- Email: [email protected]
Isolation and characterization of human cDNAs encoding a cGMP-stimulated 3′,5′-cyclic nucleotide phosphodiesterase1
Gene 191 (1997) 89–95
Isolation and characterization of human cDNAs encoding a cGMP-stimulated 3∞,5∞-cyclic nucleotide phosphodiesterase 1 Guy J. Rosman a, Timothy J. Martins a, William K. Sonnenburg b, Joseph A. Beavo b, Ken Ferguson a, Kate Loughney a,* a ICOS Corporation, 22021 20th Ave. S.E., Bothell, Bothell, WA 98021, USA b Department of Pharmacology, University of Washington, School of Medicine SJ-30, Seattle, WA 98195, USA Received 10 September 1996; received in revised form 3 December 1996; accepted 7 December 1996; Received by C.M. Kane
Abstract Human cyclic GMP-stimulated 3∞,5∞-cyclic nucleotide phosphodiesterase (PDE2A3) cDNAs were cloned from hippocampus and fetal brain cDNA libraries. A 4.2-kb composite DNA sequence constructed from overlapping cDNA clones encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. Extracts prepared from yeast expressing the human PDE2A3 hydrolyzed both cyclic AMP (cAMP) and cyclic GMP (cGMP). This activity was inhibited by EHNA, a selective PDE2 inhibitor, and was stimulated three-fold by cGMP. Human PDE2A is expressed in brain and to a lesser extent in heart, placenta, lung, skeletal muscle, kidney and pancreas. The human PDE2A3 differs from the bovine PDE2A1 and rat PDE2A2 proteins at the amino terminus but its amino-terminal sequence is identical to the bovine PDE2A3 sequence. The different amino termini probably arise from alternative exon splicing of the PDE2A mRNA. © 1997 Elsevier Science B.V. Keywords: PDE2A; EHNA; 5∞-Splice variants
1. Introduction The cyclic nucleotide phosphodiesterases (PDEs) are essential regulatory components of the cellular response to hormones, neurotransmitters and autacoids. There are seven PDE families (PDE1-PDE7) identified by their distinctive catalytic and regulatory properties, amino acid sequences, and sensitivities to PDE inhibitors (Beavo, 1995). PDEs contain an approximately 250 amino acid catalytic region in the carboxy-terminal portion of the protein (Charbonneau, 1990). The cGMP-stimulated PDEs (PDE2) hydrolyze both cAMP and cGMP, although they have a higher affinity for cGMP than for cAMP (Manganiello et al., 1990). The enzyme’s catalytic activity is regulated allosterically * Corresponding author. Tel.: +1 206 4851900; Fax: +1 206 4851961. 1 The nucleotide sequence reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession No. U67733. Abbreviations: cAMP, cyclic adenosine 3∞,5∞-monophosphate; cDNA, DNA complementary to RNA; cGMP, cyclic guanosine 3∞,5∞-monophosphate; Da, dalton; EHNA, erythro-9-(2-hydroxy-3-nonyl )adenine; kb, kilobase(s); nt, nucleotides; PDE, phosphodiesterase; pmol, picomoles; SDS, sodium dodecyl sulfate. 0378-1119/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 03 7 8 -1 1 1 9 ( 9 7 ) 0 0 0 46 - 2
by the binding of cGMP at site(s) remote from the catalytic site thereby increasing the enzyme’s affinity for cAMP. The cGMP-stimulated PDE has been purified from a number of sources (reviewed in Manganiello et al., 1990). It is a dimer of 102–105 kDa subunits and can be found in either the soluble or particulate fractions of tissue homogenates. The amino acid sequence of the bovine heart PDE2A has been determined and a cGMPbinding domain identified (Le Trong et al., 1990; Charbonneau et al., 1986). Bovine adrenal and rat brain cDNAs encoding PDE2A have been isolated (Sonnenburg et al., 1991; Tanaka et al., 1991, Repaske et al., 1992; Yang et al., 1994). Thus far the PDE2 amino acid and cDNA sequences that have been determined are derived from a single gene (PDE2A) in each species. Over most of their lengths the bovine (PDE2A1) and rat (PDE2A2) amino acid sequences were similar. However, PDE2A1 and PDE2A2 contained completely different amino-terminal regions (25 and 37 amino acids, respectively). A third amino-terminal sequence, bovine PDE2A3, has recently been reported (Genbank accession No. L49503: Juilfs, D.M, Sonnenburg, W.K., Seraji, S. and Beavo, J.A., unpublished). These amino-terminal
90
G.J. Rosman et al. / Gene 191 (1997) 89–95
differences are believed to result from alternative exon splicing and it has been suggested that the different amino termini may be responsible for particulate or soluble localization of the enzyme ( Yang et al., 1994; Sonnenburg et al., 1991; Murashima et al., 1990). In this paper we report the isolation of human PDE2A cDNAs (PDE2A3) and a preliminary characterization of the biochemical properties of the recombinant protein.
2. Results and discussion 2.1. Isolation and characterization of human PDE2A3 cDNAs A bovine PDE2A1 cDNA was used as a probe to screen human cDNA libraries. Four different cDNAs were chosen for characterization and were used to assemble a 4.2-kb composite human PDE2A3 cDNA sequence (Fig. 1). This DNA sequence encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. 2.2. Comparison of human PDE2A3 to bovine and rat PDE2A sequences The human protein is similar to the bovine PDE2A1 and rat PDE2A2 proteins although the amino termini of the three sequences diverge from each other ( Fig. 2). Excluding these divergent amino termini the three PDE2A protein sequences are 90% identical to each other. The identity is greater in the cGMP-binding domain (94% identical ) and the PDE catalytic region (98% identical ). A closer examination of the amino terminal sequences revealed that the human PDE2A3 and rat PDE2A2 sequences share an additional block of homology just upstream from the point at which they both diverge from the bovine PDE2A1 sequence. To examine the relationship between these sequences further, we turned to a comparison of the cDNA sequences. We included in the comparison cDNA sequences corresponding to each of the three protein sequences shown in Fig. 2 and also the cDNA sequence for a recently described bovine PDE2A 5∞-splice variant, PDE2A3 (Genbank accession
No. L49503: Juilfs, D.M, Sonnenburg, W.K., Seraji, S. and Beavo, J.A., unpublished ). This cDNA was isolated from a bovine brain cDNA library and is the bovine counterpart of the human PDE2A3 sequence we describe here. A comparison of the 5∞-untranslated regions and the 5∞-ends of the coding regions revealed 5 blocks of DNA sequence that we called A, B, C, common1, and common2 (Fig. 3). Common1 and common2 were blocks of cDNA sequence found in all four cDNAs whereas A, B, and C were sequences found only in a subset of the four cDNAs. The bovine PDE2A1 cDNA contained regions B, common1, C, and common2 (in this order, 5∞ to 3∞) ( Fig. 3). The rat PDE2A2 cDNA contained B, common1 and common2 (5∞ to 3∞). As has been noted ( Yang et al., 1994), the initiator methionine of the PDE2A2 cDNA differs from that of PDE2A1. A consequence of this is that different reading frames of region common1 are used which leads to the two proteins having different amino terminal regions. In summary, regions B, common1 and C encode the amino terminus unique to PDE2A1 and regions B and common1 encode the amino terminus found in PDE2A2. The human and bovine PDE2A3 cDNAs contained regions A, common1 and common2 (5∞ to 3∞) (Fig. 3). The addition of region A yielded a unique amino terminus for the PDE2A3 proteins. The PDE2A3 cDNAs used the same reading frame of region common1 as was used in rat PDE2A2. This provides the block of amino acid sequence that PDE2A3 and PDE2A2 share (and is not found in PDE2A1) (Fig. 2). The amino terminus of rat PDE2A2 is more hydrophobic than the amino termini of the bovine PDE2A1 human PDE2A3 proteins. These differences in hydrophobicity may underlie the partitioning of PDE2A activity between the membrane-associated and soluble fractions of the cell ( Yang et al., 1994). The structure of the genomic DNA encoding the splice variants is not known, except that regions B and common1 are known to be adjacent in the rat genome ( Yang et al., 1994). 2.3. Distribution of PDE2A3 mRNA A Northern blot was probed with human PDE2A to examine the tissue distribution of the human PDE2A
Fig. 1. Human PDE2A3 nucleotide and amino acid sequences. The nucleotide sequence is a composite of 4 cDNAs isolated from human cDNA libraries. The initiator methionine is at nt 162 and the stop codon is at nt 2985. The deduced amino acid sequence is positioned below the open reading frame. Nucleotide and amino acid positions are indicated at the right side of the figure. The four cDNAs align with the sequence as follows: Fbr9.2 (nt 1–3046), Hippo7.1 (nt 2646–4240), Hippo7.4 (nt 2646–4218), Hr6.1 (nt 1820–3848). The polyadenylation site in cDNA Hippo7.4 is 3 nt 5∞ to the polyadenylation site in cDNA Hippo7.1. In the 3∞-untranslated region cDNA Fbr9.2 diverged from the sequence shown here. The divergent sequences were not found in the three other human cDNAs and were not pursued. Methods: Human cDNA libraries were screened with DNA probes and PDE2A cDNAs isolated, characterized, and sequenced on both strands according to published procedures (Loughney et al., 1996). A bovine PDE2A1 probe (Sonnenburg et al., 1991) was used to screen heart (Stratagene, LaJolla, CA), hippocampus (Clontech, Palo Alto, CA), and fetal brain (Stratagene) cDNA libraries to yield the Hr, Hippo, and Fbr cDNAs, respectively.
G.J. Rosman et al. / Gene 191 (1997) 89–95
91
92
G.J. Rosman et al. / Gene 191 (1997) 89–95
Fig. 2. Comparison of human, bovine and rat PDE2A proteins. The human PDE2A amino acid sequence is shown on the top line. Positions where both the bovine (Sonnenburg et al., 1991) and rat PDE2A sequences ( Yang et al., 1994) match the human sequence are shown with dots. Dashes indicated gaps introduced to optimize the alignment. Amino acid residue numbers are indicated on the right. The cGMP binding domain (Le Trong et al., 1990) and the catalytic region (Charbonneau, 1990) are indicated.
G.J. Rosman et al. / Gene 191 (1997) 89–95
93
Fig. 3. Comparison of PDE2A cDNAs. The 5∞ portions of bovine PDE2A1, rat PDE2A2, human PDE2A3 and bovine PDE2A3 cDNAs were compared. A, B, C, common1 and common2 represent blocks of DNA sequence found in the different cDNAs. They are not drawn to scale. For each cDNA the initiator methionine is indicated with Met and represents amino acid position number 1 for each protein. The numbers under the various blocks refer to the position of the amino acids that are represented by each block; amino acid positions are numbered as in Fig. 2. Junctions between regions common1 and C in PDE2A1, regions B and common1 in PDE2A2, and regions A and common1 in PDE2A3 occur within a codon. In region common1, PDE2A1 uses one reading frame while PDE2A2 and PDE2A3 use another (see Section 2.2).
mRNA (Fig. 4). A 4.2-kb mRNA was present in all tissues represented on the filter. The highest level of PDE2A mRNA expression was found in the brain. An intermediate level of expression was detected in heart, liver, skeletal muscle, kidney and pancreas. The lowest levels of PDE2A expression were seen in the lung and placenta. This distribution pattern is similar to that found for rat and bovine PDE2A expression. In particular, of the tissues tested, PDE2A expression was highest in rat
( Repaske et al., 1993) and bovine (Sonnenburg et al., 1991) brain. The Northern blot was subsequently probed for G3PDH mRNA. A hybridizing signal was observed in each sample lane with a much stronger signal in the heart and skeletal muscle RNA (data not shown). Only a single size of PDE2A mRNA was observed. As noted in this section, multiple splice variants have been observed for this gene. It is possible that the multiple mRNAs are indistinguishable in size by Northern analysis. Alternatively, specific tissues may express only one form. Also, though RNase protection studies showed that at least two different 5∞-splice variants are expressed in bovine tissues (Sonnenburg et al., 1991), it is not known if all 3 of the splice variants ( Fig. 3) are used in a single organism. 2.4. PDE2A3 biochemical activity
Fig. 4. PDE2A Northern. A filter with 2 mg poly A+ mRNA from the tissues indicated was purchased from Clontech (Palo Alto, CA) and hybridized with an antisense riboprobe corresponding to nt 2647–2939 of human PDE2A3 (Fig. 1). Probe preparation and hybridization conditions have been described (Loughney et al., 1996). RNA size markers are shown at left. The exposure time for the autoradiogram shown was 24 h.
To determine if the human PDE2A3 encoded a PDE with characteristics typical of a PDE2 it was expressed in a strain of S. cereviseae lacking endogenous PDE activity (see legend to Fig. 5 for methods). PDE activity was measured in extracts from yeast expressing a control (empty vector) and from yeast expressing the recombinant human PDE2A3 protein. Yeast extracts were centrifuged at 14 000×g for 1 min and the supernatant was removed and assayed for PDE activity. Approximately 12% of the total PDE activity in the yeast extract is found in the 14 000×g supernatant. With a substrate concentration of 1 mM, PDE2A3expressing yeast hydrolyzed both cAMP (20.5 nmol min−1 mg−1) and cGMP (28.5 nmol min−1 mg−1). The assay background (empty vector only) was ≤0.25 nmol min−1 mg−1. A band of the predicted size (~105 kDa) was seen in the yeast extract by western blot analysis (data not shown).
94
G.J. Rosman et al. / Gene 191 (1997) 89–95
Cyclic GMP stimulates the hydrolysis of cAMP by PDE2 (Manganiello et al., 1990). Hydrolysis of 1 mM cAMP by the recombinant PDE2A3 was stimulated almost three-fold by the addition of 10 mM cGMP ( Fig. 5B). Although stimulation by cGMP is characteristic of this enzyme, a wide range of values has been reported. A recombinant bovine PDE2A1 assayed in a mammalian cell extract showed approximately 3.5-fold stimulation by cGMP (Sonnenburg et al., 1991). Purified PDE2A proteins showed 5–6-fold stimulation in one report (Martins et al., 1982) and >50-fold stimulation in another (Mumby et al., 1982). The differences in these values may reflect differences in methodology and the enzyme status. Alterations in assay temperature, assay pH, or cGMP concentration and proteolysis of the enzyme can all influence the activation state of the enzyme. Likewise, the presence of fatty acids or shortchain alcohols has been reported to inhibit the activation of cAMP hydrolysis by cGMP (Manganiello et al., 1990). The human recombinant PDE2A3 protein was not purified and one or more of the factors described above may have affected the magnitude of the cGMP stimulation. Purified recombinant protein will be used for more detailed kinetic characterizations.
3. Conclusions Fig. 5. (A) Inhibition of PDE2A3 activity by EHNA. (B) Stimulation of PDE2A3 activity by cGMP. Data from a single experiment are shown. The assays were repeated using the supernatant from a second extract and gave values that differed ≤10% from the values shown in Fig. 5. Methods: The complete open reading frame of PDE2A3 (nt 162–2989, Fig. 1) was inserted into the yeast expression vector pBNY6n (Loughney et al., 1996) to generate pCGS-6n. All new junctions and PCR generated DNA were sequenced to verify the structure of pCGS-6n. pCGS-6n was transformed into the yeast strain YKS45 (Mata his3 leu2 trp1 ura3 Dpde1::HIS3 D pde2::TRP1). Cells were grown and PDE activity was assayed as described (Loughney et al., 1996) with the following exceptions. Lysis buffer consisted of 25 mM Tris pH 8, 5 mM EDTA, 5 mM EGTA, 1 mM benzamidine (Sigma, St. Louis, MO) and 0.1 mM DTT. PBS (0.9 mM, 2.7 mM KH PO , 2 4 137.9 mM NaCl, 8.1 mM Na HPO · 7H O, pH 7.45) was used as the 2 4 2 enzyme dilution buffer. The PDE assay was performed in 40 mM Tris pH 8, 1 mM EGTA, 5 mM MgCl , and 0.1 mg/ml BSA (Fraction V, 2 ICN Biomedicals, Costa Mesa, CA). EHNA (Research Biochemicals International, Natick, MA) was dissolved in DMSO and the final concentration of DMSO in the assay was 0.7%.
EHNA is an adenosine deaminase inhibitor that is also a selective PDE2 inhibitor (Podzuweit et al., 1995). At a substrate concentration of 1 mM, EHNA (10 mM ) reduced hydrolysis in the PDE2A3 yeast extract by 76% (cAMP) and by 87% (cGMP) (Fig. 5A). These results are in reasonable agreement with the IC value of 0.8 50 mM that was observed for EHNA inhibition of soluble PDE2 activity purified by DEAE chromatography from a human myocardial extract (Podzuweit et al., 1995).
(1) The human PDE2A3 cDNA sequence, a composite of 4 cDNA clones, is 4240 nt in length and encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. (2) As expected for a PDE2 enzyme, human PDE2A3 hydrolyzes cAMP and cGMP, is inhibited by EHNA and responds to cGMP with an increase in cAMP hydrolysis. (3) A 4.2-kb message for human PDE2A was found in heart, brain, placenta, lung, skeletal muscle, kidney and pancreas, the most abundant source being brain. (4) PDE2A3 has the third type of amino terminus that has been observed for PDE2A gene products.
Acknowledgement We thank Carmen Hertel for excellent technical assistance, Dina Leviten, Christi Wood and Ernie Tolentino for synthesis of oligonucleotides and DNA sequencing and Vince Florio for critically reading the manuscript.
References Beavo, J., 1995. Cyclic nucleotide phosphodiesterases: functional implications of multiple isoforms. Physiol. Rev. 75, 725–748.
G.J. Rosman et al. / Gene 191 (1997) 89–95 Charbonneau, H., Beier, N., Walsh, K.A. and Beavo, J.A., 1986. Identification of a conserved domain among cyclic nucleotide phosphodiesterases from diverse species. Proc. Natl. Acad. Sci. USA 83, 9308–9312. Charbonneau, H., 1990. Structure-Function Relationships Among Cyclic Nucleotide Phosphodiesterases. In: Beavo, J. and Houslay, M.D. ( Eds.), Isoenzymes of Cyclic Nucleotide Phosphodiesterases. John Wiley and Sons, New York, pp. 267–296. Le Trong, H., Beier, N., Sonnenburg, W.K., Stroop, S.D., Walsh, K.A., Beavo, J.A. and Charbonneau, H., 1990. Amino acid sequence of the cyclic GMP stimulated cyclic nucleotide phosphodiesterase from bovine heart. Biochemistry 29, 10280–10288. Loughney, K., Martins, T.J., Harris, E.A.S., Sadhu, K., Hicks, J.B., Sonnenburg, W.K., Beavo, J.A. and Ferguson, K., 1996. Isolation and characterization of cDNAs corresponding to two human calcium, calmodulin-regulated, 3∞,5∞-cyclic nucleotide phosphodiesterases. J. Biol. Chem. 271, 796–806. Manganiello, V.C., Tanaka, T. and Murashima, S., 1990. Cyclic GMPstimulated cyclic nucleotide phosphodiesterases. In: Beavo, J. and Houslay, M.D. ( Eds.), Isoenzymes of Cyclic Nucleotide Phosphodiesterases. John Wiley and Sons, New York, pp. 61–85. Martins, T.J., Mumby, M.C. and Beavo, J.A., 1982. Purification and characterization of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from bovine tissues. J. Biol. Chem. 257, 1973–1979. Mumby, M.C., Martins, T.J., Chang, M.L. and Beavo, J.A., 1982. Identification of cGMP-stimulated cyclic nucleotide phosphodiesterase in lung tissue with monoclonal antibodies. J. Biol. Chem. 257, 13283–13290.
95
Murashima, S., Tanaka, T., Hockman, S. and Manganiello, V., 1990. Characterization of particulate cyclic nucleotide phosphodiesterases from bovine brain: purification of a distinct cGMP-stimulated isoenzyme. Biochemistry 29, 5285–5292. Podzuweit, T., Nennstiel, P. and Muller, A., 1995. Isozyme selective inhibition of cGMP-stimulated cyclic nucleotide phosphodiesterases by erythro-9-(2-hydroxy-3-nonyl ) adenine. Cell. Signaling 7, 733–738. Repaske, D.R., Corbin, J.G., Conti, M. and Goy, M.F., 1993. A cyclic GMP-stimulated cyclic nucleotide phosphodiesterase gene is highly expressed in the limbic system of the rat brain. Neuroscience 56, 673–686. Repaske, D.R., Swinnen, J.V., Jin, S.-L.C., Van Wyk, J.J. and Conti, M., 1992. A polymerase chain reaction strategy to identify and clone cyclic nucleotide phosphodiesterase cDNAs. J. Biol. Chem. 267, 18683–18688. Sonnenburg, W.K., Mullaney, P.J. and Beavo, J.A., 1991. Molecular cloning of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase cDNA. J. Biol. Chem. 266, 17655–17661. Tanaka, T., Hockman, S., Moos Jr., M., Taira, M., Meacci, E., Murashima, S. and Manganiello, M., 1991. Comparison of putative cGMP-binding regions in bovine brain and cardiac cGMP-stimulated phosphodiesterases. Second Messengers Phosphoproteins 13, 87–98. Yang, Q., Paskind, M., Bolger, G., Thompson, W.J., Repaske, D.R., Cutler, L.S. and Epstein, P.M., 1994. A novel cyclic GMP stimulated phosphodiesterase from rat brain. Biochem. Biophys. Commun. 205, 1850–1858.
Isolation and characterization of human cDNAs encoding a cGMP-stimulated 3∞,5∞-cyclic nucleotide phosphodiesterase 1 Guy J. Rosman a, Timothy J. Martins a, William K. Sonnenburg b, Joseph A. Beavo b, Ken Ferguson a, Kate Loughney a,* a ICOS Corporation, 22021 20th Ave. S.E., Bothell, Bothell, WA 98021, USA b Department of Pharmacology, University of Washington, School of Medicine SJ-30, Seattle, WA 98195, USA Received 10 September 1996; received in revised form 3 December 1996; accepted 7 December 1996; Received by C.M. Kane
Abstract Human cyclic GMP-stimulated 3∞,5∞-cyclic nucleotide phosphodiesterase (PDE2A3) cDNAs were cloned from hippocampus and fetal brain cDNA libraries. A 4.2-kb composite DNA sequence constructed from overlapping cDNA clones encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. Extracts prepared from yeast expressing the human PDE2A3 hydrolyzed both cyclic AMP (cAMP) and cyclic GMP (cGMP). This activity was inhibited by EHNA, a selective PDE2 inhibitor, and was stimulated three-fold by cGMP. Human PDE2A is expressed in brain and to a lesser extent in heart, placenta, lung, skeletal muscle, kidney and pancreas. The human PDE2A3 differs from the bovine PDE2A1 and rat PDE2A2 proteins at the amino terminus but its amino-terminal sequence is identical to the bovine PDE2A3 sequence. The different amino termini probably arise from alternative exon splicing of the PDE2A mRNA. © 1997 Elsevier Science B.V. Keywords: PDE2A; EHNA; 5∞-Splice variants
1. Introduction The cyclic nucleotide phosphodiesterases (PDEs) are essential regulatory components of the cellular response to hormones, neurotransmitters and autacoids. There are seven PDE families (PDE1-PDE7) identified by their distinctive catalytic and regulatory properties, amino acid sequences, and sensitivities to PDE inhibitors (Beavo, 1995). PDEs contain an approximately 250 amino acid catalytic region in the carboxy-terminal portion of the protein (Charbonneau, 1990). The cGMP-stimulated PDEs (PDE2) hydrolyze both cAMP and cGMP, although they have a higher affinity for cGMP than for cAMP (Manganiello et al., 1990). The enzyme’s catalytic activity is regulated allosterically * Corresponding author. Tel.: +1 206 4851900; Fax: +1 206 4851961. 1 The nucleotide sequence reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession No. U67733. Abbreviations: cAMP, cyclic adenosine 3∞,5∞-monophosphate; cDNA, DNA complementary to RNA; cGMP, cyclic guanosine 3∞,5∞-monophosphate; Da, dalton; EHNA, erythro-9-(2-hydroxy-3-nonyl )adenine; kb, kilobase(s); nt, nucleotides; PDE, phosphodiesterase; pmol, picomoles; SDS, sodium dodecyl sulfate. 0378-1119/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 03 7 8 -1 1 1 9 ( 9 7 ) 0 0 0 46 - 2
by the binding of cGMP at site(s) remote from the catalytic site thereby increasing the enzyme’s affinity for cAMP. The cGMP-stimulated PDE has been purified from a number of sources (reviewed in Manganiello et al., 1990). It is a dimer of 102–105 kDa subunits and can be found in either the soluble or particulate fractions of tissue homogenates. The amino acid sequence of the bovine heart PDE2A has been determined and a cGMPbinding domain identified (Le Trong et al., 1990; Charbonneau et al., 1986). Bovine adrenal and rat brain cDNAs encoding PDE2A have been isolated (Sonnenburg et al., 1991; Tanaka et al., 1991, Repaske et al., 1992; Yang et al., 1994). Thus far the PDE2 amino acid and cDNA sequences that have been determined are derived from a single gene (PDE2A) in each species. Over most of their lengths the bovine (PDE2A1) and rat (PDE2A2) amino acid sequences were similar. However, PDE2A1 and PDE2A2 contained completely different amino-terminal regions (25 and 37 amino acids, respectively). A third amino-terminal sequence, bovine PDE2A3, has recently been reported (Genbank accession No. L49503: Juilfs, D.M, Sonnenburg, W.K., Seraji, S. and Beavo, J.A., unpublished). These amino-terminal
90
G.J. Rosman et al. / Gene 191 (1997) 89–95
differences are believed to result from alternative exon splicing and it has been suggested that the different amino termini may be responsible for particulate or soluble localization of the enzyme ( Yang et al., 1994; Sonnenburg et al., 1991; Murashima et al., 1990). In this paper we report the isolation of human PDE2A cDNAs (PDE2A3) and a preliminary characterization of the biochemical properties of the recombinant protein.
2. Results and discussion 2.1. Isolation and characterization of human PDE2A3 cDNAs A bovine PDE2A1 cDNA was used as a probe to screen human cDNA libraries. Four different cDNAs were chosen for characterization and were used to assemble a 4.2-kb composite human PDE2A3 cDNA sequence (Fig. 1). This DNA sequence encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. 2.2. Comparison of human PDE2A3 to bovine and rat PDE2A sequences The human protein is similar to the bovine PDE2A1 and rat PDE2A2 proteins although the amino termini of the three sequences diverge from each other ( Fig. 2). Excluding these divergent amino termini the three PDE2A protein sequences are 90% identical to each other. The identity is greater in the cGMP-binding domain (94% identical ) and the PDE catalytic region (98% identical ). A closer examination of the amino terminal sequences revealed that the human PDE2A3 and rat PDE2A2 sequences share an additional block of homology just upstream from the point at which they both diverge from the bovine PDE2A1 sequence. To examine the relationship between these sequences further, we turned to a comparison of the cDNA sequences. We included in the comparison cDNA sequences corresponding to each of the three protein sequences shown in Fig. 2 and also the cDNA sequence for a recently described bovine PDE2A 5∞-splice variant, PDE2A3 (Genbank accession
No. L49503: Juilfs, D.M, Sonnenburg, W.K., Seraji, S. and Beavo, J.A., unpublished ). This cDNA was isolated from a bovine brain cDNA library and is the bovine counterpart of the human PDE2A3 sequence we describe here. A comparison of the 5∞-untranslated regions and the 5∞-ends of the coding regions revealed 5 blocks of DNA sequence that we called A, B, C, common1, and common2 (Fig. 3). Common1 and common2 were blocks of cDNA sequence found in all four cDNAs whereas A, B, and C were sequences found only in a subset of the four cDNAs. The bovine PDE2A1 cDNA contained regions B, common1, C, and common2 (in this order, 5∞ to 3∞) ( Fig. 3). The rat PDE2A2 cDNA contained B, common1 and common2 (5∞ to 3∞). As has been noted ( Yang et al., 1994), the initiator methionine of the PDE2A2 cDNA differs from that of PDE2A1. A consequence of this is that different reading frames of region common1 are used which leads to the two proteins having different amino terminal regions. In summary, regions B, common1 and C encode the amino terminus unique to PDE2A1 and regions B and common1 encode the amino terminus found in PDE2A2. The human and bovine PDE2A3 cDNAs contained regions A, common1 and common2 (5∞ to 3∞) (Fig. 3). The addition of region A yielded a unique amino terminus for the PDE2A3 proteins. The PDE2A3 cDNAs used the same reading frame of region common1 as was used in rat PDE2A2. This provides the block of amino acid sequence that PDE2A3 and PDE2A2 share (and is not found in PDE2A1) (Fig. 2). The amino terminus of rat PDE2A2 is more hydrophobic than the amino termini of the bovine PDE2A1 human PDE2A3 proteins. These differences in hydrophobicity may underlie the partitioning of PDE2A activity between the membrane-associated and soluble fractions of the cell ( Yang et al., 1994). The structure of the genomic DNA encoding the splice variants is not known, except that regions B and common1 are known to be adjacent in the rat genome ( Yang et al., 1994). 2.3. Distribution of PDE2A3 mRNA A Northern blot was probed with human PDE2A to examine the tissue distribution of the human PDE2A
Fig. 1. Human PDE2A3 nucleotide and amino acid sequences. The nucleotide sequence is a composite of 4 cDNAs isolated from human cDNA libraries. The initiator methionine is at nt 162 and the stop codon is at nt 2985. The deduced amino acid sequence is positioned below the open reading frame. Nucleotide and amino acid positions are indicated at the right side of the figure. The four cDNAs align with the sequence as follows: Fbr9.2 (nt 1–3046), Hippo7.1 (nt 2646–4240), Hippo7.4 (nt 2646–4218), Hr6.1 (nt 1820–3848). The polyadenylation site in cDNA Hippo7.4 is 3 nt 5∞ to the polyadenylation site in cDNA Hippo7.1. In the 3∞-untranslated region cDNA Fbr9.2 diverged from the sequence shown here. The divergent sequences were not found in the three other human cDNAs and were not pursued. Methods: Human cDNA libraries were screened with DNA probes and PDE2A cDNAs isolated, characterized, and sequenced on both strands according to published procedures (Loughney et al., 1996). A bovine PDE2A1 probe (Sonnenburg et al., 1991) was used to screen heart (Stratagene, LaJolla, CA), hippocampus (Clontech, Palo Alto, CA), and fetal brain (Stratagene) cDNA libraries to yield the Hr, Hippo, and Fbr cDNAs, respectively.
G.J. Rosman et al. / Gene 191 (1997) 89–95
91
92
G.J. Rosman et al. / Gene 191 (1997) 89–95
Fig. 2. Comparison of human, bovine and rat PDE2A proteins. The human PDE2A amino acid sequence is shown on the top line. Positions where both the bovine (Sonnenburg et al., 1991) and rat PDE2A sequences ( Yang et al., 1994) match the human sequence are shown with dots. Dashes indicated gaps introduced to optimize the alignment. Amino acid residue numbers are indicated on the right. The cGMP binding domain (Le Trong et al., 1990) and the catalytic region (Charbonneau, 1990) are indicated.
G.J. Rosman et al. / Gene 191 (1997) 89–95
93
Fig. 3. Comparison of PDE2A cDNAs. The 5∞ portions of bovine PDE2A1, rat PDE2A2, human PDE2A3 and bovine PDE2A3 cDNAs were compared. A, B, C, common1 and common2 represent blocks of DNA sequence found in the different cDNAs. They are not drawn to scale. For each cDNA the initiator methionine is indicated with Met and represents amino acid position number 1 for each protein. The numbers under the various blocks refer to the position of the amino acids that are represented by each block; amino acid positions are numbered as in Fig. 2. Junctions between regions common1 and C in PDE2A1, regions B and common1 in PDE2A2, and regions A and common1 in PDE2A3 occur within a codon. In region common1, PDE2A1 uses one reading frame while PDE2A2 and PDE2A3 use another (see Section 2.2).
mRNA (Fig. 4). A 4.2-kb mRNA was present in all tissues represented on the filter. The highest level of PDE2A mRNA expression was found in the brain. An intermediate level of expression was detected in heart, liver, skeletal muscle, kidney and pancreas. The lowest levels of PDE2A expression were seen in the lung and placenta. This distribution pattern is similar to that found for rat and bovine PDE2A expression. In particular, of the tissues tested, PDE2A expression was highest in rat
( Repaske et al., 1993) and bovine (Sonnenburg et al., 1991) brain. The Northern blot was subsequently probed for G3PDH mRNA. A hybridizing signal was observed in each sample lane with a much stronger signal in the heart and skeletal muscle RNA (data not shown). Only a single size of PDE2A mRNA was observed. As noted in this section, multiple splice variants have been observed for this gene. It is possible that the multiple mRNAs are indistinguishable in size by Northern analysis. Alternatively, specific tissues may express only one form. Also, though RNase protection studies showed that at least two different 5∞-splice variants are expressed in bovine tissues (Sonnenburg et al., 1991), it is not known if all 3 of the splice variants ( Fig. 3) are used in a single organism. 2.4. PDE2A3 biochemical activity
Fig. 4. PDE2A Northern. A filter with 2 mg poly A+ mRNA from the tissues indicated was purchased from Clontech (Palo Alto, CA) and hybridized with an antisense riboprobe corresponding to nt 2647–2939 of human PDE2A3 (Fig. 1). Probe preparation and hybridization conditions have been described (Loughney et al., 1996). RNA size markers are shown at left. The exposure time for the autoradiogram shown was 24 h.
To determine if the human PDE2A3 encoded a PDE with characteristics typical of a PDE2 it was expressed in a strain of S. cereviseae lacking endogenous PDE activity (see legend to Fig. 5 for methods). PDE activity was measured in extracts from yeast expressing a control (empty vector) and from yeast expressing the recombinant human PDE2A3 protein. Yeast extracts were centrifuged at 14 000×g for 1 min and the supernatant was removed and assayed for PDE activity. Approximately 12% of the total PDE activity in the yeast extract is found in the 14 000×g supernatant. With a substrate concentration of 1 mM, PDE2A3expressing yeast hydrolyzed both cAMP (20.5 nmol min−1 mg−1) and cGMP (28.5 nmol min−1 mg−1). The assay background (empty vector only) was ≤0.25 nmol min−1 mg−1. A band of the predicted size (~105 kDa) was seen in the yeast extract by western blot analysis (data not shown).
94
G.J. Rosman et al. / Gene 191 (1997) 89–95
Cyclic GMP stimulates the hydrolysis of cAMP by PDE2 (Manganiello et al., 1990). Hydrolysis of 1 mM cAMP by the recombinant PDE2A3 was stimulated almost three-fold by the addition of 10 mM cGMP ( Fig. 5B). Although stimulation by cGMP is characteristic of this enzyme, a wide range of values has been reported. A recombinant bovine PDE2A1 assayed in a mammalian cell extract showed approximately 3.5-fold stimulation by cGMP (Sonnenburg et al., 1991). Purified PDE2A proteins showed 5–6-fold stimulation in one report (Martins et al., 1982) and >50-fold stimulation in another (Mumby et al., 1982). The differences in these values may reflect differences in methodology and the enzyme status. Alterations in assay temperature, assay pH, or cGMP concentration and proteolysis of the enzyme can all influence the activation state of the enzyme. Likewise, the presence of fatty acids or shortchain alcohols has been reported to inhibit the activation of cAMP hydrolysis by cGMP (Manganiello et al., 1990). The human recombinant PDE2A3 protein was not purified and one or more of the factors described above may have affected the magnitude of the cGMP stimulation. Purified recombinant protein will be used for more detailed kinetic characterizations.
3. Conclusions Fig. 5. (A) Inhibition of PDE2A3 activity by EHNA. (B) Stimulation of PDE2A3 activity by cGMP. Data from a single experiment are shown. The assays were repeated using the supernatant from a second extract and gave values that differed ≤10% from the values shown in Fig. 5. Methods: The complete open reading frame of PDE2A3 (nt 162–2989, Fig. 1) was inserted into the yeast expression vector pBNY6n (Loughney et al., 1996) to generate pCGS-6n. All new junctions and PCR generated DNA were sequenced to verify the structure of pCGS-6n. pCGS-6n was transformed into the yeast strain YKS45 (Mata his3 leu2 trp1 ura3 Dpde1::HIS3 D pde2::TRP1). Cells were grown and PDE activity was assayed as described (Loughney et al., 1996) with the following exceptions. Lysis buffer consisted of 25 mM Tris pH 8, 5 mM EDTA, 5 mM EGTA, 1 mM benzamidine (Sigma, St. Louis, MO) and 0.1 mM DTT. PBS (0.9 mM, 2.7 mM KH PO , 2 4 137.9 mM NaCl, 8.1 mM Na HPO · 7H O, pH 7.45) was used as the 2 4 2 enzyme dilution buffer. The PDE assay was performed in 40 mM Tris pH 8, 1 mM EGTA, 5 mM MgCl , and 0.1 mg/ml BSA (Fraction V, 2 ICN Biomedicals, Costa Mesa, CA). EHNA (Research Biochemicals International, Natick, MA) was dissolved in DMSO and the final concentration of DMSO in the assay was 0.7%.
EHNA is an adenosine deaminase inhibitor that is also a selective PDE2 inhibitor (Podzuweit et al., 1995). At a substrate concentration of 1 mM, EHNA (10 mM ) reduced hydrolysis in the PDE2A3 yeast extract by 76% (cAMP) and by 87% (cGMP) (Fig. 5A). These results are in reasonable agreement with the IC value of 0.8 50 mM that was observed for EHNA inhibition of soluble PDE2 activity purified by DEAE chromatography from a human myocardial extract (Podzuweit et al., 1995).
(1) The human PDE2A3 cDNA sequence, a composite of 4 cDNA clones, is 4240 nt in length and encodes a 941 amino acid protein with a predicted molecular mass of 105 715 Da. (2) As expected for a PDE2 enzyme, human PDE2A3 hydrolyzes cAMP and cGMP, is inhibited by EHNA and responds to cGMP with an increase in cAMP hydrolysis. (3) A 4.2-kb message for human PDE2A was found in heart, brain, placenta, lung, skeletal muscle, kidney and pancreas, the most abundant source being brain. (4) PDE2A3 has the third type of amino terminus that has been observed for PDE2A gene products.
Acknowledgement We thank Carmen Hertel for excellent technical assistance, Dina Leviten, Christi Wood and Ernie Tolentino for synthesis of oligonucleotides and DNA sequencing and Vince Florio for critically reading the manuscript.
References Beavo, J., 1995. Cyclic nucleotide phosphodiesterases: functional implications of multiple isoforms. Physiol. Rev. 75, 725–748.
G.J. Rosman et al. / Gene 191 (1997) 89–95 Charbonneau, H., Beier, N., Walsh, K.A. and Beavo, J.A., 1986. Identification of a conserved domain among cyclic nucleotide phosphodiesterases from diverse species. Proc. Natl. Acad. Sci. USA 83, 9308–9312. Charbonneau, H., 1990. Structure-Function Relationships Among Cyclic Nucleotide Phosphodiesterases. In: Beavo, J. and Houslay, M.D. ( Eds.), Isoenzymes of Cyclic Nucleotide Phosphodiesterases. John Wiley and Sons, New York, pp. 267–296. Le Trong, H., Beier, N., Sonnenburg, W.K., Stroop, S.D., Walsh, K.A., Beavo, J.A. and Charbonneau, H., 1990. Amino acid sequence of the cyclic GMP stimulated cyclic nucleotide phosphodiesterase from bovine heart. Biochemistry 29, 10280–10288. Loughney, K., Martins, T.J., Harris, E.A.S., Sadhu, K., Hicks, J.B., Sonnenburg, W.K., Beavo, J.A. and Ferguson, K., 1996. Isolation and characterization of cDNAs corresponding to two human calcium, calmodulin-regulated, 3∞,5∞-cyclic nucleotide phosphodiesterases. J. Biol. Chem. 271, 796–806. Manganiello, V.C., Tanaka, T. and Murashima, S., 1990. Cyclic GMPstimulated cyclic nucleotide phosphodiesterases. In: Beavo, J. and Houslay, M.D. ( Eds.), Isoenzymes of Cyclic Nucleotide Phosphodiesterases. John Wiley and Sons, New York, pp. 61–85. Martins, T.J., Mumby, M.C. and Beavo, J.A., 1982. Purification and characterization of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from bovine tissues. J. Biol. Chem. 257, 1973–1979. Mumby, M.C., Martins, T.J., Chang, M.L. and Beavo, J.A., 1982. Identification of cGMP-stimulated cyclic nucleotide phosphodiesterase in lung tissue with monoclonal antibodies. J. Biol. Chem. 257, 13283–13290.
95
Murashima, S., Tanaka, T., Hockman, S. and Manganiello, V., 1990. Characterization of particulate cyclic nucleotide phosphodiesterases from bovine brain: purification of a distinct cGMP-stimulated isoenzyme. Biochemistry 29, 5285–5292. Podzuweit, T., Nennstiel, P. and Muller, A., 1995. Isozyme selective inhibition of cGMP-stimulated cyclic nucleotide phosphodiesterases by erythro-9-(2-hydroxy-3-nonyl ) adenine. Cell. Signaling 7, 733–738. Repaske, D.R., Corbin, J.G., Conti, M. and Goy, M.F., 1993. A cyclic GMP-stimulated cyclic nucleotide phosphodiesterase gene is highly expressed in the limbic system of the rat brain. Neuroscience 56, 673–686. Repaske, D.R., Swinnen, J.V., Jin, S.-L.C., Van Wyk, J.J. and Conti, M., 1992. A polymerase chain reaction strategy to identify and clone cyclic nucleotide phosphodiesterase cDNAs. J. Biol. Chem. 267, 18683–18688. Sonnenburg, W.K., Mullaney, P.J. and Beavo, J.A., 1991. Molecular cloning of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase cDNA. J. Biol. Chem. 266, 17655–17661. Tanaka, T., Hockman, S., Moos Jr., M., Taira, M., Meacci, E., Murashima, S. and Manganiello, M., 1991. Comparison of putative cGMP-binding regions in bovine brain and cardiac cGMP-stimulated phosphodiesterases. Second Messengers Phosphoproteins 13, 87–98. Yang, Q., Paskind, M., Bolger, G., Thompson, W.J., Repaske, D.R., Cutler, L.S. and Epstein, P.M., 1994. A novel cyclic GMP stimulated phosphodiesterase from rat brain. Biochem. Biophys. Commun. 205, 1850–1858.