Crystal structure of a deacetylase and inhibitors thereof

BACKGROUND OF THE INVENTION

[0003] The present invention relates to a histone deacetylase homologue from the hyperthermophilic bacterium Aquifex aeolicus, HDLP (histone deacetylase-like protein; also known as AcuCl), which shares 35.2% sequence identity with human histone deacetylase (HDAC1), that can be co-crystallized with an inhibitory ligand, and more particularly, to the detailed crystallographic data obtained from said co-crystallization which is disclosed herein. The invention also relates to methods of using the crystal structure and x-ray crystallographic coordinates of the apo-HDLP inhibitor-bound HDLP to design, isolate and screen compounds which bind to and inhibit the active site of HDLP and HDLP-related proteins, such as those proteins belonging to the HDAC family, including HDAC1.

[0004] The reversible modification of histones by-acetylation is associated with changes in nucleosome conformation and chromatin structure, and plays an important role in the regulation of gene expression (reviewed in Davie and Chadee, 1998, J. Cell Biochem. Suppl. 30-31:203-213). The histone acetylase and deacetylase enzymes that carry out these modifications are involved in many cellular processes such as cell cycle progression and differentiation, and their deregulation is associated with several types of human cancer (reviewed- in Kouzarides, 1999, Curr. Opin. Genet. Dev. 9:40-48; Hassig et al., 1997, Chem. Biol. 4:783-789; Fenrick and Heibert, 1998, J. Cell. Biochem. Suppl. 30-31:194-202).

[0005] Recently, several experimental-antitumor compounds, such as trichostatin A (TSA), trapoxin, suberoylanilide hydroxamic acid (SAHA), and phenylbutyrate have been shown to act, at least in part, by inhibiting histone deacetylases. Richon et al., 1998, Proc. Natl. Acad. Sci., USA 95:3003-3007; Yoshida et al., 1990, J. Biol. Chem. 265:17174-17179; Kijima et al., 1993, J. Biol. Chem. 268:22429-22435. Additionally, diallyl sulfide and related molecules (Lea et al., 1999, Int. J. Oncol. 2:347-352), oxamflatin (Kim et al., 1999, Oncogene 15:2461-2470), MS-27-275, a synthetic benzamide derivative (Saito et al., 1999, Proc. Natl. Acad. Sci. 96:4592-4597), butarate derivatives (Lea and Tulsyan, 1995, Anticancer Res. 15:879-883), FR901228 (Nokajima et al., 1998, Exp. Cell Res. 241:126-133), depudecin (Kwon et al., 1998, Proc. Natl. Acad. Sci. USA 95:3356-3361), and m-carboxysinnamic acid bishydroxamide (CBHA; Richon et al., Proc. Natl. Acad. Sci. USA 95:3003-3007) have been shown to inhibit histone deacetylases. In vitro, these compounds can inhibit the growth of fibroblast cells by causing cell cycle arrest in the G1 and G2 phases (Richon et al., 1996, Proc. Natl. Acad. Sci. USA 93:5705-5708; Kim et al., 1999, Oncogene 18:2461-2470; Yoshida et al., 1995, Bioessays 17:423-430; Yoshida & Beppu, 1988, Exp. Cell. Res. 177:122-131), and can lead to the terminal differentiation and loss of transforming potential of a variety of transformed cell lines. Richon et al., 1996, Proc. Natl. Acad. Sci. USA 93:5705-5708; Kim et al., 1999, Oncogene 18:2461-2470; Yoshida et al., 1987, Cancer Res. 47:3688-3691. In vivo, phenylbutyrate is effective in the treatment of acute promyelocytic leukemia in conjunction with retinoic acid. Warrell et al., 1998, J. Natl. Cancer Inst. 90:1621-1625. SAHA is effective in preventing the formation of mammary tumors in rats, and lung tumors in mice. Desai et al., 1999, Proc. AACR 40: abstract #2396; Cohen et al., Cancer Res., submitted.

[0006] Histone deacetylases catalyze the removal of acetyl groups from the e-amino groups of lysine residues clustered near the N-terminus of nucleosomal histones, and this process is associated with transcriptional repression (reviewed in Struhl, 1998, Genes Dev. 12:599-606). Deletion of the yeast histone deacetylase gene, rpd3, or its pharmacological inactivation with trichostatin A reduces the transcriptional repression in a subset of promoters, such as those of Ume6-regulated genes. -Kadosh &-Struhl, 1998, Mol. Cell. Biol. 18:5121-5127. This is accompanied by-the increased acetylatibn of H4 histones in the repressed promoter and its vicinity, but has no effect on histones at promoter distal regions. Kadosh & Struhl, 1998, Mol. Cell. Biol. 18:5121-5127; Rundlett et al., 1998, Nature 392:831-835.

[0007] Histone deacetylases are recruited to specific promoters by associating with DNA-binding transcriptional repressors, either directly or through co-repressors that bridge the deacetylase to the transcriptional repressors. For example, the Mad and Ume6-repressors bind to the co-repressor Sin3A (Laherty et al., 1997, Cell 89.349-356; Hassig et al., 1997, Cell 89:341-347; Kadosh & Struhl, 1997, Cell 89:365-371), and the nuclear receptors bind N-CoR and the related SMRT co-repressors. Nagy et al., 1997, Cell 89:373-380; Alland et al, 1997, Nature 387:49-55; Heinzel et al, 1997, Nature 20 387:43-48.

[0008] The deregulation of histone deacetylase recruitment appears to be one of the mechanisms through which these enzymes contribute to tumorigenesis. In acute promyelocytic leukemia (APL), chromosomal translocations fuse the retinoic acid receptor-α (RARα) to either PLZF or to PML. These fusion oncoproteins have aberrant transcriptional repression activity resulting, in part, through the recruitment of a co-repressor and, in turn, HDACs. Grignani et al, 1998, Nature 391:815-818; Lin et al., 1998, Nature 391:811-814. Treatment of PLZF-RARα APL cells with TSA enhances their responsiveness to retinoic acid-induced differentiation. Grignani et al, 1998, Nature 391:815-818; Lin et al., 1998, Nature 391:811-814.

[0009] The histone deacetylases comprise a large family of proteins, conserved from yeast to man, and are divided into two related classes. Class I is characterized by human HDACI, 2, 3 (Taunton et al., 1996, Science 272:408-411; Yang et al., 1996, Proc. Natl. Acad. Sci. USA 93:12845-12850; Emiliani et al., 1998, Proc. Natl. Acad. Sci. USA 95:2795-2800), and yeast RPD3 (Videl & Gaber, 1991, Mol. Cell. Biol. 11:6317-6327), and class II by the human HDAC4, 5, 6 (Grozinger et al., 1999, Proc. Natl. Acad. Sci. USA 96 :4868-4873; Fischle, et al., 1999, J. Biol. Chem. 274:11713-11720), and yeast HDA1 (Rundlett et al., 1996, Proc. Natl. Acad. Sci. USA 93:14503-14508). The two classes share a −390 amino acid region of sequence similarity, comprising the deacetylase core, but are divergent outside this region. The histone deacetylase genes belong to an even larger superfamily (Leipe & Landsman, 1997, Nucleic Acids Res. 25:3693-3697) that contains the prokaryotic acetoin utilization proteins (AcuC; 28.1%, sequence identity to HDAC1), and the prokaryotic acetylpolyamine amidohydrolases (APAH; 15.0% sequence identity to HDAC1). The enzymatic activity of AcuC is not clear, but its disruption reduces the ability of B. subtilis to breakdown acetoin and utilize it as a carbon source. Grundy et al., 1993, Mol. Microbiol. 13:259-271. APAHs catalyze the deacetylation of polyamines by cleaving a non-peptide amide bond (reviewed in Leipe & Landsman, 1997, Nucleic Acids Res. 25:3693-3697).

[0010] It is useful to address the questions of how HDACs and HDAC-related proteins catalyze the deacetylation of histones and how the above-referenced compounds, particularly those compounds with antitumor activity, inhibit this activity in order to better understand the mechanism of inhibition of HDACs and to facilitate discovery of additional useful compounds which may inhibit this activity. To this end, the present invention has determined the three dimensional structure of a HDAC1-like protein from the thermophilic bacterium Aquifex aeolicus, herein after HDLP. The determination, of the nucleic acid coding sequence of HDLP was described by Deckert et al., 1998, Nature 392:353-358. The encoded 375 residue protein, whose sequence was determined from the nucleic acid encoding sequence, shares 35.2% amino acid sequence identity with HDAC1, deacetylates histones in vitro, and is inhibited by TSA, SAHA and several other HDAC inhibitors. The determination of the three-dimensional structure of HDLP is useful in the design, identification and screening of new HDAC family inhibitory compounds which are useful for the inhibition of cell growth both in vivo and in vitro.

SUMMARY OF THE INVENTION

[0011] In general, it is the object of the present invention to provide detailed three-dimensional structural information for a family of proteins known as histone deacetylases (HDAC), and particularly a homologue from the hyperthermophilic bacterium Aquifex aeolicus HDLP (histone deacetylase-like protein) which shares 35.2% sequence identity with human histone deacetylase (HDAC1). It is also an object of the present invention to provide three-dimensional structural information of an HDLP bound to an inhibitory compound.

[0012] In one embodiment of the invention, three-dimensional structure information is obtained from a crystal of wild-type HDLP (SEQ ID NO:1) (the nucleic acid encoding wild-type HDLP is SEQ ID,NO:2). In a further embodiment of the invention, three-dimensional information is obtained from a mutant HDLP comprising two mutations (1) cysteine 75 to a serine and (2) cysteine 77 to a serine (Cys75Ser/Cys77Ser double mutant; SEQ ID NO:3) (the nucleic acid encoding-HDLP Cys75Ser/Cys77Ser double mutant is SEQ ID NO:4). The HDLP mutant of the present invention facilitates the determination of three-dimensional structural information of HDLP bound to a zinc atom at its zinc atom-binding site.

[0013] In a preferred embodiment of the invention, the three-dimensional structural information is obtained from a co-crystal of a protein-inhibitor compound complex that comprises HDLP or HDLP Cys75Ser/Cys77Ser double mutant and trichostatin A (TSA)! In another preferred embodiment of the invention the three-dimensional structural information is obtained from a co-crystal of a protein-inhibitor compound complex that comprises HDLP or HDLP Cys75Ser/Cys77Ser double mutant and suberoylanilide hydrbxamic acid (SAHA). Any HDLP or HDLP-related protein (e.g. HDAC) inhibitor compound that may be co-crystallized with HDLP may be used to form a co-crystal of the present invention.

[0014] The protein crystals and protein-inhibitory complex co-crystals of the present invention diffract to a high resolution limit of at least equal to or greater than 4 angstrom (Å). In a preferred embodiment, the protein crystals and protein-inhibitory complex co-crystals of the present invention diffract to a high resolution limit of greater than 2.5 Å.

[0015] A crystal of the present invention may take a variety of forms, all of which are contemplated by the present invention. In a preferred embodiment, the crystal has a space group of C2 with one molecule in the asymmetric unit and with unit dimensions of a=51.4 Å, b=93.8 Å, c=78.7 Å and β=96.9° (see, e.g., Example 2, below). In another preferred embodiment, the crystal has a space group of P212121 with two molecules in the asymmetric unit and with unit dimensions of a=53.4 Å, b=94.4 Å, c=156.3 Å (see, e.g., Example 2, below). The HDLP structure comprises a parallel β sheet with a helices packing against both faces. At one end of the β sheet, the HDLP has a narrow, tube-like pocket formed by several well-ordered loops. The walls of the pocket are lined with hydrophobic residues and there is a zinc binding site and several polar side chains at the bottom of the pocket. The inhibitory compounds of the present invention bind in the pocket.

[0016] The three-dimensional structural information obtained from crystals of HDLP, HDLP Cys75Ser/Cys77Ser double mutant, HDLP Cys75Ser/Cys77Ser double mutant comprising a zinc atom, HDLP comprising an inhibitory compound such as TSA or SAHA, and HDLP Cys7SSer/Cys77Ser double mutant comprising an inhibitor compound such as TSA or SAM may be employed to solve the structure of any HDLP-related protein (e.g. HDAC) crystal, or any mutant HDLP-related protein and particularly any wild type or mutant of HDLP-related protein complexed with a ligand, including a substrate or inhibitor compound. If the crystals are in a different space group than the known structure, molecular replacement may be employed to solve the structure, or if the crystals are in the same space group, refinement and difference fourier methods may be employed. The structure of HDLP-related proteins (e.g. HDAC1) comprise no greater than a 2.0 Å root mean square deviation (rmsd) in the positions of the Cα atoms for at least 50% or more of the amino acids of the full-length HDLP structure.

[0017] The present invention also provides a nucleic acid molecule encoding an HDLP Cys75Ser/Cys77Ser double mutant having the amino acid sequence of SEQ ID NO:3 and the nucleic acid sequence of SEQ ID NO:4. It is also contemplated by the invention that mutations be made in HDLP-related proteins at cysteine residues, as with the Cys75Ser/Cys77Ser double mutant, in order to facilitate the determination of the structure of said proteins bound to a zinc atom. Additionally, the present invention provides expression vectors which comprise the nucleic acid molecule encoding an HDLP Cys75Ser/Cys77ser double mutant encoded by the sequence represented by SEQ ID NO:4 operatively linked to expression control sequences.

[0018] It is another object of the present invention to provide methods for the design, identification and screening of potential inhibitor compounds of the HDLP/HDAC family. In a preferred embodiment the method for the rational design, identification and screening of potential inhibitor compounds for HDLP and HDLP-related proteins (e.g. HDACs) comprising deacetylase activity comprises the steps of: (a) using a three-dimensional structure of an HDLP as defined by the atomic coordinates of the present invention; (b) employing said three-dimensional structure to design or select said potential inhibitor compound; (c) synthesizing and/or selecting said potential inhibitor; (d) contacting said potential inhibitor compound with said enzyme in the presence of acetylated substrate; and (e) determining the percent inhibition of deacetylase activity to determine the inhibitory activity of said potential inhibitor compound. In a further preferred embodiment, the binding properties of said rationally designed inhibitory compound may be determined by a method comprising the steps of: (a) forming a complex comprising said inhibitory compound and HDLP or a HDLP-related protein, (b) co-crystallizing said inhibitory compound-HDLP complex; (c) determining said three-dimensional structure of said co-crystal through molecular replacement or refinement and difference fourier with the molecular coordinates of HDLP as defined by the present invention; and (d) analyzing the three-dimensional structure to determine the binding characteristics of said potential. inhibitor compound.

[0019] It is a further object of the present invention to identify a defined class of HDLP/HDAC family inhibitor compounds. The HDLP/HDAC family inhibitor compounds of the present invention are represented by formula (I):

[0020] wherein X comprises a cap group which binds to at least one amino acid selected from the group consisting of proline and leucine; Y comprises an aliphatic chain group which binds to at least one amino acid selected from the group consisting of leucine, phenylalanine and glycine; and Z comprises and active site binding group which binds to at least one amino acid selected from the group consisting of aspartic acid, tyrosine and histidine and may further bind to a zinc atom.

BRIEF DESCRIPTION OF THE DRAWINGS

[0021]
FIG. 1 is a table listing the statistics from the X-ray crystallographic analysis of a HDLP crystal, a HDLP-TSA co-crystal, and a HDLP-SAHA co-crystal.

[0022]
FIG. 2 shows an alignment of various HDAC homologues with percent sequence identity depicted.

[0023]
FIG. 3 shows a graph indicating the histone deacetylase activity of HDLP and HDAC1 and the inhibition of HDLP and HDAC1 by the inhibitors TSA and HC-toxin.

[0024]
FIG. 4 shows (A & B) a schematic representation of the HDLP-Zn2+-TSA complex in two approximately orthogonal views, (C) a topology diagram of HDLP indicating the regions of homology with HDAC1, and (D) a close up schematic representation of the HDLP-Zn2+-SAHA complex.

[0025]
FIG. 5 shows (A) a schematic representation of a slice through a surface representation of HDLP with the pocket internal cavities and position of the β sheet indicated, (B) a schematic representation of a close-up view of the active site looking down into the, pocket in an orientation similar to FIG. 4B.

[0026]
FIG. 6 shows (A) a space-filling representation of TSA in the active site pocket, (B) a closeup stereo view of the structure of the HDLP-Z2+-TSA complex in a similar orientation to FIG. 4B, and (C) a schematic representation of the HDLP-TSA interactions.

[0027]
FIG. 7 shows (A) a schematic representation of the regions of homology shared between HDLP and HDAC1 in an orientation similar to that of FIG. 4A, and (B) a detailed schematic representation of the homology shared in the pocket and internal cavity between HDLP and HDAC1 in an orientation similar to that of FIG. 4B.

[0028]
FIG. 8 shows a schematic representation of the proposed catalytic mechanism for the deacetylation of acetylated lysine.

[0029]
FIG. 9 shows a schematic representation of a space filling diagram showing the conserved amino acids in the active site and nearby grooves.

[0030]
FIG. 10 is the nucleic acid sequence of HDLP from Aquifex aeolicus (SEQ ID NO. 2).

[0031]
FIG. 11 is the amino acid sequence of full length HDLP from Aquifex aeolicus (SEQ ID NO. 1).

[0032]
FIG. 12 is the nucleic acid sequence of the HDLP active site mutant Tyr297Phe (SEQ ID NO. 6).

[0033]
FIG. 13 is the amino acid sequence of the HDLP active site mutant Tyr297Phe (SEQ ID NO. 5).

[0034]
FIG. 14 is the nucleic acid sequence of a double mutant of HDLP from Aquifex aeolicus comprising a Cys75Ser and Cys77Ser mutation (SEQ ID NO. 4).

[0035]
FIG. 15 is the amino acid sequence of a double mutant of HDLP from Aquifex aeolicus comprising a Cys75Ser and Cys77Ser mutation (SEQ ID NO. 3).

[0036]
FIG. 16-1 to 16-49 lists the atomic structure coordinates for HDLP as derived by X-ray diffraction from a crystal of HDLP.

[0037]
FIG. 17-1 to 17-49 lists the atomic structure coordinates for HDLP Cys75Ser/Cys77Ser double mutant comprising a zinc atom in the active site as derived by X-ray diffraction from a crystal-of the HDLP Cys7Ser/Cys77Ser double mutant.

[0038]
FIG. 18-1 to 18-99 lists the atomic structure coordinates for HDLP Cys75Ser/Cys77Ser double mutant as derived by X-ray diffraction from a co-crystal of HDLP complexed with TSA.

[0039]
FIG. 19-1 to 19-48 lists the atomic structure coordinates for HDLP Cys75Ser/Cys77Ser double mutant as derived by X-ray diffraction from-a co-crystal of HDLP complexed with SAHA.

[0040] The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of preferred embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.

DETAILED DESCRIPTION OF THE INVENTION

[0041] The present invention provides crystals of a histone deacetylase (HDAC) homologue grown in the presence and absence of a compound capable of inhibiting the histone deacetylase activity of said HDAC homologue. As referred to herein, a HDAC homologue (as well as a HDLP-related protein) is any protein molecule having (a) greater than 15% sequence identity to over the 375 amino acid residues of HDLP; (b) having no more than twenty insertions or deletions for a total of no more than 100 amino acids; and (c) deacetylase activity. Sequence identity is calculated by the program DNAstar™ using the identity matrix weighing scheme clustal method (DNAstar program, Madison, Wis.).

[0042] A HDLP/HDAC inhibitor compound, as used herein, refers to any compound represented by Formula (I):

[0043] wherein X comprises a cap group which binds to at least one amino acid selected from the group consisting of tyrosine, proline and leucine; Y comprises an aliphatic chain group from about 5 to about 10 Å, preferably 7 Å, which binds to at least one amino acid selected from the group consisting of phenylalanine and glycine; and Z comprises a active site binding group which binds to at least one amino acid selected from the group consisting of aspartic acid, tyrosine and histidine and which may further bind to a zinc atom. The HDAC inhibitory compounds of the present invention can inhibit greater than 50% of the histone deacetylase activity of a HDAC homologue or a HDLP-related protein.

[0044] To grow the crystals of the present invention, the HDAC and HDAC-inhibitory compound complex are purified to greater than 80% total protein and more preferably purified to greater than 90% total protein. For expression and purification purposes, the full-length HDLP (Genbank accession number AE000719) may be subcloned from Aquifex aeolicus chromosomal DNA preparation by the polymerase chain reaction (PCR) and inserted into an expression vector.

[0045] A large number of vector-host systems known in the art may be used. Possible vectors include, but are not limited to, plasmids or modified viruses, but the vector system must be compatible with the host cell used. Examples of vectors include E. coli bacteriophages such as lambda derivatives, or plasmids such as pBR322 derivatives or pUC plasmid derivatives, e.g., pGEX vectors (Amersham-Pharmacia, Piscataway, N.J.), pET vectors (Novagen, Madison, Wis.), pmal-c vectors (Amersham-Pharmacia, Piscataway, N.J.), pFLAG vectors (Chiang and Roeder, 1993, Pept. Res. 6:62-64), baculovirus vectors (Invitrogen, Carlsbad, Calif.; Pharmingen, San Diego, Calif.), etc. The insertion into a cloning vector can, for example, be accomplished by ligating the DNA fragment into a cloning vector which has complementary cohesive termini, by blunt end ligation if no complementary cohesive termini are available or by through nucleotide linkers using techniques standard in the art. E.g., Ausubel et al. (eds.), Current Protocols in Molecular Biology, (1992). Recombinant vectors comprising the nucleic acid of interest may then be introduced into a host cell compatible with-the vector (e.g. E. coli, insect cells, mammalian cells, etc.) via transformation, transfection, infection, electroporation, etc. The nucleic acid may also be placed in a shuttle vector which may be cloned and propagated to large quantities in bacteria and then introduced into a eukaryotic cell host for expression. The vector systems of the present invention may provide expression control sequences and may allow for the expression of proteins in vitro.

[0046] In a preferred embodiment, the full length HDLP (SEQ ID NO:2) is subcloned from Aquifex aeolicus chromosomal DNA preparation into pGEX4T3 (Amersham-Pharmacia, Piscataway, N.J.). In order, to construct a double mutant comprising a CYS75Ser and Cys77Ser mutation (SEQ ID NO:4), and to construct the HDLP active site mutant Tyr297Phe (SEQ ID NO:5 and SEQ ID NO:6), PCR site directed mutagenesis may be employed with verification by DNA sequencing by methods known to those skilled in the art (see, e.g., Example 1 below). The mutants of the present invention may be subcloned into a suitable expression vector and introduced into a host cell for protein production, as described above. The HDLP nucleic acids of the present invention may be subcloned into an expression vector to create an expression construct such that the resultant HDLP molecule which is produced comprises a fusion protein wherein said fusion protein comprises a tag for ease of purification. As referred to herein, a “tag” is any additional amino acids which are provided in a protein either c-terminally, n-terminally or internally for the ease of purification, for the improvement of production or for any other purpose which may facilitate the goals of the present invention (e.g. to achieve higher levels of production and/or purification). Such tags include tags known to those skilled in the art to be useful in purification such as, but not limited to, his tag, glutathione-s-transferase tag, flag tag, mbp (maltose binding protein) tag, etc. In a preferred embodiment, the wild-type and mutant HDLPs of the present invention are tagged with glutathione-s-transferase (see Example 1 below). In another preferred embodiment, HDAC1 is flag tagged (see Example 1 below). Such tagged proteins may also be engineered to comprise a cleavage site, such as a thrombin, enterokinase or factor X cleavage site, for ease of removal of the tag before, during or after purification. Vector systems which provide a tag and a cleavage site for removal of the tag are particularly useful to make the expression constructs of the present invention.

[0047] The tagged HDLPs and HDACs of the present invention may be purified by immuno-affinity or conventional chromatography, including but not limited to, chromatography employing the following: glutathione-sepharose™) (Amersham-Pharmacia, Piscataway, N.J.) or an equivalent resin, nickel or cobalt-purification resins, anion exchange chromatography, cation exchange chromatography, hydrophobic resins, gel filtration, antiflag epitope resin, reverse phase chromatography, etc. After purification, the HDLP and HDLP-inhibitor compound complex may be concentrated to greater than 1 mg/ml for crystallization purposes. In a preferred embodiment HDLP and HDLP-inhibitor complexes are concentrated to greater than 10 mg/ml for crystallization and in a particularly preferred embodiment, HDLP and HDLP-inhibitor complexes are concentrated to greater than 20 mg/ml.

[0048] In order to determine whether the purified HDLPs of the present invention demonstrate histone deacetylase activity, the purified HDLPs and also any HDLP-related protein may be assayed by any method known to those skilled in the art for the determination of said activity. In a preferred embodiment, the purified HDLPs of the present invention are incubated in the presence of [3H] acetyl-labeled histone substrate (Carmen et-al., 1996, J. Biol. Chem. 271:15837-15844) in a buffer suitable for detection of histone deacetylase activity (see Example 3 below); stopping the reaction; extracting the released acetate and measuring said released acetate, as described by Henzel et al. (J. Biol. Chem. 266:21936-21942 (1991); Example 3 below). In a preferred embodiment, the HDLPs of the present invention are inclubated in the presence of ZnCl2 in order to obtain histone deacetylase activity therefrom (Example 3 below).

[0049] In another embodiment, the crystals of the present invention comprise purified wild-type HDLP (SEQ ID NO:1) and are grown at room temperature by the hanging-drop vapor-diffusion method from a crystallization solution comprising one or more precipitants selected from the group consisting of isopropanol, polyethylene glycol, and tert butanol (see Example 2 below). The crystallization solution may further comprise one or more salts including salts selected from the group consisting of NaCl and KCl, and one or more buffers including buffers selected from the group consisting of Tris (tris(hydroxymethyl)aminomethane and bis-tris propane-Cl (1, 3-bis[tris(hydroxymethyl)methyl-amino] propane) (see Example 2 below). The pH of the crystallization solution is preferably between pH 5 to 9, although other pH values are also contemplated by the present invention (see Example 2 below).

[0050] Any crystallization technique known to those skilled in the art may be employed to obtain the crystals of the present invention, including, but not limited to, batch crystallization, vapor diffusion (either by sitting drop or hanging drop) and micro dialysis. Seeding of the crystals in some instances may be required to obtain X-ray quality crystals. Standard micro and/or macro seeding of crystals may therefore be used.

[0051] The crystals of the present invention may form in the space group C2 with one molecule in the asymmetric unit and with unit dimensions of a=51.4 Å, b=93.8 Å, c=78.7 Å and β=96.9° (see Example 2 below). The crystals of the present invention may also form in the space group P212121 with two molecules in the asymmetric unit and with unit dimensions of a=53.4 Å, b=94.4 Å, c=156.3 Å (see Example 2 below). However, the present invention contemplates crystals which form in any space group including, but not limited to, C2, P21, P212121, P3121, P432121, and C2221. The crystals diffract to a resolution greater than 4 Å, preferably greater than 2.5 Å.

[0052] To collect diffraction data from the crystals of the present invention, the crystals may be flash-frozen in the crystallization buffer employed for the growth of said crystals, however with preferably higher precipitant concentration (see, e.g., Example 2 below). For example, but not by way of limitation, if the precipitant used was 28% PEG 1500, the crystals may be flash frozen in the same crystallization solution employed for said crystal growth wherein the concentration of the precipitant is increased to 35% (see Example 2 below). If the precipitant is not a sufficient cryoprotectant (i.e. a glass is not formed upon flash-freezing), cryoprotectants (e.g. glycerol, low molecular weight PEGs, alcohols, etc) may be added to the solution in order to achieve glass formation upon flash-freezing, providing the cryoprotectant is compatible with preserving the integrity of the crystals. The flash-frozen crystals are maintained at a temperature of less than −110° C. and preferably less than −150° C. during the collection of the crystallographic data by X-ray diffraction. The X-ray diffraction data may be processed with DENZO and SCALEPACK (Otwinowski & Minor, 1997, Method Ensemble. 276:307-326) but any method known to those skilled in the art may be used to process the X-ray diffraction data.

[0053] In order to determine the atomic structure of HDLP according to the present invention, multiple isomorphous replacement (MIR) analysis, model building and refinement may be performed. For MIR analysis, the crystals may be soaked in heavy-atoms to produce heavy atom derivatives necessary for MIR analysis. As used herein, heavy atom derivative or derivitization refers to the method of producing a chemically modified form of a protein or protein complex crystal wherein said protein is specifically bound to a heavy atom within the crystal. In practice a crystal is soaked in a solution containing heavy metal atoms or salts, or organometallic compounds, e.g., lead chloride, gold cyanide, thimerosal, lead acetate, uranyl acetate, mercury chloride, gold chloride, etc, which can diffuse through the crystal and bind specifically to the protein. The location(s) of the bound heavy metal atom(s) or salts can be determined by X-ray diffraction analysis of the soaked crystal. This information is used to generate MIR phase information which is used to construct the three-dimensional structure of the crystallized HDLPs and HDLP-related proteins of the, present invention. In a preferred embodiment, the heavy atoms comprise thimerosal, KAu(CN)2 and Pb(Me)3OAc (see Example 2 below). The MIR phases may be calculated by any program known to those skilled in the art and preferably with the program MLPHARE (The CCP4 suite: Programs for computational crystallography, 1994, Acta Crystallogr. D. 50:760-763) and may also use the anomalous diffraction signal from the thimerosal derivative. In a preferred embodiment, the MIR phases were calculated at 2.5 Å and have a mean figure of merit of 0.55 (see FIG. 19 and Example 2 below). The phases may be improved where necessary by solvent flattening by methods known to those skilled in the art including, but not limited to, through the use of the program DM (The CCP4 suite: Programs for computational crystallography, 1994, Acta Crystallogr. D 50:760-763).

[0054] Thereafter, an initial model of the three-dimensional structure may be built using the program O (Jones et al., 1991, Acta Crystallogr. A 47:110-119). The interpretation and building of the structure may be further facilitated by use of the program CNS (Brunger et al., 1998, Acta Crystalloqr. D 54:9.05-921).

[0055] For the determination of the HDLP-inhibitor compound complex structure, if the space group of the HDLP-inhibitor compound complex crystal is different, molecular replacement may be employed using a known structure of apo-HDLP ( as referred to herein, apo-HDLP or apo-HDAC is-the enzyme which is not complexed with an inhibitor compound) or any known HDLP/inhibitor complex structure whose structure may be determined as described above and below in Example 2. If the space group of the HDLP-inhibitor compound crystals is the same, then rigid body refinement and difference fourier may be employed to solve the structure using a known structure of apo-HDLP (as referred to herein, apo-HDLP or apo-HDAC is the enzyme which is not complexed with an inhibitor compound) or any known HDLP/inhibitor complex structure.

[0056] The term “molecular replacement” refers to a method that involves generating a preliminary model of the three-dimensional structure of the HDLP crystals of the present invention whose structure coordinates are unknown prior to the employment of molecular replacement. Molecular replacement is achieved by orienting and positioning a molecule whose structure coordinates are known (in this case the previously determined apo-HDLP) within the unit cell as defined by the X-ray diffraction pattern obtained from an HDLP or HDLP-related protein crystal whose structure is unknown so as to best; account for the observed diffraction pattern of the unknown crystal. Phases can then be calculated from this model and combined with the observed amplitudes to give an approximate Fourier synthesis of the structure whose coordinates are unknown. This in turn can be subject to any of several forms of refinement to provide a final, accurate structure.

[0057] Any method known to the skilled artisan may be employed to determine the structure by molecular replacement. For example, the program AMORE (The CCP4 suite: Programs for computational crystallography, 1994, Acta Czystallogr. D. 50:760-763) may be employed to determine the structure of an unknown histone deacetylase +/− an inhibitor by molecular replacement using the apo-HDLP coordinates (FIG. 16). For the structure determination of the inhibitory compound TSA, the structure of TSA was obtained from the Cambridge Structural Database (Refcode TRCHST, <<http://www.ccdc.cam.ac.uk >>) may be employed to define the stereochemical restraints used in the refinement with the program CNS (Brunger et al., 1998, Acta Crystallogr. D 54:905-921).

[0058] The three-dimensional structural information and the atomic coordinates associated with said structural information of HDLP are useful for solving the structure of crystallized proteins which belong to the HDAC family by molecular replacement. Similarly, any structure of a crystallized protein which is thought to be similar in structure based on function or sequence similarity or identity to HDLP may be solved by molecular replacement with the HDLP structural information of the present invention. The structure of HDLP-related proteins as determined by molecular replacement as described above and in Example 2 below, comprise a root mean square deviation (rmsd) of no greater than 2.0 Å in the positions of Cα atoms for at least 50% or more of the amino acids of the structure over the 375 residues of full length HDLP. Such a rmsd may be expected based on the amino acid sequence identity. Chothia & Lesk, 1986, Embo J. 5:823-826.

[0059] The refined three-dimensional HDLP structures of the present invention, specifically apo-HDLP, Cys75Ser/Cys77Ser double mutant HDLP comprising a zinc atom in the active site, HDLP/TSA complex comprising a zinc atom in the-active site, and HDLP/SAHA complex comprising a zinc atom in the active site, are represented by the atomic coordinates set forth in FIGS. 16 to 19 respectively. The refined model for apo-HDLP comprising amino acids 1-375 consists of wild-type HDLP residues 2 to 373 with residues 1, 374 and 375 not modeled and presumed disordered and was determined to a resolution of 1.8 Å. Similarly, the refined model for Cys75Ser/Cys77Ser double mutant HDLP comprising a zinc atom in the active site also consists of residues 2 to 373 with residues 1, 374 and 375 not modeled and presumed disordered and was determined to a resolution of 2.0 Å. The refined model for the HDLP/TSA complex comprising a zinc atom in the active site consists of the Cys75Ser/Cys77Ser double mutant HDLP residues 2 to 373 with residues 1, 374 and 375 not modeled and presumed disordered, has TSA in the binding pocket and was determined to a resolution of 2.1 Å. The HDLP/SAHA complex is similar to the HDLP/TSA complex but has SARA in the binding pocket and was determined to a resolution of 2.5 Å.

[0060] For the purposes of further describing the structure of HDLP and HDLP-related proteins, including, but not limited to, HDACs, from the data obtained from the HDLP crystals of the present invention, the definition of the following terms is provided:

[0061] The term “β sheet” refers to two or more polypeptide chains (or β strands) that run alongside each other and are linked in a regular manner by hydrogen bonds between the main chain C═O and N—H groups. Therefore all hydrogen bonds in a beta-sheet are between different segments of polypeptide. Most β-sheets in proteins are all-parallel (protein interiors) or all-antiparallel (one side facing solvent, the other facing the hydrophobic core), Hydrogen bonds in antiparallel sheets are perpendicular to the chain direction and spaced evenly as pairs between strands. Hydrogen bonds in parallel sheets are slanted with respect to the chain direction and spaced evenly between strands.

[0062] The term “α helix” refers to the most abundant helical conformation found in globular proteins. The average length of an α helix is 10 residues. In an α helix, all amide protons point toward the N-terminus and all carbonyl oxygens point toward the C-terminus. The repeating nature of the phi, psi pairs ensure this orientation. Hydrogen bonds within an α helix also display a repeating pattern in which the backbone C═O of residue X (wherein X refers to any amino acid) hydrogen bonds to the backbone HN of residue X+4. The α helix is a coiled structure characterized by 3.6 residues per turn, and translating along its axis 1.5 Å per amino acid. Thus the pitch is 3.6×1.5 or 5.4 Å. The screw sense of alpha helices is always right-handed.

[0063] The term “loop” refers to any other conformation of amino acids (i.e. not a helix, strand or sheet). Additionally, a loop may contain bond interactions between amino acid side chains, but not in a repetitive, regular fashion.

[0064] Amino acid residues in peptides shall herein after be abbreviated as follows: Phenylalanine is Phe or F; Leucine is Leu or L; Isoleucine is Ile or I; Methionine is Met or M; Valine is Val or V; Serine is Ser or S; Proline is Pro or P; Threonine is Thr or T; Alanine is Ala or A; Tyrosine is Tyr or Y; Histidine is His or H; Glutamine is Gln or Q; Asparagine is Asn or N; Lysine is Lys or K; Aspartic Acid is Asp or D; Glutamic Acid is Glu or E; Cysteine is Cys or C; Tryptophan is Trp or W; Arginine is Arg or R; and Glycine is Gly or G. For further description of amino acids, please refer to Proteins: Structure and Molecular Properties by Creighton, T. E., W. H. Freeman & Co., New York 1983.

[0065] The term “positively charged amino acid” refers to any amino acid having a positively charged side chain under normal physiological conditions. Examples of positively charged amino acids are Arg, Lys and His. The term “negatively charged amino acid” refers to any amino acid having a negatively charged side chain under normal physiological conditions. Examples of negatively charged amino acids are Asp and Glu. The term “hydrophobic amino acid” refers to any amino acid having an uncharged, nonpolar side chain that is relatively insoluble in water. Examples of hydrophobic amino acids are Ala, Leu, Ile, Gly, Val, Pro, Phe, Trp and Met. The term “hydrophilic amino acid” refers to any amino acid having an uncharged, polar side chain that is relatively soluble in water. Examples of hydrophilic amino acids are Ser, Thr, Tyr, Asp, Gln, and Cys. The term “aromatic amino acid” refers to any amino acid comprising a ring structure. Examples of aromatic amino acids are His, S Phe, Trp and Tyr.

[0066] The term “charge relay system” refers to a His-Asp arrangement as described by Fersht & Sperling, 1973, J. Mol. Biol. 74:137-149; Blow et al., 1969, Nature 221:337-340.

[0067] The information obtained from the three-dimensional structures of the present invention reveal that HDLP has a single-domain structure that belongs to the open α/β class of folds (see, e.g., Branden, 1980, Q. Rev. Biophys. 13:317-38). Two orthogonal views of the overall three-dimensional structure of HDLP are depicted in FIG. 4A and 4B. The HDLP structure has a central eight-stranded parallel β sheet (strands arranged as β2-β1-β3-β8-β7-β4-β5-β6), and sixteen α helices (labeled α1 through α16 respectively). See FIG. 4C. Four of the helices pack on either face of the β sheet (α7, α8, α9, α10 and α11, α12, α13, α14) forming the core α/β structure characteristic of this class of folds. Most of the remaining eight, helices are positioned near one side of the β sheet, near stands β2-β1-β3-β8. Large, well defined loops, (Loops L1-L7; FIG. 4C) originate from the C-terminal ends of the β-strands. The extra helices and the large L1-L7 loops are associated with a significant extension of the structure beyond the core α/β motif. This extension of the structure gives rise to two prominent architectural features: a deep, narrow pocket and an internal cavity adjacent to the pocket. These two architectural features comprise the active site (see FIG. 5A). The structure of HDLP-related proteins (e.g. HDA˜s) may also comprise the conserved α/β structure characteristic.

[0068] The term “active site” comprises any or all of the following sites in HDLP, the substrate binding site, the site where the cleavage of an acetyl group from a substrate occurs or the site where an inhibitor of the HDAC family or, more particularly, HDLP binds. The active site, as referred to herein, comprises Asp166, Asp258, His170, Tyr297, His131, His132, Asp168, Asp173, Phe141, Phe198, Leu265, Pro22 and Gly140, and also a metal bound at the bottom of the pocket by Asp173, Asp168 and His defined by the coordinates listed in FIGS. 16 to 19 with an rmsd of 2.0 Å. The metal which binds at the bottom of the pocket will be a divalent cation selected from the group consisting of zinc, cobalt or manganese.

[0069] The deep narrow pocket has a tube-like shape with a depth of ˜11 Å. The pocket opening constricts half way down to ˜4.5 by 5.5 Å, and becomes wider at the bottom (see FIG. 5A). The pocket and its immediate surroundings are made up of loops L1 through L7.

[0070] The walls of the pocket are covered with side chains of hydrophobic and aromatic residues (Pro22, Tyr91near the entrance; and Gly140, Phe141, Phe 198, Leu265 and Tyr297 further down; FIG. 5B). For numbering of amino acids please refer to SEQ ID NO:1. of particular interest are Phe141 and Phe198, whose phenyl groups face each other in parallel at a distance of 7.5 Å, marking the most slender portion of the pocket (see FIG. 5B). Of particular interest is that only one pocket residue differs in HDAC1 when the sequences are aligned (alignment may be accomplished using DNAstar™ MegAlign™ program, Madison, Wis.), this residue is Glu98 of HDAC1 which is Tyr91 in HDLP. The structure reveals that this residue in HDLP is mostly solvent exposed.

[0071] Near the bottom of the pocket of the active site at its narrowest point, is located a zinc ion (see FIG. 6A). In order to obtain the zinc in the structure, the crystals may be soaked in zinc (e.g. ZnCl2) or co-crystalized in the presence of zinc, The zinc ion is. coordinated by Asp168 (Oδ1, 2.1 Å) His170 (Nδ1, 2.1 Å), Asp258 (Oδ1, 1.9 Å) and a water molecule (2.5 Å). See FIG. 5B and 6B. The amino acid residues that coordinate zinc are arranged in a tetrahedral geometry, but the position of the water molecule, which is also hydrogen bonded to His131, deviates from this geometry by ˜25°.

[0072] In addition to the zinc ligands, the bottom of the pocket contains two histidine (His131 and His132), two aspartic acids (Asp166 and Asp173) and a tyrosine (Tyr297). See FIG.5B and 10B. Each of the histidines makes a hydrogen bond through its Nδ1 to an aspartic acid carboxylate oxygen, with the oxygen located in the plane of the imidizole ring (FIG. 5B). This His-Asp arrangement is characteristic of the charge relay system present in the active sites of serine proteases, where it serves to polarize the imidizole Ne and increase its basicity. Fersht & Sperling, 1973, J. Mol. Biol. 74:137-149; Blow et al., 1969, Nature 2:337-340.

[0073] The Asp166-His131 charge pair relay (hereafter referred to as “buried charged relay”) is positioned even deeper in the pocket and more buried compared to the Asp173-His132 charge relay (hereafter referred to as “exposed charge relay”) which is partially solvent exposed. The buried charge relay makes a hydrogen bond (2.6 Å) to the zinc-bound water molecule referred to above, and this hydrogen bond could contribute to the deviation of the water-zinc coordination from ideal geometry (FIG. 5B). The exposed charge relay is directed to a point ˜2.5 Å away from the water molecule and closer to the surface.

[0074] Tyr 297 is positioned next to the zinc, opposite from where the two charge relay systems are located. The Tyr hydroxyl group lies 4.4 Å away from the zinc atom and has no interactions with the rest of the protein (FIG. 5B). Next to Tyr297, there is an opening in the pocket wall, which leads to the adjacent internal cavity.

[0075] The floor of the internal cavity is made up of portions of the L3 and L7 loops as they emerge from the β strands, and the roof is made up by the α1-L1-α2 segment. The L1 loop appears more flexible than other loops in the structure. This may allow the transient exchange of the cavity contents with the bulk solvent.

[0076] The cavity is lined primarily with hydrophobic residues and is particularly rich in glycine residues (Ala127, Gly128, Gly129,Met130, and Phe141of L3; Gly293, Gly294, Gly295 and Gly296 of L7; and Tyrl7, Pro22 and Leu23 of L1). There are only two charged residues in the cavity,(Arg27 and His 21) and these are contributed by the L1 loop.

[0077] The cavity may provide space for the diffusion of the acetate product away from the catalytic center, which may otherwise be crowded and shielded during deacetylation from the solvent when the substrate is bound. Such a role for the cavity is supported by the observation that the cavity contains three water and two isopropanol molecules (from the crystallization buffer) in the 1.8 Å apo-protein structure. The cavity may also bind another cofactor, in addition to zinc, for the facilitation of the enzymatic activity of the HDLP. A proposed catalytic mechanism for deacetylation is provided in FIG. 8.

[0078] The structure of HDLP as defined by the present invention, in conjunction with the HDAC1 sequence homology, shows that the 375-amino acid HDLP protein corresponds to the histone deacetylase catalytic core which is conserved across the HDAC family (see FIG. 2). The 35.2% HDLP-HDAC1-sequence identity predicts structural similarity with a rmsd in Cα positions of ˜1.5 Å. Chothia and Lesk describe the relation between the divergence of sequence and the structure of proteins in Embo J. 5:823-826 (1986). The 40-residue C-terminus of HDLP is likely to have a divergent structure since this region has lower homology to HDAC1, although the α16 helix in this region is part of the conserved open α/β core fold and HDAC1 is likely to comprise a similar helix. However divergent this C-terminal region may be, this region is outside the active site and is likely to not effect the structure of the active site. Beyond the C-terminus of the histone deacetylase catalytic core, HDAC family members are divergent in length and sequence. In the HDAC family, this region (amino acid residues ˜390-482) is highly polar, populated with acidic residues, and is likely to be flexible or loosely folded.

[0079] The HDLP-HDAC homology maps primarily to the hydrophobic core and to the L1-L7 loops, with portions of the loops that make up the pocket and adjacent cavity having the highest level of amino acid residue sequence conservation (FIG. 9A and 9B). specifically, all of the polar residues in the active site (the zinc ligands, the two charge relay systems, and Tyr297) and the hydrophobic residues that make up the walls of the pocket (Gly140, Phe141, Phe198 and Leu265) are identical. Among the residues that make up the internal cavity, the ones closest to the active site are either identical or conservatively substituted (for example, Leu23 → Met and Metl30 → Leu). Surface residues around the pocket are conserved to a lesser extent, but are still above 35% average sequence identity.

[0080] The information obtained from the inhibitor-bound HDLP complex crystal structures of the present invention reveal detailed information which is useful in the design, isolation, screening and determination of potential inhibitor compounds which may inhibit HDLP/HDAC family members. As described above, the HDLP structure consists of a parallel β sheet with a helices packing against both faces (FIG. 4A, 4B, and 4C). At one end of the β sheet, 7 loops (L1-L7) form a narrow, tube-like pocket which are lined with hydrophobic residues and which comprise a zinc binding site, several polar side chains, including two Asp-His charge relay systems. Mutation of the zinc ligands and other polar residues at the pocket bottom reduces or eliminates the catalytic activity.

[0081] The present inventors found that mutation at the Tyr297Phe site reduced activity. See also, Hassig et al., 1998, Proc. Natl. Acad. Sci. USA 95:3519-3524; Kadosh & Struhl, 1998, Genes Dev. 12:797-805. The elimination of activity by mutation of these residues indicates that this region is the enzyme active site. Adjacent to the active site, there is an internal cavity that may provide space for the diffusion of the acetate reaction product. Homology at the active site between HDLP and HDAC1, as described above, indicates that they share structural and functional homology.

[0082] The inhibitor compound, trichostatin A (TSA) (Tsuji et al., 1976, J. Antibiotics 29:1-6) binds HDLP by inserting its long aliphatic chain, which has a hydroxamic acid-group at one end, into the pocket (FIG. 6A, 6B and 6C). The aliphatic chain makes multiple contacts in the well-like, hydrophobic portion of the pocket. The hydroxamic acid reaches the polar bottom of the pocket, where it coordinates the zinc in a bidentate fashion and also forms hydrogen bonds with the polar residues in the active site, including the two charge relay system histidines. The aromatic dimethylamino-phenyl group at the other end of the TSA chain makes contacts at the pocket entrance and serves to cap it. The amino acid residues of HDLP which contact TSA are conserved in HDAC, indicating that TSA binds and inhibits HDAC in a similar fashion to HDLP.

[0083] In the complex, the hydroxamic acid, most of the aliphatic chain and part of the dimethylamino-phenyl group of TSA are buried (60% of TSA's surface area; FIG. 6A). The hydroxamic acid group binds the zinc in a bidentite fashion forming bonds through its carbonyl (2.4 Å) and hydroxyl groups (2.2 Å) resulting in a penta-coordinated Zn2+(FIG. 6B and 6C). The hydroxamic acid hydroxyl group replaces the water molecule that binds to the zinc in the apo-HDLP structure described above. The hydroxamic acid also hydrogen bonds with both charge relay system histidines (hydroxyl oxygen to His132Ne2, 2.8 Å; and nitrogen to His132 Ne2, 2.8 Å), and the Tyr297 hydroxyl group (2.4 Å; FIG. 6B and 6C).

[0084] The 5-carbon long branched alkene chain of TSA fits snugly in the narrow portion of the pocket making multiple van der Waals contacts with all of the hydrophobic groups lining the pocket (FIG. 6B and 6C). Near its center, the chain contains a methyl substituted carbon-carbon double bond which is sandwiched between the phenyl groups of the Phe141 and Phe98 at the tightest point of the pocket (FIG. 6A and 6B). The length of the alkene chain appears optimal for spanning the length of the pocket, and allowing contacts both at the bottom and at the entrance of the pocket, although, the cap group of Formula (I) may provide length to span the pocket allowing for a shorter alkene chain (aliphatic chain).

[0085] At the entrance of the pocket, one face of the planar structure formed by the dimethylamino-phenyl and adjacent carbonyl groups of TSA makes contacts at the rim of the pocket (Pro22, Tyr91, Phe141; FIG. 6B and 6C). This packing is facilitated by the roughly 110° angle in the overall structure of TSA at the junction of the aliphatic chain and the dimethylamino-phenyl group (occurring at the Sp3 hybridized C8 carbon). Upon TSA binding, the side chain of Tyr91, which is mostly solvent exposed, changes conformation to make space for the dimethylamino-phenyl group. This is the only change near the active site observed upon TSA binding.

[0086] The hydroxamic acid group is a common motif in zinc metalloprotease inhibitors. See U.S. Pat. Nos. 5,919,940 and 5,917,090; See also, Grams et al., 1995, Biochemistry 34:14012-14020; Lovejoy et al., 1999, Nat. Struct. Biol. 6:217-221; and Holmes & Matthews, 1981, Biochemistzy 20:6912-6920. Like TSA, these inhibitors also coordinate the active site zinc in a bidentate fashion using their hydroxamate hyroxyl and carbonyl oxygens, replace the nucleophilic water molecule with their hydroxamate hydroxyl groups and form hydrogen bonds to the general base (Grams et al., 1995, Biochemistry 34:14012-14020; Lovejoy et al., 1999, Nat. Struct. Biol. 6:217-221; and Holmes & Matthews, 19l8, Biochemistry 20:6912-6920).

[0087] SAHA, which has a ˜30-fold weaker inhibitory activity than TSA (Richon et al., 1998, Proc. Natl. Acad. Sci. USA 95:3003-3007), binds HDLP similarly to TSA (see, e.g., FIG. 4D). The SARA hydroxamic acid group makes the same contacts to the zinc and active site residues, and the importance of these interactions is underscored by the loss of activity of SAHA derivatives lacking the hydroxamic group (Richon et al., 1998, Proc. Natl. Acad. Sci. USA 95:3003-3007)

[0088] The six-carbon long aliphatic chain of SAHA packs in the tube-like hydrophobic portion of the pocket. Compared to TSA however, SAHA's aliphatic chain packs less snugly and makes fewer van der waals contacts, in part, because SAHA lacks TSA's C15 methyl group branch. SAHA also lacks TSA's double bonds in this region, and this may lead to increased flexibility of the aliphatic chain. The cap group of SAHA consists of a phenyl-amino ketone group. In the crystal structure, the phenyl group has weak electron density, suggesting that it does not pack as well as the cap group of TSA. This may be due to the larger separation between the hydroxamic and cap groups of SAHA compared to TSA (compare TSA, Formula (II) and SAHA, Formula (III), below).

[0089] The determination of the structure of HDLP and HDLP bound to an inhibitory compound has enabled, for the first time, the identification of the active site of HDLP and of related HDLP proteins, such as proteins belonging to the HDAC family.

[0090] The three-dimensional structural information and the atomic coordinates associated with said structural information of HDLP bound to an inhibitory compound is useful in rational drug design providing for a method of identifying inhibitory compounds which bind to and inhibit the enzymatic activity of HDLP, HDAC family proteins and other histone deacetylase-like proteins related to HDLP. Said method for identifying said potential inhibitor for an enzyme comprising deacetylase activity comprises the steps of (a) using a three-dimensional structure of HDLP as defined by its atomic coordinates listed in FIG. 16 to 19; (b) employing said three-dimensional structure to design or select said potential inhibitor; (c) synthesizing said potential inhibitor; (d) contacting said potential inhibitor with said enzyme in the presence of an acetylated substrate; and (e) determining the ability of said inhibitor to inhibit said deacetylase activity.

[0091] The potential HDLP and HDLP-related (e.g. HDAC) inhibitors identified by the method of the present invention are represented by formula (I)

[0092] wherein X comprises a cap group which binds to at least one amino acid selected from the group consisting of proline and leucine; Y comprises an aliphatic chain group which binds to at least one amino acid selected from the group consisting of leucine, phenylalanine and glycine; and Z comprises an active site binding group which binds to at least one amino acid selected from the group consisting of aspartic acid, tyrosine and histidine and wherein Z may further bind to a zinc atom and with the provision that the compound of Formula (I) is not TSA, trapoxin, SAHA, SAHA derivatives described in U.S. Pat. Nos. 5,608,108; 5,700,811; 5,773,474; 5840,960 and 5,668,179.

[0093] The present invention permits the use of molecular design techniques to design, identify and synthesize chemical entities and compounds, including inhibitory compounds, capable of binding to the active site of HDLP and HDLP-related proteins. The atomic coordinates of apo-HDLP and inhibitor-bound HDLP may be used in conjunction with computer modeling using a docking program such as GRAM, DOCK, HOOK or AUTODOCK (Dunbrack et al., 1997, Folding& Design 2:27-42) to identify potential inhibitors of HDLP and HDLP-related proteins (e.g. HDAC1). This procedure can include computer fitting of potential inhibitors to the active site of HDLP to ascertain how well the shape and the chemical structure of the potential inhibitor will complement the active site or to compare the potential inhibitors with-the binding of TSA or-SAHA in the active site. See Bugg et al, 1998, Scientific American December:92-98; West et-al., 1995, TIPS 16:67-74. The potential inhibitors designed by modeling with a docking program conform to the general formula (I) as described above. Computer programs may also be-employed to estimate the attraction, repulsion and stearic hindrance of the HDLP and potential inhibitor compound. Generally, the tighter the fit, the lower the stearic hindrances, the greater the attractive forces, and the greater the specificity which are important features for a specific inhibitory compound which is more likely to interact with HDLP and HDLP-related proteins rather than other classes of proteins. These features are desired particularly where the inhibitory compound is a potential antitumor drug.

[0094] The compounds of the present invention may also be designed by visually inspecting the three-dimensional structure to determine more effective deacetylase inhibitors. This type of modeling may be referred to as “manual” drug design. Manual drug design may employ visual inspection and analysis using a graphics visualization program such as “0” (Jones, T. A., Zhou, J. Y., Cowan, S. W., and Kjeldgaard, M., Improved method for building protein models in electron density maps and the location of errors in these models, Acta Crystallog., A47, 110-119.

[0095] Initially potential inhibitor compounds can be selected for their structural similarity to the X, Y and Z constituents of formula (I) by manual drug design. The structural analog thus designed can then be modified by computer modeling programs to better define the most likely effective candidates. Reduction of the number of potential candidates is useful as it may not be possible to synthesize and screen a countless number of variations compounds that may have some similarity to known inhibitory molecules. Such analysis has been shown effective in the development of HIV protease inhibitors (Lam et al., 1994, Science 263:380-384; Wlodawer et al., 1993, Ann. Rev. Biochem. 62:543-585; Appelt, 1993 Perspectives in Drug Discovery and Design 1:23-48; Erickson, 1993, Perspectives in Drug Discovery and Design 1:109-128. Alternatively, random screening of an small molecule library could lead to potential inhibitors whose inhibitory activity may then be analyzed by computer modeling as described above to better determine their effectiveness as inhibitors.

[0096] The compounds designed using the information of the present invention may be competitive or noncompetitive inhibitors. These designed inhibitors may bind to all or a portion of the active site of HDLP and may be more potent, more specific, less toxic and more effective than known inhibitors for HDLP and HDLP-related proteins, and particularly HDACs. The designed inhibitors may also be less potent but have a longer half life in vivo and/or in vitro and therefore be more effective at inhibiting histone deacetylase activity in vivo and/or in vivo for prolonged periods of time. Said designed inhibitors are useful to inhibit the histone deacetylase activity of HDLP and HDLP-related proteins (e.g. HDAC1), to inhibit cell growth in vitro and in viva and may, be particularly useful as antitumor agents.

[0097] The present invention also permits the use of molecular design techniques to computationally screen small molecule data bases for chemical entities or compounds that can bind to HDLP in a manner analogous to the TSA and SAHA as defined by the structure of the present invention. Such computational screening may identify various groups which may be defined as “X”, “Y” or “Z” of formula (I) above and may be employed to synthesize the potential inhibitors of the present invention comprising formula (I) Such potential inhibitors may be assayed for histone deacetylase inhibitory activity in a histone deacetylase activity assay (see Example 3 below), may be co-crystallized with HDLP to determine the binding characteristics through X-ray crystallography techniques defined above (e.g. said co-crystal structure may be determined by molecular replacement to assess the binding characteristics of said potential inhibitor), or may be assessed based on binding activity by incubating said potential inhibitor with said HDLP, performing gel filtration to separate any free potential inhibitor to HDLP-bound inhibitor, and determining the amount of histone deacetylase activity of the inhibitor-bound HDLP. To measure binding constants (e.g., Kd), methods known to those in the art may be employed such as Biacore™ analysis, isothermal titration calorimetry, Elisa with a known drug on the plate to show competitive binding, or by a deacetylase activity assay.

[0098] The design of potential inhibitors of the present invention is further facilitated by reference to FIG. 9, which is a surface representation figure that depicts the surface grooves. Analysis of such grooves gives insight into the constituents of the cap group of formula (I). The surface grooves are labeled groove A, groove A′, groove B and groove C, into which additional cap groups may bind. The structure of HDLP bound to either TSA or SAHA shows that the cap groups of TSA and SAHA bind in groove A. By analysis of the amino acid sequence identity of HDLP and HDACs, Groove A is well conserved in HDACs, has a significant hydrophobic component, appears deep enough to allow for significant interactions and is also the largest of the four grooves. In addition to the dimethylamino phenyl group of the TSA, the A groove can fit approximately 200 daltons worth of groups (e.g. groove A could accommodate a naphthalene-like group after an appropriate spacer, etc.). Groove A, as referred to herein, is characterized by the following conserved residues of HDLP: His 21, Pro22, Lys24, Phe141, Leu265 and Phe335. The periphery of groove A comprises unconserved residues. Additionally, Groove A′, as referred to herein, comprises primarily unconserved residues.

[0099] Groove B is immediately adjacent to the pocket. Of significance is that the bottom of groove B comprises the N-epsilon nitrogen of His170, which coordinates the zinc through its N-delta nitrogen. Significant binding energy may be achieved by contacting the Ne proton of His170with a carboxylic acid or sulfate group. In addition, groove B may be large enough to fit a phenyl group, the face of which may comprise a partial negative charge which may pack over the N-epsilon proton of His170. The conserved residues of groove B, as referred to herein are: His170, Tyr196 and Leu265.

[0100] Groove C is not as well conserved as the other two grooves and the amino acid residues which comprise groove C are mostly polar and solvent exposed. Groove C, as referred to herein comprises the following conserved residues: Asn87, Gly140 and Phe198.

[0101] The compounds of the present invention are represented by formula (I):

[0102] Examples for suitable X constituents wherein X comprises a cap group may be described in three categories, depending upon which surface of groove A, A′, B and/or C they are targeted to. The cap group may comprise all three categories on the same compound. Of particular benefit may be replacing the cap group of TSA or SARA with a large, rigid structure. Nonlimiting examples for suitable cap groups (X) of formula (I) which may bind in-groove A are:

[0103] (1) attaching a 1-3 methyl linker followed by a phenyl or naphthalene group from the para or meta position of SAHA's phenyl group represented by formula (IV):

[0104] (2) attaching a 2-3 methyl linker followed by a phenyl or naphthalene group from the meta position of TSA's phenyl cap group, or from TSA's dimethyl amino group represented by formula (V):

[0105] and which may bind in groove B is a 1-3 methyl group spacer followed by a carboxylate, sulfate or phenyl group as represented by formula (VI):

[0106] With respect to the aliphatic (Y) group, the diameter of the pocket suggests that one more methyl “side chain” could fit, in addition to the C15 methyl group on the C10 carbon. Nonlimiting suitable examples for Y constituents wherein Y comprises an aliphatic chain group are as follows: (1) add a methyl group to TSA on the C12 carbon (with or without a methyl group on the C10 carbon and with or without double bonds and with or without substituting the X and/or Z constituents of formula (I) as represented by formula (VII):

[0107] (2) add a methyl group to TSA on the C9 carbon (with or without a methyl group on the C10 carbon; with or without both or either of the double bonds, and with or without substituting the X and/or Z constituents of formula (I) as represented by formula (VIII);

[0108] (3) replace the two alkalene double bonds of TSA with only one between C10 and C11, which may free the C11 and C12 torsion to allow for a better fit, the X and/or Z groups may also be substituted as represented by formula (IX):

[0109] (4) cyclize C15 and C12 carbons of TSA through a sulphur atom (or nitrogen atom), the-X and/or Z groups may also be substituted as represented by formula (X):

[0110] (5) extend from the C9 carbon of TSA such that the extension approaches and/or enters groove B (see FIG. 9); making C9 sp3 so that it can have some freedom; attach to C9 a 1-3 methyl group spacer which may include a double bond and they attaching thereto a sulfate, carboxylate, sulfate, hyroxyl, or phenyl group which may make an interaction with the N-epsilon proton of His170 which may coordinate the zinc atom as represented by formula (XI):

[0111] (6) extend off the C8 carbon (replacing C14) of TSA such that the extension approaches or enters groove B; attach a 1-3 methyl group spacer (which may include a double bond) and then link thereto a carboxylate, sulfate, hydroxyl or phenyl group such that an interaction is made with the N-epsilon proton of His170 that coordinates the zinc atom; the X and/or Z constituents may also be substituted as represented by formula (XII):

[0112] (7) substitute the C8 carbon at the end of the aliphatic chain such that the substitution may contact groove A, A′, B and or C, in such an example, a cap group (X) may or may not be required and the X and Z constituents may be substituted as well, as represented by formula (XIII):

[0113] (8) formulas VII through XIII above wherein the aliphatic chain further comprises a methyl group between the active site binding group (Z) and the C8 carbon, and preferably just before the C8 carbon, increasing the distance between X and Z, (9) make the connection between the aliphatic chain and the cap group more rigid (e.g., by closing a 6-membered ring which may or may not comprise oxygen, the X and Z group may also be substituted as represented by formula (XIV):

[0114] and (10) combining two or more of the changes depicted by formulas (VII-XIV).

[0115] Additionally, nonlimiting examples for suitable Z groups wherein Z comprises an active site binding group are as follows: (1) hydroxamic acid, (2) carboxylic acid, (3) sulfonamide, (4) acetamide, (5) epoxyketone, (6) an ester with a methyl linker and a hydroxyl of acetate ester group to lead into the cavity and interact with a conserved arginine (Arg27) as represented by formula (XV):

[0116] and (7) an alphaketone as represented by formula (XVI):

[0117] Additionally, other suitable X, Y and Z constituents may be envisioned by the skilled artisan given the three-dimensional structural information of the present invention.

[0118] After having determined potential suitable X, Y and Z constituents, the constituents are combined to form a compound of formula (I) using combinatorial chemistry techniques. This may be achieved according to U.S. Pat. Nos. 5,608,108; 5,700,811; 5,773,474; 5,840,960 and 5,668,179, incorporated herein by reference. Any methods known to one of skill in the art may be employed to synthesize compounds of formula (I) comprising X, Y and Z constituents as determined by the methods described above.

[0119] As mentioned above, the compounds of formula (I) are useful to inhibit the histone deacetylase activity of HDLP and HDAC-related proteins. Such inhibition may allow for a reduction or cessation of cell growth in vitro and in vivo.

[0120] For in vitro use, such reduction or cessation of cell growth is useful to study the role of histone deacetylation and differentiation during the cell cycle and also to study other mechanisms associated with cell cycle arrest and particularly how the repression of transcription is involved in cell cycle progression which may be studies in a yeast model system such as that described by Kadosh & Struhl, 1998, Mol. Cell. Biol. 18:5121-5127. In vitro model systems which may be employed to study the effects of potential inhibitors on cell cycle progression and also tumor growth include those described by: Richon et al, 1998, Proc. Natl. Acad. Sci. USA 95:3003-3007; Yoshida et al., 1995, Bioessays 17:423-430; Kim et al., 1999, Oncogene 18:2461-2470; Richon et al., 1996, Proc. Natl. Acad. Sci. USA 93:5705-5708; and Yoshida et al., 1987, Cancer Res. 47:3688-3691.

[0121] For in vivo use, such a reduction or cessation of cell growth is useful to study the effect of said inhibitor compounds in non-human animal model systems of cancer and is also useful for the treatment of cancer in a recipient in need of such treatment. Non-limiting examples of animals which may serve as non-human animal model systems include mice, rats, rabbits, chickens, sheep, goats, cows, pigs, and non-human primates. See, e.g., Desai et al., 1999, Proc. AACR 40: abstract #2396; Cohen et al., 1999, Cancer Res., submitted. The compounds of the present invention may be administered to a transgenic non-human animal wherein said animal has developed cancer such as those animal models in which the animal has a propensity for developing cancer (e.g. animal model systems described in U.S. Pat. Nos. 5,777,193, 5,811,634, 5,709,844, 5,698,764, and 5,550,316). Such animal model systems may allow for the determination of toxicity and tumor reduction effectiveness of the compounds of the present invention.

[0122] A preferred compound of the present invention may comprise high specific activity for HDLP and HDAC-related proteins, good bioavailability when administered orally, activity in reducing or ceasing cell growth in tumor cell lines, and activity in reducing or ceasing tumor growth in animal models of various cancers.

[0123] Accordingly, another aspect of this invention is a method of eradicating or managing cancer in a recipient, which may be an animal and is preferably a human. Said method comprises administering to said recipient a tumor reducing amount of a compound as defined by formula (I) above, or a physiological acceptable salt thereof.

[0124] In a further aspect of the invention, there is provided a composition comprising the compound of formula (I) and an excipient or carrier. Administration of the foregoing agents may be local or systemic. Such carriers include any suitable physiological solutions or dispersant or the like, The physiologic solutions include any acceptable solution or dispersion media, such as saline, or buffered saline. The carrier may also include antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like. Except insofar as any conventional media, carrier or agent is incompatible with the active ingredient, its use in the compositions is contemplated.

[0125] Routes of administration for the compositions containing the delivery vehicle constructs of the present invention include any conventional and physiologically acceptable routes, such as, for example, oral, pulmonary, parenteral (intramuscular, intraperitoneal, intravenous (IV) or subcutaneous injection), inhalation (via a fine powder formulation or a fine mist), transdermal, nasal, vaginal, rectal, or sublingual routes of administration and can be formulated in dosage forms appropriate for each route of administration.

[0126] The following examples are provided to more clearly illustrate the aspects of the invention and are not intended to limit the scope of the invention.

EXAMPLES

Example 1

Protein Production and Purification

[0127] Full-length wild-type HDLP (Genbank accession number AE000719) was subcloned from an Aquifex aeolicus chromosomal DNA preparation (provided by Robert Huber of Universitaet of Regensburg, Germany) into the pGEX4T3 (Amersham-Pharmacia, Piscataway, N.J.) vector using the polymerase chain reaction (PCR). The cysteine-to-serine and active site mutants were constructed by PCR site directed mutagenesis and were sequenced. The HDLP-glutathione S-transferase (GST) fusion protein was produced in Escherichia coli, purified by affinity chromatography using a column of glutathione-sepharose resin (Amersham-Pharmacia, Piscataway, N.J.), and by anion-exchange chromatography (Q-sepharosel™; Amersham-Pharmacia, Piscataway, N.J.). HDLP was cleaved from the fusion protein with thrombin at 4° C., was purified by anion-exchange (Q-sepharose™; Amersham-Pharmacia, Piscataway, N.J.) and gel filtration chromatography (Superdex™200; Amersham-Pharmacia, Piscataway, N.J.), and was concentrated to typically 25 mg/ml in a buffer of 25 mM bis-tris propane (BTP), 500 mM NaCl, 5 mM dithiothrietiol (DTT), 2% isopropanol, pH 7.0.

[0128] Although, it is not known what metal cofactor HDLP contains in vivo, it is presumed to be zinc because of the arrangement of the ligands and the similarities in the active site to the zinc proteases. The lack of metal in the purified HDLP is presumed due, in part, to the use of DTT during purification. HDLP was reconstituted with Zn2+by mixing the Cys75Ser/Cys77Ser double mutant at 10 mg/ml with a 5-fold molar excess of ZnCl2 in a buffer of 25 mM bis-tris propane, 200 mM NaCl, 1% isopropanol, pH 7.0. Unbound ZnCl2 was removed by fractionating HDLP through a G25 desalting column (Amersham-Pharmacia, Piscataway, N.J.). The HDLP-Zn2+-TSA complex was prepared by incubating the Zn2+reconstituted HDLP mutant with 1 mM TSA for 45 minutes, followed by gel filtration chromatography (Superdex™200; Amersham-Pharmacia, Piscataway, N.J.) to remove excess TSA, and concentration to typically 25 mg/ml in a buffer of 25 mM bis-tris propane, 500 mM NaCl, 1% isopropanol, pH 7.0.

[0129] FLAG epitope tagged human HDAC1 was overexpressed using a baculovirus expression system in Hi5 (Invitrogen, Carlsbad, Calif.) insect cells grown in suspension in serum-free media (Sf900, Gibco, Grand Island, N.Y.). The fusion protein was purified by anion exchange and affinity chromatography using Anti-FLAG M2 affinity resin (Sigma, St. Louis, Mo.) and FLAG Peptide (Sigma, St. Louis, Mo.).

Example 2

Crystallization and Data Collection

[0130] Crystals of apo-HDLP were grown at room temperature by-the hanging-drop vapor-diffusion method, from 7.5% isopropanol, 28% PEG 1500, 425 mM NaCl, 100 mM Tris-Cl, pH 7.0. They form in space group C2 with a=51.4 Å, b=93.8 Å, c78.7 Å, β=96.9 Å, and contain one HDLP molecule in the asymmetric unit. Diffraction data were collected with crystals flash-frozen in a buffer of 7.5% isopropanol, 35% PEG 1500, 75 mM NaCl, 100 mM Tris-Cl, pH 8.0, at -170° C.

[0131] The structure of the HDLP-Zn2+complex was determined from HDLP Cys75Ser/Cys77Ser double mutant crystals grown from 23% tert-butanol, 27% PEG 1500, 400 mM KC1, 100 mM bis-tris propane-Cl, pH 6.8. Space group and cell dimensions were identical to the apocrystals. The HDLP-Zn2+crystals were harvested and frozen in 27% tert-butanol, 22% PEG 1500, 50 mM KCl, 20 mM NaCl, 0.2 mM ZnCl2, 100 mM bis-tris propane, pH 6.8, at −170° C.

[0132] Crystals of the HDLP-Zn2+-TSA complex comprised HDLP Cys75Ser/Cys77Ser double mutant and were grown from 23% tert-butanol, 27% PEG 1500, 600 mM KC1, 100 mM bis-tris propane-Cl, pH 6.8, by microseeding. The crystals were grown in the presence of zinc. They form in space group

[0133] P212121 with a=53.4 Å, b=94.4 Å, c156.3 Å and contain two HDLP-Zn2+-TSA complexes in the asymmetric unit. The HDLP-Zn2+-TSA crystals were harvested and frozen in the same cryobuffer as the HDLP-Zn2+crystals except that 0.5mM TSA was added. Data were processed with DENZO and SCALEPACK (Otwinowski & Minor, 1997, Method. Ensemble. 276:307-326). MIR analysis, model building and refinement.

[0134] The HDLP-Zn2+-SAHA complex crystals were grown and evaluated the same as the HDLP-Zn2+-TSA crystals. However, the restraints for the SAHA structure were constructed based on stereochemical parameters from TSA. Like the apo-HDLP crystals, the SAHA/HDLP co-crystals grew in space group C2.

[0135] Heavy-atom soaks were performed with the apo-HDLP crystals in a buffer of 7.5% isopropanol, 30% PEG 1500, 75 mM NaCl, 100 mM Tris-Cl, pH 8.0, supplemented with 1.0 mM thimerosal for 2h, 5 mM KAu(CN)2 for 1h, and 1 mM Pb(Me)3OAc for 2h. MIR phases were calculated with the program MLPHARE (The CCP4 suite: Programs for computational crystallography, 1994, Acta Crystallogr. D 50:760-763) at 2.5 Å using the anomalous diffraction signal from the thimerosal derivative, and had a mean figure of merit of 0.55. The phases were improved by solvent flattening with the program DM (The CCP4 suite: Programs for computational crystallography, 1994, Acta Crystallogr. D 50:760-763) , and were used to build the initial model with the program O (Jones et al., 1991, Acta Crystallogr. A 47:110-109). Successive rounds of rebuilding and simulated annealing refinement with the program CNS (Brunger et al., 1998, Acta Crystallogr. D 54:905-921) allowed interpretation of HDLP from residues 2 to 373. Residues 1, 374, and 375 were not modeled and are presumed to be disordered.

[0136] The structure of the HDLP-Zn2+-TSA and HDLP-Zn2+-SAHA complex were determined by molecular replacement with the program AMORE (The CCP4 suite: Programs for computational crystallography, 1994, Acta Crystallogr. D 50:760-763) using the apo-HDLP structure as a search model. The initial electron density maps had strong and continuous difference density for the entire TSA molecule. However the SAHA molecule was not as well ordered in the cap group region. The structure of TSA was obtained from the Cambridge Structural Database (Refcode TRCHST) and was used to define stereochemical restraints used in the refinement with the program CNS. The restraints of SAHA were constructed based on stereochemical parameters from TSA and surrounding amino acid residues. The dimer interface in the HDLP-Zn2+-TSA and HDLP-Zn2+-SAHA crystals primarily involves Phe200 on the protein surface. The Phe200 side chain contacts Tyr91, whose side chain conformation changes on TSA binding, and part of the dimethyl amino phenyl group of TSA from the second protomer. The HDAC family does not contain a phenylalanine residue at the equivalent position.

EXAMPLE3

Histone Deacetylase Assays

[0137] Purified proteins were assayed by incubating 10 μg of [3H] acetyl-labeled murine erythroleukemia histone substrate and HDAC assay buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10% glycerol) for 30-60 minutes at 37° C. in a total volume of 30 μl. The final concentrations of HDLP and HDAC1-FLAG were 3.6 μM and 0.24 μM, respectively. Assays were performed in duplicate. The reactions were stopped and the released acetate was extracted and assayed as described (Hendzel et al., 1991, J. Biol. Chem. 266:21936-21942). [3H] acetyl-labeled murine erythroleukemia histones were prepared essentially as described (Carmen et al., 1996, J. Biol. Chem. 271:15837-15844). Inhibitors were added in the absence of substrate and incubated on ice for 20 minutes, substrate was added, and the assay performed as described above. HDLP was inclubated with 20 μM ZnCl2 and 20 μM MnCl2(H2O)4 in HDAC buffer and tested for activity.

[0138] Only HDLP dialyzed against ZnCl2 had activity. HDAC1-FLAG was dialyzed against 20 μM ZnCl2 in HDAC buffer which had no effect on activity. Therefore, HDAC1-FLAG contains a metal as purified.

[0139] The in vivo substrate of HDLP is not known. HDLP may have a role in acetoin utilization like the B. subtilis AcuC gene product, and it has been annotated as such in the genome sequence, but the reaction catalyzed by AcuC is also not known. Furthermore, the A. aeolicus genome appears to lack the acua and acuB genes that are part of the acuABC operon of B. subtilis (Deckert et al., 1998 Nature 2:353-358), and HDLP is as similar to human HDAC1 (35.2% identity) as it is to B. subtilis AcuC (34.7% identity).

[0140] Throughout the application, various publications are referenced by author, date and citation. The disclosures of these publication in their entireties are hereby incorporated by reference.

[0141] While this invention has been particularly shown and described with references to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims.

	Number	Date	Country
Parent	PCT/US00/24700	Sep 2000	US
Child	10095109	Mar 2002	US

Crystal structure of a deacetylase and inhibitors thereof

Information

Publication Number

Date Filed

Date Published

Inventors

CPC

US Classifications

International Classifications

Abstract

Description

Claims

RELATED APPLICATIONS

GOVERNMENT SUPPORT

Provisional Applications (1)

Continuations (1)