Potential Roles of Conserved Amino Acids in the Catalytic Domain of the cGMP-binding cGMP-specific Phosphodiesterase (PDE5)*

The known mammalian 3′:5′-cyclic nucleotide phosphodiesterases (PDEs) contain a conserved region located toward the carboxyl terminus, which constitutes a catalytic domain. To identify amino acids that are important for catalysis, we introduced substitutions at 23 conserved residues within the catalytic domain of the cGMP-binding cGMP-specific phosphodiesterase (cGB-PDE; PDE5). Wild-type and mutant proteins were compared with respect toK m for cGMP, k cat, and IC50 for zaprinast. The most dramatic decrease ink cat was seen with H643A and D754A mutants with the decrease in free energy of binding (ΔΔG T ) being about 4.5 kcal/mol for each, which is within the range predicted for loss of a hydrogen bond involving a charged residue. His643 and Asp754 are conserved in all known PDEs and are strong candidates to be directly involved in catalysis. Substitutions of His603, His607, His647, Glu672, and Asp714 also produced marked changes in k cat, and these residues are likely to be important for efficient catalysis. The Y602A and E775A mutants exhibited the most dramatic increases inK m for cGMP, with calculated ΔΔG T of 2.9 and 2.8 kcal/mol, respectively, that these two residues are important for cGMP binding in the catalytic site. Zaprinast is a potent competitive inhibitor of cGB-PDE, but the key residues for its binding differ significantly from those that bind cGMP.

The 3Ј:5Ј-cyclic nucleotide phosphodiesterase (PDE) 1 superfamily catalyzes the hydrolysis of 3Ј:5Ј-cyclic nucleotides to the corresponding nucleoside 5Ј-monophosphates. On the basis of their structural, kinetic, and regulatory characteristics, they have been recently classified into seven major families (1). Comparison of the reported PDE sequences reveals a conserved region of approximately 270 amino acids located toward the COOH terminus of PDE molecules (2). This region is more conserved within an individual PDE family (65-80% amino acid identity) than among different PDE families (25-40% identity). Studies using limited proteolysis of the different PDEs (3)(4)(5)(6), deletion mutagenesis (7)(8)(9), and point mutations targeting conserved residues (7) strongly support the assertion that this region constitutes a catalytic domain of all PDEs. In addition to the conserved residues that play a role in catalysis and substrate binding, the catalytic domain is likely to contain determinants that confer cyclic nucleotide specificity of different PDEs.
cGMP-binding cGMP-specific PDE (cGB-PDE; PDE5A) is an enzyme with high selectivity for cGMP as substrate. In addition to the site of cGMP hydrolysis, cGB-PDE contains two allosteric cGMP-binding sites that are located toward the NH 2 terminus of the cGB-PDE molecule (10). Our ultimate aim is to construct a comprehensive structure-function map of the cGB-PDE using site-directed mutagenesis as a tool. Recently, we replaced several conserved residues in the high affinity allosteric site a (11) and proposed a role of each residue in the putative NKX n D motif, which constitutes a new class of cGMPbinding sites. Detailed analysis of the sequence alignment of the catalytic region of all known PDEs to date reveals two blocks of conserved residues (10). One of these blocks has some sequence similarity to the allosteric binding sites (12), which might suggest some evolutionary relationship between cGMP binding in the allosteric and catalytic sites. However, the cGMP-binding properties and the function of the allosteric sites are quite different from those of the catalytic site. Another block of the conserved residues possesses sequence similarity to Zn 2ϩ -binding sites of the different Zn 2ϩ -dependent hydrolases, and could be a part of the PDE catalytic mechanism (13).
In the present study, scanning mutagenesis has been used to examine the importance of 23 conserved amino acids in the catalytic domain of the cGB-PDE in maintaining catalytic function. Each of these 23 conserved residues has been substituted individually. After expressing and partially purifying the mutant proteins, we have assessed the effect of these mutations on substrate binding, catalysis, and specific inhibitor binding by measuring the K m value for cGMP, k cat , and IC 50 for zaprinast, respectively. 3 H]cGMP was purchased from Amersham Corp. cGMP, histone VIII-S, Crotalus atrox snake venom, 3-isobutyl-1-methylxanthine, and zaprinast were obtained from Sigma. Hydroxyapatite was from Bio-Rad.
E. coli XL1-blue cells were used for all transformations. DNA fragments were purified by the freeze squeeze method from agarose slices using SPIN-X™ centrifuge filter units (Costar). DNA was purified from large scale vector preparations using a QIAGEN Plasmid Maxi kit according to the manufacturer's protocol (QIAGEN). All DNA segments subjected to mutagenesis, and subcloning reactions, were sequenced in their entirety to ensure the presence of the desired mutation and proper in-frame subcloning.
Expression of Wild-type and Mutants-Sf9 cells were cotransfected with Bsu36I-digested BacPAK6 viral DNA (CLONTECH) and one of the mutated cGB-8/14 clones in the pBacPAK9 expression vector by the lipofection method according to the protocol from CLONTECH. At 3 days post-infection, the cotransfection supernatant was collected, amplified twice in Sf9 cells, and then used directly as virus stock for expression without additional purification of recombinant viruses. High Five cells (Invitrogen) grown at 27°C in complete Grace's insect medium (Invitrogen) with 10% fetal bovine serum (Intergen) and 10 g/ml gentamycin (Life Technologies, Inc.) in T-185 flasks were infected by 5 ml of virus stock/flask. The culture medium was harvested at 96 h post-infection. Recombinant enzyme production was calculated from Equation 1.
cGB-PDE in sample ϭ ͑total protein in sample͒͑specific enzyme activity in sample͒͑n͒ ͑specific enzyme activity of purified cGB-PDE͒ (Eq. 1) Specific enzyme activity for purified cGB-PDE was taken as 2.5 mol/ mg/min, and n is defined as k cat,wild-type /k cat,mutant .
Purification of Wild-type and Mutant cGB-PDEs-The culture medium (ϳ250 ml) was fractionated by sequential ammonium sulfate precipitation at 4°C. The fraction precipitated by 25-40% saturation was resuspended in 30 ml of 10 mM sodium phosphate buffer, pH 7.2, and centrifuged at 48,000 ϫ g for 30 min at 4°C. The supernatant was loaded onto a hydroxyapatite (Bio-Rad) column (1.5 ϫ 15 cm) equilibrated with 10 mM sodium phosphate buffer, pH 7.2. The column was washed with 100 ml of 70 mM sodium phosphate buffer, pH 7.2, and then eluted with 120 mM sodium phosphate buffer, pH 7.2, at a flow rate of 5 ml/h. The pool containing cGB-PDE activity was diluted with six volumes of ice-cold deionized water and concentrated to approximately 1 ml using an Amicon filtration cell equipped with a PM-30 membrane. All purification steps were performed at 4°C. The final preparation was stored in 20% glycerol at Ϫ70°C.
Catalytic Activity of cGB-PDE-PDE activity was measured using a modification of the assay procedure described previously (12). Incubation mixtures contained 40 mM MOPS, pH 7.5, 0.5 mM EGTA, 15 mM magnesium acetate, 0.15 mg/ml bovine serum albumin, 20 M cGMP (unless otherwise stated), [ 3 H]cGMP (100,000 -150,000 cpm/assay), and one of the cGB-PDE samples, in a total volume of 250 l. The incubation time was 10 min at 30°C. The reaction was stopped by placing the tubes in a boiling water bath for 3 min. After cooling, 20 l of 10 mg/ml C. atrox snake venom was added, followed by a 20-min incubation at 30°C. Nucleoside products were separated from unreacted nucleotides on the columns with DEAE Sephadex A-25 equilibrated with 20 mM Tris-HCl buffer, pH 7.5, and counted. In all studies, less than 15% of the total [ 3 H]cGMP was hydrolyzed during the reaction. The apparent K m and V max values were determined from Lineweaver-Burk plots after assaying PDE activity in duplicate at 1-250 M cGMP. k cat was obtained by dividing V max by the molar enzyme concentration. The molar enzyme concentration was calculated as described below under "Other Methods." To determine IC 50 values for zaprinast, the PDE activity was assayed in duplicate in the presence of 0.5-30 M zaprinast. All values determined represent at least three measurements using at least two different PDE preparations.
cGMP Saturation Binding-The cGMP saturation binding assay was conducted in a total volume of 60 l containing 10 mM sodium phosphate buffer, pH 6.8, 1 mM EDTA, 0.2 mM 3-isobutyl-1-methylxanthine, 0.5 mg/ml histone VIII-S, and 0.5-25 M [ 3 H]cGMP. The reaction was initiated by addition of an aliquot of enzyme. Following a 60-min incubation on ice, assay mixtures were filtered onto premoistened Millipore HAWP filters (pore size, 0.45 m), which were then rinsed four times with a total of 4 ml of cold 10 mM sodium phosphate buffer, pH 6.8, with 1 mM EDTA, and then dried and counted.The data were corrected by subtraction of nonspecific binding, which was defined as either the [ 3 H]cGMP bound in the absence of cGB-PDE or the [ 3 H]cGMP bound in the presence of a 100-fold excess of unlabeled cGMP. A similar 2-4% of nonspecific binding was obtained with each method. The data were subjected to nonlinear least squares analysis using the program MINSQ II (Micromath Scientific Software, Salt Lake City, UT) to obtain the dissociation constant (K d ).
Other Methods-SDS-electrophoresis in 10% polyacrylamide gels and Western blot analysis were done as described previously (12). Total protein concentrations were determined by the method of Bradford (14) using bovine serum albumin as the standard. To determine the cGB-PDE protein concentration, the Coomassie Brilliant Blue-stained SDSpolyacrylamide gels of wild-type and mutant enzymes were scanned using an E-C Apparatus Corp. densitometer equipped with GS370 v.3.0 software from Hoeffer. The cGB-PDE protein concentration was calculated from the fraction of the cGB-PDE band times the total protein concentration determined by Bradford assay. To convert the cGB-PDE protein concentration into the molar cGB-PDE concentration, the value of the molecular weight of cGB-PDE of 98.5 kDa (calculated from the amino acid sequence of cGB-PDE) was used.

RESULTS
Mutagenesis Strategy-The sequence alignments of the conserved catalytic domain of different PDEs have been published (2,10,15,16). These studies revealed two blocks of conserved amino acid residues (Tyr 596 -His 675 and Asp 754 -Glu 783 in the case of cGB-PDE) separated by a variable sequence containing two invariant residues (Thr 713 and Asp 714 in the case of cGB-PDE) located approximately in the middle of this sequence (Fig. 1). It has been suggested that the first block is responsible for Zn 2ϩ binding and could be part of the catalytic machinery of PDEs (13). Mutational studies on one of the PDE4 isozymes have shown that replacement of invariant His 278 , His 311 , or Thr 349 (corresponding to His 643 , His 675 , or Thr 713 in cGB-PDE) decreased the V max of this enzyme, but K m measurements for substrate were not reported (7). The second block possesses some general sequence similarity with the allosteric cGMPbinding sites (12) and could be involved in substrate binding. These findings prompted us to systematically assess the functional role of individual conserved amino acids using scanning mutagenesis.
Twelve amino acid residues in the first block were substituted singly by alanine. Thr and Asp of the TD dyad and seven residues in the second block were also replaced by alanine. Lys and Phe in the second block were replaced by Met and Leu, respectively.
Expression and Purification of Wild-type and Mutated Forms of cGB-PDE-Wild-type and mutants of the bovine lung cGB-PDE were expressed in High Five cells as described under "Experimental Procedures." The levels of expression of most of the mutants were comparable to that of the wild-type enzyme. The total production of recombinant cGB-PDEs was approximately 1-6 mg/100 ml of culture. The wild-type and mutant cGB-PDEs were partially purified similarly from culture medium using ammonium sulfate precipitation and hydroxyapatite chromatography as described under "Experimental Procedures." There was no noticeable difference in binding to and subsequent elution of these proteins from the hydroxyapatite column compared with that for the wild-type enzyme. Fig. 2 shows a Coomassie Blue-stained SDS-polyacrylamide gel of partially purified mutants obtained following the hydroxyapatite column step. All mutated cGB-PDEs migrated with essentially the same mobility as that of the wild-type enzyme. The identity of the recombinant proteins was verified by Western blot analysis (data not shown).
Kinetic Analysis of Mutants-The kinetic parameters, K m for cGMP and k cat (Table I), were determined from Lineweaver-Burk plots. The contribution of the substituted amino acid side chain to binding energy in enzyme-transition state complexes was calculated from values of the catalytic efficiency (k cat /K m ) using Equation 2.
⌬⌬G T is the change in the free energy of binding in enzymetransition state complexes attributable to the substituted group (17). R, the ideal gas constant, is equal to 1.98 ϫ 10 Ϫ3 kcal/degree/mol, and T, the temperature at which the assay was done, is equal to 303 K. The effect of substitution of the amino acid side chain that interacts with a substrate may be manifested in terms of k cat , K m , or both (17). In the present study, the binding of substrate in the transition state was chosen because the intrinsic binding energy of groups on the enzyme and substrate may not be fully realized until the enzyme-transition state complex is formed, whereby some of the binding energy may be diverted to stabilize the transition state. Previous studies have determined the magnitude of the changes in transition state binding expected for the disruption of particular interactions between enzymes and substrates (18,19). Deletion of a charged group to disrupt a hydrogen bond between the enzyme and a substrate weakened the binding energy by 3.5-4.5 kcal/mol (19), whereas the disruption of an electrostatic interaction between a charged group in the enzyme and substrate weakened binding by 2.0 kcal/mol (18). It is important to emphasize that the values for the calculated ⌬⌬G T are maximum values that include any loss of binding energy due to small perturbations of the overall conformation of the enzyme. Therefore, only the amino acid positions whereby substitutions cause large loss of function can be considered essential. Alternatively, the residues whose substitution lead to moderate changes may be involved in the general arrangement of the catalytic site.
Based on calculated ⌬⌬G T (Table I), the mutants could be arbitrarily placed into two groups. The first group includes Y596A, E632A, H674A, H675A, S756A, K760M, F776L, G780A, D781A, and E783A mutants. The changes found for these mutants are not sufficient to suggest an essential role for these residues. The second group of 13 mutants have ⌬⌬G T values in the range expected for important roles for these amino acid residues in the wild-type enzyme. This group could be divided into three categories: those defective mainly in k cat , those defective mainly in K m , and those defective in both of these parameters.
Mutants Defective in k cat -Nine mutants have k cat that is less than 15% of the wild-type value (Fig. 3), including two mutants (H643A and D754A) that retain only 0.4% of wild-type k cat . Substitution within the second block of conserved amino acid residues (Fig. 1) had little effect on k cat value, except for substitution of the invariant Asp 754 . Mutations with markedly decreased k cat were primarily clustered around the conserved HX 3 HX n E motifs of the putative Zn 2ϩ -binding site (13). The mutation of the Asp 714 in the invariant TD dyad also displayed significantly reduced PDE activity.
Five mutants (H603A, N604A, H607A, D644A, and D714A) were defective in k cat only (Table I). These residues may be directly involved in catalysis, or they may provide important structural features that allow for effective catalysis. Substitution of these residues may perturb the configuration of the active site. H643A and D754A mutants have a 7-fold increase in K m ; however, this defect is insignificant in comparison to the large decrease (270-and 280-fold, respectively) in k cat . One possible interpretation of such large changes by mutation of His 643 and Asp 754 is that these residues represent a catalytic dyad.
Mutants Defective in k cat and K m -H647A and E672A mutants possess a 10-and 14-fold increase in K m for cGMP, and an 8-and 13-fold decrease in k cat , respectively. The role of these residues cannot be interpreted unambiguously. They may be involved in catalysis, important for recognition of substrate, or provide a structural role.
Mutants Defective in K m -Four mutants (Y602A, T713A, E775A, and Q779A) were defective mainly in K m ( Table I). Two of these (Thr 713 and Gln 779 ) are uncharged amino acids and, despite the moderate changes in K m when these are substituted with alanine, the ⌬⌬G T for each of these mutants was in the range that is predicted for loss of a hydrogen bond between an enzyme polar side chain and the substrate (19). Two mutants, Y602A and E775A, exhibited profound losses in affinity for cGMP with K m values of 65 and 70 M, respectively, compared with a K m of 2 M for wild-type cGB-PDE. The ⌬⌬G T for each of these mutants is 2.9 or 2.8 kcal/mol, respectively, which is within the range expected for the loss of a salt bridge (electrostatic interaction) (18) and approaching the range (3.5-4.5 kcal/mol) expected for the loss of a hydrogen bond involving a charged residue (19).
To further probe the possible function of Tyr 602 and Glu 775 in cGMP binding to the substrate site, three additional mutants (Y602F, E775D, and E775Q) were generated, expressed and partially purified using the experimental procedures described for the major set of mutants. Y602F, E775D, and E775Q possessed the same level of expression, the same chromatographic behavior on hydroxyapatite columns, and exhibited the same mobility on the SDS-polyacrylamide gel as did wild-type enzyme. The identity of these mutants was verified by Western blot analysis (data not shown). Kinetic parameters of the Y602F mutant were indistinguishable from those of wild-type enzyme (Table II). Similarly, insignificant effect (3-fold) on K m for cGMP and on k cat was found for the E775D mutant. In contrast, the E775Q mutant was clearly defective in substrate binding.
Zaprinast-binding Site-Zaprinast selectively inhibits cGMP-specific PDEs (20). Fig. 4 shows that double reciprocal plots for wild-type cGB-PDE at various concentrations of zaprinast intersect at 1/V max ; this indicates a competitive inhibitory mechanism for zaprinast. By determining the values of K m for cGMP at different concentrations of zaprinast, the value for K i was 0.15 M under the experimental conditions used. As a competitive inhibitor, zaprinast should directly compete with cGMP for the active site of cGB-PDE. Thus, zaprinast may be used to probe the structure of the cGB-PDE active site. IC 50 values for zaprinast using the wild-type and all of the mutant enzymes are shown in Table I. To generate each IC 50 value, the cGMP concentration in the assay was one-third the K m for each mutant tested. When using low substrate concentrations, IC 50 values approach the K i for a competitive inhibitor. To decipher how the replacement of conserved residues in the catalytic domain of cGB-PDE affects the affinity for zaprinast in com-  Table I) is expressed as a percentage of k cat of wild-type cGB-PDE. The identity of the protein preparation is indicated under each bar. parison with cGMP, the IC 50 for zaprinast and the K m for cGMP of wild-type enzyme were taken as 1.0 and the corresponding values for mutants were calculated as a fold change for each parameter. Fig. 5 shows that many mutants (Y596A,  H603A, N604A, H607A, E632A, D644A, H675A, D714A,  S756A, K760M, F776L, D781A, and E783A) with small changes in K m for cGMP had small changes in IC 50 for zaprinast, suggesting that residues replaced in these mutants do not participate directly either in cGMP binding or in interaction with zaprinast. Some mutants (for example, H643A or H647A) had moderate changes in the both parameters. It is possible that residues replaced in these mutants are important for catalysis or for structural integrity of the catalytic site rather than for direct interaction with cGMP or zaprinast. The fact that the largest changes in K m (Y602A, E672A, T713A, E775A, and Q779A) did not correlate with the largest changes in IC 50 (D754A and G780A) was unexpected. The greatest increase in the IC 50 value was found for the D754A mutant, which also exhibited a profound change in catalytic activity. This residue could be a major target of zaprinast action, but the selectivity of zaprinast inhibition of the cGMP-specific PDEs must be provided by other components of the catalytic domain since aspartic acid in this position is conserved in all mammalian PDEs.
Structural Integrity of Mutants-To address structural integrity of the mutant enzymes, the deviation of all three kinetic parameters (K m , k cat , and IC 50 ) for each mutant from the same parameters of wild-type enzyme was calculated. If the value for any mutant exhibited a small change (7-fold or less), this would imply that this mutant preserved the overall structure typical for wild-type enzyme, and this mutant was not included for further characterization. By following this rule, Y602A, H643A, H647A, E672A, D754A, E775A, and E775Q mutants were selected for additional characterization. cGB-PDE contains two allosteric cGMP-binding sites that are located toward the NH 2 terminus of the protein molecule, and which are distinct from the site of cGMP hydrolysis. The [ 3 H]cGMP filterbinding assay was used to examine the affinity of cGMP binding to these sites. The [ 3 H]cGMP-binding concentration curves were almost indistinguishable (data not shown) for these proteins, and the data subjected to nonlinear least squares analysis did not show a significant difference in dissociation constant (K d ). The K d values of wild-type and Y602A, H643A, H647A, E672A, D754A, E775A, and E775Q mutants were found to be approximately equal (1.3, 1.8, 1.3, 1.3, 1.5, 1.3, 2.0, and 1.8 M, respectively). The binding stoichiometry for these seven mutants was in the range 0.5-0.6 mol of [ 3 H]cGMP/ monomer, which approximates the value for native cGB-PDE reported previously (6).
Taken together, these data imply that differences in K m , k cat , and IC 50 values (Tables I and II) are not due to nonspecific conformational effects induced by the mutations, and that all mutants preserved the overall structure typical for wild-type enzyme.

DISCUSSION
The strategy of scanning mutagenesis is widely used to identify potentially important residues involved in protein function. Once identified, these residues can be analyzed more extensively to further examine their functional and structural role.  Table I) are expressed as a fold increase with respect to the wild-type cGB-PDE. The identity of the protein preparation is indicated under each bar. In this study, the possible role of 23 amino acid residues that are conserved in all known catalytic domain of PDEs has been evaluated by systematic substitution of these residues in cGB-PDE. All mutants were expressed as full-length enzymes with similar levels of production and the same chromatographic properties. This indicates that the substitutions did not cause gross perturbations to the tertiary conformation and subsequent destabilization or proteolysis of the enzymes. An interesting finding was that the residues that are most important for catalysis (His 643 and Asp 754 ) and for substrate binding (Tyr 602 and Glu 775 ) are not located in two different blocks of conserved residues. Furthermore, the invariant TD dyad, which is located in the highly variable sequence between these two blocks, contributes to catalysis (Asp 714 ) and substrate binding (Thr 713 ). Such distribution of the functionally important residues assumes that the active site may be formed at the interface between these blocks of conserved residues. Thus, the catalytic and substrate-binding components are overlapping.
Nine residues (His 603 , Asn 604 , His 607 , His 643 , Asp 644 , His 647 , Glu 672 , Asp 714 , and Asp 754 ) have been found to be important for catalysis in cGB-PDE. Substitution of either of two of them, His 643 or Asp 754 , is accompanied by the largest decrease in catalytic efficiency. His 643 of cGB-PDE corresponds to His 278 of a cAMP-specific PDE (RNPDE4D). The latter residue was shown to be critical for PDE4 activity (7), and our results are consistent with this observation. In addition to His 278 , two more residues (His 311 and Thr 349 ) of PDE4 were found to be critical for catalytic activity in PDE4 (7). Substitution of the corresponding positions of cGB-PDE (His 675 and Thr 713 ) produced little effect on catalysis. Instead of a possible His/Thr catalytic dyad in PDE4, a possible His/Asp catalytic dyad was uncovered in cGB-PDE (PDE5). This could be an interesting point of discrimination between cAMP-specific PDEs and cGMP-specific PDEs.
A full understanding of PDE catalysis is complicated by the fact that the activities of PDEs are supported by a number of divalent metal cations (13,21,22). The catalytic activity of cGB-PDE is supported by Zn 2ϩ , Mn 2ϩ , Co 2ϩ , and Mg 2ϩ , but Zn 2ϩ promotes cGB-PDE activity at lower concentrations than do other cations (13). Sequence alignment reveals two putative Zn 2ϩ -binding motifs which are conserved in all mammalian PDEs (13). These motifs include His 603 , His 607 , and Glu 632 followed by His 643 , His 647 , and Glu 672 residues of cGB-PDE. All of these residues were replaced with alanine in the present study, and five of six of these mutants exhibited a marked decrease in catalytic efficiency (Table I). The only residue in this group which seems nonessential for catalysis is Glu 632 . When histidine residues of HSPDE4A at positions 433, 437, 473, and 477, which correspond to His 603 , His 607 , His 643 , and His 647 of cGB-PDE, respectively, were changed independently to serine residues, cAMP hydrolyzing activity of HSPDE4A was substantially reduced (23). Our results are consistent with this observation. Asn 604 and Asp 644 are flanking residues for His 603 and His 643 , respectively. In traditional Zn 2ϩ -binding sites of different hydrolases, the flanking position of the first histidine is occupied by glutamic acid (His-Glu-X 2 -His-X n -Glu). Mutagenesis of this glutamic acid led to a drastic decrease of the catalytic activity of aminopeptidase A (24). Replacement of Asn 604 or Asp 644 by alanine in cGB-PDE decreased the k cat approximately 9-fold. Unfortunately, the basal activity of cGB-PDE in the metal-free assay was strongly diminished, and for mutants defective in k cat (in the range of 10 -250-fold) it was technically difficult to measure the enhancement of activity due to Zn 2ϩ . Nevertheless, our data emphasize an important role for His 603 , His 607 , His 643 , His 647 , Glu 672 , Asp 714 , and Asp 754 for catalysis, and one possible function of this combination of residues is to provide a Zn 2ϩ -binding site(s). However, it appears that a single Zn 2ϩ -binding motif is insufficient to support normal catalysis in PDEs. It should also be mentioned that a His/His/Glu site has been described for binding Mn 2ϩ in the active site of 3,4-dihydroxyphenylacetate 2,3-dioxygenase (25), and several enzymes contain multinuclear metal-binding sites that are created by multiple histidines and acidic residues. The PDE2 exhibits highest activity with Mn 2ϩ (21), and cGB-PDE activity is also supported by Mn 2ϩ (13). It is possible that the same site of cGB-PDE, which tightly binds Zn 2ϩ , can bind other divalent cations as well. However, in studies of the rod outer segment PDE (PDE6), prolonged treatment with chelators inactivates the enzyme even in the presence of Mn 2ϩ , Mg 2ϩ or Co 2ϩ , and catalytic activity is restored by addition of Zn 2ϩ . 2 Y602A, T713A, E775A, and Q779A mutants were defective in K m with small or negligible changes in k cat . The k cat values of these mutants are critical for interpretation, because it is assumed that Tyr 602 , Thr 713 , Glu 775 , and Gln 779 residues of cGB-PDE are important for binding the cGMP substrate. Unfortunately, the magnitude of K m changes for T713A and Q779A mutants cannot be interpreted unambiguously, because the values for the calculated ⌬⌬G T are maximum values that include any loss of binding energy due to small perturbations of the overall conformation of the enzyme. Only the amino acid positions whereby substitutions cause large loss of function can be considered essential. Alternatively, the residues whose substitution lead to moderate changes in substrate binding may be involved in the general arrangement of the substrate-binding site, and do not necessarily interact directly with substrate. For this reason, only two residues (Tyr 602 and Glu 775 ), for which substitution exhibited the largest changes in K m (Table I) were selected for further analysis, and additional mutations (Y602F, E775D, and E775Q) were generated and analyzed (Table II). The possible roles of Tyr 602 and Glu 775 in cGMP binding in the catalytic site of cGB-PDE are discussed below.
Tyrosine 602-There are several options for the tyrosine residue to interact with cGMP: 1) a hydrogen bond to an oxygen atom on the phosphoryl group, 2) a hydrogen bond to the 2Ј-OH group of the ribose ring, and 3) stacking interaction with the base. These types of interactions for a tyrosine residue have been found in the complex of ribonuclease T1 with 2Ј-GMP (26), in the nonphysiological complex of nucleoside diphosphate kinase with cAMP (27), in the AMP-binding site of fructose-1,6bisphosphatase (28), and in site B of the regulatory subunit of protein kinase A (29). To scrutinize the possible functions of Tyr 602 in the cGB-PDE catalytic site, results of cyclic nucleotide analogs as competitive inhibitors of cGB-PDE (30) can also be considered. The published results indicate that the 2Ј-OH group of cGMP is not a major requirement for binding to the catalytic site of cGB-PDE (30,31). Tyr 602 is invariant in all mammalian PDEs, but in Drosophila dunce PDE (32) and in Saccharomyces cerevisiae PDE (33), the corresponding Tyr 602 position is occupied by Phe. Phe could mimic Tyr for possible participation in stacking interactions with the guanine base. The Y602F mutant was indistinguishable from wild-type enzyme in terms of K m and k cat values (Table II). This is a strong argument that Tyr 602 stacks with the purine base of cGMP in the substrate-binding site of cGB-PDE and is likely to serve a similar role in other PDEs.
Glutamic Acid 775-There are two major chemical options for the glutamic acid residue to interact with cGMP: 1) a hydrogen bond interaction with the 2Ј-OH group of the ribose ring, such as that found in the catabolite gene activator protein