The gamma subunit of rod cGMP-phosphodiesterase blocks the enzyme catalytic site.

Cyclic GMP phosphodiesterase (PDE) is the effector enzyme in the visual transduction cascade of vertebrate photoreceptor cells. In the dark, the activity of the enzyme catalytic alpha and beta subunits (Palphabeta) is inhibited by two gamma subunits (Pgamma). Previous results have established that approximately 5-7 C-terminal residues of Pgamma comprise the inhibitory domain. To study the interaction between the Pgamma C-terminal region and Palphabeta, the Pgamma mutant (Cys68 --> Ser, and the last 4 C-terminal residues replaced with cysteine, Pgamma-1-83Cys) was labeled with the fluorescent probe 3-(bromoacetyl)-7-diethylaminocoumarin (BC) at the cysteine residue (Pgamma-1-83BC). Pgamma-1-83BC was a more potent inhibitor of PDE activity than the unlabeled mutant, suggesting that the fluorescent probe in part substitutes for the Pgamma C terminus in PDE inhibition. HoloPDE (Palphabetagamma2) had no effect on the Pgamma-1-83BC fluorescence, but the addition of Palphabeta to Pgamma-1-83BC resulted in an approximately 8-fold maximal fluorescence increase. A Kd for the Pgamma-1-83BC-Palphabeta interaction was 4.0 +/- 0.5 nM. Zaprinast, a specific competitive inhibitor of PDE, effectively displaced the Pgamma-1-83BC C terminus from its binding site on Palphabeta (IC50 = 0.9 microM). cGMP and its analogs, 8-Br-cGMP and 2'-butyryl-cGMP, also competed with the Pgamma-1-83BC C terminus for binding to Palphabeta. Our results provide new insight into the mechanism of PDE inhibition by showing that Pgamma blocks the binding of cGMP to the PDE catalytic site.

In the visual transduction cascade of rod photoreceptor cells, the photoexcited visual receptor, rhodopsin, interacts with the rod G-protein, transducin, and stimulates the exchange of GTP for bound GDP. The GTP-bound ␣ subunit of transducin dissociates from rhodopsin and the transducin ␤␥ subunits and activates the effector enzyme, cGMP phosphodiesterase (PDE), 1 by relieving the inhibitory constraint imposed by two identical inhibitory subunits of PDE (P␥) on the enzyme ␣␤ catalytic subunits (P␣␤) (for review see Refs. [1][2][3][4]. Insights into the P␥-P␣␤ interaction are critical for understanding the mechanisms of PDE inhibition by P␥ and PDE activation by transducin. Approximately 5-7 C-terminal amino acid residues of P␥ are involved in the inhibitory interaction with P␣␤ (5)(6)(7)(8)(9). Recently, using a cross-linking approach we have identified a site on P␣ for binding of the P␥ C terminus as a region P␣-751-763 within the PDE catalytic domain (10). The finding suggests that the P␥ C terminus either occupies the site for binding and catalysis of cGMP or induces local conformational changes of the PDE catalytic site that block cGMP hydrolysis. Here, we study the interaction between the C terminus of P␥ and P␣␤ using a novel fluorescence assay to elucidate the mechanism of PDE inhibition by P␥.

EXPERIMENTAL PROCEDURES
Materials-cGMP was obtained from Boehringer Mannheim. 3-(bromoacetyl)-7-diethylaminocoumarin (BC) was purchased from Molecular Probes, Inc. Trypsin and soybean trypsin inhibitor were from Worthington. Zaprinast and all other reagents were purchased from Sigma.
Preparation of Trypsin-activated PDE, P␥, and P␥ Mutants-Bovine rod outer segment (ROS) membranes were prepared by the method of Papermaster and Dreyer (11). PDE was extracted from ROS membranes as described in Ref. 12. PDE and trypsin-activated PDE (tPDE) were prepared and purified as described previously (6). The purified proteins were kept in 40% glycerol at Ϫ20°C. The P␥ subunit was expressed in Escherichia coli and purified as described in Ref. 9. The P␥ mutants P␥Cys 68 3 Ser and P␥-1-83Cys (Cys 68 3 Ser, and the last 4 C-terminal residues, Tyr-Gly-Ile-Ile, replaced with a single cysteine) were obtained as described in Ref. 10.
Preparation of P␥-1-83BC, P␥BC, and P␥-24 -45BC-To obtain P␥-1-83BC and P␥BC, a 5 mM stock solution of BC (final concentration, 200 M) was added to either 100 M P␥-1-83Cys or 100 M P␥, each in 20 mM HEPES buffer (pH 7.6), and the mixture was incubated for 30 min at room temperature. P␥-1-83BC and P␥BC were then passed through a PD-10 column (Pharmacia Biotech Inc.) equilibrated with 20 mM HEPES buffer (pH 7.6) containing 100 mM NaCl and purified by reversed-phase HPLC on a C-4 column Microsorb-MW (Rainin) using a 0 -100% gradient of acetonitrile, 0.1% trifluoroacetic acid. The preparations of P␥-1-83BC and P␥BC contained no free BC. Using ⑀ 445 ϭ 53,000 for BC, the molar ratio of BC incorporation into P␥ and P␥-1-83BC was greater than 0.8 mol/mol. P␥-24 -45BC was prepared by labeling of peptide P␥-24 -45Cys and purified as described in Ref. 13. A P␥ mutant, P␥Cys 68 3 Ser, and a peptide, P␥-24 -45, that contain no cysteine were not derivatized with BC under similar conditions, suggesting the selectivity of cysteine labeling.
Fluorescent Assays-Fluorescent assays were performed on a F-2000 Fluorescence Spectrophotometer (Hitachi) in 1 ml of 80 mM Tris-HCl buffer (pH 7.6) containing 2 mM MgCl 2 . Fluorescence of P␥-1-83BC, P␥BC, or P␥-24 -45BC was monitored with excitation at 445 nm and emission at 495 nm. The assays were carried out at equilibrium, which was typically reached less than 3 s after mixing of the components. The concentration of labeled polypeptides was determined using ⑀ 445 ϭ 53,000. Where indicated, zaprinast was added from 1 mM stock solution to an assay buffer. The K d values in Figs. 2 and 4 were calculated by fitting the data to Equation 1, where F 0 is a basal fluorescence of P␥-1-83BC or P␥-24 -45BC, F is the fluorescence after additions of tPDE, F/F 0max is the maximal relative increase of fluorescence, and X is a concentration of free tPDE. The IC 50 values in Figs. 3 and 5 were calculated by fitting the data to the one site competition equation with variable slope: * This work was supported by National Eye Institute Grant EY-10843. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
where X is a concentration of a competing ligand (zaprinast, cGMP, or cGMP analogs) and H is a Hill slope. Analytical Methods-The PDE activity was measured using the proton evolution assay of Liebman and Evanczuk (14). The assay was performed at room temperature in 200 l of 10 mM HEPES (pH 7.8) containing 100 mM NaCl and 1 mM MgCl 2 . The reaction was initiated by the addition of cGMP. The pH was monitored with a pH microelectrode (Microelectrode, Inc.). Protein concentrations were determined by the method of Bradford (15) using IgG as a standard or using calculated extinction coefficients at 280 nm. SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli (16) in 12% acrylamide gels. Fitting of the experimental data was performed with nonlinear least squares criteria using GraphPad Prizm Software. The K i , K d , and IC 50 values are expressed as the means Ϯ S.E. of three independent measurements.

Inhibition of tPDE Activity by P␥-1-83Cys, P␥-1-83BC, and
P␥BC-To study the inhibitory interaction between the P␥ C terminus and P␣␤, we have replaced C-terminal amino acid residues of P␥ with a fluorescent probe by labeling the P␥-1-83Cys mutant with BC. To determine if the fluorescent probe can substitute for the P␥ C terminus, the ability of the labeled and unlabeled mutant to inhibit tPDE was tested. Limited proteolysis of holoPDE with trypsin removes intrinsic P␥ subunits and small farnesylated and geranyl-geranylated C-terminal fragments of P␣ and P␤, respectively, leaving mainly intact active catalytic P␣␤ subunits (17,18). The P␥-1-83Cys mutant was, as expected, a weak inhibitor of tPDE activity. It inhibited maximally only ϳ50% of tPDE activity with an apparent K i value of 13 nM (not shown). The labeled mutant, P␥-1-83BC, was a significantly more potent inhibitor of tPDE than P␥-1-83Cys. P␥-1-83BC almost completely inhibited tPDE activity with an apparent K i of 5.4 Ϯ 0.5 nM (Fig. 1A). Free BC had no effect on tPDE activity at tested concentrations up to 20 M (not shown). This suggests that when the P␥ C-terminal residues, Gly-Ile-Ile, are replaced with the fluorescent probe BC, the probe binds into a complimentary hydrophobic pocket on P␣␤ and substantially restores the inhibitory potential of the P␥-1-83Cys mutant. Furthermore, the apparent K i for tPDE inhibition by P␥-1-83BC was influenced by the cGMP concentration in the assay. Decrease in cGMP concentration from 3 to 0.3 mM resulted in a reduction of the apparent K i value from 5.4 to 2.1 nM (Fig. 1A). In control experiments, labeling of wild-type P␥ with BC at Cys 68 did not notably affect the ability of P␥ to inhibit tPDE (not shown). Both P␥ and P␥BC fully inhibited tPDE with K i values less than 0.25 nM (not shown), and the K i values were not affected by the concentration of cGMP in the assay (not shown). Next we examined effects of P␥-1-83BC and P␥ on K m values of cGMP hydrolysis by tPDE. A K m of 90 M was calculated from the Michaelis-Menten plot for tPDE (Fig.  1B). This K m is consistent with earlier estimates (17). In the presence of 3.5 and 8 nM of P␥-1-83BC, the apparent K m values were 215 and 480 M, and V max values were 95 and 80%, respectively (Fig. 1B). The inhibitory interaction between P␥-1-83BC and P␣␤ was not purely competitive with cGMP, because the V max values were also affected by P␥-1-83BC (Fig.  1B). In agreement with previous data (17), the addition of P␥ did not significantly change the K m for cGMP hydrolysis by tPDE (not shown).
Binding of P␥-1-83BC to P␣␤-Addition of tPDE to P␥-1-83BC increased the fluorescence of the probe many fold in a dose-dependent manner (Fig. 2). The binding curve shows a single class of binding sites with K d ϭ 4.0 Ϯ 0.5 nM and a maximal fluorescence enhancement F/F 0 ϭ 8.4 Ϯ 0.2. No change in the fluorescence of P␥-1-83BC was detected upon the addition of holoPDE or tPDE reconstituted with P␥. Furthermore, addition of P␥ to the P␥-1-83BC-P␣␤ complex readily reduced fluorescence to a basal F 0 level, suggesting that the fluorescent increase reflects a specific interaction between P␣␤ and P␥-1-83BC (not shown). Fluorescence of P␥BC was not notably affected in the presence of tPDE. It appears that in the P␥BC-P␣␤ complex the probe is oriented away from the P␣␤ subunits, whereas in the P␥-1-83BC-P␣␤ complex the probe occupies the pocket for binding of the P␥ C terminus.
Effects of Zaprinast on the Interaction between P␥-1-83BC and P␣␤-Zaprinast is a well characterized competitive inhibitor of photoreceptor PDEs and cGMP-binding, cGMP-specific PDE (19,20). We have investigated effects of zaprinast on the interaction between P␥-1-83BC and P␣␤. Zaprinast had no effect on the basal fluorescence of P␥-1-83BC. Addition of increasing concentrations of zaprinast resulted in a complete reversal of the fluorescent enhancement of P␥-1-83BC bound to P␣␤ (IC 50 of 0.9 M) (Fig. 3). Zaprinast was effective in blocking the P␥-1-83BC-P␣␤ interaction within a pharmacologically relevant range of concentrations. A K d value of 140 nM for zaprinast binding to rod PDE catalytic sites was calculated based on inhibition of PDE activity by zaprinast (19). Our assay does not allow us to calculate the true K d value for the zapri-

Mechanism of Rod cGMP Phosphodiesterase Inhibition 11687
nast binding from the curve in Fig. 3 because zaprinast does not compete for the P␥-24 -45BC-P␣␤ interaction (see below) and cannot completely displace P␥-1-83BC from P␣␤. A K d value for the zaprinast-PDE interaction lower than an IC 50 value of 0.9 M (Fig. 3) can be predicted.
To examine the effects of zaprinast on the apparent K d of P␥-1-83BC binding to P␣␤, the binding curves were obtained in the presence of different concentrations of zaprinast (Fig. 4). Increasing concentrations of zaprinast reduced the fluorescent enhancement of P␥-1-83BC by P␣␤ and increased the apparent K d of the P␥-1-83BC binding to P␣␤. The apparent K d calculated from the binding curves in the presence of 2 and 4 M of zaprinast were 14.7 Ϯ 0.9 and 22.6 Ϯ 1.3 nM, respectively. The decrease in fluorescence of the P␥-1-83BC-P␣␤ complex and the increase in apparent K d values suggest that zaprinast competitively displaces the fluorescently labeled P␥-1-83BC C terminus from the binding pocket on P␣␤. Dipyridamole, another potent competitive inhibitor of photoreceptor PDEs (19), was unsuitable for studies using our assay. Dipyridamole is highly fluorescent with maximal emission at 480 nm.
Effects of Zaprinast on the Interaction between P␥-24 -45BC and P␣␤-To investigate if zaprinast can compete for binding between the polycationic region of P␥, P␥-24 -45, and P␣␤, we utilized an assay of interaction between a synthetic peptide, P␥-24 -45Cys, labeled with BC and tPDE (13). Zaprinast at concentrations that completely reversed the fluorescent enhancement of the P␥-1-83BC-P␣␤ complex had no effect on the fluorescent increase of P␥-24 -45BC caused by its binding to tPDE (not shown). At a higher concentration (10 M), zaprinast reduced the maximal increase in fluorescence of the P␥-24 -45BC-P␣␤ complex by only ϳ15% without affecting the K d . The K d values for P␥-24 -45BC binding to P␣␤ in the presence of 10 M of zaprinast or with no zaprinast added were indistinguishable (ϳ26 nM) (not shown). Interestingly, the apparent K d values for P␥-1-83BC binding to P␣␤ at higher concentrations of zaprinast (Fig. 4) approached the K d for the P␥-24 -45BC-P␣␤ complex. This supports the notion that zaprinast competes with the P␥-1-83BC C terminus for binding to P␣␤ and does not affect the interaction of the polycationic region, P␥-24 -45, with P␣␤.
Effects of cGMP and Its Analogs on the Interaction between P␥-1-83BC and P␣␤-The relative potency of zaprinast in competition with the P␥-1-83BC C terminus for binding to P␣␤ indicated that cGMP and its analogs might be effective as well. Effects of cGMP on the P␥-1-83BC-P␣␤ interaction were tested in the presence or the absence of Mg 2ϩ . Mg 2ϩ participates in binding of cGMP to the PDE catalytic site and is critical for cGMP hydrolysis. In the presence of Mg 2ϩ , cGMP reversed the fluorescent enhancement of P␥-1-83BC bound to P␣␤ with an apparent IC 50 of 0.77 mM (Fig. 5A). Presumably, cGMP would be significantly more effective in the absence of cGMP hydrolysis. In the presence of EDTA, cGMP was ϳ2.5-fold less potent (IC 50 of 1.9 mM, Fig. 5A). Because cGMP is not hydrolyzed in the presence of EDTA, the data suggest that Mg 2ϩ enhances affinity of cGMP for the catalytic site by more than 2.5-fold. We also tested two cGMP analogs, 8-Br-cGMP and 2Ј-butyryl-cGMP, because they block ROS-PDE activity with a relatively high affinity (21). Both cGMP analogs competed with the P␥-1-83BC C terminus for binding to P␣␤ (Fig. 5B). The IC 50 values from the competition curves for 8-Br-cGMP and 2Јbutyryl-cGMP were 0.42 mM and 0.58 mM, respectively. Perhaps, an IC 50 for 2Ј-butyryl-cGMP was higher than an IC 50 for 8-Br-cGMP because tPDE hydrolyzed 2Ј-butyryl-cGMP with a rate of ϳ15% of cGMP hydrolysis, whereas 8-Br-cGMP was resistant to the hydrolysis (not shown). DISCUSSION Interaction between the inhibitory ␥ subunits of PDE and the catalytic P␣␤ subunits is essential for blocking PDE activity in Mechanism of Rod cGMP Phosphodiesterase Inhibition 11688 the dark and for inactivation of the enzyme upon recovery of the photoreceptor cell from light stimulation. The P␥ subunits bind to P␣␤ with very high affinity (K d Ͻ 100 pM) (22). The high affinity of the P␥-P␣␤ interaction is provided by two major binding sites on P␥, the central polycationic region, P␥-24 -45, and the C-terminal 5-7 amino acid residues (5)(6)(7)(8)(9). The main role of the P␥-24 -45 region is to enhance the affinity of P␥ interaction with P␣␤. The C terminus of P␥ is critical for PDE inhibition. Truncations of the P␥ C-terminal residues lead to a loss of the P␥ inhibitory function (5,8,9). Peptides corresponding to the C-terminal region of P␥ can fully inhibit PDE activity (6,7,9). Recently, we have shown that the C-terminal region of P␥ binds within the catalytic domain of PDE (10). This finding raised the possibility that P␥ may inhibit PDE activity by physically blocking the binding site for cGMP. An alternative mechanism of PDE inhibition by P␥ would be a local conformational change of the PDE catalytic site that prevents cGMP hydrolysis. Standard analysis for competitive (noncompetitive) inhibition of cGMP hydrolysis may not discriminate between the two mechanisms because P␥ binds to P␣␤ very tightly (K d Ͻ100 pM) compared with cGMP binding (K m for cGMP is within 17-80 M range) (17,19,22). The large differences in affinity and very slow off-rates of P␥ from P␣␤ (Ͼ10 min) (8,22) may not allow cGMP to compete with P␥ bound to P␣␤. It has been shown previously that the addition of P␥ to tPDE caused very little change in the apparent K m value (17). Indeed, in our experiments the K m value of 90 M was unaffected by the addition of P␥.
To study the mechanism of PDE inhibition by P␥, we developed an assay that reports binding of the P␥ C terminus to P␣␤. The assay utilizes a P␥ mutant with the C-terminal amino acid residues replaced with a fluorescent probe, BC. The fluorescently labeled mutant, P␥-1-83BC, was a more potent inhibitor of PDE activity than the unlabeled mutant, P␥-1-83Cys, suggesting that the probe interacts with the inhibitory pocket on P␣␤. Addition of P␥-1-83BC to P␣␤ led to a dose-dependent increase of the apparent K m values for cGMP hydrolysis. Binding of P␥-1-83BC to P␣␤ produced a large ϳ8-fold increase in the probe fluorescence. Zaprinast, a specific competitive inhibitor of photoreceptor PDEs, effectively competed for the interaction between P␥-1-83BC and P␣␤, but had no effect on binding of the polycationic region, P␥-24 -45, to P␣␤. Perhaps the fact that P␥-1-83BC binds to P␣␤ (K d of 4 nM) less tightly than P␥ has helped zaprinast to compete for the P␥-1-83BC-P␣␤ interaction. cGMP and its analogs, 8-Br-cGMP and 2Јbutyryl-cGMP, were also effective in blocking the interaction between the P␥-1-83BC C terminus and P␣␤ using the fluorescent assay. Effects of cGMP and its analogs on the P␥-1-83BC binding to P␣␤ cannot be attributed to the noncatalytic cGMP-binding sites of PDE, because bovine rod PDE contains two molecules of tightly bound cGMP with an extremely slow off-rate (t1 ⁄2 ϭ ϳ4 h) (23).
Overall, our data strongly suggest that P␥ inhibits PDE activity by physically blocking access of the substrate, cGMP, to the PDE catalytic site. The region of P␣, P␣-751-763, that interacts with the C terminus of P␥ (10) is adjacent to the NKXD motif. In G-proteins, the NKXD motif specifies binding of the GTP guanine ring (24,25). Based on our results, it is likely that the NKXD motif is involved in the binding of cGMP by photoreceptor PDEs.

Mechanism of Rod cGMP Phosphodiesterase Inhibition
11689