FIELD OF THE INVENTION The present invention relates generally to a family of phosphodiesterases designated PDE8A and uses thereof.

BACKGROUND OF THE INVENTION Phosphodiesterases (PDEs) hydrolyze 3', 5 cyclic nucleotides to their respective nucleoside 5 monophosphates. The cyclic nucleotides cAMP and cGMP are synthesized by adenylyl and guanylyl cyclases, respectively, and serve as second messengers in a number of cellular signaling pathways The duration and strength of the second messenger signal is a function of the rate of synthesis and the rate of hydrolysis of the cyclic nucleotide.

Multiple families of PDEs have been identified. The nomenclature system includes first a number that indicates the PDE family. To date, seven families (PDE1-7) are known which are classified by: (i) primary structure ; (ii) substrate preference ; (iii) response to different modulators ; (iv) sensitivity to specific inhibitors ; and (v) modes of regulation [Loughney and Ferguson. in Phosphodiesterase Inhibitors, Schudt, et al. (Eds.), Academic Press: New York, New York (1996) pp. 1-19]. The number indicating the family is followed by a capital letter, indicating a distinct gene, and the capital letter followed by a second number, indicating a specific splice variant or a specific transcript which utilizes a unique transcription initiation site.

The amino acid sequences of all mammalian PDEs identified to date include a highly conserved region of approximately 270 amino acids located in the carboxy terminal half of the protein [Charbonneau. et al., Proc. Natl. Acad. Sci. (USA) 83: 9308- 9312 (1986)]. The conserved domain includes the catalytic site for cAMP and/or cGMP hydrolysis and two putative zinc binding sites as well as family specific determinants [Beavo, Physiol. Rev. 75 : 725-748 (1995) ; Francis, et al., J. Biol. Chem. 269: 22477-

22480 (1994)]. The amino terminal regions of the various PDEs are highly variable and include other family specific determinants such as: (i) calmodulin binding sites (PDE1); (ii) non-catalytic cyclic GMP binding sites (PDE2, PDE5, PDE6); (iii) membrane targeting sites (PDE4); (iv) hydrophobic membrane association sites (PDE3); and (v) phosphorylation sites for either the calmodulin-dependent kinase 11 (PDE1), the cAMP- dependent kinase (PDE1, PDE3, PDE4), or the cGMP dependent kinase (PDE5) [Beavo, <BR> <BR> Physiol. Rev. 75: 725-748 (1995) : Manganiello, et al., Arch. Biocheni. Acta 322: 1-13<BR> <BR> (1995); Conti, eral., Physiol. Rev. 75: 723-748 (1995)].

Members of the PDE 1 family are activated by calcium-calmodulin. Three genes have been identified ; PDE 1 A and PDEIB preferentially hydrolyze cGMP while PDE1C has been shown to exhibit a high affinity for both cAMP and cGMP The PDE2 family is characterized as being specifically stimulated by cGMP [Loughney and Ferguson, supra]. Only one gene has been identified, PDE2A, the enzyme product of which is specifically inhibited by erythro-9- (2-hydroxy-3-nonyl) adenine (EHNA). Enzymes in the PDE3 family are specifically inhibited by cGMP Two genes are known, PDE3 A and PDE3B, both having high affinity for both cAMP and cGMP, although the V. for cgmp hydrolysis is low enough that cGMP functions as a competitive inhibitor for cAMP hydrolysis. PDE3 enzymes are specifically inhibited by milrinone and enoximone [Loughney and Ferguson, szrpra]. The PDE4 family effects cAMP hydrolysis and includes four genes, PDE4A, PDE4B, PDE4C, and PDE4D, each having multiple splice variants.

Members of this family are specifically inhibited by the anti-depressant drug rolipram.

Members of PDE5 family bind cGMP at non-catalytic sites and preferentially hydrolyze cGMP. Only one gene, PDE5A, has been identified. The photoreceptor PDE6 enzymes specifically hydrolyze cGMP [Loughney and Ferguson, supra]. Genes include PDE6A and <BR> <BR> PDE6B (the protein products of which dimerize and bind two copies of a smaller y inhibitory subunit to form rod PDE), in addition to PDE6C which associates with three smaller proteins to form cone PDE. The PDE7 family effects cAMP hydrolysis but, in contrast to the PDE4 family, is not inhibited by rolipram [Loughney and Ferguson, supra].

Only one gene, PDE7A, has been identified.

1. Given the importance of cAMP and cGMP in intracellular second messenger signaling, there thus exists an ongoing need in the art to identify addition PDE species. Identification of heretofore unknown families of PDEs, and genes and splice variants thereof, will provide additional pharmacological approaches to treating conditions in which cyclic nucleotide pathways are aberrant as well as conditions in which modulation of intracellular cAMP and/or cGMP levels in certain cell types is desirable.

SUMMARY OF THE INVENTION In brief, the present invention provides polypeptides and underlying polynucleotides for a novel PDE family designated PDE8. The invention includes both naturally occurring and non-naturally occurring PDE8 polynucleotides and polypeptide products thereof. Naturally occurring PDE8 products include distinct gene and polypeptide species within the PDE8 family (i. e., PDE8A); these species include those which are expressed within cells of the same animal and well as corresponding species homologs expressed in cells of other animals. Within each PDE8 species, the invention further provides splice variants encoded by the same polynucleotide but which arise from distinct mRNA transcripts (i. e., PDE8A1 and PDE8A2). Non-naturally occurring PDE8 products include variants of the naturally occurring products such as analogs (i. e., wherein one or more amino acids are added, substituted, or deleted) and those PDE8 products which include covalent modifications (i. e., fusion proteins, glycosylation variants, Met- 'PDE8s, Met-2-Lys'- PDE8s, GIy''PDE8s and the like). The PDE8 family is distinguished from previously known PDE families in exhibiting high affinity for hydrolysis of both cAMP and cGMP but relatively low sensitivity to enzyme inhibitors specific for other PDE families. In a preferred embodiment, the invention provides a polynucleotide comprising the sequence set forth in SEQ ID NO: 1. The invention also embraces polynucleotides encoding the amino acid sequence set out in SEQ ID NO: 2. A presently preferred polypeptide of the invention comprises the amino acid sequence set out in SEQ ID NO: 2. The invention provides two splice variant cDNAs which give rise to two polypeptides designated

PDE8A1 and PDE8A2. PDE8A1 and PDE8A2 polypeptides, and the polynucleotides encoding the polypeptides, are discussed herein as representative of the PDE8 enzyme family embraced by the invention.

The present invention provides novel purified and isolated polynucleotides (e. g., DNA sequences and RNA transcripts, both sense and complementary antisense strands, including splice variants thereof) encoding the human PDE8s. DNA sequences of the invention include genomic and cDNA sequences as well as wholly or partially chemically synthesized DNA sequences. "Synthesized,"as used herein and is understood in the art, refers to purely chemical, as opposed to enzymatic, methods for producing polynucleotides."Wholly"synthesized DNA sequences are therefore produced entirely by chemical means, and"partially"synthesized DNAs embrace those wherein only portions of the resulting DNA were produced by chemical means. A preferred DNA sequence encoding a human PDE8 polypeptide is set out in SEQ ID NO: 1. Also preferred are polynucleotides encoding the PBE8 polypeptide of SEQ ID NO: 2 and the PDE8A1 and PDE8A2 splice variant polypeptides set out in SEQ ID NOs: 6 and 4, respectively. Preferred polynucleotides encoding PDE8A1 and PDE8A2 are set out in SEQ ID NOs: 5 and 3, respectively. The invention further embraces species, preferably mammalian, homologs of the human PDE8 DNA.

The invention also embraces DNA sequences encoding PDE8 species which hybridize under moderately stringent conditions to the non-coding strands, or complements, of the polynucleotides in SEQ ID NOs: 1,3 and 5. DNA sequences encoding PDE8A polypeptides which would hybridize thereto but for the redundancy of the genetic code are contemplated by the invention. Exemplary moderate hybridization conditions are as follows : hybridization at 65°C in 3X SSC, 0.1% sarkosyl, and 20 mM sodium phosphate, pH 6.8, and washing at 65°C in 2X SSC with 0.1% SDS. It is understood in the art that conditions of equivalent stringency can be achieved through variation of temperature and buffer, or salt concentration as described Ausebel, et al.

(Eds.), Protocols in Molecular Bioloav, John Wiley & Sons (1994), pp. 6.0.3 to 6.4.10.

Modifications in hybridization conditions can be empirically determined or precisely calculated based on the length and the percentage of guanosine/cytosine (GC) base pairing

of the probe. The hybridization conditions can be calculated as described in Sambrook, et al., (Eds.), Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press: Cold Spring Harbor, New York (1989), pp. 9.47 to 9.51.

Autonomously replicating recombinant expression constructions such as plasmid and viral DNA vectors incorporating PDE8 sequences are also provided.

Expression constructs wherein PDE8-encoding polynucleotides are operatively linked to an endogenous or exogenous expression control DNA sequence and a transcription terminator are also provided.

According to another aspect of the invention, host cells are provided, including procaryotic and eukaryotic cells, either stably or transiently transformed with DNA sequences of the invention in a manner which permits expression of PDE8 polypeptides of the invention. Host cells of the invention are a valuable source of immunogen for development of antibodies specifically immunoreactive with PDE8. Host cells of the invention are also conspicuously useful in methods for large scale production of PDE8 polypeptides wherein the cells are grown in a suitable culture medium and the desired polypeptide products are isolated from the cells or from the medium in which the cells are grown by, for example, immunoaffinity purification.

Knowledge of PDE8 DNA sequences allows for modification of cells to permit, or increase, expression of endogenous PDE8. Cells can be modified (e. g., by homologous recombination) to provide increased PDE8 expression by replacing, in whole or in part, the naturally occurring PDE8 promoter with all or part of a heterologous promoter so that the cells express PDE8 at higher levels. The heterologous promoter is inserted in such a manner that it is operatively-linked to PDE8 encoding sequences. See, for example, PCT International Publication No. WO 94/12650, PCT International Publication No. WO 92/20808. and PCT International Publication No. 91/09955. The invention also contemplates that, in addition to heterologous promoter DNA, amplifiable marker DNA (e. g, ada, dhfr, and the multifunctional CAD gene which encodes carbamyl phosphate synthase, aspartate transcarbamylase, and dihydroorotase) and/or intron DNA may be inserted along with the heterologous promoter DNA. If linked to the PDE8

coding sequence, amplification of the marker DNA by standard selection methods results in co-amplification of the PDE8 coding sequences in the cells.

The DNA sequence information provided by the present invention also makes possible the development through, e. g homologous recombination or"knock-out" strategies [Capecchi, Scie71ce 244 : 1288-1292 (1989)], of animals that fail to express functional PDE8 or that express a variant of PDE8. Such animals are useful as models for studying the iii vivo activities of PDE8 and modulators of PDE8.

The invention also provides purified and isolated mammalian PDE8 polypeptides. Presently preferred PDE8A polypeptides are set out in SEQ ID NOs: 4 and 6. Most preferred is a PDE8 polypeptide comprising the amino acid sequence set out in SEQ ID NO: 2. PDE8 polypeptides of the invention may be isolated from natural cell sources or may be chemically synthesized, but are preferably produced by recombinant procedures involving host cells of the invention. Use of mammalian host cells is expected to provide for such post-translational modifications (e. g., glycosylation, truncation, lipidation, and phosphorylation) as may be needed to confer optimal biological activity on recombinant expression products of the invention. PDE8 products of the invention may be full length polypeptides, biologically active fragments, or variants thereof which retain specific PDE8 biological activity. Variants may comprise PDE8 polypeptide analogs wherein one or more of the specified (i. e.. naturally encoded) amino acids is deleted or replaced or wherein one or more non-specified amino acids are added: (1) without loss of one or more of the biological activities or immunological characteristics specific for PDE8; or (2) with specific disablement of a particular biological activity of PDE8.

Variant products of the invention include mature PDE8A products, i. e., PDE8 products wherein leader or signal sequences are removed, having additional amino terminal residues. PDE8 products having an additional methionine residue at position-1 (Met-'-PDE8) are contemplated, as are PDE8 products having additional methionine and lysine residues at positions-2 and-I (Met-2-Lys-'-PDE8). Variants of these types are particularly useful for recombinant protein production in bacterial cell types.

The invention also embraces PDE8 variants having additional amino acid residues which result from use of specific expression systems. For example, use of

commercially available vectors that express a desired polypeptide such as a glutathione-S- transferase (GST) fusion product provide the desired polypeptide having an additional glycine residue at position-1 as a result of cleavage of the GST component from the desired polypeptide. Variants which result from expression in other vector systems are also contemplated.

The invention further embraces PDE8 products modified to include one or more water soluble polymer attachments. Particularly preferred are PDE8 products covalently modified with polyethylene glycol (PEG) subunits. Water soluble polymers may be bonded at specific positions, for example at the amino terminus of the PDE8 products. or randomly attached to one or more side chains of the polypeptide.

Also comprehended by the present invention are antibodies (e. g., monoclonal and polyclonal antibodies, single chain antibodies, chimeric antibodies, CDR- grafted antibodies and the like) and other binding proteins specific for PDE8 products or fragments thereof Specific binding proteins can be developed using isolated or recombinant PDE8 products, PDE8 variants, or cells expressing such products. Binding proteins are useful for purifying PDE8 products and detection or quantification of PDE8 products in fluid and tissue samples using known immunological procedures. Binding proteins are also manifestly useful in modulating (i. e., blocking, inhibiting or stimulating) biological activities of PDE8, especially those activities involved in signal transduction.

Anti-idiotypic antibodies specific for anti-PDE8 antibodies are also contemplated.

The scientific value of the information contributed through the disclosures of DNA and amino acid sequences of the present invention is manifest. As one series of examples, knowledge of the sequence of a cDNA for PDE8A makes possible through use of Southern hybridization or polymerase chain reaction (PCR) the identification of genomic DNA sequences encoding PDE8 and PDE8 expression control regulatory sequences such as promoters, operators, enhancers, repressors, and the like. DNA/DNA hybridization procedures carried out with DNA sequences of the invention under moderately to highly stringent conditions are likewise expected to allow the isolation of DNAs encoding allelic variants of PDE8A, allelic variants are known in the art to include structurally related proteins sharing one or more of the biochemical and/or immunological

properties specific to PDE8A. Similarly, non-human species genes encoding proteins homologous to PDE8A can also be identified by Southern and/or PCR analysis. As an alternative, complementation studies can be useful for identifying other human PDE8 products as well as non-human proteins, and DNAs encoding the proteins, sharing one or more biological properties of PDE8A.

Polynucleotides of the invention are also useful in hybridization assays to detect the capacity of cells to express PDE8. Polynucleotides of the invention may also be the basis for diagnostic methods useful for identifying a genetic alteration (s) in a PDE8 locus that underlies a disease state or states.

Also made available by the invention are anti-sense polynucleotides which recognize and hybridize to polynucleotides encoding PDE8. Full length and fragment anti-sense polynucleotides are provided Anti-sense polynucleotides are particularly relevant to regulating expression of PDE8 by those cells expressing PDE8 mRNA.

The DNA and amino acid sequence information provided by the present invention also makes possible the systematic analysis of the structure and function of PDE8s. DNA and amino acid sequence information for PDE8 also permits identification of molecules with which PDE8A will interact. Agents that modulate (i. e., increase, decrease, or block) PDE8 activity may be identified by incubating a putative modulator with PDE8 and determining the effect of the putative modulator on PDE8 phosphodiesterase activity. The selectivity of a compound that modulates the activity of the PDE8 can be evaluated by comparing its activity on the PDE8 to its activity on other PDE enzymes. Cell based methods, such as di-hybrid assays and split hybrid assays, as well as in vitro methods, including assays wherein a polypeptide or its binding partner are immobilized, and solution assays are contemplated by the invention.

Selective modulators may include, for example, antibodies and other proteins or peptides which specifically bind to the PDE8 or PDE8 nucleic acid, oligonucleotides which specifically bind to the PDE8 or PDE8 nucleic acid, and other non- peptide compounds (e. isolated or synthetic organic molecules) which specifically react with PDE8 or PDE8-encoding nucleic acid. Mutant forms of PDE8 which affect the enzymatic activity or cellular localization of the wild-type PDE8 are also contemplated by

the invention. Presently preferred targets for the development of selective modulators include, for example: (1) regions of the PDE8 which contact other proteins and/or localize the PDE8 within a cell, (2) regions of the PDE8 which bind substrate, (3) allosteric cyclic nucleotide-binding site (s) of PDE8, (4) phosphorylation site (s) of PDE8 and (5) regions of the PDE8 which are involved in multimerization of PDE8 subunits.

Modulators of PDE8 activity may be therapeutically useful in treatment of a wide range of diseases and physiological conditions in which PDE activity is known to be involved.

The invention further contemplates small molecule modulators of PDE8A enzyme activity. There are at least three different types of libraries used for the identification of small molecule modulators These include : (I) chemical libraries, (2) <BR> <BR> natural product libraries, and (3) combinatorial libraries comprised of random peptides, oligonucleotides or organic molecules.

Chemical libraries consist of structural analogs of known compounds or compounds that are identified as"hits"or"leads"via natural product screening. Natural product libraries are collections of microorganisms, animals, plants, or marine organisms which are used to create mixtures for screening by: (1) fermentation and extraction of broths from soil, plant or marine microorganisms or (2) extraction of plants or marine organisms. Combinatorial libraries are composed of large numbers of peptides, oligonucleotides or organic compounds as a mixture. They are relatively easy to prepare by traditional automate synthesis methods, PCR, cloning or proprietary synthetic methods. Of particular interest are peptide and oligonucleotide combinatorial libraries.

Still other libraries of interest include peptide, protein, peptidomimetic, multiparallel synthetic collection, recombinatorial, and polypeptide libraries. For a review of combinatorial chemistry and libraries created therefrom, see Myers, Curr. Opion.

Biotechnol. 8: 701-707 (1997).

Identification of modulators through use of the various libraries described herein permits modification of the candidate"hit'' (or"lead") to optimize the capacity of the"hit"to modulate activity.

The invention further provides methods to identify a specific binding partner compound of a PDE8A polypeptide of the invention comprising the steps of a)

contacting the PDE8A polypeptide with a compound under conditions which permit binding between the compound and the PDE8A polypeptide ; b) detecting binding of the compound to the PDE8A polypeptide; and c) identifying the compound as a specific binding partner of the PDE8A polypeptide. Binding partner identified in the methods of the invention preferably modulate PDE8A enzyme activity, either through inhibition or activation, or enhancement, of the enzyme.

The invention also provides methods to identify a specific binding partner compound of a PDE8A polynucleotide of the invention comprising the steps of : a) contacting the PDE8A polynucleotide with a compound under conditions which permit binding between the compound and the PDE8A poiynucleotide, b) detecting binding of the compound to the PDE8A polynucleotide ; and c) identifying the compound as a specific binding partner of the PDE8A polynucleotide. The binding partner of the PDE8A polynucleotide preferably modulates expression of the PDE8A polypeptide encoded by the PDE8A polynucleotide, either through inhibiting expression or enhancing expression.

The invention also provides compounds identified by a method of the invention, as well as compositions comprising a compound identified and a pharmaceutically acceptable carrier.

DETAILED DESCRIPTION OF THE INVENTION The present invention is illustrated by the following examples which relate to the isolation of polynucleotides encoding PDE8 polypeptides as well as expression and characterization of the encoded polypeptides. Example 1 describes methods for searching expressed sequence tag (EST) databases in order to identify probes potentially useful for isolating DNAs of the invention. Example 2 relates to identification of PDE8A-encoding polynucleotides. Example 3 addresses sequence analysis of the isolated polynucleotides.

Example 4 describes analysis of polypeptides encoded by the PDE8A polynucleotides. Example 5 addresses expression of recombinant PDE8A polypeptides. Example 6 relates to Northern analysis of PDEBA expression. Example 7 describes chromosome mapping of the gene encoding PDE8A. Example 8 describes confirmation that PDE8A1 and PDE8A2 are splice variants. Example 9 addresses expression and characterization of

recombinant PDE8A. Example 10 details production of anti-PDE8A monoclonal antibodies. Example 11 describes an analysis of PDE8A expression by in situ hybridization.

Example 1 Identification of an EST Related to a Human PDE Using the sequences of known human, 3', 5'cyclic nucleotide phosphodiesterases, a search of the National Center for Biotechnology Information (NCBI) Expressed Sequence Tags (EST) database was undertaken in order to identify cDNA fragments that could potentially be useful for the identification of novel phosphodiesterase (PDE) genes. This database contains DNA sequences representing one or both ends of cDNAs collected from a variety of tissue sources. A single sequencing run is performed on one or both ends of the cDNA and the quality of the DNA sequence varies tremendously. At the time the PDE searches were performed, the EST sequence database contained more than 600, 000 cDNA sequences from a variety of organisms.

The search for novel PDE sequences included three steps. First the BLASTN program available through NCBI was used to identify DNA sequences in the <BR> <BR> <BR> <BR> EST sequence database with homology to cDNA sequences encoding known human PDEs. The program compares a nucleotide query sequence against a nucleotide sequence database. The cDNA sequences of the fifteen known human PDEs were submitted and fifteen BLASTN searches were performed ; the query PDE sequences included PDE1A3 <BR> <BR> <BR> <BR> [Loughney, et al., J. Biol. Chem. 271 : 796-806 (1996)], PDE 1 B1 [Yu, et al., Cell<BR> <BR> <BR> <BR> <BR> <BR> <BR> Signaling, in press (1997)], PDE 1 C2 [Loughney, et al., J. Biol. Chem. 271: 796-806 (1996)], PDE2A3 al.,Gene191:89-95(1997)],PDE3A[Meacci,etal.,et Proc. Natal. Acad. Sci. (USA) 89: 3721-3725 (1992)], PDE3B [Miki et al., Genomics 36: 476-485 (1996)], PDE4A5 [Bolger,et al., Mol. Cell. Biol. 13 : 6558-6571 (1993)], <BR> <BR> <BR> PDE4B2 [Bolger, etal., Mol. ('ell. Biol. 13: 6558-6571 (1993)], PDE4C [Bolger, etal.,<BR> <BR> <BR> <BR> <BR> <BR> <BR> Mol. CelL Biol. 13: 6558-6571 (1993)], PDE4D I and PDE4D3 [Bolger, et al., Mol. Cell.<BR> <BR> <BR> <BR> <BR> <BR> <BR> <P>Biol. 13: 6558-6571 (1993)], PDE5A, PDE6A [Pittler, et al., Genomics 6: 272-283<BR> <BR> <BR> <BR> <BR> <BR> <BR> (1990)], PDE6B [Collins, et al.. Geromic. s 13: 698-704 (1992)], PDE6C [Piriev, et al.,

Genomics 28: 429-435 (1995), and PDE7A1 [Michaeli, et al., J. Biol. Chem. 17: 12925- 12932 (1993)]. The BLASTN results were examined and EST sequences that were judged as corresponding to each of the fifteen known PDE cDNAs were identified and collected into a table. The PDE6A and PDE6B sequences used as queries were truncated at 3'end (removing a portion of the 3'untranslated region) due to the presence of repetitive elements in the 3'untranslated region of the cDNAs.

Secondly, the NCBI TBLASTN program was used to examine the homology between the protein sequence of the fifteen known human PDEs (as above) and the six different possible proteins encoded by each of the EST DNA sequences. In this search, the EST sequences are translated in six frames and the amino acid sequences generated are compared to the query PDE amino acid sequences Sequences identified as homologous at the amino acid level were examined and any EST sequences positively identified as corresponding to a known PDE during the BLASTN search described above were discarded.

The third step of the search involved analyzing the sequences that were not known PDEs. These amino acid sequences were homologous to a known PDE but were not identified as one of the 15 known PDE genes during the BLASTN searches.

The BLAST searches identified an EST sequence (designated W04835) from a human fetal lung cDNA library as encoding an amino acid sequence having homology to the catalytic region of PDE2A, PDE3A, PDE3B, PDE4A, PDE4B, PDE4C, PDE5A, rod alpha PDE6A, rod beta PDE6B, cone alpha PDE6C, and PDE7A. The database sequence for W04835 is set out in SEQ ID NO: 7. Results from the database analysis as discussed below are exemplified using the PDE4D sequence.

W04835 cDNA was obtained from American Type Culture Collection (Rockville, MD) which maintains and makes publicly available deposits of ESTs identified and sequenced by I. M. A. G. E., Lawrence Livermore National Laboratory, Livermore, CA). The W04835 DNA was sequenced upon receipt to confirm its identity and determined to be consistent with SEQ ID NO: 7.

The amino acid sequence encoded by the-1 reading frame of EST sequence W04835 was recognized by all of the PDE query cDNA sequences except

PDE1A, 1B and 1C. Using the TBLASTN results with PDE4D3 as an example, two regions of similarity were detected. The first region showed 15/37 exact matches or 40% identity (19/37 similar amino acids) and included the HD (X) 2HXG (X) l3A (SEQ ID NO: <BR> <BR> <BR> <BR> 8) motif found in all of the query sequences. [Charboneau, Mol. Pharmacol. CellRegul.

2: 267-298 (1990)]. The second region showed 9/20 exact matches or 45% identity and included the YHNxxHA motif found in most of the query sequences. BLASTN analysis of the W04835 sequence revealed that it was unique in that it was not identical to any other human DNA sequences in the Genbank database. The EST database entry for W04835 identified the sequence as being similar to PIR: A48719, the bovine cC3MP binding, cGMP hydrolyzing PDESAI sequence. Comparison of the protein sequence of W04835 frame-1 to the bovine PDE5AI sequence revealed 58/153 matches for an overall identify of 38%. Within this region were small regions of greater homology; one region showed a 12/14 identical amino acids. Given the unique nature of the W04835 sequence, its relatively low homology to bovine PDE5A1, and the presence of the amino acid motifs found in most other known human PDE amino acid sequences, W04835 represents a novel human PDE cDNA.

Example 2 Isolation of Putative PDE cDNA W04835 cDNA insert was digested from the pT7T3D vector into two <BR> <BR> <BR> <BR> fragments with the restriction enzymes EcoRI and HindIII and the two fragments were purified using two sequential low melting agarose gels. Both fragments were used as probes to screen cDNA libraries derived from human heart (Stratagene, La Jolla, CA), and human fetal brain (Stratagene) using procedures routinely practiced in the art.

Approximately 5 x 105 phage from each library were screened. Hybridization was carried out overnight in buffer containing 3X SSC, 0.1% Sarkosyl, 20 mM sodium phosphate, pH <BR> <BR> <BR> <BR> 6.8,10X Denhardt's solution, and 50 Llfli salmon sperm DNA at 65 ° C. The filters were washed at 65°C in buffer containing 2X SSC and 0.1% SDS prior to autoradiography.

Nine clones from the fetal brain cDNA library and two from the heart cDNA library hybridized to the W04835 probe. Partial sequencing and mapping led to

the selection of one clone from the fetal brain library designated FB66a for further characterization.

A second screening of approximately 7.5 x 105 phage from the fetal brain cDNA library under conditions used in the first screening using the 1.3 kb EcoRI/HindIII fragment from the 5'portion of W04835 yielded nineteen additional cDNA clones. Six of these cDNAs also hybridized to a Hi7zdIII/KpnI fragment of W04835 which includes a 256 nucleotide region at the 5'end of W04835. Partial sequencing and mapping of five of the clones led to the selection of a second clone designated FB85c-2 for further analysis.

Example 3 DNA Sequence Analysis of FB66a and FB85c-2 The DNA sequence of FB66a was determined for both strands using DNA oligonucleotide primers set out below in SEQ ID NOs: 9 to 31 and a Perkin Elmer Applied Biosystems Division 373A DNA Sequencer according to the maunfacturer's suggested protocol. The amount of PCR product used as template was calculated based on the size of the PCR product and was sequenced using ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit with ApliTaq DNA Polymerase, FS (Perkin Elmer, Foster City. CA) and asymmetric PCR. The reaction product was purified on a AGCT spin column (Advanced Genetic Technologies Corp., Gaithersburg, MD) and dried.

Loading buffer was added to each purified sample and the mixture heated at 90°C for two minutes. The solution was transferred to ice until being loaded onto a 4% polyacrylamide gel. Data was automatically collected once the Data Collection program was initiated and was automatically analyzed and read by the Sequence Analysis program. All editing was performed manually and the resulting sequences were aligned where the consensus sequence was determined.

M13Rev. l GGAAACAGCTATGACCATG SEQ ID NO: 9 W48A2 ACTCTCCAAGGAAATACAG SEQ ID NO: 10 W48A9 CTGTCTCTGCACTAACAC SEQ ID NO: 1 I W48A4 TTGGCAAGGCCTCTGCAT SEQ ID NO: 12 W48S1 CCTCTATGAACTGAGCAG SEQ ID NO: 13

W48A1 GAAGGCACTGCCACTGAT SEQ ID NO: 14 W48S6 TCGAGCTGTATCGGCACT SEQ ID NO: 15 W48A5 AGCGTGTGATTGTTCTGAA SEQ ID NO : 16 W48S7 TGCTGGCCAAGTAGCAAG SEQ ID NO: 17 W48A6 AAGGTCACAGGCAGTCAT SEQ ID NO: 18 W48S2 GAAGAGTGGCAAGGTCTC SEQ ID NO: 19 W48S3 TCATGACCTGGACCACCAG SEQ ID NO : 20 W48A8 CCTTCTTGAAGAGGTTTGC SEQ ID NO: 21 W48S4 ATGACTGCCTGTGACCTT SEQ ID NO: 22 W48S5 CTGCTATACAACCCTTACC SEQ ID NO : 23 W48S8 GCTAATATTGCTGAGGCC SEQ ID NO: 24 W48A7 TAAGTGAGAGGTGACTGC SEQ ID NO: 25 W48S9 CCTAAAGGGCTGAGATCA SEQ ID NO: 26 W48S10 CGCAGTCACCTCTCACTT SEQ ID NO: 27 M13 TGTAAAACGACGGCCAGT SEQ ID NO: 28 W48A11 ACAAAACGCCTATGGTGG SEQ ID NO: 29 W48A10 TTGATCTCAGCCCTTTAGC SEQ ID NO: 30 W48S11 TCATGTGGCAGGAAACTG SEQ ID NO: 31 The FB66a cDNA, set out in SEQ ID NO 3, is 4389 nucleotides in length and, from nucleotide 3 to nucleotide 2411 encodes a protein of 803 amino acids with a predicted molecular weight of approximately 90,775 Da The deduced amino acid sequence for FB66a is set out in SEQ ID NO : 4 The first methionine is encoded at nucleotide 45; the absence of an upstream in frame stop codon makes it unclear whether this residue is an internal methionine or the beginning of the open reading frame.

The DNA sequence of FB85c-2 (SEQ ID NO: 5) was similarly determined using primers M13Rev. 1, W48A2, W48A9, W48A4, W48S 1, W48A1, W48S6, W48A5, W48A6, W48S2, W48S3, W48S4, W48S5, W48S7, W48A8, and M13. FB85c-2 appeared to include two distinct DNA inserts, only one of which was homologous to W04835. The region homologous to W04835 was approximately 2.8 kb in length. The precise sequence at the 5'end of the insert could not be determined and thus a few hundred bases of sequence in what may be a 5'-untranslated region are not included in the 2573 nucleotide sequence set out in SEQ ID NO: 5. Nucleotide 67 to nucleotide 2406 encodes a protein having 779 amino acid protein (SEQ ID NO: 6) having a predicted

molecular weight of 88,353 Da. An in frame upstream stop codon makes it likely that the methionine encoded at nucleotide position 67 is the initiation methionine.

The proteins encoded by FB66a and FB85c-2 have different amino terminal sequences which may be due to alternative splicing. The DNA sequences diverge from each other 5'of nucleotide 112 in FB66a and nucleotide 104 in FB85c-2. Thus, FB85c-2 has 13 amino acids at the amino terminus that are not found in the FB66a protein. The FB66a protein includes 23 unique amino terminal residues if the initiating methionine at presumed to be encoded at nucleotide 35 ; the protein includes more than 37 unique amino terminal residues if the open reading frame in the FB66a clone is incomplete.

BLASTN analysis, wherein a query nucleotide sequence is compared against a nucleotide sequence database, of the FB66a sequence revealed no identity with sequences in Genbank, NCBI STS. NCBI HTGS, or NCBI GSS databases. However, two identical sequences were identified in the NCBI EST database.

One sequence was the W04835 EST which was used to identify the cDNA clone. The second, AA307865 (SEQ ID NO: 32), derived from a colon cancer cell line KM12C (HCC) showed sequence identity with the 3'untranslated region of the FB66a and FB85c-2 clones. During the search in which AA307865 was identified, additional EST DNAs were identified presumably encoding putative mouse (EST AA386789, SEQ ID NO: 38) and rat (EST H32 ; 34. SEQ ID NO: 33) homologs to the human proteins encoded by FB66a and FB85c-2. The mouse sequence was 86% identical to the human sequences and the rat sequence was 81 %.

Example 4 Analysis FB85c-2 and FB66a Protein The PDEs encoded by clones FB85c-2 and FB66a were designated PDE8A1 and PDE8A2, respectively. Both PDE8A proteins. having complete amino acid sequence identity beyond the point of divergence discussed above, are most similar to human PDE2A, PDE5A, PDE6A, PDE6B. and PDE6C. Tables 1 and 2 show percent amino acid identity between PDE8A and PDE2A, PDE5A and PDE6A.

PDE8A1 and PDE8A2 share homology with other PDEs over the catalytic region (amino acids 492 through 748 in PDE8A1) and with the putative cGMP binding domain conserved in the amino terminal region of the PDE2A, PDE5A, PDE6A, PDE6B, AND PDE6C. The potential cGMP binding domain of PDE8A extends from amino acids 75 to amino acid 445 in the PDE8A1 polypeptide. Within the cGMP binding domains of PDE2A, PDE5A, PDE6A, PDE6B, and PDE6C, there are two internal repeats designated "a"and"b,"and each repeat contains a series of conserved amino acids [McAllister-Lucas, etal., J. Biol. Chem. 268: 22863-22873 (1993)]. In the corresponding"b"repeat region of PDE8A, all of the conserved amino acids are found ; in the corresponding"a"repeat region, only some of the conserved residues were detected. An aspartate residue, shown to be essential for the cGMP binding by bovine PDE5A [McAllister-Lucas, et al., J. Biol.

Chem. 270: 1-9 (1995)] is not present in the"a"repeat region of PDE8A. It is therefore uncertain whether this region in PDE8A functions to bind cGMP Table 1 PDE8A Identitv in the Entire Protein PDE 2A 5A 6A 8A 2A 100 19 16 28 5A 100 23 28 6A 100 21 8A 100 Table 2 PDE8A ldentity in the Catalytic Domain PDE 2A 5A 6A 8A 2A 100 38 33 41 5A 100 42 46 6A 100 37 8A 100

Example 5 Expression of Recombinant PDE8A An expression construct for PDE8A was generated that included DNA sequences 3'from the point of divergence of PDE8A1 and PDE8A2 through the stop codon. The expression construction included DNA encoding an eight amino acid epitope tag. The so-called"FLAG tag,"comprising the peptide sequence set out in SEQ ID NO: 34, was added to the amino terminus in order that the protein could be identified by Western blotting techniques using an anti-FLAG M2 antibody (Eastman Kodak, Rochester, NY) which specifically recognized the peptide of SEQ ID NO: 34.

SEQ ID NO: 34 Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys Sequences encoding an initiating methionine at the proteins amino terminus was also added.

As a first step in constructing the expression plasmid, PCR was performed using FB66a DNA as a template using primers set out in SEQ ID NOs : 35 (below) and W48A2 (SEQ ID NO: 10, p. 14) in a reaction mixture containing 2 ul each primer (stock 100 ug/ml), 2 ul 10X PCR buffer 11 (Perkin Elmer), 2 u ! 10X stock of each nucleotide (stock 2 mM), 1.2 ut MgCl2 (stock 25 mM). 0. 09 l 5 Units/i taq polymerase (Perkin Elmer), FB66a DNA and water to bring the reaction mixture to 20 u !. In the 5'primer (SEQ ID NO: 35), an NcoI site is in bold and the FLAG tag encoding region is underlined.

SEQ ID NO: 35 CAGTCAGCTAGCCGCCATGGACTACAAGGAC- GACGATGACCAAGTTGACTGATGAAAAGGTG PCR was carried out in a Perkin Elmer DNA Thermal Cycler under the following conditions: 94°C for 4 minutes followed by 30 cycles ouf 94C for one minute, 50°C for one minute, and 72°C for two minutes.

The resulting PCR product was digested with Ncol and Kpnl, gel purified, and subcloned into Bluescript SKH vector previously digested with the same enzymes.

The Bluescript vector had previously been modified to include a SacI/NcoI alcohol

dehydrogenase 2 (ADH2) promoter fragment removed from a YEpC-PADH2d vector [Price, etal., Meth. Enzymol. 185: 308-315 (1990)]. The resulting plasmid was designated W48pcrl.

AKpnIlSstI fragment containing the 3'portion of the open reading frame was isolated from a FB66a cDNA and inserted into W48pcrl previously digested with Kpnl and EcoRV. The resulting plasmid was designated W485.1.

A SaclIKpnl fragment containing the ADH2 promoter and the 5'portion of the PDE8A gene was isolated from W49pcrl. A KpnIlSaA fragment containing the 3' region of PDE8A was isolated from W485.1. The two fragments were ligated into the yeast expression vector YEpC-PADH2d that had been previously digested with SacI and Sall. The resulting plasmid was designated W48-2ADH2 and was deposited on October 2,1997 under the terms of the Budapest Treaty with the American Type Culture Collection (A. T. C. C.), 12301 Parklawn Drive, Rockville, MD 20852. The bacterial strain bearing plasmid W48-2ADH2 was assigned accession number ATCC 98552. The DNA sequences generated by PCR and the DNA sequences at the PDE8/vector junctions were determined to insure proper plasmid construction. Upon confirmation of the sequence, the plasmid was transformed into a yeast strain BJ2-54 lacking endogenous PDE activity (ura3-52; trp 1; leu2; cir °; gal2, pep4-3 ; prb 1-1122 ; prc 1-402; Z1PDE1 :. URA 3 ; HIS3 ; SPDE2:: TRPI).

The host cells were grown overnight in SC-leu selective media including 2% glucose, diluted to 1-2 x 105 cells/ml and subsequently grown to a density of 10' <BR> <BR> cells/ml in the same media. The presence of the expression plasmid appeared to increase the doubling time for cell growth two-to three-fold even under non-inducing conditions.

The cells were collected by centrifugation, washed with YEP media including 3% glycerol, resuspended in YEP/3% glycerol at a density of 10'cells/ml, and grown for 24 hours prior to harvest. Cells were frozen until use.

Frozen cell pellets (0.06 ml) were thawed and suspended in 0.2 ml lysis buffer containing 100 mM MOPS, pH mM NaCl, 2 uM ZnS02,2 mM dithiothreitol, and 10 llg/ml each protease inhibitors pepstatin, leupeptin, and aprotinin.

Approximately 0.2 ml of 0.5 mm glass beads were added to the cells which were then

lysed with four 30-second cycles of vortexing. The lysate was aspirated and the beads were washed twice with 0.3 ml lysis buffer. The lysate was combined with the washes to generate the yeast extract. In some experiments the lysate was fractionated by centrifugation at 105,000 x g for thirty minutes.

Western analysis was carried out on yeast extract containing the recombinant protein as follows. Proteins were first separated on SDS-PAGE and transferred to Immobilon-P (Millipore) using standard methods. The protein blots were blocked using 5% non-fat dry milk in 20 mM Tris-HCI, pH 7.4,150 mM NaCl, 0.05% Tween-20 (TBST buffer plus milk) for one hour at room temperature. The blots were incubated with anti-FLAG M2 antibody (discussed above) at a concentration of 1 ug/ml in TBST buffer plus milk for one hour, after which the blots were washed four times with TBST buffer. The blots were then incubated for one hour with blotting grade affinity purified goat anti-mouse IgG antibody conjugated to horse radish peroxidase (HRP) (BioRad). The goat IgG was previously diluted 1: 10,000 in TBST buffer plus milk. The blots were washed four times with TBST and treated, according to the manufacturer's suggested protocol, with the RenaissanceX system (New England Nuclear Life Sciences Products) for enhanced chemiluminescence prior to autoradiography. The majority of the protein detected by the antibody was the size expected for the recombinant protein.

PDE activity was assayed by detection of 32P-phosphate released from 32P- <BR> <BR> cAMP or 32P-cGMP as described previously [Loughney eí al., J. Biol. Chem. 271: 796- 806 (1996)]. The yeast extract was diluted in 0.5X lysis buffer also containing 0.5 mg/ml bovine serum albumin. Twenty 1ll of the yeast extract, or diluted yeast extract, was assayed in a 100 ul reaction volume which included an additional 50 mM Tris-HCI (pH 8.0), 5 mM MgCl2 1 FM Zn SO2, and 0.1 mg/ml bovine serum albumin. Protein concentration was assayed by the method of Bradford.

PDE8A was observed to hydrolyze both cAMP and cGMP In unfractionated lysats, the specific activity for cAMP was 3.9 nmol/min/mg and for cGMP was 7.6 nmoUmin/mg. Fractionation revealed that 20-40% of the total activity was associated with the high speed supernatant fraction. Kinetic analysis of the activity with cAMP as substrate suggested the presence of both low and high Km forms of the enzyme

in a 1: 1 activity ratio. The estimated K", values were 0.2 uM and 350 hum. Analysis of the high speed pellet suggested that the same species were present but in a high K.: low Km activity ratio of 1: 4. Kinetic analysis with cGMP as substrate also suggested the presence of low and high forms of the enzyme. In these analyses, K. values were estimated to be 3 uM and 300 HM.

The IC50 values for inhibition of PDE8A activity were determined using a set of isozyme-selective PDE inhibitors and the non-selective inhibitor isomethyl butyl xanthine (IBMX). Since these assays were performed at a cAMP concentration of 60 nM, the IC ; o values reflect inhibition of the low K", form only. The results are set out in Table 3 with values shown in micromolar units.

Table 3 PDE8 Inhibition with Isozyme-specific PDE Inhibitors Compound Target PDE ICso for IC.,, for Fold Family Target Family PDE8 Difference IC224 PDE1 0.08-0.008 2.7 38-338 EHNA PDE2 2 65 31 Cilostamide PDE3 0.02 12 750 IC197 PDE4 0.02 14 714 DMPPO PDE5 0.016 1.1 66 IBMX Non-selective 1-40 4.6 0.12-4.6 The ICso values for each of the selective inhibitors were at least 30 times higher against PDE8 than against their target isozymes which suggests that the inhibitory profile of PDE8 is distinct from that of PDEs 1-5. The hydrolysis of cAMP and cGMP clearly distinguishes the enzymatic activity of PDE8A from that of PDE6 and PDE7A. The IC^o of the non-selective inhibitor IBMX for PDE8 was in the range observed for known human PDEs suggesting that the catalytic site of PDE8 resembles those of other human and mammalian PDEs and is distinct from lower eukaryotic forms that are insensitive to IBMX.

Example 6 Northern Analysis of PDE8A Expression Northern analysis of PDE8A expression was carried out using a human multiple tissue blot (Clontech, Palo Alto, CA). The 327 base probe was extended from nucleotide 1767 to nucleotide 2293 in SEQ ID NO: 3. Riboprobe preparation and hybridization conditions were as previously described [Loughney, et al. supra].

Results showed a 9.5 kb mRNA in all tissues examined but band intensity varied. The signal was strongest in heart, brain, and kidney; the signal was weaker in liver, placenta, pancreas, and skeletal muscle. The signal was weakest in lung.

Example 7 Chromosome Mapping of Human PDE8A Yeast artificial chromosomes (YACs) containing the human PDE8A gene were isolated from a panel of human YACs purchased from Research Genetics and screened by PCR as follows.

The YAC super-pools were screened with two nested pairs of primers.

In the first screening reaction, sense primer W48S8 (SEQ ID NO: 36) was paired with the anti-sense primer W48A10 (SEQ ID NO: 37). PCR was carried out with 10 mM Tris-HCI, pH 8.3,50 mM KC1,2 mM MgSO4,0.2 mM of each dNTP, 10, ug/ml of each primer, 0.5 units of Taq polymerase (Perkin-Elmer) and 1.5 jul of YAC pool DNA as template. Reactions were carried out for 30 cycles, each cycle consisting of one minute at 94°C, two minutes at 60°C, and four minutes at 72°C. After the first round of amplification, the reaction products were reamplified with the internal pair of primers W48S12 (SEQ ID NO: 36) and W48A12 (SEQ ID NO: 37).

W48S12 SEQ ID NO: 36 CCAGAAGGGGTACTTTTCC W48A12 SEQ ID NO: 37 CATTGTCCTGAGGCTGTGG

The reactions were carried out as described above except that the template was 1, ul of a 1: 10 dilution (in water) of the first round reaction. Super-pools yielding the correct size PCR product were identified and the corresponding sub-pools were screened with the same nested pairs of primers under the same conditions to identify unique addresses for YACs containing PDE8A.

Yeast strains harboring the relevant YACs were purchased from Research Genetics. In order to verify the presence of the PDE8A gene in the various YACs, DNA was prepared from each strain and analyzed by PCR with primers W48S8 and W48A10. DNA was prepared from each strain according to a method previously described [Hoffman and Winston, Gene 57 : 267-272 (1987)] but modified as follows.

Strains were grown overnight at 30°C in YEP media containing glucose. Ten ml of culture was pelleted by centrifugation and resuspended in 200, ul of aqueous buffer containing 10 mM Tris-HCI, pH 8.0,100 mM NaCl, 1 mM Na2EDTA, 1 % SDS, and 2% Triton-X100. The cells were lysed by vortexing in the presence of 200 il of phenol/chloroform (1: 1 mixture) and 100 IAI of glass beads (425-600, um). Following lysis, 200, ul of TE Buffer (10 mM Tris, pH 8.0,1 mM Na2EDTA) was added and the sample was centrifuged to separate the phases. The organic phase was extracted again with 200, ul of aqueous buffer. The pooled aqueous phase was treated with 100 units of bovine pancreatic RNase (Boehringer Mannheim) for 1 hour at 37°C and the sample was extracted with phenol/chloroform, re-extracted with chloroform, and ethanol precipitated according to established methods. The resultant pellet was resuspended in 50, ul TE Buffer. PCR was carried out as described above except that the reaction volume was 25 Itl and the template consisted of 1, ul of the relevant yeast DNA preparation.

Three human YACs containing the PDE8 gene were identified with addresses 805B6,919H10 and 920A3 (as per the CEPH designation). According to information in the Center for Genome Research database (Whitehead), the three YACs overlap one another and are part of a singly-linked contig (WC6.16) on human chromosome 6. Two sequence tagged sites within this contig (D6S305 and D6S411)

have been placed on the chromosomes 6 genetic map at a position 167 cM from the end of 6p in work at the Center for Genome Research; D6S305 has been mapped to a position 173 cM from the end of 6p in work at CEPH-Genethon. Three other YACs within the WC6.16 contig (932F1,956B1 and 947D5) have been mapped by florescence in situ hybridization at CEPH-Genethon. The hybridization signals fall between 0.94-and 0.99 fractional length units from the end of 6p. According to the <BR> <BR> <BR> <BR> CEPH integrated summary map [Chumakov et al., Nature 377 (Supp): 175-297 (1995)], this region corresponds to the cytogenetic region 6q26-27.

Heritable defects that have been associated with this region of the human genome include retinal cone degeneration (OMIM database), Insulin-dependent diabetes <BR> <BR> <BR> <BR> mellitus [Davies et al. Nature 371: 130-136 (1994); Luo et al. Am. J. Hum. Genet.<BR> <BR> <BR> <BR> <BR> <P>57 : 911-919 (1995)] and juvenile onset parkinsonism [Matsumine et al. Am. J. Hum.

Genet. 60: 588-596 (1997)]. In addition, loss of heterozygosity (LOH) is frequently observed in this region in a variety of different cancer cells, including Burkitt's <BR> <BR> <BR> <BR> lymphoma [Parsa et al. Genes, Chromosomes & Cancer 9 : 13-18 (1994)], astrocytoma<BR> <BR> <BR> <BR> <BR> [Liang et al. Neurology 44: 533-536 (1994)], gastric carcinoma [Queimado et al.

Genes, Chromosomes & Cancer 14: 28-34 (1995)], parathyroid adenoma [Tahara et al.

Cancer Res. 56: 599-605 (1996)] and ovarian carcinoma [Cooke et al. Genes, Chromosomes & Cancer 15: 223-233 (1996); Saito et al. Cancer Res. 56: 5586-5589 (1996)]. LOH has been suggested to indicate the presence of a tumor suppressor gene in the affected region [Weinberg, Science 254: 1138-1146 (1991)]. Due to its widespread expression, it is possible that mutation of the PDE8 gene may be involved in all or some of these genetic abnormalities.

Example 8 Verification that PDE8A1 and PDE8A2 Represent Splice Variants and Efforts to Extend the 5'Sequence of PDE8A2 To verify that PDE8A1 and PDE8A2 represent 5'splice variants, two approaches were taken. First, PCR analysis revealed that, in genomic DNA, neither PDE8A1 nor PDE8A2 sequences were adjacent the DNA sequence of the common

region. The genomic sequences upstream of the common region were present in a third PDE8A cDNA, FB74b, which was identified in the group of six original clones that hybridized to the 5'end of probe W04835 described in Example 2. The partial sequence (755 nucleotides at the 3'end) of clone FB74b is set out in SEQ ID NO: 39. The FB74b cDNA diverged from FB85c-2 and FB66a at the same position as FB85c-2 and FB66a diverged from each other, but the FB74b clone did not maintain the open reading frame.

In the FB74b sequence 5'to the point of sequence divergence from the FB66a and FB85c-2 clones, an in-frame stop codon was closer to the point of divergence than an initiating methionine codon indicating that, if FB74b represented a cDNA rather than an unspliced precursor, the initiating methionine would necessarily be located in the sequence common to both FB66a and FB85c-2.

PCR analysis was performed using one primer designated FB74bS 1 (SEQ ID NO: 40) within the FB74b upstream sequences and a second primer designated W48A9 (SEQ ID NO: 11) within the sequences common to FB74b, FB66a, and FB85c-2.

FB74bS1 GTTAGATGAGAGGTTGCTGG SEQ ID NO: 40 Using 1 llg of human genomic DNA as template, a band was amplified having the same size as the one amplified using FB74b as template, indicating that the sequences unique to FB74b and the common region were adjacent in genomic DNA. Thus, the FB74b sequence may represent an unspliced intron or may represent a third splice variant that would encode a protein with an initiating methionine within the common region. In either case, the FB85c-2 and FB66a sequences are presumably generated by splicing.

Secondly, 5'RACE analysis was performed using RNA isolated from human cortex, cerebellum, heart, liver and lung tissues. RNA was isolated from frozen tissue fragments as described [Loughney et al, J. Viol. Chem. 271: 796-806 (1996)] and poly As mRNA was selected using the Fast Track Tm mRNA isolation system (Invitrogen).

Double stranded cDNA was prepared using 5 llg poly A+ mRNA and a cDNA synthesis

kit (Boehringer Mannheim). The cDNA was ligated to a linker formed by annealing oligonucleotides L15 (SEQ ID NO: 41) and L30 (SEQ ID NO: 42).

L 15 GTATGCTAATCTCAG SEQ ID NO: 41 L30 CAACTCGAATTCCTTGACAGATTAGCATAC SEQ ID NO: 42 For the 5'RACE, the linker-ligated cDNA was amplified by PCR using oligonucleotides L18 (SEQ ID NO: 43) and W48A13 (SEQ ID NO: 44).

L18 CAACTCGAATTCCTTGAC SEQ ID NO: 43 W48A13 GTTGTTCTTCCTCTTCAGCC SEQ ID NO: 44 The reaction contained 10 mM Tris-HCI, pH 8.3,50 mM KCI, 1.5 mM MgCL ;, 0.2 mM of each dNTP, 10 pg/ml of each primer and 1 ul of linker-ligated cDNA in a reaction volume of 25 pl. Following heating step at 94°C, PCR was initiated by the addition of 0.1 unit of Taq polymerase (Boehringer Mannheim) and continued with 30 cycles of one minute at 94°C, two minutes at 60°C, and four minutes at 72°C.

The products of the PCR reaction were diluted ten-fold with water and used as template in a second PCR reaction with oligonucleotides L21 (SEQ ID NO: 45) and W48A9S (SEQ ID NO: 46) under the same conditions described above.

L21 CAACTCGAATTCCTTGACAGA SEQ ID NO : 45 W48A9S GATCGTCGACCTGTCTCTGCACTAACAC SEQ ID NO: 46 DNA amplified in the second PCR reaction was cleaved with EcoRI and Sall and ligated into the vector Bluescript (Stratagene) previously digested with the same enzymes.

Initially, DNA sequences in five plasmids from each tissue source were examined and both PDE8A1 and PDE8A2 5'sequences were found among the cDNAs isolated. FB74b 5'sequences were also obtained, as were several sequences, each isolated only once, that could represent yet additional splice variants or unrelated DNA sequences.

Because none of the PDE8A2-like cDNAs extended further 5'than did the original FB66a cDNA, additional PDE8A2 RACE clones were analyzed in an attempt to extend the 5'end sequence. An additional five lung PDE8A2 cDNAs were identified and sequenced, but none extended the PDE8A2 sequence.

A second round of RACE PCR was repeated using the L21 primer (SEQ ID NO: 45) with primer W48A14S (SEQ ID NO: 47).

W48A14S SEQ ID NO: 47 GATCGTCGACAAGCACTCGGTCAGCCTTCG The resultant clones were screened by PCR and the longest ones were chosen for sequencing. Only two clones were longer than the original FB66a cDNA and they extended the 5'sequence 8 and 12 bp, respectively, in the untranslated region. The FB66a sequences were extended with 5'-CCCAGGGCGCCA. The extreme 5'end of FB66a is very GC rich which may contribute to the difficulty in isolating full length cDNAs.

Example 9 Expression and Characterization of PDE8A The recombinant PDE8A described in Example 5 existed in both low affinity and high affinity forms in yeast extract. Because of the possibility that the low affinity form represented partially inactive enzyme, PDE8A expression was carried out in sf9 and COS cells in an attempt to either obtain a homogeneous enzyme or determine if the two kinetic forms are always expressed from the cDNA.

The PDE8 sf9 expression construct was generated with a 3'KpnI-SalI fragment from plasmid W485.1 (described in Example 5) and a 5'fragment generated by PCR as follows. The primers FLAG-1 (SEQ ID NO: 48) and W48A4 (SEQ ID NO: 12) were used in PCR with PDE8 COS-1 DNA (described below) as template.

FLAG-1 GATCGGATCCACCATGGACTACAAGG SEQ ID NO: 48

PCR was performed as described in Example 8 except that 2 mM MgSO4 was used in place of MgCl2 and 0.02 U Taq polymerase was used. Following a four minute initial incubation at 94°C, 30 cycles were performed with one minute at 94°C, one minute at 50°C, and two minutes at 72°C. The 5'amplification product was cleaved with BamHI and KpnI, gel purified, and ligated with the 3'fragment into vector pFASTBAC (Gibco BRL, Gaithersburg, MD) previously digested with BamHI and SalI. The resulting plasmid was designated pFBRPDE8. All PCR amplification products and all new junctions were verified by sequencing.

Recombinant viral stocks were produced using the FastBac system (Gibco BRL) according to the manufacturer's suggested protocol and protein expression was carried out as follows. Sf9 cells were grown at 27°C in CCM3 media (Hyclone, Logan, UT) containing 50 U/ml penicillin and 50, ug/ml streptomycin sulfate (Gibco).

Exponentially growing cells were infected at a multiplicity of approximately two virus per cell and incubated for 48 hours. Cells were collected by centrifugation, washed with CMF-PBS (2.7 mM KCl, 1.5 mM KH., P041 137 mM NaCl, 8.1 mM Na2PO4), and the pellets were frozen and stored at-80°C until use. Cells were lysed in buffer (50 mM MOPS pH 7.2,10, ut zinc sulfate, 1 mM DTT, 2 mM benzamidine, 10 Itg/ml each pepstatin, leupeptin, and aprotinin, and 20 yg/ml each calpain I and calpain II inhibitors) by vortexing in the presence of an equal volume of glass beads (acid washed, 0.5 mm, Sigma) and PDE activity was determined as described in Example 5.

In the sf9 extract, 45.4 nmol/min/mg PDE activity was detected for cAMP hydrolysis (100 AM substrate) and 69.4 nmol/min/mg for cGMP hydrolysis (100, ut substrate). The background PDE activity was negligible. The PDE8A activity appeared to be a mixture of high and low affinity forms as detected in yeast extracts as described in Example 5.

For expression in COS cells, PDE8 COS-1 was generated by combining <BR> <BR> <BR> <BR> a 3'KpnIlSalI fragment from plasmid W485.1 (Example 5) and a NheIlKpnI fragment obtained by cleavage of a PCR amplification product from a reaction including FB66a

cDNA as a template with primers W48A2 (SEQ ID NO: 10) and ATG (SEQ ID NO: 35). Conditions for the PCR included an initial incubation for four minutes at 94°c followed by 30 cycles of one minute at 94°C, one minute at 50°C and two minutes at 72°C in a Perkin Elmer Cetus DNA thermal cycler. The resulting 5'fragment and the 3'fragment described above were ligated into vector pClneo (Promega, Madison, WI) which had been previously digested with NheI and Sal ; I.

Semi-confluent COS cells growing in 15 cm dishes were washed once with 25 ml DMEM (Dulbecco's Modified Eagle Media, 100 U/ml penicillin and 100 jMg/ml streptomycin sulfate, GIBCO), after which 14 ml of DMEM/DEAE- dextran/chloroquine was added per plate. DMEM/DEAE dextran/chloroquine is comprised of 75 ml DMEM and 30 Iti 0.25 M chloroquine in PBS (2.7 mM KCl, 1.5 mM KH2PO4,137 mM NaCI, 8.1 mM Na2PO4,0.9 mM CaCI2 0.5 mM MgCl), together with 0.75 ml 50, ug/ml DEAE-dextran (Pharmacia, Uppsala, Sweden).

Twenty, ug of plasmid DNA in 135 fi » l Tris/EDTA buffer (TE) added per plate and the plates were incubated for two hours at 37°C in 5 % CO2. The media was removed and 12 ml of 10% DMSO/PBS was added for one minute and removed. The cells were washed once with 25 ml DMEM, after which another 25 ml of DMEM containing 10% fetal calf serum (Hyclone, Logan, UT) was added and the cells were incubated overnight at 37°C in 5% CO2. The media was removed and the monolayer was washed with 25 ml of CMF-PBS. Six ml of a solution containing 0.05% trypsin/0.5 mM EDTA (Gibco) was added and the cells were incubated five minutes at 37°C.

Cells were removed from the plates by trituration and transferred to conical centrifuge tubes. The plates were washed with six ml of complete DMEM to harvest any remaining cells and the wash solution was added to the centrifuge tubes. Cells were pelleted by centrifugation for five minutes at approximately 340 x g, resuspended in five ml complete DMEM, removed to a 15 cm tissue culture dish containing 20 ml complete DMEM, and incubated overnight in 5 % CO2.

The monolayer was washed two times with CMF-PBS, incubated five minutes at 37°C in versene (0.5 mM Na2EDTA 2H2O, 137 mM NaCl, 2.68 mM KCl,

8.1 mM Na2HPO4,1.1 mM glucose, pH 7.4), and harvested as described above.

Pelleted cells were washed with CMF-PBS, frozen in dry ice, and stored at-80°C until use. Cells were lysed in buffer (50 mM MOPS, pH 7.2,10 juM zinc sulfate, 1 mM DTT, 2 mM benzamidine, 10, ag/ml each pepstatin, leupeptin, and aprotinin, and 20 , ug/ml each calpain I and calpain II inhibitors) by passage through a French pressure cell (SLM Instruments) at 20,000 psi and PDE activity was determined as described in Example 5.

PDE8A expression was low in the COS cell extract and could not be accurately characterized due to the high level of background activity from endogenous PDEs. In order to more fully characterize the COS cell expression product, the enzyme including a FLAG tag at the amino terminus (Example 5) is purified from a 100,000 x g supernatant of cell extract using an anti-FLAG M2 affinity column (Sigma) according to the manufacturer's suggested protocol. In order to more accurately characterize yeast PDE8A activity, expression of a recombinant protein that is truncated at the amino terminus but retains the catalytic region is carried out as described in Example 5 in an attempt to obtain a homogenous protein.

Example 10 Production of Anti-PDE8A Antibodies A GST fusion protein was produced in E. coli to provide an antigen for <BR> <BR> <BR> <BR> generation of monoclonal antibodies to PDE8A. An EcoRI fragment from FB70a (a PDE8A cDNA that includes nucleotides 182-1330 of FB85c-2 and which was one of the nine clones originally identified which hybridized to the full length W04835 probe <BR> <BR> <BR> <BR> described in Example 2) was inserted into the EcoRI site of pGEXSXl (Pharmacia) and the resultant construct was transformed in the E. coli strain XL1 Blue. A GST-PDE8A fusion protein including 382 amino acids from PDE8A was produced from this construct following induction with IPTG. The fusion protein was isolated using SDS-PAGE, the band of appropriate size excised from the gel following staining with cold 0.4 M KC1, and the protein obtained from the acrylamide by electroelution. The

elution product was dialyzed against PBS and concentrated using Centriprep 10 and Centricon columns (Amicon, Beverly MA) prior to being injected into mice.

On day 0, four Balb/c mice were pre-bled and injected subcutaneously with a panel of antigens including 30, ug/mouse GST-PDE8 fusion protein in complete Freund's adjuvant in 200 nul total volume. The same injections were repeated at weeks three and nine in incomplete Freund's adjuvant. Ten days after the last immunization, test bleeds were obtained and screened by antigen capture ELISA and Western analysis.

In the ELISA, Immulon 4 plates (Dynex, Cambridge, Massachusetts) were coated at 4°C with 50, I/well of a solution containing 2, ug/ml GST-PDE8 in 50 mM carbonate buffer, pH 9.6. Plates were blocked with 0.5% fish skin gelatin (Sigma) for 30 minutes and 50, ul serum diluted in PBS with 0.5 % Tween 20 (PBST) was added. Serum dilutions ranged from 1: 100 to 1: 102,400 and were obtained by a series of doubling dilutions. After incubation at 37 ° C for 30 minutes and washing three times with PBST, 50 itl of horseradish peroxidase-conjugated goat anti-mouse IgG (fc) antibody (Jackson) (diluted 1: 10000 in PBST) was added. Plates were incubated as above and washed four times with PBST. Antibody was detected with addition of tetramethyl benzidine (Sigma Chemical, St. Louis, Missouri) and the color reaction was stopped after five minutes with the addition of 50 µl of 15 % H2SO4.

Absorbance at 450 nM was measured on a plate reader.

For Western analysis. SDS-PAGE gels were run with approximately 10 itg yeast PDE8 extract and approximately 200 ng of gel-purified GST-PDE8 and the proteins were transferred to Immobilon-PVDF. A standard enhanced chemiluminescence (ECL) Western blot protocol was performed using BioRad goat anti-mouse IgG horseradish peroxidase as the secondary antibody.

In preparation of hybridomas, splenocytes from mice giving a positive result from the ELISA and/or Western blotting protocols above, were fused to NS-1 cells in a ratio of 5: 1 by standard methods using polyethylene glycol 1500 (Boehringer Mannheim) (Harlow and Lane, Antibodies, a Laboratory Manual, Cold Spring Harbor Laboratory, 1988). The fused cells were resuspended in 200 ml RPMI containing 15 %

FBS, 100 mM sodium hypoxanthine, 0.4 mM aminopterin, 16 mM thymidine (HAT) (Gibco), 25 units/ml IL-6 (Boehringer Mannheim) and 1.5 x 106 murine thymocytes/ml and dispensed into ten 96-well flat bottom tissue culture plates (Corning, United Kingdom) at 200 ul/well. Cells were fed on days 2,4, and 6 days post fusion by aspirating approximately 100 ttl from each well with an 18 G needle (Becton Dickinson) and adding 100, ul/well plating medium described above except containing 10 units/ml IL-6 and lacking thymocytes. On days 9 to 12, supernatants from the fusion wells were screened by antigen capture ELISA using GST and GST-PDE8 and by ECL Western analysis as described above.

A positive signal of the expected size was obtained on both lanes of the Western blot using mouse blood and a monoclonal antibody with very weak reactivity to the yeast recombinant protein was obtained in the subsequent fusion. The entire procedure is repeated using 50 lAg antigen/mouse to obtain more strongly immunoreactive monoclonal antibodies.

Example 11 Analysis of PDE8A Expression by in situ Hybridization Expression of PDE8A was examined in tissue sections by in situ hybridization as described below.

Preparation of probe An XhollEcoRI restriction enzyme fragment from the cDNA FB70a (corresponding to nucleotides 571 to 1226 of SEQ ID NO: 1) was subcloned into a Bluescript vector (Stratagene, La Jolla, CA) to generate an expression plasmid designated PDE8XR2A. The plasmid was cleaved with Xhol and transcribed (see below) with T3 polymerase to generate an antisense probe. A sense probe was <BR> <BR> <BR> generated by cleaving PDE8XR2A with EcoRI and transcribing with T7 polymerase.

The PDE8A templates were transcribed using a RNA Transcription kit (Stratagene, La Jolla, CA) in a reaction containing 5, ul of 5X transcription buffer (Stratagene), 30 mM DTT (Stratagene), 0.8 mM each ATP, CTP, GTP (10 mM (Stratagene), 40 U RNase

Block II (Stratagene), 12.5 U T3 or T7 polymerase (Stratagene), and 300 ng linearized plasmid template, 50 juCi 35S-UTP (greater than 1000 Ci/mmol, Amersham, Arlington Heights, IL). The mixture was incubated at 37°C for one hour after which the template DNA was removed by addition of 1 il ouf RNase-free RNase-free DNase (Stratagene) and incubation for 15 minutes at 37°C. The probe was hydrolyzed by adding 4 ul 1 M NaHCO3 and 6 µl 1 M Na2CO3 for 22 minutes at 60°C and the reaction mixture was neutralized by addition of 25 il of a solution containing 100 lul 3 M sodium acetate, 5 µl acetic acid (VWR, So. Plainfield, NJ), and 395 µl dH2O. A Quick Spin G50 RNA column (5'-3'Inc., Boulder, CO) was prepared according to the manufacturer's suggested protocol. The probe was placed in the center of the column and the column centrifuged for four minutes at 1,000 rpm in a desk top centrifuge. The column flow- through was mixed with 50 » 1 dH2O, 2 Shl of a 10 mg/ml tRNA solution, 10 µl 3 M sodium acetate, and 200/1 100% ethanol (VWR) and the resulting mixture was incubated at-20°C overnight. The probe solution was microfuged for 15 minutes at 4°C, the supernatant was removed, and the pellet was resuspended in 40 Itl 1X TBE containing 1, ul of 0.1 M DTT. The probe was stored at-70°C until the in situ hybridization assay was performed.

Preparation of tissue samples and in situ hybridization Tissues (National Disease Research Interchange, Philadelphia, PA and Cooperative Human Tissue Network, Philadelphia, PA) were sectioned at 6 item and placed on Superfrost Plus slides (VWR). Sections were fixed for 20 minutes at 4°C in 4% paraformaldehyde (Sigma, St. Louis, MO). The slides were rinsed in three changes of 1X CMF-PBS, dehydrated with three successive washes with 70% ethanol, 95% ethanol and 100% ethanol, and dried for 30 minutes at room temperature. The slides were placed in 70% formamide (J. T. Baker) in 2X SSC for two minutes at 70°C, rinsed in 2X SSC at 4°C, dehydrated through 70%, 95% and 100% ethanol washes, and dried for 30 minutes at room temperature.

A prehybridization step was performed by placing the slides in an Zht box containing a piece of filter paper saturated with box buffer containing 50% formamide (J. T. Baker) in 4X SSC. Each section was covered with 100 1 of 2 butter consisting of 1096 dextran sulfate (Sigma), 50% formamide (J. T. Baker, Phillpsburg, NJ), 100 mM DIT (Boehringer Mannheim, Indianapolis, IN), 0.3 M <BR> <BR> <BR> <BR> NaCl (Sigma), 20 mM Tris, pH 7.5.5 mM EDTA (Sigma), and 1X Denhardt's solution (Sigma) and the slides were incubated at 42°C for 1 hour. The probe, as described above, was prepared by mixing 4 x 105 cpm/tissue section with S g1 of a 10 mg/ml tRNA solution per section and heating the mixture at 95°C for three minutes.

Ice cold rHB2 buffer was added to bring the final volume to 20 µl/section. The probe- <BR> <BR> <BR> <BR> containing solution (20 µl/section) was added to 100 FI rHB2 buffer previously applied. The slides were incubated at 55°C for 12 to 16 hours. Following hybridization, the slides were washed once in 4X SSC containing 10 mM DTT for one hour at room temperature, once in 50' deionized fonnamide (J. T. Baker), 1X SSC, <BR> <BR> <BR> and 1 mM DTT for 40 minutes at 60 C. once in 2X SSC for 30 minutes at room oncein0.1Xtemperature;and SSC minutesatroomtemperature.Thesections30 were dehydrated through 70%, 95%, and 100% ethanol washes and air dried for 30 minutes. The slides were dipped in Kodak NTB2 nuclear emulsion, dried for one to three hours at room temperature in! he dark and stored in the dark at 4°C with d until time of development. The slides were developed in 4°C Kodak Dektol developer for four minutes, dipped four times in 4°C dH2O, and placed in 4°C Kodak fixer for four minutes. The slides were rinsed in dH, O and a standard H&B étain was ars as follows.

The slides were rinsed in dH2O and stained with hematoxylin and eosin <BR> <BR> <BR> <BR> by tender of the slides through a series of the following step: five minutes in formaldehydelalcohol (100 ml formaldehyde, 900 ml 80% ethanol); three rinses in <BR> <BR> <BR> <BR> water for a total of two minutes: five minutes in 0. 75% Harris hematoxylin (Sigma) ; three rinses in water for a total of two minutes; one dip in 1 % HCl/50% ethanol; one rince in water; four dips in 1 % lithium carbonate; ten minutes in tap water; two

minutes in 0.5 % eosin (Sigma); three rinses in water for a total of two minutes; two minutes in 70 % ethanol; three one minute rinses in 95 % ethanol; two one minute rinses in 100% ethanol; and two two minutes rinses in xylene. Slides were mounted with cytoseal 60 (Stephens Scientific, Riverdale, NJ).

The signals obtained with an antisense PDE8A probe were compared to the control signals generated by a sense PDE8A probe and any signal specific to the antisense probe was assumed to represent PDE8A expression. PDE8A signal was detected throughout much of the cerebellum, in a subset of cells in the seminiferous tubules of the testes, on scattered cells of yet undetermined origin in skeletal muscle, in granulosa cells and ovarian stroma in the ovary, in epithelial cells in the loop of Henle in the kidney and on the smooth muscle of some arterioles in the heart.

These results differ from those obtained by Northern blotting and described in Example 6 in that a moderate signal was detected in heart by Northern blot while the in situ data using this heart sample gave a weak signal. The inconsistency could reflect differences in the tissues from different individuals or level of detection differences inherent in the two methods. The signal in the ovary and the signal in the kidney may indicate that PDE8A is involved in ovulation or in salt and/or water homeostasis, respectively.

Numerous modifications and variations in the invention as set forth in the above illustrative examples are expected to occur to those skilled in the art.

Consequently only such limitations as appear in the appended claims should be placed on the invention.