Structural Requirements of the Photoreceptor Phosphodiesterase γ-Subunit for Inhibition of Rod PDE6 Holoenzyme and for Its Activation by Transducin*

The central enzyme of the visual transduction cascade, cGMP phosphodiesterase (PDE6), is regulated by its γ-subunit (Pγ), whose inhibitory constraint is released upon binding of activated transducin. It is generally believed that the last four or five C-terminal amino acid residues of Pγ are responsible for blocking catalysis. In this paper, we showed that the last 10 C-terminal residues (Pγ78–87) are the minimum required to completely block catalysis. The kinetic mechanism of inhibition by the Pγ C terminus depends on which substrate is undergoing catalysis. We also discovered a second mechanism of Pγ inhibition that does not require this C-terminal region and that is capable of inhibiting up to 80% of the maximal cGMP hydrolytic rate. Furthermore, amino acids 63–70 and/or the intact α2 helix of Pγ stabilize binding of C-terminal Pγ peptides by 100-fold. When PDE6 catalytic subunits were reconstituted with portions of the Pγ molecule and tested for activation by transducin, we found that the C-terminal region (Pγ63–87) by itself could not be displaced but that transducin could relieve inhibition of certain Pγ truncation mutants. Our results are consistent with two distinct mechanisms of Pγ inhibition of PDE6. One involves direct interaction of the C-terminal residues with the catalytic site. A second regulatory mechanism may involve binding of other regions of Pγ to the catalytic domain, thereby allosterically reducing the catalytic rate. Transducin activation of PDE6 appears to require interaction with both the C terminus and other regions of Pγ to effectively relieve its inhibitory constraint.

The photoreceptor cyclic nucleotide phosphodiesterase (PDE6) 2 is the central enzyme in the vertebrate visual signaling pathway in rods and cones. Phototransduction is initiated when light induces the isomerization of the 11-cis-retinal chromophore of rhodopsin, which leads to activation of the photoreceptor-specific G-protein, transducin. Transducin binds GTP and releases its activated ␣-subunit (T␣*-GTP) to activate membrane-associated rod PDE6 by relieving the inhibition of the ␥-subunit at the active sites. The activation of PDE6 results in rapid lowering of cGMP levels, closure of cGMP gated ion channels, and hyperpolarization of the cell membrane (1)(2)(3)(4)(5).
The PDE6 holoenzyme consists of a catalytic dimer of ␣and ␤-subunits (P␣␤) and two inhibitory ␥-subunits (P␥) that are tightly bound to P␣␤. Considering its small size, the P␥ subunit of rod and cone PDE6 serves a remarkable number of functions (reviewed in Refs. (6 and 7): 1) a primary function of the P␥ subunit is to inhibit catalysis of cGMP by binding to the catalytic domain of PDE6; 2) the P␥ subunit also enhances the binding affinity of cGMP to the regulatory GAF (cGMP-regulated PDEs, certain adenylate cyclases, and the transcription factor Fh1A of bacteria) domain; 3) activated transducin binds to the P␥ subunit, leading to deinhibition of PDE6 at its active site; and 4) during deactivation, the ␥-subunit participates in a protein complex with the RGS9 and other proteins to accelerate the GTPase activity of activated transducin.
Despite a wealth of biochemical and structural data, there is currently no consensus in the literature on the exact structural element(s) responsible for inhibition of cyclic nucleotide hydrolysis at the PDE6 active site. For example, C-terminally truncated P␥ mutants lacking the last 5-10 amino acids have been reported to result in no inhibition (9,22,23) up to 50% inhibition (15) of catalytic activity. There are also reports that inhibition of PDE6 catalysis can occur without an absolute requirement of the extreme C-terminal amino acids (8,(15)(16)(17).
The kinetic mechanism of P␥ inhibition of PDE6 catalysis is generally believed to occur by a simple competition of P␥ and substrate for access to the active site. Although this model is supported by kinetic studies showing competition between cAMP hydrolysis and P␥ binding at the active site (12), similar experiments using cGMP as the substrate are not consistent with a simple competitive inhibition mechanism (24). Pharmacological studies with active site inhibitors also reveal complex kinetic behavior of P␥, PDE inhibitors, and substrate at the PDE6 active site (25)(26)(27), suggesting that a more complex mechanism of inhibition by P␥ may be occurring.
The same two functional domains of P␥ that interact with PDE6 have also been shown to bind to activated transducin ␣-subunit. The polycationic, GAF domain-interacting region of P␥ (amino acids 24 -45) interacts with transducin, although the functional significance of this interaction for the mechanism of transducin activation is not clear (28 -31). The C-terminal region of P␥ is bound by activated transducin during PDE6 activation (8,9,20,32,33). Furthermore, during the deactivation process, this same C-terminal region of P␥ forms a complex with the RGS9-1 complex to help accelerate the GTPase activity of transducin (20) and speed the deactivation process.
In this paper, we documented two distinct mechanisms of P␥ inhibition: one using the last 10 C-terminal residues of P␥ to "block" the active site and a second one in which amino acid residues between positions 61 and 76 of P␥ can "induce" an inhibitory conformational change in the catalytic domain. Furthermore, we defined a similar region of P␥ (beginning approximately at amino acid 61 and extending to the end of the ␣2 helical region) that appears able to "sense" conformational changes in the catalytic domain. The ability of transducin to relieve inhibition of some-but not all-of our P␥ constructs suggests that transducin requires multiple sites of interaction with P␥, including interactions distinct from its role in displacing the P␥ C terminus from the enzyme active site.

EXPERIMENTAL PROCEDURES
Materials-Bovine retinas were purchased from W. L. Lawson, Inc. Synthetic peptides P␥63-87, P␥71-87 P␥74 -87, P␥78 -87, and P␥81-87 were purchased from New England Peptide and repurified to Ͼ95% purity by reverse phase high pressure liquid chromatography. Ultima Gold scintillation fluid was from PerkinElmer Life Sciences. The filtration membranes were from Millipore, the bicinchoninic acid protein assay reagents were from Pierce, and all of other chemicals were from Sigma-Aldrich. The primers for the construction of C-terminally truncated P␥ mutants were from Invitrogen. The plasmid mini and midi preparation kits were from Qiagen.
Construction of P␥ Mutants-The P␥ truncation mutants lacking various lengths of the C-terminal amino acids were generated using PCR; P␥1-83 and P␥1-86 were constructed as previously described (15). For the mutants P␥1-60, P␥1-70, and P␥1-80, the constructs were chemically synthesized (Blue Heron Biotechnology). For PCR constructs, the primers were designed based on the sequence of rod bovine P␥ (primer sequences available on request), and the PCR products were inserted into the NdeI and BamHI sites of the pET11a vector (Novagen). The sequences of all P␥ mutants were confirmed by DNA sequencing using Applied Biosystems BigDye terminator cycle sequencing kits in the Hubbard Center for Genome Studies at the University of New Hampshire.
Purification of P␥ Truncation Mutants-Recombinant P␥ truncation mutants were expressed in Escherichia coli BL21/ DE3. The bacterial extract was partially purified by cation exchange chromatography using SP Sepharose, followed by C4 reverse phase high pressure liquid chromatography (34). The purity (Ͼ95%) and size of these proteins were evaluated by SDS-polyacrylamide gel electrophoresis. Most mutants were also examined by analytical ultracentrifugation using sedimentation velocity analysis to confirm the absence of aggregates or self-association of the constructs. Protein concentrations were determined by the bicinchoninic acid protein assay (35) using bovine ␥-globulin as a standard.
PDE6 and P␣␤ Purification and Functional Assays-Bovine rod PDE6 was purified from bovine retinas as described (36). P␣␤ catalytic dimers lacking P␥ were prepared by limited trypsin proteolysis and repurified by Mono Q anion exchange chromatography to remove proteolytic fragments of P␥ (36). The PDE6 concentration was estimated based on the rate of cGMP hydrolysis of trypsin-activated PDE6 and a knowledge of the k cat of the enzyme (5600 mol cGMP hydrolyzed per P␣␤ per s) (37); this estimate was validated by determining stoichiometric binding of [ 3 H]cGMP to PDE6 with a filter binding assay (38).
To prepare reconstituted PDE6 with cGMP bound to the regulatory GAF domains, purified P␣␤ was incubated with 10 mM EDTA for 2 h at room temperature, followed by the addition of 1 M cGMP in the presence of truncation mutants (P␥1-80); the inclusion of P␥ constructs containing the proline-rich and polycationic region (see Fig. 1A) with P␣␤ served to sufficiently stabilize cGMP binding to the noncatalytic binding sites so that these sites remained liganded for the duration of the activity assay. Occupancy of the GAP domains with bound cGMP was verified experimentally as described previously (39). After adding 10 mM MgCl 2 to restore the ability of PDE6 to undergo catalysis, cyclic nucleotide hydrolytic rates were measured in 20 mM Tris, 10 mM MgCl 2 , 0.5 mg/ml bovine serum albumin, and the indicated concentration of cyclic nucleotides using either a phosphate release microplate assay or a radiotracer assay (40).
Purification of Activated Transducin ␣-Subunit (T␣*) and Transducin Activation of Reconstituted P␣␤ and P␥ Mutants-Transducin ␣-subunits were extracted from the PDE6-depleted bovine ROS membranes by the addition of 50 M GTP␥S. The extracted T␣*-GTP␥S was purified on a Blue Sepharose column as described (41,42), followed by gel filtration chromatography to completely remove PDE6. The concentration of T␣*-GTP␥S was determined by a colorimetric protein assay. Purified T*␣-GTP␥S was stored at 4°C and used for a few weeks. For the transducin activation measurement, purified P␣␤ was preincubated with P␥ mutants or peptides at the indicated concentrations to maximally inhibit PDE activity. For the case of P␥63-87, the concentration (10 M) was chosen to suppress ϳ70% of PDE activity, enabling us to detect the ability of transducin to stimulate PDE activity. 9 M activated transducin (supplemented with 50 M GTP␥S) was added to above mixture and incubated for 5 min. The PDE activity was measured using radiotracer assay using 0.1 mM [ 3 H]cAMP as substrate.
Data Analysis-Except where noted, the experiments were repeated at least three times, and the average results were reported as the means Ϯ S.D. The curve fitting was performed with the computer program SigmaPlot (SPSS, Inc., Chicago, IL). The affinity of peptides (IC 50 ) was determined by fitting experimental data to a three-or four-parameter logistic dose response function (43).

RESULTS
The Last 10 Amino Acids of P␥ Are Sufficient to Fully Inhibit Catalysis-Although it is known that the C-terminal region of P␥ interacts with the catalytic pocket and blocks catalysis (see the Introduction), it is surprising that the structural require-ments for complete suppression of PDE6 catalytic activity have never been determined. To better understand the structure/function relationship of P␥ to block catalysis of the PDE6 active sites, we synthesized a series of synthetic peptides differing in the length of the C-terminal amino acid sequence (Fig. 1B) and examined their ability to inhibit PDE6 catalytic activity.
Although the last few C-terminal residues of P␥ have been reported to physically interact with the catalytic domain to block catalysis (9,(13)(14)(15)(16)(17), the shortest peptide we tested (P␥81-87) showed no ability to inhibit the catalytic activity of PDE6 under our experimental conditions (IC 50 Ͼ 2 mM; Fig. 2). However, a synthetic peptide containing three additional amino acids to include the ␣3 helix (P␥78 -87) was sufficient to completely suppress the catalytic activity of PDE6 (IC 50 ϭ 21.4 Ϯ 1.2 M). This result demonstrates the central importance of the ␣3 helical region of P␥ to stabilize binding at least 100-fold so that this segment of the P␥ structure can bind to and inhibit catalysis at the active site of PDE6.
Compared with the dramatic stabilizing interactions conferred by the ␣3 helical region of P␥, we observed only modest additional stabilizing interactions when longer P␥ synthetic peptides were tested. As shown in Fig. 2, a 4-fold increase in inhibitory potency was observed for P␥74 -87 compared with P␥78 -87. Inclusion of part (P␥71-87) or all (P␥63-87) of the ␣2 helical region of P␥ had an even smaller effect on the IC 50 of inhibition of these P␥ peptides (Fig. 2). We conclude that the region between amino acids 63 and 77 of the P␥ sequence contain minor stabilizing interactions with the PDE6 catalytic domain, compared with the important stabilizing influence of the ␣3 helical region that comprises amino acids 78 -83 of P␥.
The C-terminal Region of P␥ Is a Classical Noncompetitive Inhibitor of cGMP Hydrolysis, but a Competitive Inhibitor When cAMP Is the Substrate-While investigating the doseresponse relationship of the C-terminal synthetic peptides of P␥ shown in Fig. 2, we were surprised to find that the IC 50 for inhibition was independent of the cGMP substrate concentration used for the activity measurements (data not shown). This is inconsistent with the commonly held model that the C terminus of P␥ is a competitive inhibitor of the active site of PDE6 (see the Introduction).
To examine this further, we assayed the inhibitory effects of our C-terminal synthetic peptides over a range of cGMP concentrations to determine the kinetic mechanism of inhibition. Surprisingly, a double-reciprocal plot of these data visually demonstrates that P␥63-87 was a simple noncompetitive inhibitor of catalysis (K I ϭ 1.8 Ϯ 0.7 M) when cGMP was the substrate (Fig. 3A). When cAMP was used as the substrate, the  (19); the functional interaction of these putative structural elements with PDE6 catalytic dimer is a focus of the experiments in this paper. B, some of the P␥ truncation mutants and synthetic peptides used in this study are depicted.  FEBRUARY 12, 2010 • VOLUME 285 • NUMBER 7 mechanism of inhibition by P␥63-87 changed to that of a simple competitive inhibitor ( Fig. 3B;

Mechanisms of P␥ Inhibition of PDE6
To evaluate whether the inhibition mechanism was dependent on the length of the P␥ synthetic peptide, we repeated this analysis with the shorter P␥78 -87 peptide. We observed the same double-reciprocal pattern indicative of noncompetitive inhibition when cGMP was the substrate (Fig. 3C, K I ϭ 28.1 Ϯ 3.2 M) and competitive inhibition when cAMP was the substrate ( Fig. 3D; K I ϭ 48.5 Ϯ 2.1 M). Although shortening the P␥ C-terminal peptide by 15 residues reduced the inhibitory potency ϳ10-fold in both cases, the mechanism of inhibition of the P␥ peptide (noncompetitive with cGMP, competitive with cAMP) was unchanged by reducing its size.
Because cGMP is a ligand for the noncatalytic cGMP-binding site within the regulatory GAF domains, but cAMP is not (44), we attempted to test the hypothesis that occupancy of the GAF domain by cGMP allosterically alters the catalytic domain and how it interacts with P␥. Unfortunately, we were unable to find experimental conditions that retained bound cGMP in the GAF domain-binding site in the absence of the GAF-interacting region of P␥ for the duration of time needed to evaluate the kinetic mechanism of inhibition by P␥.
Inhibition of Catalysis Can Occur by a Distinct Mechanism Induced by the First 76 Amino Acids of P␥-Although the above results demonstrate that the last 10 amino acids are sufficient to suppress 100% of the catalytic activity of PDE6, previous work suggested that other regions of P␥ could modulate catalytic activity (see the Introduction). Recently, we showed that P␥1-60 had no inhibitory activity by itself but was able to induce an allosteric change that was communicated between the regulatory GAF domains and the catalytic domains (39). To examine evidence for an alternative mechanism of P␥ inhibition, we used a set of longer truncation mutants (Fig. 1B) to identify structural elements that induce a reduction in catalytic activity of the PDE6 catalytic dimer.
Surprisingly, the truncation mutant P␥1-76 that lacks the last 11 amino acids (including the entire ␣3 helix) showed 80% inhibition of cGMP hydrolysis (Fig. 4A). The extent to which truncated P␥ was able to inhibit cGMP hydrolysis depended on the length of the protein. Whereas P␥1-60 had no detectible inhibitory activity (data not shown), P␥1-70 was capable of partially inhibiting cGMP hydrolysis (37%), and adding an additional three amino acids (P␥1-73) led to inhibition of 56% of the total catalytic activity (Fig. 4A). We conclude that P␥ truncation mutants containing the region between amino acids 61-76 ("inducer" region; see "Discussion") are able to inhibit cGMP hydrolysis by a different mechanism than that used by the last 10 C-terminal residues of P␥.
Because this result was entirely unexpected, we were concerned that contaminating fragments of P␥ (resulting from trypsin proteolysis that was used to prepare P␣␤; see "Experimental Procedures") might account for this behavior. To rule out this possibility, we added the C-terminal peptide (P␥63-87) in an amount stoichiometric with the P␣␤ concentration used in these experiments (10 pM). We observed no inhibition of catalysis by this amount of P␥63-87 in the absence (Fig. 2,  filled circles) or presence of the P␥ truncation mutants (data not shown).
The effectiveness of this so-called inducer region of P␥ (i.e. amino acids 61-76) to inhibit catalysis was found to depend on the substrate being tested. Fig. 4B shows that most P␥ truncation mutants were less effective in inhibiting cAMP hydrolysis than cGMP hydrolysis. Furthermore, the extent of maximal inhibition of cAMP hydrolysis was similar for the P␥ truncation mutants varying in length from P␥1-73 to P␥1-83 (Fig. 4C).
To test whether the greater effectiveness of P␥ truncation mutants to inhibit cGMP hydrolysis resulted from allosteric regulation of the catalytic domains by cGMP binding to the regulatory GAF domains, we prepared PDE6 catalytic dimers reconstituted with the P␥1-80 truncation mutant to which cGMP was either present or absent at the regulatory GAF domains (see "Experimental Procedures") and then measured the rate of cAMP hydrolysis. We found no difference in cAMP hydrolytic rates as a function of cGMP occupancy of the GAF domains (data not shown), indicating that the occupation of the GAF domain by cGMP was not responsible for this effect.
We also noted that the effectiveness of P␥ truncation mutants to inhibit catalysis was dependent on the cGMP substrate concentration. When several different P␥ truncation mutants were tested with millimolar levels of cGMP as substrate (Table 3 of Ref. 15; see also supplemental Table S1), lower maximal extents of inhibition were seen compared with the results in Fig. 4A using 1 M cGMP as substrate. Although the mechanism by which high cGMP concentrations reduces the maximum inhibitory effect of these P␥ truncation mutants is not understood, we speculate that high cGMP concentrations may induce a conformational change in the PDE6 catalytic dimer, similar to that recently reported for the related PDE5 enzyme (45), that alters the effectiveness of P␥ truncation mutants to inhibit catalysis.

A Short Segment within the C-terminal Region of P␥ Can Detect Conformational Changes Induced by the Binding of P␥ Truncation Mutants-
The previous sections defined two different regions of P␥ that are capable of inhibiting catalysis. The C-terminal 10 amino acids ("blocking" region) are, by themselves, sufficient to fully inhibit catalysis, whereas the interval between amino acids 61 and 76 (inducer region) can inhibit most of the PDE6 activity in the absence of the blocking region. To examine whether these two different inhibitory activities interact with each other in a cooperative manner, we tested the hypothesis that the N-terminal portion of P␥ would enhance the binding affinity of C-terminal peptides.
Having determined that the inhibitory activity of P␥63-87 was potentiated by the presence of P␥1-70, we next asked whether shorter C-terminal peptides of P␥ would also be sensitive to the presence of N-terminal truncation mutants. As seen in Fig. 5B, shortening the C-terminal fragment by eight amino acids (P␥71-87) abolished the ability of P␥1-70 or P␥1-83 to enhance the binding of the C-terminal peptide; in the latter case P␥1-83 actually reduced by ϳ4-fold the effectiveness of P␥71-87 to inhibit catalysis. The ability of P␥63-87, but not P␥71-87, to be stabilized by the P␥1-70 N-terminal fragment defines a region of 7-10 amino acid residues starting  Table S1. C, summary of the maximum inhibition of the truncation mutants when cGMP (F) or cAMP (E) was used as substrate. The x axis represents the position of the last amino acid at the C-terminal end of the P␥ mutants.
at position 63 that is capable of detecting the presence of the N-terminal fragment and/or a conformational change of the catalytic domain caused by this fragment ("sensor" region; see "Discussion").
The fact that the potentiation effect of the N-terminal truncation mutant was only observed when combined with a C-terminal peptide with which it shared 8 amino acids (residues 63-70) raised a concern about a potential artificial interaction between P␥1-70 and P␥63-87 being responsible for this effect. To directly address this, sedimentation equilibrium measurements were performed with the two individual P␥ proteins and with an equimolar mixture of them. Even at the highest concentration tested (50 M), no evidence for self-association or interaction of the two P␥ fragments could be detected (data not shown). Furthermore, we have conducted an additional experiment in which a 100-fold enhancement of the inhibitory potency of P␥67-87 could be observed with the N-terminal mutant P␥1-68 (data not shown); in this instance, the overlap was reduced to 2 amino acids. Together, these two observations greatly reduce the likelihood that the observed potentiation of the inhibitory effectiveness of the C-terminal region of P␥ by these N-terminal P␥ truncation mutants can be attributed to an artificial interaction of the overlapping sequence of the two P␥ constructs.
Transducin Activation of PDE6 Holoenzyme Requires Additional Sites of Interaction besides the C-terminal Blocking Region of P␥-It is known that transducin ␣-subunit has multiple sites of interaction with PDE6 holoenzyme, and the prevailing view is that T␣* binds to the C-terminal region of P␥ to displace these blocking residues and thereby relieve its inhibitory action. Having documented above two different mechanisms of regulation of PDE6 activity by P␥, we questioned whether T␣* interacts with one or both regions responsible for these different mechanisms of inhibition. To study this, we mixed P␣␤ with various fragments of P␥ and tested the ability of the reconstituted PDE6 to become activated upon exposure to saturating amounts of T␣*-GTP␥S. Because P␣␤ binds with high affinity to the N-terminal truncation mutants of P␥ that we tested for this experiment (Fig. 4 and supplemental Table S1), we were able to evaluate the ability of activated transducin to displace the P␥ mutants bound to the P␣␤ dimer and thereby relieve inhibition of catalysis. Mutants missing the last one or two C-terminal amino acids were also able to be activated and in fact showed a greater extent of activation than wild-type P␥ bound to P␣␤ (Fig. 6). In contrast, even high concentrations of T␣*-GTP␥S were incapable of activating P␣␤ that had been reconstituted with the C-terminal peptide P␥63-87 (Fig. 6). The inclusion of P␥1-70 (to enhance 100-fold the C-terminal peptide binding affinity; see Fig. 5A) with P␥63-87 was also ineffective in restoring the ability of T␣*-GTP␥S to activate the reconstituted PDE6 enzyme (Fig. 6). This result indicates that the multiple sites of interaction observed in the crystal structure complex of T␣*-GTP␥S with the C-terminal half of P␥ (20) were not sufficient to effectively displace the C-terminal blocking region from the PDE6 catalytic pocket under our experimental conditions.
When we tested the ability of T␣*-GTP␥S to relieve the partial inhibition imparted by P␥1-73 reconstituted with P␣␤, we observed that activated transducin could activate PDE6 to nearly the same extent as the P␣␤ catalytic dimer (Fig. 6). This indicates that in the absence of the C-terminal blocking region of P␥, T␣*-GTP␥S can interact with sufficient affinity to reverse the inhibition that is induced by the region of P␥ ranging from amino acids 61 to 76. It is reasonable to conclude from Fig. 6 that activated transducin can bind to full-length P␥ at multiple sites and that the N-terminal interacting sites confer binding stability that permits relief of inhibition by both reversing conformational changes induced by amino acids 61-76 as well as displacing the C-terminal blocking peptide from the catalytic pocket.

DISCUSSION
Three major conclusions emerge from this paper: 1) the minimum structural requirement for complete inhibition of catalysis is the last 10 amino acids (including the ␣3 helical segment) of the P␥ subunit; 2) a novel mechanism for inhibiting catalysis results from an apparent conformational change in the catalytic domain that can be induced by a region of P␥ consisting of amino acids 61-76 (including the ␣1 and ␣2 helical regions); and 3) activated transducin relieves both types of P␥ inhibition, but to displace the extreme C-terminal blocking residues the N-terminal domain of P␥ must be linked to the C-terminal region.
Two Distinct Inhibition Mechanisms Utilize Different Regions of the P␥ Molecule-Previous work established that the last several amino acids of the P␥ sequence are critical for inhibiting catalysis of PDE6 (9, 14 -17), and the preponderance of evidence has suggested a classical competitive inhibition mechanism between the P␥ C terminus and substrate (12,17,26). Our results demonstrate that in addition to the last several amino acids of P␥, an intact ␣3 helical region is also required, because P␥81-87 was completely ineffective in inhibiting catalysis, whereas P␥78 -87 could completely suppress catalytic activity (Fig. 2). Our functional definition of the blocking region of P␥ in this study (Fig. 7, region B) is consistent with structural studies revealing that P␥ residues 78 -87 assume a nearly identical conformation when bound to the ␣-subunit of transducin (20) or when complexed with a PDE5/6 chimeric catalytic domain (21).
However, close examination of the kinetic mechanism by which this blocking region of P␥ inhibits catalysis does not sup-port a simple direct competitive inhibition between P␥ and substrate except when cAMP is the substrate (Fig. 3). When cGMP is the substrate, a pattern consistent with noncompetitive inhibition is found for both the minimal blocking region (P␥78 -87) as well as for a larger C-terminal peptide containing additional sites of interaction with the catalytic domains. We suggest that cGMP binding either to the noncatalytic binding sites in the regulatory GAF domain (39) or to the catalytic domain induce a conformational change that alters the interaction of P␥ with the PDE6 catalytic domain. The idea that cGMP binding to the catalytic domains of PDE6 might be responsible for some of the novel observations in this study is consistent with a recent study of the closely related PDE5 enzyme in which large conformational changes are observed upon cGMP binding to the catalytic domain (45). Future experiments will be needed to discriminate the locus of this proposed effect of cGMP binding, as well as its physiological relevance for PDE6 regulation during phototransduction.
In addition to the ability of the blocking region of P␥ to inhibit PDE6 activity, we have defined a second, allosteric mechanism of inhibition and have localized the region of P␥ capable of inducing partial inhibition of catalysis to P␥ amino acids 61-76 (Fig. 7, segment I). The comparison between P␥1-60 and P␥1-70 (Fig. 4) clearly demonstrated that the additional 10 amino acids in the longer truncation mutant induced a partial inhibition of catalytic activity in the absence of the blocking C-terminal segment. This region encompasses the ␣1 helical region previously defined in structural studies (19,20). Furthermore, the ability of the region comprising amino acids 71-76 to further increase the inhibition of cGMP hydrolysis (Fig. 4C) suggests additional inhibitory interactions of the ␣2 helical region of P␥ with the PDE6 catalytic dimer.
Our identification of a novel allosteric inhibitory mechanism that occurs upon binding of the so-called inducer region of P␥ (Fig. 7, segment I) to the catalytic domain of PDE6 may be consistent with two recent structural determinations of related  Interacting sites of P␥ subunit with PDE6 catalytic subunits and transducin ␣-subunit. Above the schematic of the P␥ subunit domain organization, two interacting regions of transducin ␣-subunit with P␥ are depicted, one between the polycationic region of P␥ with the ␣3-␤5 region of T␣* (32) and a second between the C-terminal region of P␥ and the switch II/␣3 region of T␣* (20). The black bars represent the known interacting amino acid residues of P␥ with transducin ␣-subunit based on structural and crosslinking studies (14,20,33). Below the P␥ domain organization are shown the two major regions of interaction with the PDE6 catalytic dimer: the GAF-interacting region and the catalytic interacting region. Based on the work in this paper, we define the following functional attributes within the catalytic domain interacting region: the inducer region, amino acids 61-76 (I), serves to induce an inhibited state of the enzyme that likely involves conformational changes in the catalytic domain; the sensor region, amino acids 63-70 (S) and/or an intact ␣2 helical segment (amino acids 69 -73), is responsible for sensing conformational changes within the catalytic domain that result in increased affinity of the C-terminal peptide for the catalytic domain; and the blocking region (B), amino acids 78 -87, directly interacts with the catalytic domains of PDE6 to suppress cyclic nucleotide hydrolysis.
PDEs. Of immediate relevance is the structural determination of a PDE5/6 chimeric catalytic domain complexed with P␥70 -87 (21) that shows that P␥ residues 71 and 73 interacting with the PDE5/6 H-loop and its adjacent ␣12 helix within the catalytic domain. Intriguingly, this H-loop (found in all PDE family members whose crystal structures have been determined) is believed to be a major structural element responsible for allosteric regulation of the cGMP-stimulated PDE2 enzyme (46). Although the relevance of the PDE5/6 chimera and PDE2 structures to the structure and regulation of photoreceptor PDE6 remains to be determined, our biochemical study supports the idea that P␥ can interact with the H-loop to modulate catalytic activity distinct from the blocking action of the extreme C terminus of P␥.
Distinct Segments of the P␥ Sequence Induce and Detect Conformational Changes in the Catalytic Domain of PDE6-The 100-fold increase in binding affinity of P␥63-87 when P␥1-70 is bound to PDE6 catalytic subunits (Fig. 5A) reveals a much larger allosteric effect of P␥ on the conformation of the catalytic domain than the 2-fold allosteric effect previously reported (39) for the shorter P␥1-60 (Fig. 5A). This requirement for amino acids 61-70 suggests that an intact ␣1 helix on the N-terminal P␥ fragment is essential for causing this large enhancement of binding affinity of the C-terminal P␥ fragment.
The observation that the shorter C-terminal P␥ fragment, P␥71-87, fails to show this 100-fold enhancement of binding affinity (Fig. 5B) that occurs with the P␥63-87 peptide suggests that amino acids 63-70 contribute to sensing the conformational change in the catalytic domain that is induced by P␥1-70. Alternatively, the fact that P␥63-87 contains an intact ␣2 helical segment (amino acids 69 -73), whereas P␥71-87 does not, suggests the potential importance of the ␣2 helical structure in sensing the conformational changes induced by binding of the ␣1 helical region of P␥ to the catalytic domain of P␣␤. Although our current results do not pinpoint which amino acids within this region are responsible for this effect, the conformational changes inferred from these experiments reflect a previously unappreciated aspect of PDE6 regulation that warrants further study.

Transducin Interaction with the N-terminal Half of P␥ Is Required for Efficient Relief of Inhibition of the PDE6
Holoenzyme-The C-terminal region of P␥ has been documented to interact with transducin (8,9,20,32,33); however, our functional assay demonstrated that these interaction sites alone were not sufficient to relieve the inhibition of the reconstituted enzyme (Fig. 6). In addition, activated transducin failed to relieve the inhibition of P␥63-87 in the presence of the truncation mutant (P␥1-70), demonstrating the requirement of a physical linkage of the N-and C-terminal regions of P␥ for effective transducin activation of PDE6. Furthermore, the ability of transducin to reverse the inhibition of the P␥ truncation mutants lacking the C-terminal blocking region (Fig. 6) demonstrates an effective interaction of transducin with the N-terminal fragment of P␥ to displace P␥ from the catalytic subunit and allow catalysis to occur. This functional assay of P␥-transducin interaction agrees well with previous studies that localize an important site of interaction of transducin with the polycationic region of P␥ (see the Introduction).
Interestingly, the polycationic region (and the adjacent proline-rich region) of P␥ (Fig. 1) also interact with the GAF domains of PDE6 with high affinity and enhance the binding affinity of cGMP to the GAF domain (12). Our results suggest that the activation of PDE6 by transducin is a competition between transducin and PDE6 for binding of P␥ in two distinct regions, namely its C-terminal blocking region and the central region of P␥. Moreover, the fact that the binding affinity of the P␥ central region depends on cGMP occupancy of the PDE6 GAF domains opens the possibility that the effectiveness of transducin to compete with PDE6 to activate catalysis may be regulated by cGMP levels in the photoreceptor outer segment. However, it is important to recognize that our functional assays of transducin activation of PDE6 catalytic dimers associated with P␥ fragments are conducted in solution-not on the rod outer segment disk membrane-with purified, reconstituted components, and caution is required in relating these results to the situation in living photoreceptors.
Conclusion-This paper demonstrates the functional importance of the three ␣-helical domains in the C-terminal region of P␥ to bind to the catalytic domains of PDE6 to directly block catalysis at the active site and to induce and sense conformational changes in the catalytic domain that alter its interactions with P␥. We further show the functional relevance of these P␥ structural elements to the mechanism of transducin activation of PDE6.
Mutations in P␥ have been shown to lead to defects in phototransduction as well as retinal degeneration in animal models (reviewed in Refs. 7 and 47) but has not been seen in humans thus far. Importantly, some of these mutations are located in the C-terminal region (48 -51) that we have investigated in this study. Furthermore, a single amino acid substitution in the transducin ␣-subunit (i.e. G38D) disrupts the ability of transducin to bind P␥ (52,53) and is known to cause the Nougaret form of dominant stationary night blindness (54). The results in this study may provide a molecular basis for understanding how alterations in different regions of the P␥ molecule may adversely affect the transducin-mediated regulation of PDE6, as well as the ability of PDE6 to regulate cGMP homeostasis in the dark-adapted state and/or during visual excitation, recovery, and adaptation of photoreceptors.