Partial reconstitution of photoreceptor cGMP phosphodiesterase characteristics in cGMP phosphodiesterase-5.

Photoreceptor cGMP phosphodiesterases (PDE6) are uniquely qualified to serve as effector enzymes in the vertebrate visual transduction cascade. In the dark-adapted photoreceptors, the activity of PDE6 is blocked via tight association with the inhibitory gamma-subunits (Pgamma). The Pgamma block is removed in the light-activated PDE6 by the visual G protein, transducin. Transducin-activated PDE6 exhibits an exceptionally high catalytic rate of cGMP hydrolysis ensuring high signal amplification. To identify the structural determinants for the inhibitory interaction with Pgamma and the remarkable cGMP hydrolytic ability, we sought to reproduce the PDE6 characteristics by mutagenesis of PDE5, a related cyclic GMP-specific, cGMP-binding PDE. PDE5 is insensitive to Pgamma and has a more than 100-fold lower k(cat) for cGMP hydrolysis. Our mutational analysis of chimeric PDE5/PDE6alpha' enzymes revealed that the inhibitory interaction of cone PDE6 catalytic subunits (PDE6alpha') with Pgamma is mediated primarily by three hydrophobic residues at the entry to the catalytic pocket, Met(758), Phe(777), and Phe(781). The maximal catalytic rate of PDE5 was enhanced by at least 10-fold with substitutions of PDE6alpha'-specific glycine residues for the corresponding PDE5 alanine residues, Ala(608) and Ala(612). The Gly residues are adjacent to the highly conserved metal binding motif His-Asn-X-X-His, which is essential for cGMP hydrolysis. Our results suggest that the unique Gly residues allow the PDE6 metal binding site to adopt a more favorable conformation for cGMP hydrolysis.

cGMP phosphodiesterases (PDE6) 1 play the role of effector enzymes in the vertebrate visual transduction cascade. In retinal rod cells, photoexcited rhodopsin induces GDP/GTP exchange on the visual G protein, transducin (Gt), and liberated Gt␣GTP activates PDE6. A homologous cascade operates in cone photoreceptors. cGMP hydrolysis by active PDE6 results in the closure of cGMP-gated channels in the plasma membrane (1,2). The key attributes of the visual cascade, low noise and high gain signal amplification, place specific requirements on PDE6. The enzyme must have a very low basal cGMP hydrolytic rate in the dark-adapted photoreceptors and a very high catalytic rate in the transducin-activated state. This is achieved through two unique features of PDE6: the inhibitory interaction of the catalytic subunits with the ␥-subunit and an exceptionally high k cat value for cGMP hydrolysis when the inhibition is turned off.
The lack of a practical expression system for PDE6 (3)(4)(5) has stalled the progress in determining the structural basis of PDE6 function. We have begun to study the structure and function relationship of PDE6 by constructing chimeras between cone PDE6␣Ј and cGMP binding cGMP-specific PDE (PDE5 family) (5,6). PDE5 and PDE6 display a high degree of identity (45-48%) between the catalytic domains, a strong substrate selectivity for cGMP, and similar sensitivity to a common set of competitive inhibitors (7)(8)(9)). Yet, the reported maximal rate of cGMP hydrolysis by PDE5 catalytic dimers is only ϳ10 moles of cGMP per mole of PDE⅐sec, which is ϳ400 -550-fold lower than the k cat estimates for PDE6 (5, 10 -15). Furthermore, the activity of PDE5 is unaffected by the PDE6 ␥-subunit (5,6). This, and a robust functional expression of PDE5 using the baculovirus/insect cell system (16), makes PDE5 a valuable tool for "gain of PDE6 function" experiments. Recently, we have shown that a substitution of the segment PDE5-(773-820) by the corresponding PDE6␣Ј-(737-784) sequence in the wild-type PDE5 or in a PDE5/PDE6␣Ј chimera containing the catalytic domain of PDE5 results in chimeric enzymes capable of inhibitory interaction with P␥ (6). Alaninescanning mutational analysis of the previously identified P␥ cross-linking site, PDE6␣Ј-(750 -760) (17), revealed a critical P␥-interacting residue, Met 758 (6). In a model of the PDE6␣Ј catalytic domain, Met 758 faces the opening of the catalytic cavity (6). We then hypothesized that P␥ may interact with additional nonconserved residues located at the perimeter of the cavity, thus allowing P␥ to serve as a lid on the catalytic pocket. In this study, we mutated three candidate P␥ contact residues identified from the model of PDE6␣Ј and examined these mutants for inhibition by P␥.
The rationale for our search of the catalytic determinants of PDE6 was based on biochemical evidence and the crystal structure of the PDE4 catalytic domain (18 -20), which suggests the critical role of the two highly conserved metal binding motifs, His-Asn-X-X-His (I) and His-Asp-X-X-His (II), in the hydrolysis of cyclic nucleotides. We replaced PDE6␣Ј domains containing motifs I and II into PDE5. Resulting chimeric PDEs and corresponding mutants have been analyzed to test our hypothesis. probe (H-15) antibodies were purchased from Santa Cruz Biotechnology. Zaprinast and all other reagents were purchased from Sigma.

Materials
Cloning of P␥ Mutants-P␥ mutants were generated based on the pET11a-P␥ expression vector (21,22). Residues Ile 86 and Ile 87 were substituted for alanine using PCR-directed mutagenesis. PCR products were obtained using a forward primer containing a NdeI site and a reverse primer containing the mutations and a BamHI site. The fragments were digested with NdeI/BamHI and subcloned into the pET11a-P␥ digested with the same enzymes.
Preparation of P␥ and P␥ Mutants-The P␥-subunit and its mutants were expressed in Escherichia coli and purified on a SP-Sepharose fast flow column and on a C 4 HPLC column (Microsorb-MW, Rainin) as described (22). Purified proteins were lyophilized, dissolved in 20 mM HEPES buffer, pH 7.5 and stored at Ϫ80°C until use.
Site-directed Mutagenesis of PDE5 and Chi16 -Site-directed mutagenesis of PDE5 was performed using a QuikChange TM kit. A pair of complementary oligonucleotides encoding for the Ala 608 3 Gly and Ala 612 3 Gly substitutions (PDE5A608G/A612G) was used to PCR-amplify the pFastBacHTb-PDE5 vector. The PCR product was treated with DpnI to eliminate the template and was transformed into E. coli DH5␣. Chi16 mutants with single substitutions of residues Lys 769 , Phe 777 , and Phe 781 by Ala were constructed using PCR-directed mutagenesis. A unique NheI site (PDE5 codons for Pro 661 -Leu 662 ) was introduced into Chi16 using a QuikChange TM kit. The 5Ј-primer sequence included the NheI recognition site. Reverse primers contained a desired mutation and the StuI site. The PCR products were digested with NheI/StuI and subcloned into the modified Chi16 vector cut with the same enzymes. Sequences of all mutants were verified by automated DNA sequencing at the University of Iowa DNA Core Facility.
Expression and Purification of Recombinant PDEs and their Mutants-Sf9 cells were harvested at 60 h after infection, washed with 20 mM Tris-HCl buffer, pH 7.8 containing 50 mM NaCl, and resuspended in the same buffer containing a protease inhibitor mixture (10 g/ml pepstatin, 5 g/ml leupeptin, and 0.2 mM phenylmethylsulfonyl fluoride). The cell suspensions were sonicated using 30-s pulses for a total duration of 3 min. The supernatants (100,000 ϫ g, 45 min) were loaded onto a column with a His-Bind resin (Novagen) equilibrated with 20 mM Tris-HCl buffer, pH 7.8, containing 10 mM imidazole. The resin was washed with a 5ϫ volume of the buffer containing 500 mM NaCl and 25 mM imidazole. Proteins were eluted with the buffer containing 250 mM imidazole. ␤-mercaptoethanol (2 mM) was added to the eluate. PDE5, Chi20, Chi21, and PDE5A608G/A612G were additionally purified using ion-exchange chromatography on a Mono Q ® HR 5/5 column (Amersham Pharmacia Biotech). Purified proteins were dialyzed against 40% glycerol and stored at Ϫ20°C.
Other Methods-PDE activity was measured using [ 3 H]cGMP as described (23,24). Less than 15% of cGMP was hydrolyzed during these reactions. The K i values for inhibition of PDE activity by P␥ and zaprinast were measured using 0.5 M cGMP (i.e. Ͻ35% of the K m value for chimeric and mutant PDEs). Protein concentrations were determined by the method of Bradford (25) using IgG as a standard or by using calculated extinction coefficients at 280 nm. The molar concentrations of Chi20, Chi21, and mutatnt PDEs, [PDE], were calculated based on the fraction of PDE protein in preparations, and the molecular mass of 93.0 kDa. The fractional concentrations of PDE were determined from analysis of the Coomassie Blue-stained SDS gels using a HP ScanJet II CX/T scanner and Scion Image Beta 4.02 software. A typical fraction of Chi16 mutants in partially purified preparations was 10 -15%. A typical fraction of purified Chi20, Chi21, and PDE5A608G/ A612G was 65-70%. The k cat values for cGMP hydrolysis were calculated as V max /[PDE]. SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli (26) in 10 -12% acrylamide gels. For Western immunoblotting, proteins were transferred to nitrocellulose (0.1 m, Schleicher & Schuell) and analyzed using rabbit His-probe (H-15) or sheep anti-PDE6␣Ј antibodies (5,6,27). The antibody-antigen complexes were detected using anti-rabbit or anti-goat/sheep IgG conjugated to horseradish peroxidase and ECL reagent (Amersham Pharmacia Biotech.). Fitting the experimental data to equations was performed with nonlinear least squares criteria using GraphPad Prizm Software. The K i , K m , and IC 50 values are expressed as mean Ϯ S.E. for three independent measurements.

Mutational Analysis of the P␥ Binding Site of PDE6␣Ј-
Previously, we demonstrated that PDE5/PDE6␣Ј chimeras containing a PDE6␣Ј sequence, PDE6␣Ј-(737-784), are effectively inhibited by P␥, and two residues, Met 758 and Gln 752 , participate in the inhibitory interaction (6). Based on the model structure of PDE6␣Ј (6), three solvent-exposed nonconserved PDE6␣Ј residues, Lys 769 , Phe 777 , and Phe 781 , were chosen for further mutational analysis of the P␥ binding region (Fig. 1A). A PDE5/PDE6␣Ј chimera, Chi16 (6), served as a template for single substitutions of these residues by Ala. The Chi16 mutants were expressed in Sf9 insect cells and partially purified. Expression of the K769A, F777A, and F781A mutants have yielded similar amounts of soluble protein (50 -100 g/100 ml of culture). Neither of these mutations has significantly affected the catalytic properties of chimeric PDE. The K m and k cat values for cGMP hydrolysis for all three mutants were in the 3-10 M range, and the 5-10 s Ϫ1 range, respectively (Table I). As an additional control for the structural integrity of the catalytic site, mutants of Chi16 were tested for the PDE activity inhibition by zaprinast, a specific competitive inhibitor of PDE5 and PDE6. The largest change, a 2-fold increase in the IC 50 value, was caused by the F781A substitution (Table I). Nonetheless, such a change represents an insignificant loss of affinity to zaprinast.
The test of the ability of Chi16 mutants to be inhibited by P␥ showed that the K769A mutation had no effect on the inhibitory interaction with P␥ (K i 2.9 nM) ( Table I). Two other mutants, F777A and F781A, displayed significant impairments in the inhibition by P␥. The F777A substitution reduced both the maximal inhibition of PDE activity by P␥ (ϳ45%) and the K i value (K i of 19 nM). The inhibition of F781A mutant by P␥ also was incomplete (ϳ65%) and associated with an increase in the K i value (K i of 31 nM) ( Fig. 2A and Table I).
Effects of the C-terminal P␥ Mutants on the Catalytic Activity of Mutant Chi16 -C-terminal P␥ mutants were designed based on the evidence for the critical role of the P␥ C terminus in PDE6 inhibition (21,28). The two extreme C-terminal P␥ residues, Ile 86 and Ile 87 , were replaced by Ala to obtain the P␥I86A and P␥I87A mutants, respectively. The P␥ mutants were analyzed for their ability to inhibit trypsin-activated PDE6␣Ј (tPDE), Chi16, and the M758A, F777A, and F781A mutants ( Fig. 2; Table I). P␥I86A and P␥I87A fully inhibited tPDE activity. However, the potency of the inhibition was reduced ϳ4 -5-fold (K i of 0.75 nM for P␥I86A and K i of 0.65 nM for P␥I87A, compared with K i of 0.15-0.2 nM for P␥). A similar increase in the K i values was observed from the inhibition of Chi16 activity by P␥I86A (K i of 13 nM) and P␥I87A (K i of 7 nM) (Fig. 2, B and C; Table I). Yet, P␥I86A and P␥I87A did not fully inhibit Chi16, maximal inhibition was 65 and 70%, respectively. (Fig. 2, B and C; Table I). No appreciable inhibition of M758A by either P␥ mutant was seen even at inhibitor concentrations as high as 5 M. The inhibition of F777A by P␥I86A was partial (45%) with the K i value of 96 nM, whereas P␥I87A inhibited this Chi16 mutant with an even smaller maximal effect (25%, K i of 64 nM). The F781A mutant was inhibited by P␥I86A and P␥I87A with K i values of 49 and 32 nM and maximal effects of 40 and 55%, respectively (Fig. 2, B and C; Table I).
Catalytic Properties of PDE5/PDE6␣Ј Chimeras Containing the PDE6␣Ј Metal Binding Sites-Two conserved metal binding motifs found in all PDEs are absolutely critical for cyclic nucleotide hydrolytic activity (18 -20). To identify the structural elements responsible for the unique catalytic properties of PDE6, chimeric PDE5/PDE6␣Ј have been generated by introduction into PDE5 of PDE6␣Ј domains containing metal binding motifs, I and II. A replacement of the PDE6␣Ј-(562-617) segment into PDE5 yields a chimeric PDE5/PDE6␣Ј, Chi20, that incorporates both PDE6␣Ј metal binding sites and the connecting sequence (Fig. 1A). Chi20 was expressed in Sf9 cells as a functional enzyme at ϳ400 g/100 ml and purified to ϳ 65-70% purity (Fig. 1B). The catalytic characteristics of Chi20 were examined in comparison to those of PDE5 and native PDE6␣Ј. PDE6␣Ј has reported K m (17-25 M) and k cat (3500 -4500 moles of cGMP per mole of PDE⅐s) values for cGMP hydrolysis that are ϳ5 and ϳ400-fold higher than the respective constants for PDE5 (5, 10 -11, 14). The catalytic parameters of Chi20 were significantly different from those of PDE5. Chi20 hydrolyzed cGMP with the K m value of 12 M, which is ϳ4-fold higher than the K m value for PDE5 but similar to that of PDE6␣Ј (Table I). The maximal activity of 116 moles of cGMP per mole of PDE⅐s for Chi20 is ϳ10-fold higher than that of PDE5. Chi20 was inhibited by zaprinast with the IC 50 value of 0.35 M, which is comparable with that of PDE5 (Table I).
To determine the role of individual metal binding motifs and their adjacent regions in cGMP hydrolysis by PDE6, we inserted a PDE6␣Ј fragment corresponding to the helix-␣6 (20), PDE6␣Ј-(562-574), into PDE5 (Chi21) (Fig. 1). The catalytic properties of Chi21 and the inhibition by zaprinast (K m of 17 M, k cat of 110 moles of cGMP per mole of PDE⅐s, and IC 50 0.39 M) were similar to those of Chi20.
Catalytic Properties of the PDE5A608G/A612G Mutant-The alignment of sequences from different PDE families corresponding to the ␣6 helix shows a glycine residue, PDE6␣ЈGly 562 , conserved only in photoreceptor PDEs (Fig. 3A). A second Gly residue, PDE6␣ЈGly 566 , is conserved in PDE6␣Ј and PDE6␣, but substituted by Ala in PDE6␤ and PDE5 (Fig.  3A). To test the hypothesis that Gly 562 and Gly 566 of PDE6␣Ј are responsible for the differences in catalytic properties of Chi21 and PDE5, a doubly substituted mutant of PDE5, A608G and A612G, was expressed and purified from Sf9 cells. Similar to Chi20 and Chi21, PDE5A608G/A612G hydrolyzed cGMP with a K m value of 14 M and a k cat value of 105 moles of cGMP per mole of PDE⅐s (Table I). DISCUSSION An interaction between PDE6 catalytic and inhibitory P␥subunits keeps the visual effector enzyme inhibited in the dark. Previous biochemical studies have established that the ␥-subunit of photoreceptor PDE inhibits the enzyme activity by blocking its catalytic site (29). The major inhibitory domain has been localized to the P␥ C terminus (21,28). Recently, we have demonstrated that P␥ inhibits the activity of PDE5/PDE6␣Ј chimera, Chi 16, containing residues PDE6␣Ј-(737-784) (6).  Essential P␥ binding residues, Gln 752 and Met 758 , of PDE␣Ј have been identified via mutagenesis of Chi16 (6). A model of the PDE6␣Ј catalytic domain places Met 758 at the opening of the catalytic pocket (6). Hypothetically, to ensure an effective catalytic block, the P␥ C terminus may lie over or might be inserted into the catalytic cavity. The former appears more likely because the catalytic pockets of different cyclic nucleotide PDEs are made up of highly conserved residues, whereas the inhibition by P␥ is a unique attribute of PDE6. We speculated that to cover the catalytic pocket, the P␥ C terminus, besides Met 758 , interacts with additional nonconserved residues located at the perimeter of the entrance to the active site. The fact that the introduction of PDE6␣Ј-(737-784) into PDE5/ PDE6␣Ј chimera leads to a full inhibition of the PDE activity by P␥ suggests the PDE6␣Ј-(737-784) segment contains most if not all residues interacting with the P␥ C terminus. In the PDE6␣Ј model, PDE6␣Ј-(737-784) comprises about half of the catalytic cavity mouth. Residues at three positions within PDE6␣Ј-(737-784) (Lys 769 , Phe 777 , and Phe 781 ) are conserved among photoreceptor PDEs but have nonhomologous substitutions in PDE5. Supporting our hypothesis, replacement of two residues, Phe 777 and Phe 781 , by Ala in Chi16 has resulted in mutant PDEs that in comparison with Chi16 were less potently and incompletely inhibited by P␥. Phe 777 and Phe 781 are located next to each other, opposite to the Met 758 side of the catalytic opening (Fig. 3, B and C). Thus, it appears that the P␥ C terminus makes a bridge over the catalytic pocket. Such a model provides an interesting explanation to the results of an earlier study that examined inhibition of PDE6 by C-terminally truncated P␥ mutants (21). Truncations of one or two of the C-terminal Ile 86 -Ile 87 residues led to substantial increases in the K i value, whereas further truncations, up to 8 -11 C-terminal residues, reduced the maximal inhibition of PDE6 activity without significantly affecting the K i value (21). A plausible interpretation is that P␥Ile 86 -Ile 87 interact with residues on one side of the catalytic pocket and other residues, perhaps P␥-(77-85), stretch over the catalytic cavity until P␥ reaches the opposite side. Accordingly, removal of P␥Ile 86 -Ile 87 decreases the affinity of P␥ for the PDE6 catalytic subunit, whereas progressive removal of P␥-(77-85) residues gradually facilitates access of cGMP to the catalytic site. To determine the orientation of the P␥ C terminus against the catalytic site and identify point-to-point interactions with PDE6␣Ј, we examined the inhibition of Chi16 and the M758A, F777A, and F781A mutants of Chi16 by two P␥ mutants, P␥I86A and P␥I87A. The simplest prediction is that if a C-terminal Ile of P␥ interacts with one of the three PDE6␣Ј residues, the corresponding mutant PDE would be inhibited comparably by P␥ and by the P␥ mutant. Complicating this prediction, side chains of Phe 777 and Phe 781 make a hydrophobic contact and thereby may support each other in the interaction with P␥. The analysis of inhibition of Chi16 mutants by P␥ mutants indicates that Ile 86 and Ile 87 of P␥ interact with Phe 777 and Phe 781 of PDE6␣Ј. Moderate increases in the K i values and reductions in the maximal inhibition of F777A and F781A caused by the P␥I86A substitution suggest that Ile 86 probably contacts one or both the PDE6␣Ј residues. The failure of P␥I86A to inhibit M758A is consistent with the notion that Ile 86 binds Phe 777/781 , but not Met 758 . The lack of inhibition is likely caused by the inability of M758A and P␥I86A to establish at least two of the three critical contacts involving Met 758 , Phe 777 , and Phe 781 . The P␥I87A mutant did not appreciably inhibit the activity of the M758A mutant PDE. P␥I87A inhibited F781A stronger than F777A pointing to a probable contact between P␥Ile 87 and Phe 781 of PDE6␣Ј. The incomplete inhibition of mutant PDEs by P␥ or P␥ mutants most likely reflects equivalent partial inhibition of both active sites of the catalytic dimer, rather than the loss of inhibition at one site.

Structure-Function Determinants of PDE6
The analysis of P␥ secondary structure predicts an ␣-helical structure for the C-terminal residues P␥-(75-84) (30). The C terminus of P␥, P␥-(75-87), manually docked to the PDE6␣Ј catalytic site is shown in Fig. 3, B and C. The model assumes the helical structure of P␥-(75-84) and the contacts between P␥Ile 86 -Ile 87 and PDE6␣ЈPhe 777 -Phe 781 . This orientation of P␥ is also consistent with Gln 752 of PDE6␣Ј (6) making a contact with a P␥ residue located N-terminally to P␥-(75-87).
The remarkable ability of photoreceptor PDEs to hydrolyze cGMP with a catalytic rate constant of ϳ4000 -5500 moles of cGMP per mole of PDE⅐s (12)(13)(14)(15) is essential to the signal amplification in the visual cascade. All catalytic subunits of cyclic nucleotide PDEs contain two strictly conserved metal binding motifs, His-Asn-X-X-His (motif I) and His-Asp-X-X-His (motif II). In PDE6␣Ј these motifs are as follows: 557 His-Asn-Trp-Arg-His 561 and 597 His-Asp-Ile-Asp-His 601 . The crucial role of the metal ions and the binding motifs for PDE catalytic activity has been recently supported by a crystallographic study of the PDE4 catalytic domain (20). Rather than forming separate metal binding sites, both motifs are involved in coordination of two bound metal ions, ME1 and ME2 (20). For example, ME1, most likely a tightly bound Zn 2ϩ , is coordinated by the His residue (His 561 of PDE6␣Ј) from motif I, and the His and Asp residues from motif II (His 597 -Asp 598 ). A model of cAMP docked in the PDE4 active site demonstrates that ME1 and ME2 bind the cyclic phosphate, position a potential water molecule for the nucleophilic attack, and would serve to stabilize the transition state (20). In view of the role of metal binding sites in hydrolysis of cyclic nucleotides, we have considered the motifs I and II as probable structural determinants of the catalytic properties of PDE6. Motifs I and II are practically identical in PDE5 and PDE6. Therefore, a spatial orientation of these sites might be a potential key factor for cGMP hydrolysis. Motif I comprises the N-terminal potion of the helix-␣6, and motif II is in the loop connecting helices 7 and 8. A PDE5/ PDE6␣Ј chimera, Chi20, was generated by replacing a PDE6␣Ј domain containing helices ␣6-␣8 into PDE5. The analysis of Chi20 revealed a more than 10-fold increase in the maximal catalytic rate accompanied by a ϳ5-fold increase in the K m value. Subsequent chimeric PDE, Chi21, containing only helix ␣6 of PDE6␣Ј displayed catalytic properties similar to those of Chi20. An alignment of sequences of photoreceptor PDEs and PDE5 corresponding to the helix-␣6 shows a high degree of homology with the notable exception of residues at two positions corresponding to PDE6␣Ј Gly 562 and Gly 566 . Gly 562 of PDE6␣Ј is conserved only in the PDE6 family, but substituted by Ala in PDE5 (Fig. 3A). Importantly, Gly 562 immediately follows His 561 from motif I. His 561 , by analogy to PDE4, is involved in coordination of ME1, and in the positioning of His 557 to accomplish the protonation of the O3Ј leaving group (20). To probe the role of the Gly residues, a doubly substituted PDE5 mutant, A608G/A612G, has been made. The k cat value of the A608G/A612G mutant was comparable with those of Chi20 and Chi21, and ϳ10-fold higher then that of PDE5. These results suggest that the Gly residues are in part responsible for the catalytic characteristics of PDE6. Most likely, they allow for a positioning of motif I that is most favorable for cGMP hydrolysis. Other yet to be defined determinants contribute to the unique catalytic power of PDE6, because the achieved k cat value is still ϳ40 -50-fold lower than k cat described for native activated PDE6. Overall, our results suggest that a progressive incorporation of PDE6 domains or residues into PDE5 not only allows a structure-function analysis of PDE6, but also repre-sents a realistic approach to generate a chimeric enzyme that would be functionally indistinguishable from PDE6.