38&ggr;. The ten amino acids immediately preceding G113 of P38&ggr; are given starting with T103. Thus, for example, Gly113 of P38&ggr; corresponds or is equivalent to the following: Gly110 of P38&agr; (MP38_HUMAN), Glu107 of mouse ERK2, and Asp150 of human JNK3. The last column of Table 1 shows the Swiss-Prot database accession number. 2 TABLE 1 Sequence Alignments for Selected Proteins Corresponding Amino Acid Sequences Using Swiss-Prot Access Protein Amino Acid Numbering Number ERK6_HUMAN T103 D F Y L V M P F M112 G113 P53778 MP38_HUMAN N100 D V Y L V T H L M109 G110 Q16539 ERK2_HUMAN K99 D V Y I V Q D L M108 E109 P28482 ERK2_MOUSE K97 D V Y I V Q D L M106 E107 P27703 JNK3_HUMAN Q140 D V Y L V M E L M149 D150 P53779 KAPA_MOUSE S114 N L Y M V M E Y V123 A124 P05132 INSR_HUMAN Q1097 P T L V V M E L M1106 A1107 P06213 LCK_HUMAN E309 P I Y I I T E Y M318 E319 P06239 ZA70_HUMAN E408 A L M L V M E M A417 G418 P43403 PKD1_DICDI T107 K I H F I M E Y A116 G117 P34100 KPCI_YEAST N898 R I Y F A M E F I907 G908 P24583 CLK1_HUMAN G235 H I C I V F E L L244 G245 P49759 CLK2_HUMAN G237 H M C I S F E L L246 G247 P49760 DOA_DROME G243 H M C I V F E M L252 G253 P49762 DSK1_SCHPO A160 H V C M V F E V L169 G170 P36616 MKK1_YEAST S293 S I Y I A M E Y M302 G303 P32490 MKK2_YEAST S286 S I Y I A M E Y M295 G296 P32491 NIMA_EMENI Q83 D L Y L Y M E Y C92 G93 P11837 KMOS_HUMAN S133 L G T I I M E F G142 G143 P00540 KC1A_HUMAN D84 Y N V L V M D L L93 G94 P48729 KC1B_BOVIN D84 Y N V L V M D L L93 G94 P35507 KC1D_HUMAN D76 Y N V M V M E L L85 G86 P48730 CK11_YEAST L136 H N I L V I D L L145 G146 P23291 CK12_YEAST L143 H N I L V I D L L152 G153 P23292 HR25_YEAST 576 Y N A M V I D L L85 G86 P29295 KNS1_YEAST N387 H I C L V T D L Y396 G397 P32350 KYK1_DICDI D1360 H H C I V T E W M1369 G1370 P18160 CKI1_SCHPO L79 H N V L V I D L L88 G89 P40233 CDK2_HUMAN N74 K L Y L V F E F L83 H84 P24941 KPBG_HUMAN T97 F F F L V F D L M106 K107 Q16816 KCC1_HUMAN G89 H L Y L I M Q L V98 S99 Q14012

[0119] As noted above, the conformation of Gly113 is very different from the conformations reported for corresponding amino acids in other protein kinases. For Gly113 of the P38&ggr;-AMPPNP complex, &psgr;112 was found to be about 24 degrees and &phgr;113 was found to be about 96 degrees. Table 2 shows the dihedral angles for Met112 and Gly113 of P38&ggr;-AMPPNP complex and how these angles compare to those of the corresponding amino acids in other MAP kinases whose crystal structures have been reported. The protein names for the known proteins are provided as their Protein Data Bank™ (pdb) accession numbers. The Protein Data Bank is an international repository for three dimensional structures and can be located at www.rcsb.org/pdb/. 3 TABLE 2 Dihedral Angles (in degrees) for Met112 and Gly113 and Equivalents in P38_ and Other Protein Kinases Met 112 Gly 113 Protein &phgr; &psgr; &phgr; &psgr; P38&ggr;-AMPPNP −106.2 23.8 96.24 −90.6 P38&agr;-liganda −80.8 −26.5 95.7 −22.5 1ERKb −119.1 131.7 −51.6 −55.6 2ERKc −99.5 130.3 −42.7 −49.9 1p38d −92.7 128.4 −82.1 −103.2 1ATPe −96.6 89.1 −56.1 −30.1 1JNKf −105.3 170.6 −92.2 −22.8 1IR3g −112.7 87.9 −44.2 −38.4 1IRKh −85.6 109.9 −40.7 −38.4 3LCKi −121.7 105.9 −53.3 −38.2 ain-house structure of complex with a designed inhibitor; bunphosphorylated ERK, reported in Nature, 367, 704, (1994); cphosphorylated ERK, Cell, 90, 859 (1997); dunphosphorylated p38&ggr;, Proc. Nat. Acad. Science, 94, 2327 (1997); ecyclic AMP dependent protein kinase or cAPK, Acta Crys. Sec. D, 49, 362 (1993); funphosphorylated JNK3, Structure, 6, 983 (1998); ginsulin receptor tyrosine kinase, Embo J., 16, 5572 (1997); hinsulin receptor tyrosine kinase, Nature, 372, 786, (1994); ilymphocyte-specific kinase, Nature, 368, 764, (1994)

[0120] It is well-recognized that there will be some variability in the conformations of corresponding amino acids in similar or identical proteins when the protein crystallization and structure determination are repeated. This variability in the &phgr; and &psgr; dihedral angles may be approximated by reference to Ramachandran plots comparing the conformations obtained for two or more identical or similar proteins [Blundell and Johnson, Protein Crystallography, Academic Press, London, 1976]. It may be expected that the dihedral angles of equivalent amino acid residues in identical or similar proteins will vary as much as about 45° or more.

[0121] It should be noted that the amino acid numbering defined in the Protein Data Banks may be offset from the numbering given in the Swiss-Prot database. This offset, when it occurs, will be readily understood by those skilled in the art. Thus, the sequences of those proteins that are listed in both databases may be easily compared despite offsets in amino acid numbering that may occur. Examples of such offsets occur for INSR_HUMAN where A1107 according to Swiss-Prot numbering is the same as A1080 in the PDB database and for LCK_HUMAN where E319 according to Swiss-Prot numbering is the same as E320 by PDB numbering.

[0122] The &psgr;112 and &phgr;113 dihedral angles of the P38&ggr;-AMPPNP complex shown in Table 2 indicate that the conformation of Gly113 in this complex is “flipped” or rotated considerably relative to corresponding amino acids in other MAP kinases. Therefore, the structure coordinates of P38&ggr; set forth in FIG. 1 represent, inter alia, what is believed to be a conformation at Met 112 and Gly113 that had not been observed for other crystalline protein kinases, especially other MAP kinases.

[0123] Accordingly, another embodiment of this invention relates to a crystalline protein kinase-ligand complex, said kinase comprising amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 is in the range of about 30° to 150°. Preferably, the &psgr; angle of the crystalline protein kinase-ligand complex is in the range of about −45° to 45° and most preferably in the range of about −30° to 30°. Preferably, the &phgr; angle is in the range of about 45° to 135°, and most preferably is in the range of about 60° to 120°. Examples of kinases that may provide such a crystalline protein kinase when complexed with a ligand are described by Hanks et al., Science, 241, 42 (1988) and Hanks and Quinn, Methods in Enzymology, 200, 38 (1991). Other examples of such kinases may be found at www.sdsc.edu/Kinases/pkr/pk_catalytic/pk_hanks_seq_align_long.html, where the kinases are listed with their corresponding sequence alignments.

[0124] Another embodiment of this invention relates to a crystalline protein kinase-ligand complex, said kinase selected from the proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 is in the range of about 30° to 150°. Preferably, the &psgr; angle of the crystalline protein kinase-ligand complex is in the range of about −45° to 45° and most preferably in the range of about −30° to 30°. Preferably, the &phgr; angle is in the range of about 45° to 135°, and most preferably is in the range of about 60° to 120°.

[0125] Structural information regarding the conformation of the Met112 and Gly113 residues of the crystalline P38&ggr; complex may be encoded in a machine-readable data storage medium as described above for encoding the other structural coordinates of the protein. Accordingly, another embodiment of this invention relates to a computer for producing a three-dimensional representation of an ATP binding site of a protein kinase-ligand complex, or portion thereof, wherein said computer comprises:

[0126] a) a machine-readable data storage medium comprising a data storage material encoded with machine-readable data, wherein said machine-readable data comprises the structure coordinates of a kinase, or portion thereof, said kinase or portion thereof comprising amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 is in the range of about 30° to 150°;

[0127] b) a working memory for storing instructions for processing said machine-readable data;

[0128] c) a central-processing unit coupled to said working memory and to said machine-readable data storage medium, for processing said machine readable data into said three-dimensional representation; and

[0129] d) an output hardware coupled to said central-processing unit, for receiving said three-dimensional representation. Preferably, the machine-readable data comprises the structure coordinates of a kinase, or portion thereof, said kinase comprising amino acid residues corresponding to the Met112 and Gly113 amino acids of P38&ggr; or corresponding to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, wherein the &psgr; angle is in the range of about −45° to 45° and most preferably in the range of about −30° to 30°, and the &phgr; angle is in the range of about 45° to 135°, and most preferably in the range of about 60° to 120°. In a more preferred embodiment of this computer, the machine readable data comprises the structure coordinates of a crystalline protein kinase-ligand complex, or portion thereof, where said kinase is selected from a protein listed in Table 1.

[0130] For designing new compounds that associate with a protein kinase binding pocket, it is useful to employ information that includes the conformations of the Met112 and Gly113 residues, or their equivalents, along with other structural information regarding amino acids in the binding pocket. For example, to evaluate the ability of a chemical entity to bind to a protein kinase, the conformations of Met112 and Gly113, or equivalents, may be used along with structure coordinates of the backbone atoms of amino acids in the protein kinase binding pocket. These structure coordinates and the structure coordinates of the p38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp170, Gly173, and Leu174 according to FIG. 1 should not differ by more than about 3.0 angstroms in root mean square deviation, preferably the root mean square deviation is within about 2.7 angstroms, and most preferably within about 2.5 angstroms. For example, the root mean square deviation between the structure coordinates of the p38&ggr; amino acids and those of a p38&ggr; complex (see Table 2) was found by applicants to be 2.41 angstroms. Resolution error may account for variation in the root mean square deviation of a few tenths of an angstrom.

[0131] Accordingly, another embodiment of this invention provides a method for evaluating the ability of a chemical entity to associate with a protein kinase binding pocket, said method comprising the steps of:

[0132] a) creating a computer model of the binding pocket using structure coordinates wherein:

[0133] (i) the root mean square deviation between said structure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp170, Gly173, and Leu174 according to FIG. 1 is within about 3.0 angstroms,

[0134] (ii) said binding pocket model depicts amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, and

[0135] (iii) said binding pocket model depicts the &psgr; angle of the residue corresponding to Met112 to be in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 to be in the range of about 30° to 150°;

[0136] b) employing computational means to perform a fitting operation between the chemical entity and the binding pocket model; and

[0137] c) analyzing the results of said fitting operation to quantify the association between the chemical entity and the binding pocket model.

[0138] A useful root mean square deviation between the structure coordinates of a particular binding pocket and the structure coordinates of the binding pocket of another protein kinase may be readily determined by one skilled in the art. For example, when the protein kinase is selected from a protein listed in Table 1, the root mean square deviation is preferably within about 2.7 angstroms, and is more preferably within about 2.5 angstroms.

[0139] This invention also provides a method for identifying a potential agonist or antagonist of a molecule comprising a P38&ggr;-like binding pocket, comprising the steps of:

[0140] a) creating a computer model of the binding pocket using structure coordinates wherein:

[0141] (i) the root mean square deviation between said scructure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp170, Gly173, and Leu174 according to FIG. 1 is within about 3.0 angstroms,

[0142] (ii) said binding pocket model depicts amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, and

[0143] (iii) said binding pocket model depicts the &psgr; angle of the residue corresponding to Met112 to be in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 to be in the range of about 30° to 150°;

[0144] b) employing said model of the binding pocket to design or select said potential agonist or antagonist;

[0145] c) synthesizing said agonist or antagonist; and

[0146] d) contacting said agonist or antagonist with said molecule to determine the ability of said potential agonist or antagonist to interact with said molecule.

[0147] A preferred embodiment of this method uses the structure coordinates of the Met112 and Gly113 amino acids of p38&ggr; or the Met112 and Gly113 equivalent residues of a protein listed in Table 1.

[0148] In order that this invention be more fully understood, the following examples are set forth. These examples are for the purpose of illustration only and are not to be construed as limiting the scope of the invention in any way.

EXAMPLE 1

Expression and Purification of P38&ggr; Protein

[0149] P38 with a His6 tag was overexpressed in E.Coli, and then purified by using metal affinity resin followed by MonoQ resin. The purified material was phosphorylated with constituitively active MKK6, and purified again with MonoQ resin (Fox, T. et al., manuscript in preparation). Size-exclusion chromatography was performed to determine the apparent molecular weights of unphosphorylated and phosphorylated P38&ggr; as follows. A Superdex 75 HR 10/30 column (Pharmacia, Uppsala) was equilibrated in 12.5 mM HEPES, pH 7.3, containing 6.25% (v/v) glycerol and 100 mM KCl. Bovine serum albumin (67 kDa), ovalbumin (43 kDa), chymotrypsinogen (25 kDa), ribonuclease A (13.7 kDa) were used to calibrate the column prior to P38&ggr; analyses. A flow rate of 0.25 ml/min was used for chromatographic runs and samples were loaded in a volume of 100-200—1 at 0.7-4 mg/ml.

EXAMPLE 2

Crystallization of P38&ggr;

[0150] Crystals of phosphorylated P38&ggr; complexed with AMP-PNP were grown by vapor diffusion. Clusters of rods appeared after 3 to 7 days when protein (0.5 mM P38&ggr; with 5 mM AMP-PNP and 0.02% C12E9) was mixed with an equal volume of reservoir (100 mM NaOAc, 100 mM Tris 8.5, 27% PEG 4000, 10 mM MgCl2, and 5 mM DTT) and allowed to stand at room temperature. Single crystals with 100 mM maximum thickness were separated from their parent cluster, cryoprotected by adding ethylene glycol to a final concentration of 15% over 15 min in three equal steps, and flash cooled to −170° C. in a stream of gaseous nitrogen.

EXAMPLE 3

X-Ray Data Collection and Structure Determination

[0151] The diffraction pattern displayed symmetry consistent with space group P212121, with unit cell dimensions a=63.50 Å, b=66.82 Å, and c=206.02 Å. Diffraction extended to 4.0 Å in the a*, b* direction and 3.0 Å in the c* direction. Data collection at NSLS X25 allowed a significant improvement in the observed diffraction limit: data were collected to 3.0 Å in the a*, b* direction and at least 2.4 Å in the c* direction. Data were integrated to 2.4 Å [Otwinowski, Z. in CCP4 Study Weekend (eds. Sawyer, L., Isaacs, N. & Bailey, S.) 56-62 (SERC Daresbury Laboratory, England) (1993); Minor, W. XDISPLAYF Program, Purdue University, (1993)]. The overall R-merge for the data was 6.7%, with I/sig (I)=2.0 at 2.4 Å resolution. The X-ray data comprised 31732 unique reflections derived from 118429 intensity measurements. The data were 90% complete overall and 76.5% complete in the 2.49-2.40 Å resolution shell. Data incompleteness, particularly in the highest resolution shell, reflects the anisotropic nature of the diffraction.

[0152] The volume of the asymmetric unit indicated the presence of two P38&ggr; molecules. The self-rotation function calculated with POLARAFN [Acta Crys D50, 760-763 (1994)] revealed a noncrystallographic peak with intensity half of the origin at Kappa=180°, omega=90°, and Phi=44°.

[0153] Coordinates for the structure of phosphorylated ERK2 were not initially available from the protein data bank and could not be-used for molecular replacement. Several different models for P38&ggr; were constructed based on the X-ray coordinates of P38&ggr; or unphosphorylated ERK2 with either all side chains truncated to alanine, or with only the nonconserved side chains truncated to alanine or glycine [Zhang et al., Nature 367, 704-711 (1994); Wilson and Su, J Biol Chem 271, 27696-27700 (1996)]. No rotation function solutions were obtained using these models with either the X-plor or AMORE molecular replacement packages. The anisotropy of the data, as well as the presence of two molecules in the asymmetric unit, could be reasons for the lack of a successful molecular replacement solution. Variability in the orientation between the large and small kinase domains may have been an additional complicating factor.

[0154] To position correctly an initial P38&ggr; model, experimental phases at low resolution were obtained from two derivatives. Crystals were soaked with 0.2 mM ethylmercurychloride (EMP) for 5 days, and with 2 mM EuCl3 overnight. Diffraction data were collected on the in house RaxisIIc, and integrated to 5.0 Å [see Owinowski and Minor, supra]. Difference Patterson maps were interpreted by using SHELXS-97 [Acta Crys A46, 467-473 (1990)]. The EMP derivative yielded four sites and the Europium derivative yielded two sites. These heavy atom positions were refined by using ML-PHARE [Acta Crys D50, 760-763 (1994)] which yielded an overall figure of merit of 0.53 to 5 Å. The resulting electron density maps showed clear solvent and protein regions. Six heavy atom sites were identified within a continuous envelope of protein density and grouped into two sets of three sites. These two sets were related to one another by a two-fold axis, which was consistent with the self-rotation function. Each set of three sites was assumed to correspond to a monomer of P38&ggr;, and the two-fold operation was used to improve the experimental electron density by noncrystallographic symmetry (NCS) averaging. Solvent flattening combined with two-fold averaging using Dm (final correlation coefficient of averaging of 0.851) produced an electron density map at 5.0 Å that allowed placement of the P38&ggr; model. The N-terminal domain had to be rotated by several degrees with respect to the C-terminal domain in order to fit both domains into the experimental density. At this stage the model was refined against the high resolution synchrotron data. Rigid body refinement and torsional dynamics refinement yielded an initial Rfree of 42%.

[0155] The quality of the model was improved by cycles of model building, positional refinement, and thermal factor refinement, interspersed with torsional dynamics runs using data from 50.0 to 2.4 Å. All stages of model refinement were carried out using the new program CNS [Acta Crys D54, 905-921 (1988)] with bulk solvent correction and anisotropic scaling. NCS restraints were applied throughout the refinement. The current P38&ggr; model contains two monomers, each with 329 protein residues, one bound AMP-PNP molecule, and two Mg2+ ions. A total of 186 water molecules were included in the entire asymmetric unit. The current Rwork is 23.2% (Rfree=28.3%) versus all data with |F|>2_(F) between 50-2.4 Å resolution (27841 reflections). PROCHECK was used to analyze the model stereochemistry [Acta Crys D50, 760-763 (1994)]. All of the residues were in the most favored and additional allowed regions of the Ramachandran plot. One residue per monomer (Val187) from the phosphorylation loop was in the disallowed region. The P38&ggr; model has deviations from ideal bond lengths and angles of 0.010 Å and 1.63° respectively. No electron density was observed for amino acids 1-7, 34-39, 316-321, 330-334, and 354-end, therefore these residues were not included in the model. The eight residue histidine tag and 21 residues at the C-terminus are also disordered. Subsequent to the structure refinement, the phosphorylated ERK2 coordinates were released, and the final refined P38&ggr; structure was compared with that structure.

EXAMPLE 4

Overall Structure

[0156] The P38&ggr; structure was solved with a combination of low resolution MIR and molecular replacement using a model of the unphosphorylated form of P38&agr; [Wilson and Su, J Biol Chem 271, 27696-27700 (1996)]. The current structure includes two P38&ggr; molecules per asymmetric unit, each with 329 amino acids, a bound AMP-PNP, and two Mg2+ ions. A total of 186 water molecules were modeled in the asymmetric unit. The current Rfree and Rwork are 28.3% and 23.2%, respectively. The refined model has deviations from ideal bond lengths and angles of 0.01 Å and 1.6°. The two P38&ggr; molecules in the asymmetric unit superimpose with an overall r.m.s.d. of 0.013 Å using all C&agr; atoms, and thus represent two independent but highly similar structures of activated P38&ggr;.

Comparison of Kinase Structures

[0157] Electron density for the main chain atoms of P38&ggr; is visible from residue 8 to 353, with breaks at residues 34-39, 316-321 and 330-334 (FIG. 1). The glycine rich loop, which contains the consensus Gly-X-Gly-X-X-Gly sequence (residues 34-39 in P38&ggr;) is mobile, and residues 34-39 could not be modeled. The homologous region of P38&agr; is also flexible, and has average B-values equal to 61 Å. In contrast, the AMP-PNP ligand is well ordered, as are all nearby secondary structural elements. Strong electron density for the residues at the N- and C-terminal ends of the glycine rich loop is also observed. The C-terminal 40 residues of both P38&ggr; molecules in the asymmetric unit are not as well ordered as the rest of the structure. Helix &agr;L16 can be modeled, but contains several disordered side chains. The region just before helix &agr;L16 is poorly ordered and does not form the 3/10 helix L16 observed in the structure of phosphorylated ERK2. Helix &agr;L16 and 3/10 helix L16 are involved in dimer formation in the structure of phosphorylated ERK2 [Canagarajah et al., Cell 90, 859-869 (1997)].

[0158] Activated P38&ggr; contains a small amino terminal domain comprised mainly of &bgr;-strands, and a large carboxyl terminal domain that consists mostly of &agr;-helices (FIG. 1). This fold is common among kinases [Taylor & Radzio-Andzlem (1994); Structure 2, 345-355; Kultz J Mol Evol 46, 571-588 (1998)]. A deep cleft at the interface between the domains forms the binding site tor ATP and Mg2+. The two domains are connected by a hinge, located at a point adjacent to the adenine base and near residue 113 (FIG. 1).

[0159] Whereas the sequence, fold, and topology of P38&ggr; is similar to P38&agr; (FIGS. 1, 2), the domains of activated P38&ggr; are closed relative to P38&agr; Independent superimpositions of the domains of P38&ggr; onto the P38&agr; structure yield r.m.s. deviations of 1.2 Å for the N-terminal domain (P38&ggr; C&agr; carbons from residues 10-16, 19-33, and 40-113), and 0.62 Å for the C-terminal domain (P38&ggr; C&agr; carbons from residues 125 to 160, 206 to 238 and 282 to 297). Greater differences between P38&ggr; and P38&agr; are observed when the whole proteins are compared. Superimposition of the C-terminal domain of P38&ggr; onto the corresponding lobe of P38&agr; revealed a rotation of the N-terminal domain of P38&ggr; by 20° relative to the orientation seen in P38&agr; (FIG. 2). Other differences between the structure of phosphorylated P38&ggr; and P38&agr; occur in the conformation of &agr;1L14, &agr;2L14, &agr;1L12, the phosphorylation loop, and &agr;L16.

[0160] Inter-domain rotation, or domain closure, is common in MAP kinase structures, and is observed to different extents. The structures of unphosphorylated and phosphorylated ERK2 show a 5° difference in domain closure. The structure of unphosphorylated JNK3 reveals that a 10° domain rotation would be needed to superimpose both domains with the structure of phosphorylated P38&ggr; or phosphorylated ERK2. P38&agr; MAP kinase is more open in its unphosphorylated state than ERK2 or JNK3. Despite a large difference in the conformations of the unphosphorylated proteins, the domains of the activated forms of P38&ggr; and ERK2 can be superimposed with a rotation of only 3°. Comparison to solved kinase structures indicates that the relative positions of the domains in activated P38&ggr; is most similar to activated ERK2 MAP kinase.

[0161] The structures of phosphorylated P38&ggr; and phosphorylated ERK2 are similar, with a few significant differences. One conformational difference is a movement of the &agr;1L14, &agr;2L14 helical region. With the large domains superimposed, the difference in &agr;1L14, &agr;2L14 orientation between the two structures is about 6 Å, when measured at the most extreme portion of the helices. Another difference between the two structures is that the P38&ggr; activation loop is six residues shorter than the activation loop in ERK2. Excluding these two regions allows one to superimpose P38&ggr; C&agr; carbons 19-33, 40-58, 61-94, 97-113, 117-177, 182-243, and 269-315 with the corresponding ERK2-P2 atoms to yield an r.m.s.d. of 1.1 Å. This reflects the high similarity between the two structures. A comparison of the activation loops, using P38&ggr; C&agr; carbons 173-177 and 182-188 yields an r.m.s.d. of 0.3 Å.

[0162] The structure of the phosphorylation loop differs between phosphorylated P38&ggr; and unphosphorylated P38&agr; (FIG. 2). The phosphorylation loop contains the TGY sequence present in all P38 MAP kinases. Phosphorylation of Thr183 and Tyr185 results in a movement of the activation loop, and produces changes in the P38&ggr; structure.

[0163] Phospho-Thr183 sits at the interface between the two domains. The Thr183 phosphate group interacts with Arg70, Arg73 and Lys69 from the N-terminal domain, and Arg152 and Arg176 from the C-terminal domain (FIGS. 1 and a point adjacent to the adenine base and near residue 113. The hinge-point and residue pThr183 are located at opposite ends of the interface between the two domains. The network of interactions between pThr183 and these basic residues pulls the domains together. As a result, the relative orientations of the amino acids, including the catalytic residues, located between the hinge and pThr183 are changed. A similar set of interactions between the phospho-threonine and nearby basic residues was reported for the structure of phosphorylated ERK2 [Canagarajah et al., Cell 90, 859-869 (1997)]. Phosphorylated P38&ggr; is in a conformation consistent with activity. The active site of phosphorylated P38&ggr; is shown in detail in FIG. 4, and compared with the active sites of P38&agr; and cAPK in FIGS., 5a and 5b. The interactions between the non-hydrolyzable nucleotide analog AMP-PNP and P38&ggr; (FIG. 4) are very similar to those made between bound nucleotide and cAPK [Zheng et al., Acta Cryst. D49, 362-365 (1993); Bossemeyer et al., EMBO Journal 12, 849-859 (1993); Narayana et al., Structure 5, 921-935 (1997)].

[0164] The N1 and N6 nitrogen atoms of AMP-PNP form hydrogen bonds to the backbone amide nitrogen atom of Met112 and the backbone carbonyl oxygen atom of Pro110, respectively. Interactions between the glycine rich loop and the nucleotide are not observed in the P38&ggr; structure. The relative positions of catalytic residues Lys53, Glu74 and Asp153 provide information about the state of activation of the kinase [Kumar et al., J. Biol. Chem. 270, 29043-29046 (1995); Robinson et al., Curr Opin Cell Biol 9, 180-186 (1997)]. Comparison of P38&ggr; with cAPK after superimposing the nucleotides from the two structures (FIG. 5b), reveals that the active site residues in the two structures are in almost the same conformation. The cAPK structure also contains a bound peptide inhibitor, and the complex is believed to represent a bioactive conformation of cAPK [Zheng et al, supra; Bossemyer et al., supra; Narayana et al., supra]. The nucleotides in both structures adopt almost the same conformations, and the relative positions of the catalytic residues Lys-56, Glu-74 and Asp-153 are conserved. There are also two bound metal ions in each complex. After superimposition, metal I in cAPK is separated from the corresponding metal in P38&agr; by 0.5 Å, and metal II from P38&ggr; is 1.4 Å removed from metal II in cAPK. Because the conformation and relative orientation of the catalytic residues and cofactors in the active sites of the two kinases are almost the same, the structure of phosphorylated P38&ggr; described here is likely to represent an active conformation.

[0165] Comparing the phosphorylated P38&ggr; to the known, unphosphorylated P38&agr; one finds that the active site residues of P38&agr; are significantly displaced relative to their orientation in P38&ggr;. This presumably reflects the inactive state of unphosphorylated P38&agr; (FIG. 5a). Two types of structural differences are observed between unactivated P38&agr; and activated P38&ggr;. A rigid body motion occurs between the two domains, and secondary structure elements and residues move as a result of phosphorylation and AMP-PNP binding.

[0166] Using the newly-determined structure of P38&ggr;, the structure of unphosphorylated P38&agr; could be altered to properly position its catalytic residues in an active conformation. Without the P38&ggr; structural information, it was not known whether domain movement alone would be enough to properly position the catalytic in an active conformation or whether activation would also require other changes [Johnson et al., Curr. Opin. Struct. Biol. 6, 762-769 (1996); Yamaguchi et al., Nature 384, 484-489 (1996); Johnson et al., Cell 85, 149-158 (1996); Russo et al., Nature Struc Biol 3, 696-700 (1996)].

[0167] To address this question, the structure of unphosphorylated P38&agr; was altered to resemble phosphorylated P38&ggr;. Only a rigid-body movement, centered on the hinge residue 113, was used to change the relative orientation of the two domains in P38&agr;. The resulting model maintains the detailed secondary structure features present in non-phosphorylated P38&agr;, but has the same domain closure as P38&ggr;. The positions of catalytic residues in the active site of this modified P38&ggr; model match well to those observed in the structure of activated P38&ggr;. The rigid body movement shifts P38&agr; residue Lys-53 2.9 Å closer to its counterpart in P38&ggr; (from 4.4 Å to 1.5 Å separation). Glu-71 (P38&agr;) moves 2.8 Å nearer to its equivalent residue in P38&ggr; (from 3.2 Å to 0.4 Å separation). Thus, the structures of P38&agr; and P38&ggr; suggest that a simple domain rotation accounts for most of the rearrangement of catalytic residues necessary for activation of P38&ggr;.

[0168] Other movements may contribute to activation of P38&ggr;. For example, phosphorylation of Tyr185 leads to a rearrangement of surrounding secondary structure elements that may effect substrate binding. Arg192 interacts with the pTyr185 phosphate group in the P38&ggr; structure, and is shifted more than 5 Å relative to its position in the apo-P38&agr; structure. Such coordination of Arg 192 and its effect on substrate binding have been discussed with regard to ERK2and JKN3 [Zhang, Nature 367, 704-711 (1994); Xie and Su, supra; Canagarajah, supra]. In the P38&ggr; structure, pTyr185 interacts directly with Arg189 and Arg192 (FIG. 3). Comparison of the P38&ggr; pTyr185 conformation, as well as the backbone conformation with the corresponding residue of phosphorylated ERK2, shows that the two residues are in nearly the same conformation.

Activated P38_is Monomeric

[0169] The two P38&ggr; proteins in the crystallized complex show no evidence of dimeric interaction, as evidenced by the examination of the activation loops of the two proteins. This is unlike the activated, phosphorylated ERK2, which is believed to reveal a dimer interface that is not observed in the non-phosphorylated form [Zhang et al., supra; Canagarajah et al., supra; Khokhlatchev et al., supra]. The dimer interface in phosphorylated ERK2 reportedly buries a total of 1470 Å2 of surface area, and is formed in part by an ion pair between His176 from one molecule and Glu343 from the other molecule. In addition, Leu333, Leu336, and Leu344 are reported to further stabilize the dimer interface.

[0170] The entire surface of each P38&ggr; molecule in the asymmetric unit was examined in search of any dimer interface. The crystal of P38&ggr; belongs to space group P212121, which contains only two-fold screw axes, but no crystallographic two-fold axes. The only two-fold axis in the crystal is the non-crystallographic axis which relates the two molecules within the asymmetric unit. This dimeric interaction involves Pro282, Asn286, Lys290, Leu283, Pro309, and Glu312. This non-crystallographic dimer interface buries only 680 Å2 of surface area, less than half of the 1470 Å2 buried in the phosphorylated ERK2 dimer interface.

[0171] To characterize further the oligomeric state of activated P38&ggr; in solution, size-exclusion chromatography was performed to determine the apparent molecular weights of unphosphorylated and phosphorylated P38&ggr;. To facilitate comparison with the phosphorylated ERK2 results [Khokhlatchev et al., supra], the same column resin, buffer, and loading conditions were used. The chromatographic profiles of unphosphorylated and phosphorylated P38&ggr; showed that both proteins eluted with a similar retention time, corresponding to a molecular. weight of 44.5 kDa as determined from the protein calibration curve. The absence of dimer formation of phosphorylated P38&ggr; in solution is consistent with the absence of dimer formation in the crystal structure of P38&ggr;. It is also consistent with the absence of dimer formation in ERK2 mutants where His176 is deleted [Khokhlatchev et al., supra].

[0172] Conformations of Activation Loops of Kinases

[0173] The number of residues in the activation loops of different kinases varies, ranging from 8 amino acids in calmodulin dependent DAP-kinase to 37 residues in LIMK2 [Deiss et al., Genes Dev. 9, 15-30 (1995); Okano et al., J. Biol. Chem. 270, 31321-31330 (1995)]. The P38&ggr; activation loop consists of residues Gly173-Thr188. The phosphorylation loop of ERK2 is six residues longer in sequence and spans amino acids Gly167-Thr188. The loop region of cAPK is the same length as P38&ggr;, and spans amino acids Gly186-Thr201. FIG. 6 highlights the loop regions from P38&ggr;, ERK2-P2, and cAPK. Except for a longer loop size for ERK2, the structures of the loop regions of activated P38&ggr; and activated ERK2are nearly identical. The distance between the phosphate moieties from Thr183 in P38&ggr; and ERK2 is only 0.4 Å, and separation between the Tyr185 phosphate from P38&ggr; and ERK2 is 1.6 Å. The phosphorylation loop of cAPK does not superimpose as well with the two MAP kinase phosphorylation loops, although the Thr phosphate is only 2.0 Å away from the P38&ggr; Thr183 phosphate. The phosphorylation loop regions from P38&ggr;, ERK2 and cAPK have different lengths, but in their phosphorylated states adopt almost the same conformations.

[0174] FIGS. 1a-6 further depict the structure of the phosphorylated P38&ggr;/MgAMP-PNP complex. Thus, FIG. 1a depicts an overview of the phosphorylated P38&ggr; structure. The large and small domains are pulled together by interactions mediated by phospho-Thr183. Ribbon diagrams of the activated P38&ggr; structure with the amino-terminal small domain are colored light orange and the carboxy-terminal large domain colored blue. The interface between the two domains (residue 113) can be thought of as a hinge point through which domain movement occurs. Four Arg residues and one Lys residue are explicitly shown coordinated to the phosphate of pThr183. Arg70, Arg73 and Lys69 anchor the small domain to pThr183, and Arg152 and Arg176 anchor the large domain to pThr183. PThr183 pulls the domains together. All figures were made with RIBBONS [Carson et al., J. Mol. Graphics 4, 121-122 (1986)].

[0175] FIG. 2 is a superimposition of the structures of unphosphorylated P38&agr; and phosphorylated P38&ggr;. P38&agr; is shown in light blue and dark blue (activation loop), and P38&ggr; is shown in light orange and dark orange (activation loop). The C&agr; atoms from residues 125 to 160, 206 to 238 and 282 to 297 were used to superimpose the two proteins with an r.m.s.d. of 0.62 Å. Also shown is the AMP-PNP and two Mg2+ ions from the P38&ggr; structure. All atoms of the phosphorylated Thr183 and Tyr185 from the P38&ggr; structure are shown. Major changes upon phosphorylation are a significant domain closure and a rearrangement of the activation loop.

[0176] FIG. 3 is a detailed stereo view of activation loop. All atom stereo view of the P38&ggr; activation loop (residues 174 to 189). Residues that coordinate pThr183 and pTyr185 are also shown. Hydrogen bonds are indicated with dashed grey lines. The phosphate atoms are shown in pink.

[0177] FIG. 4 is a stereo view of AMP-PNP. All major interactions with protein side chains are indicated with dashed grey lines. The bound Mg2+ ions are indicated by black spheres. The phosphate atoms are shown in pink. Met109 can be seen behind the adenine base, blocking the hydrophobic pocket. Water molecules have been removed for clarity.

[0178] FIGS. 5a and 5b are a comparison of the active site of activated P38&ggr; with P38&agr; and cAPK. P38&ggr; is shown in orange, P38&agr; in blue, and cAPK in red. In all three structures a salt bridge is observed between Lys56 and Glu74 (P38&ggr; numbering). a) Comparison of the active sites of P38&ggr; with P38&agr; by superimposition of their carboxyl terminal large domains. Catalytic residues are misaligned. The distance between Asp153 and Lys53 is 12.6 Å in the P38&agr; structure compared with 8.5 Å in the phosphorylated P38&ggr; structure. b) Comparison of the active sites of P38&ggr; with cAPK (Protein Data Base code: 1ATP, ref. 22) by superimrposition of all atoms of their bound AMP-PNP molecules. All catalytic residues align to within a fraction of an Å. The distance between Asp153 and Lys53 is 8.5 Å in the activated P38&ggr; structure. This distance is very close to the distance of 7.8 Å observed in activated cAPK, suggesting that the structure reported here is of the activated kinase. Asp171 is excluded from these figures for clarity because it is obscured by AMP-PNP and Mg2+ ions.

[0179] FIG. 6 is a comparison of activated phosphorylation loops from P38&ggr; (dark orange), ERK2 (dark blue), and cAPk (red). Superimposition of these three structures was with the C&agr; atoms of residues 125 to 160, 206 to 238 and 282 to 297 of P38&ggr;.In order to ensure an unbiased comparison of the lip regions, these residues were omitted from the calculation. All three lip regions have different lengths, but have surprisingly similar conformation. Comparison of P38&ggr; and ERK2 superimposes the two phorphorylated amino acids almost exactly, despite a six amino acid difference in length. The phosphorylated Thr197 of cAPK also superimposes well with the two MAP kinase structures. This comparison suggests that the phosphorylated lip structures observed in P38&ggr; and ERK2 may be representative of all MAP kinases.

EXAMPLE 5

The Use of P38&ggr;/MgAMP-PNP

Coordinates for Inhibitor Design

[0180] The coordinates of FIG. 1 are used to design compounds, including inhibitory compounds, that associate with P38&ggr; or homologues of P38&ggr;. This process may be aided by using a computer comprising a machine-readable data storage medium encoded with a set of machine-executable instructions, wherein the recorded instructions are capable of displaying a three-dimensional representation of the P38&ggr;/MGAMP-PNP complex or a portion thereof. The graphical representation is used according to the methods described herein to design compounds. Such compounds associate with the P38&ggr; at the active site.

EXAMPLE 6

P38&ggr; Activity Inhibition Assay

[0181] To determine the IC50 of compound binding to P38&ggr;, the kinase activity of P38Y was monitored by coupled enzyme assay. In this assay, for every molecule of ADP generated by the P38Y kinase activity one molecule of NADH is converted to NAD which can be conveniently monitored as an absorbance decrease at 340 nm. The following are the final concentrations of various reagents used in the assay: 100 mM HEPES buffer, pH 7.6, 10 mM MgCl2, 30 &mgr;M ATP, 2 mM phosphoenolpyruvate, 2 &mgr;M pyruvate kinase, 2 &mgr;M lactate dehydrogenase, 200 &mgr;M NADH, 200 &mgr;M EGF receptor peptide KRELVEPLTPSGEAPNQALLR, and 10 nM activated P38&ggr;. First, all of the above reagents with the exception of ATP were mixed and 175 &mgr;l aliquots were placed per well of 96-well plate. A 5 &mgr;l DMSO solution of the compound was added to each well, mixed, and allowed to stand at 30° C. for 10 minutes. Typically about 10 different concentrations of the compound were tested. The reactions were initiated with the addition of 20 &mgr;l of ATP solution. Absorbance change at 340 nm were monitored as a function of time. IC50 is obtained by fitting the rates vs. compound concentration data to a simple competitive inhibition model.

[0182] While we have described a number of embodiments of this invention, it is apparent that our basic constructions may be altered to provide other embodiments which utilize the products, processes and methods of this invention. Therefore, it will be appreciated that the scope of this invention is to be defined by the appended claims, rather than by the specific embodiments which have been presented by way of example.

Claims

1. A crystalline composition comprising a phosphorylated P38 protein-ligand complex.

2. The crystalline composition of claim 1 wherein the complex is capable of being resolved at 2.4 Å resolution, the complex comprising:

a) a purified enzyme selected from phosphorylated P38&agr;, phosphorylated P38&bgr;, phosphorylated P38&dgr;, phosphorylated P38&ggr;, or a phosphorylated isoform of any of the foregoing;

b) a ligand; and

c) magnesium ions.

3. The crystalline composition according to claim 2, wherein said enzyme is P38&ggr;.

4. A crystalline protein kinase-ligand complex, said kinase comprising a binding pocket defined by the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1, or a homologue of said kinase, wherein said homologue comprises a binding pocket that has a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å.

5. A crystalline protein kinase-ligand complex, said kinase selected from the proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 of p38&ggr; is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 of p38&ggr; is in the range of about 30° to 150°.

6. The crystalline protein kinase-ligand complex of claim 5 wherein the &psgr; angle is in the range of about −45° to 45°.

7. The crystalline protein kinase-ligand complex of claim 6 wherein the &psgr; angle is in the range of about −30° to 30°.

8. The crystalline protein kinase-ligand complex of any of claims 5 to 7 wherein the &phgr; angle is in the range of about 45° to 135°.

9. The crystalline protein kinase-ligand complex of any of claims 5 to 7 wherein the &phgr; angle is in the range of about 60° to 120°.

10. A crystalline protein kinase-ligand complex, said kinase comprising amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 is in the range of about 30° to 150°.

11. A method for evaluating the ability of a chemical entity to associate with a molecule or molecular complex comprising a binding pocket, said method comprising the steps of:

a) creating a computer model of the binding pocket using structure coordinates wherein the root mean square deviation between said structure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1 is not more than about 1.15 Å;

b) employing computational means to perform a fitting operation between the chemical entity and said computer model of the binding pocket; and

c) analyzing the results of said fitting operation to quantify the association between the chemical entity and the binding pocket model.

12. The method according to claim 11, wherein said binding pocket is further defined by the structure coordinates of P38&ggr; amino acids Pro32, Cys42, Ser43, Val53, Ile55, Lys57, Leu58, Thr59, Arg70, Glu74, Gly88, Leu107, Val108, Leu116, Gly117, Pro156, Leu159, Val161, Lys168, Phe172, Ala175, and Thr188 according to FIG. 1.

13. The method according to claim 12 wherein said molecule or molecular complex is defined by the set of structure coordinates for all P38&ggr; amino acids according to FIG. 1.

14. A method of utilizing molecular replacement to obtain structural information about a molecule or a molecular complex of unknown structure, comprising the steps of:

a. crystallizing said molecule or molecular complex;

b. generating an X-ray diffraction pattern from said crystallized molecule or molecular complex;

c. applying at least a portion of the structure coordinates set forth in FIG. 1 to the X-ray diffraction pattern to generate a three-dimensional electron density map of at least a portion of the molecule or molecular complex whose structure is unknown.

15. A computer for producing a three-dimensional representation of a molecule or molecular complex, wherein said computer comprises:

a) a machine-readable data storage medium comprising a data storage material encoded with machine-readable data, wherein said machine-readable data comprises the structure coordinates of P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1, or structural coordinates having a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å;

b) a working memory for storing instructions for processing said machine-readable data;

c) a central-processing unit coupled to said working memory and to said machine-readable data storage medium, for processing said machine readable data into said three-dimensional representation; and

d) an output hardware coupled to said central-processing unit, for receiving said three-dimensional representation.

16. A method for identifying a potential agonist or antagonist of a molecule comprising a P38&ggr;-like binding pocket, comprising the steps of:

a. using the atomic coordinates of Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1±a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, to generate a three-dimensional structure of molecule comprising the P38&ggr;-like binding pocket;

b. employing said three-dimensional structure to design or select said potential agonist or antagonist;

c. synthesizing said agonist or antagonist; and

d. contacting said agonist or antagonist with said molecule to determine the ability of said potential agonist or antagonist to interact with said molecule.

17. The method according to claim 16, wherein the atomic coordinates of Pro32, Val33, Ala40, Val41, Cys42, Ser43, Val53, Ala54, Ile55, Lys56, Lys57, Leu58, Thr59, Arg70, Glu74, Ile87, Gly88, Leu107, Val108, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Leu116, Gly117, Lys118, Asp153, Lys155, Pro156, Gly157, Asn158, Leu159, Ala160, Val161, Lys168, Leu170, Asp171, Phe172, Gly173, Leu174, Ala175, and Thr188 according to FIG. 1±a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, are used to generate said three-dimensional structure of the molecule comprising a P38&ggr;-like binding pocket.

18. The method according to claim 17, wherein the atomic coordinates of all the amino acids of P38&ggr; according to FIG. 1±a root mean square deviation from the backbone atoms of said amino acids of not more than 1.15 Å, are used to generate a three-dimensional structure of molecule comprising a P38&ggr;-like binding pocket.

19. A computer for producing a three-dimensional representation of a protein kinase or a protein kinase-ligand complex, or portion thereof, wherein said computer comprises:

a) a machine-readable data storage medium comprising a data storage material encoded with machine-readable data, wherein said machine-readable data comprises the structure coordinates of said kinase, or portion thereof, said kinase or portion thereof comprising amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, wherein the &psgr; angle of the residue corresponding to Met112 is in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 is in the range of about 30° to 150°;

b) a working memory for storing instructions for processing said machine-readable data;

c) a central-processing unit coupled to said working memory and to said machine-readable data storage medium, for processing said machine readable data into said three-dimensional representation; and

d) an output hardware coupled to said central-processing unit, for receiving said three-dimensional representation.

20. The computer of claim 19 wherein the &psgr; angle is in the range of about −45° to 45°.

21. The computer of claim 20 wherein the &psgr; angle is in the range of about −30° to 30°.

22. The computer of any of claims 19 to 21 wherein the &phgr; angle is in the range of about 45° to 135°.

23. The computer of any of claims 19 to 21 wherein the &phgr; angle is in the range of about 60° to 120°.

24. The computer of claim 19 wherein the machine-readable data comprises the structure coordinates of a protein kinase, or portion thereof, said kinase selected from a protein listed in Table 1.

25. A method for evaluating the ability of a chemical entity to associate with a protein kinase binding pocket, said method comprising the steps of:

a) creating a computer model of the binding pocket using structure coordinates wherein:

(i) the root mean square deviation between said structure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1 is within about 3.0 angstroms,

(ii) said binding pocket model depicts amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, and

(iii) said binding pocket model depicts the &psgr; angle of the residue corresponding to Met112 to be in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 to be in the range of about 30° to 150°;

b) employing computational means to perform a fitting operation between the chemical entity and the binding pocket model; and

c) analyzing the results of said fitting operation to quantify the association between the chemical entity and the binding pocket model.

26. The method of claim 25 wherein the root mean square deviation is within about 2.7 angstroms.

27. The method of claim 26 wherein the root mean square deviation is within about 2.5 angstroms.

28. The method of claim 25 wherein the &psgr; angle is in the range of about −45° to 45°.

29. The method of claim 26 wherein the &psgr; angle is in the range of about −45° to 45°.

30. The method of claim 27 wherein the &psgr; angle is in the range of about −45° to 45°.

31. The method of claim 25 wherein the &psgr; angle is in the range of about −30° to 30°.

32. The method of claim 26 wherein the &psgr; angle is in the range of about −30° to 30°.

33. The method of claim 27 wherein the &psgr; angle is in the range of about −30° to 30°.

34. The method of any of claims 25 to 33 wherein the &phgr; angle is in the range of about 45° to 135°.

35. The method of claim 25 to 33 wherein the &phgr; angle is in the range of about 60° to 120°.

36. The method of claim 25 wherein the protein kinase is selected from a Table 1 protein.

37. A method for identifying a potential agonist or antagonist of a molecule comprising a P38&ggr;-like binding pocket, comprising the steps of:

a) creating a computer model of the binding pocket using structure coordinates wherein:

(i) the root mean square deviation between said structure coordinates and the structure coordinates of the P38&ggr; amino acids Val33, Ala40, Val41, Ala54, Lys56, Ile87, Met109, Pro110, Phe111, Met112, Gly113, Thr114, Asp115, Lys118, Asp153, Lys155, Gly157, Asn158, Ala160, Leu170, Asp171, Gly173, and Leu174 according to FIG. 1 is within about 3.0 angstroms,

(ii) said binding pocket model depicts amino acid residues that correspond by functional and/or sequence alignment to the Met112 and Gly113 residues of P38&ggr; or that correspond by functional and/or sequence alignment to the Met112 and Gly113 equivalent residues of one or more proteins listed in Table 1, and

(iii) said binding pocket model depicts the &psgr; angle of the residue corresponding to Met112 to be in the range of about −60° to 60° and the &phgr; angle of the residue corresponding to Gly113 to be in the range of about 30° to 150°;

b) employing said model of the binding pocket to design or select said potential agonist or antagonist;

c) synthesizing said agonist or antagonist; and

d) contacting said agonist or antagonist with said molecule to determine the ability of said potential agonist or antagonist to interact with said molecule.

38. The method of claim 37 wherein the root mean square deviation is within about 2.7 angstroms.

39. The method of claim 38 wherein the root mean square deviation is within about 2.5 angstroms.

40. The method of claim 37 wherein the &psgr; angle is in the range of about −45° to 45°.

41. The method of claim 38 wherein the &psgr; angle is in the range of about −45° to 45°.

42. The method of claim 39 wherein the &psgr; angle is in the range of about −45° to 45°.

43. The method of any of claims 37 to 42 wherein the &psgr; angle is in the range of about −30° to 30°.

44. The method of any of claims 37 to 42 wherein the &phgr; angle is in the range of about 45° to 135°.

45. The method of any of claims 37 to 42 wherein the &phgr; angle is in the range of about 60° to 120°.