11“” {itlglfttlkl'n’qllll‘tft , gcéiittvth ntlt‘fvlr} , Ilia it) u. I: Si ‘ ioé . . f§i£\tf§£l‘ is; ‘ ‘ [it'll s g 31.5» it; » 151111.111]. 0.! i ‘ $360.11 $731.") Illinlfiullhk.ttl« 1 .5559!» , ‘éné‘fa . ‘§£A¢A,o§‘ pt A 15“: \g“: ll‘ik.’ 31“}.[1 14 a.§~i¥‘t¢l§%§li . i I _ I53 . .. . \ll‘ fi§\}'1535§3\l §z$1§iso '1‘. .‘ I . .111:§u\.\w$ is [ww~5\5§\z§itll. 1395...... I 5 1r 1 1:135:15!!!» 3'41. 511111.521; iii»; .. 3" 11...: ...«\\1.......HH}2¢(4§;\1\i\$111}lrfilil1:cljflm)§:u§1 kc. ‘ v . . - a It 17 \u I I I «Nxhv‘igfiikvt 9:! , 1 x. 35515)):- (23 (1.12.115 r}\\1‘ v oil-irons reiinol ociive PDE /Co“"" come/TNGMP 7 NOT+ \ /' " \__X_>1 ‘ (1’ > T Not No+ V Figure I. 6 identified calmodulin, a calcium—mediated regulatory protein, and calmodulin binding activity in the ROS E 18 J although its function remains to be uncovered. In fibroblasts, calmodulin has been reported to alter Na+ flux E 19 J, an observation which merits pursuit in the case of the ROS. There are others, however, who have raised some doubts about the Ca2+ model. Szuts enployed 45Ca to study the effect of light on compartmentation and release of Ca2+ in excised retinas and disks E 20, 21 J. No dark uptake or light induced release was detected. Cavaggioni and Sorbi also could not detect a dark uptake of Ca2+ into the disks [ 22 3. Others have studied the rate and stoichiometry of light induced Ca2+ release, particularly from reconstituted rhodopsin/phospholipid vesicles E 23, 24 J. The observed rates of about I Ca2+ ion released per bleached rhodOpsin per sec were inadequate to meet the requirements of the model. Recently, several groups have argued that Ca2+ is sequestered by light sensitive binding sites in the extradiskal space, rather than inside disks E 22, 25 - 27 J. Experiments utilizing ionOphores to enhance Caz+ release, however, seem to refute that possibility E 23 J. cGMP Model. The second model of visual excitation arose from the observation of a drop in cyclic nucleotide levels in the ROS in response to light E 28 J. In this model, bleached rhodopsin interacts through some intermediates to lower cGMP levels in the ROS. The change in cGMP concentration then acts to affect membrane permeability to Na+ (Figure 1). As with the Ca?+ model, it must be demonstrated that light lowers cGMP levels, that cGMP levels alter Na+ flux, and that the rate of cGMP concentration changes is adequate to block Na+ on a millisecond time scale. The mechanisms of this model are better described than is the case for Ca2+, yet direct evidence for its function remains elusive. Cyclic nucleotides in other tissues are frequently involved in membrane permeability E 29 J. One study has described a depolarization of ROS membranes induced by cGMP injection into rods of excised retinas that occurs within msecs after injection E 30 J. Another group superfused retinas with exogenous cGMP and obtained the same results E 31 J. Light has been shown to alter cGMP levels through activation of ROS cGMP phosphodiesterase (PDE) E 32 ~ 35 J. cGMP dependent protein kinase activity has been reported in the ROS E 36 - 39 J although no function has yet been uncovered. Several researchers have argued that PDE activation is both rapid and large enough to lower cGMP levels sufficiently to be involved in visual excitation E 34, 40, 41 J. Others have argued that the measured decline in cGMP levels and increase in PDE activity are delayed and insufficient for a role in initiating a hyperpolarized signal E 42 - 44 J. Clearly, both models require further clarification. However, there are other vision related activities that occur in the ROS in which either or both Ca and cGMP could be involved. Rhodopsin regeneration E 45 J, adaptation of response over several orders of magnitude of illumination E 46 J, and restoration of metabolites to dark conditions after light stimulation are possible roles. In addition, there seems to be a close interaction between Ca and cGMP in various tissues E 47 J. Although the PDE has been shown to be ca2+ ' t' ' . v ' insensitive [ 18 J, CaZT has been reported to depress 8 ROS cGMP levels E 41, 48 J and conversely, cGMP alters Ca2+ levels E 20, 31 J. External Ca2+ concentration is reported to regulate GTPase activity E 49, 50 J in the ROS. cGMP dependent protein kinase activity in skinned cardiac fibers decreases Ca2+ sensitivity of contraction E 51 J. Fatt and Skulachev make similar proposals that cGMP plays a comparable role in the ROS, decreasing Ca2+ sensitivity of ROS hyperpolarization by changing Ca2+ binding affinities E 52, 53 J. Resolution of the roles in ROS of cGMP and Ca2+ and their possible interaction, awaits further research. PDE Characterization. We have studied the cGMP PDE in order to determine the mechanism of its activation in the ROS. It is activated by light and has a further requirement for GTP E 34, 42 J whilch results in an increase in activity on the order of loo-fold E 34, 54 J. Liebman and Pugh have reported that activation occurs within msec of rhodopsin bleaching E 40 J. The enzyme behaves as a peripheral membrane protein, renaining associated with the ROS under isotonic conditions, but is freely soluble hypotonically E 35 J. This solubilization does not appear to result from release of the enzyme from a membrane compartment by hypotonic rupture, but rather an ionic strength dependent membrane association/dissociation E 33 J. O'Brien has shown that PDE binds to phospholipid bilayers in isotonic buffer, but not under hypotonic conditions E 55 J, suggesting that membrane association of PDE involves principally phospholipids. PDE was first purified from frog ROS by Miki_§t_al. E 56 J and subsequently from bovine ROS by several laboratories E 57 - 59 J. Miki et al. reported that the purified enzyme was no longer 9 responsive to either light or GTP E 56 J. Baehr.et_al. E 57 J utilized the salt dependence of membrane binding to purify the PDE from bovine ROS, and, in the process, also purified an '80K‘ protein. We improved on Baehr's procedure of separating PDE and '80K‘, which we called G protein, and noted a special relationship between the two proteins E 59 J. In the absence of reducing agents, PDE and G proteins behaved as a single Species by native gel electrophoresis and gel filtration chromatography, with a concomitant increase and stabilization of PDE activity E 59 J. KUhn, noting that G underwent a light induced binding to the ROS E 60 J, utilized that characteristic to purify both PDE and G E 56 J. PDE consists of three non~identical subunits with molecular weights on $03 gels of 88,000 0x), 84,000 (6), and 13,000 (i) E 57 J. Using purified PDE, G, and ROS depleted of soluble proteins (DROS), we reported reconstitution of the GTP and light dependent activation of PDE E 54 J. Reconstitution of PDE activation was linearly dependent on G protein concentration and PDE, requiring GTP and DROS as well as light E 54, 61 J. The GTP concentration dependence and level of activation of the reconstituted system clearly resembled that of intact ROS. The reconstituted system provided a means to study the mechanisms of PDE activation and the roles of the various components involved. We specifically noted that trypsin treatment of DROS increased their ability to activate PDE. Sulfhydryl reduction with dithiothreitol (DTT), or modification with N-ethylmaleimide, on the other hand, reduced that ability E 54 J. Miki 33:3fl, observed that PCMB inhibited activation of PDE in intact frog ROS E 62 J. 10 G Protein Characterization. G protein behavior also varied depending on the presence of reducing agents during its isolation. Baehr et al purified '80K’ in the presence of DTT E 57 J. We found their procedure to be inconsistent in resolving PDE and G, and inclusion of EDTA increased the resolution of these proteins E 59 J. G protein isolated in this way was functional in reconstituting PDE activation E 54 J. In the absence of reducing agents and EDTA, a fraction of G protein did not resolve from PDE. As G is in excess relative to PDE in the ROS E 54, 63 J, there is a fraction of G which nevertheless is resolved from PDE under these conditions. However, this G protein is nonfunctional in reconstituting PDE activation E 54 J. The G complexed with the PDE apparently stabilized PDE activity, but was not adequate to reconstitute the activation system. Curiously, although the PDE:G complex behaved as a single species as described above, its apparent molecular weight in gel filtration and sedimentation velocity experiments was not different from that of PDE alone E 59 J. This suggests a possible subunit rearrangement. Alternatively, there could be multimers of G subunits that c0purify with PDE under these conditions. Baehr gt_al. reported the formation of multimers of Gd and G3: in isotonic salt E 63 J. G consists of 3 non-identical subunits with relative molecular weights from $03 gels of 40,000 (as), 35,000 (as), and 9,000 (er) I 59 J. The stoichiometry of the subunits, based on the molecular weight of the native protein is 1:1:1. Clearly, the interactions between PDE, G and the ROS membranes can be rather complex perhaps involving subunit rearrangements. This possibility will be discussed further in _ relation to the mechanism of PDE activation. G was first detected in ROS as a GTP binding protein with GTPase 11 activity E 64 J. This activity is dependent upon bleaching of ROS E 65 J. Godchaux and Zimmerman reported two distinct classes of GTP binding sites E 64 J. One was specific for accepting GTP, which could then be hydrolysed by a slow GTPase activity. The second could accept either GTP or GDP E 64 J. These two sites seemed to exist in a 1:1 ratio. Subsequent studies have failed to address the issue of two binding sites. Fung and Stryer noted a GTP—GOP exchange catalysed by G protein in the presence of bleached ROS E 66 J. KUhn has probed a light induced binding of G to the ROS E 67 J. This activity required a protease sensitive membrane component and was reversed by GTP. Fung_gt_al. have described the separation of G°(from Gex'by the use of the GTP analog, 5‘-guanylyl-B;i-imido-diphosphate (GMPPNP) E 68 J. Following chromatography, the GMPPNP was found bound to the G«_ subunit. This GM-GMPPNP complex, separated from 633, in the presence of unbleached ROS, could reconstitute PDE activation E 68 J. Uchida §t_al. reported similar findings in frog ROS by GMPPNP-bound G protein, except that they did not attempt to separate an active subunit E 69 J. Baehr_gt_al. have recently perfonned extensive characterization of the native purified G protein E 63 J. They reported that G binds GMPPNP in a 1:1 ratio with a K0 of less than 1 pM. They also reported that light induced binding to membranes saturates at a ratio of 1:4 (GzrhodOpsin), but that the light dependent GTPase activity of G did not saturate up to at least a 10:1 ratio of G protein to rhodopsin E 60 J. This suggests that the light signal can be passed on from one rhodOpsin to many G proteins. Fung_et_al. estimate that one rhodopsin can induce 500 G protein~GMPPNP binding events E 68 J. 12 phosphoryloied hv rhodopsin ‘ ADP \l/ rhodopsin ATP ociivuting rhodopsiniw— > opsin / , GTP P' G-GDP' 0- GTPG P04 \ active PDE inactive PDE Figure 2. 13 Thus, it can be seen that G is a complex, multifunctional, multisubunit protein. In order to clearly understand its role in PDE activation, the functions of G and the involvement of each subunit will need to be described from both a mechanistic and a structural perSpective. Model of PDE Activation. This model is a synthesis of data from many researchers. Figure 2 depicts the flow of events from light (hv) absorption to PDE activation. Following absorption of light, rhodopsin undergoes a series of conformational changes E 70 J. These changes expose, either on rhodopsin or some other membrane component, a site to which G binds Specifically E 71, 72 J. G then exchanges bound GDP for GTP and slowly hydrolyses that to GDP and inorganic ph05phate E 64 J. Alternatively, some G, upon binding GTP, is released from the membrane and may dissociate into G¢_and GBK'E 68 J. The G«;GTP complex then interacts with PDE to generate an activated form. Hurley and Stryer have pr0posed that the mode of interaction involves either removal of PDEK or a rearrangement in the internal subunit relationship of PDE E 73 J. They argue that PDEK interacts with PDE«@,in an inhibitory manner which GdeTP perturbs. This is based on reconstitution of trypsin-activated PDE with boiled and purified PDEK, which reverses that trypsin activation E 73 J. Several researchers have also addressed the issue of 'turn off' signals for PDE activation. Both Bitensky's group and Liebman and Pugh have pr0posed that GTP hydrolysis terminates PDE activation E 42, 74 J. Rhodopsin phosphorylation by ATP has also been pr0posed to prevent the formation of additional G:GTP complex E 40, 75 J. Finally, since Opsin has been shown to be ineffective in 14 reconstitution of PDE activation E 76, 77 J, the transience of the rhodopsin intermediates, Specifically metarhodOpsin II E 72, 78 J, may limit the extent and duration of increased PDE activity, and therefore the extent of the change in cGMP concentrations. This model is far from complete and the details are not universally accepted. There is no role for calcium in this model although several groups have reported an effect of calcium on cGMP and GTP levels in the ROS E 40, 45 J. The PDE catalytic site(s) has not been assigned to any particular subunit, although non—catalytic cGMP binding sites have been reported E 79 J. Reconstitution of the PDE activation system with GapGMPPNP required unbleached ROS E 68 J. Whether the ROS provided a convenient site for protein interaction, other G subunits (G and x), or some other role is unknown. The role of sulfhydryls and subunit rearrangements suggested by our work E 54 J, has not been further defined. Takemoto gt_al. E 80 J have reported that 8«azido-GTP, a photoaffinity analog of GTP E 81 J, binds to Ge when incubated with soluble native G protein. In addition to the binding site on Ga for GMPPNP, this may represent a second class of GTP binding sites. Alternatively, it may reflect a difference in confonnation between the soluble G protein, and membrane associated protein. It was not reported whether GB binding of 8-azido—GTP resulted in any functional changes in the protein. As binding of GTP to soluble G protein has not been previously reported, Takemoto's binding of 8-azido-GTP may not be directly relevant to physiological situations. The work to be reported here is designed to more clearly define the G protein subunits involved in GTP binding. We have utilized K-3ZP-8-azido-GTP to examine GTP binding, but unlike -. 15 Takemoto, who used soluble G protein E 80 J or Fung, who used G protein reconstituted with rhodopsin/phospholipid vesicles E 68 J, we have performed the experiments in intact ROS. While we do not expect to describe exact functional interactions in this way, we do expect to more precisely describe the phenomenon of GTP binding in the ROS. This will enable subsequent reconstitutional work aimed at those functional interactions. We report that G subunit labeling with 8—azido-GTP depends upon complex relationships between ROS membranes, POE, G protein, and the solvent conditions. We have addressed the physiological relevance of 8—azido—GTP binding both by using intact ROS and by examining the ability of 8-azido-GTP to function as a GTP analog in the ROS. MATERIALS AND METHODS ROS Isolation. ROS were isolated under dim red light on sucrose density gradients according to Kohnken et al. E 59 J within twelve hours of slaughter. Aliquots were stored at ~160° until use. G Protein Isolation. G protein was purified from ROS for labeling by X—32P~8~azido~GTP by two different procedures. The first employed low salt extraction of soluble proteins from unbleached ROS. These were resolved into pure G and POE proteins by DEAE-cellulose and gel filtration chromatography in the presence of 2-mercaptoethanol (BME) and EDTA as previously described E 59 J. The second procedure employed selective elution of G protein from bleached ROS by 80 PM GTP in 5 mM Tris—Cl pH 7.4, 1 mM dithio- threitol (DTT) according to KUhn E 58 J. The supernatant containing G protein was further purified on a 15 x 0.7 cm hexylagarose column according to Fung gt_al. E 68 J. The column was equilibrated in 10 mM 4~morpholino~pr0pane sulfonic acid (MOPS) pH 7.5, 2 mM MgC12, 0.1 mM phenylmethylsulfonylfluoride (PMSF), 1 mM DTT. After applying the supernatant to the column, GTP was eluted by 75 mM NaCl in the equilibration buffer. G protein was eluted with 300 mM NaCl in the same buffer. MnClZ was added to 4 mM and the protein was concentrated to 1 mg/ml by positive pressure ultrafiltration. This 16 17 procedure was also employed by Takemoto et al. E 80 J to study G protein labeling. Labeling reaction methods using purified G protein are described in Results. Sodium Dodecyl Sulfate (SOS) Polyacrylamide Gel ElectrOphoresis. Gels of 11% acrylamide were prepared according to Kohnken §t_al. E 59 J. Protein samples were applied and electrophoresis performed as described previously E 59 J. Gels were either rinsed with water and dried immediately or stained with Coomassie Brilliant Blue and destained briefly prior to drying. Dried gels were exposed to Kodak XAR-S film for 48 to 120 hrs at -200 with an intensifying screen. 8—Azido-GTP Photolysis. Photolysis of 20 pM 8-azido-GTP in 100 mM NaH2P04 pH 7.6, 5 mM MgCl2 was carried out on ice in 1.5 ml conical polypropylene centrifuge tubes. Ultraviolet (UV) light was directed on the sample at a distance of 3 cm from above using a Mineral light, model UVS-ll. Change in UV absorption spectrum induced by photolysis was measured on a Cary 14 recording spectrophotometer using an unphotolysed sample of 8-azido-GTP as a reference solution. DE 81 Chromatography. Samples containing 1 pl of 1 mM X-3ZP- 8—azido-GTP (0.2 pCi/nmol) were spotted on Whatman 0E 81 paper. Ascending chromatography was performed in 50 mM sodium acetate, 50 mM citric acid, pH 3.75 and completed in 1.5 hr at room temperature. Papers were air dried and either exposed to x-ray film or sliced into 5 mm strips. Radioactivity in the strips was detected by liquid scintillation counting. 18 Protein Labeling. In a typical assay, 2‘pl unbleached ROS (about 301pg protein) were incubated with ZO‘pM X—3ZP-8—azido-GTP (0.5 pCi/nmol) in 100 mM NaH2P04 pH 7.6, 5 mM MgCl2, ZOO‘pM ATP in a final volume of 50Iul. Reagents were mixed under dim red light on ice in 1.5 ml conical polypropylene centrifuge tubes. Room and UV lights were turned on for 7 min. The UV light source was positioned as described under '8-Azido-GTP Photolysis'. The lights were then turned off and the samples were transferred to Beckman Airfuge tubes. Supernatant proteins and membrane associated proteins were separated by centrifugation at 130,000 x g for 5 min. SDS and BME were added to both the supernatants and resuspended pellets to a final concentration of 1% each. Samples were kept at 40 for up to 3 hr before boiling and application to 11% acrylamide gels. Permutations of this procedure are described in Results. Miscellaneous Methods. Soluble proteins were labeled with X-32P-8-azido-GTP according to Takemoto et_§l, E 80 J except UV bleaching was allowed to proceed for 7 min rather than 1 min. Protein concentrations were determined according to Lowry et_al. E 85 J. Rhodopsin concentration was determined according to McConnell_et"al. E 86 J. ROS cGMP PDE activation by light and GTP were perfonmed according to Kohnken gt_al. E 54 J. 8-azido-GTP was substituted for GTP in the apprOpriate assays. Materials. 8-azido—GTP and {-32P-8-azido~GTP were purchased from Schwarz/Mann. All other reagents were obtained through Sigma. RESULTS Properties of 8-Azido-GTP. Geahlen and Haley E 81 J reported that 8—azido-CGMP exhibited an absorbance maximum at 279 nm which was diminished by UV bleaching. He found that 8-azido-GTP also had an absorbance maximum at 279 nm. UV bleaching of 20‘pM 8-azido—GTP in 100 mM NaH2P04 PH 7.6, SmM MgCl2 With a UVS-Il mineral light produced a decrease in absorbance at 279 nm and a corresponding increase at approximately 240 nm. The decrease in absorbance was half-maximal at 4 min and was essentially complete by 30 min (Figure 3). The maximum change in absorbance at 279 nm was 0.176. Inclusion of 0.15 mg/ml bovine serum albumin (BSA) in the buffer did not alter the time course nor the extent of the spectral change. 1.5 cm of plate glass and 0.5 cm 10% BHT in ethanol located between the light source and 8-azido-GTP prevented UV photolysis. This method, which prevents UV bleaching but allows visible light transmission, was employed in the remainder of the non—UV experiments in this report. Geahlen and Haley also reported that the purity of 8—azido-GTP could be assessed by chromatography on Whatman DE 81 paper E 81 J. We used DE 81 paper to chromatographically resolve photolytic products of X-32P-8—azido-GTP from unreacted material. In Figure 4a, llpl of 15 phllb32P—8-azido-GTP (0.2 pCi/nmol) was spotted on 19 20 Figure 3. UV Bleaching of 8-Azido-GTP. Aliquots of 20‘pM 8-azido- GTP in 100 mM NaH2P04 pH 7.6, 5mM MgClZ Were photolysed for the indicated times with a Mineral light UVS-Il as described in Methods. The UV spectrum of each aliquot was compared to that of an unphotolysed sample with the difference at 279 nm plotted. 21 1/ // .92.. Cu E T U law M :4 “0.15 r 6.. mu eemew dog “0.05 - q 0. Figure 22 Figure 4. DE 81 Chromatography of ‘6-333-8-Azido-olr. Samples containing 15pm {-32m8—azido-GTP (0.2 luCi/nmo ) in 100 mu NaH2P04 pH 7.6, 5 mM MgCl2 were bleached with UV light for 5 min. Aliquots of llpl were spotted on 0E 81 paper. Ascending chromatography and analysis were perfonned as described in Methods. Fraction 1 includes the origin while fraction 14 includes the solvent front. (a) no additions, with (O--O), and without (X--X) UV bleaching. (b) 0.15 mg/ml BSA included in the sample with (C)--()) and without (X--X) UV bleaching. (c) 0.75 mg/ml ROS protein was included in the sample (C)——())er#lROS and UV bleaching, (X-u-X) without ROS but with UV bleaching. I l T O 0"“0 +uv me -uv Figure 4a. Fraction 24 b l,OOO- - O——O +uw x " . X—-->< -uv s o 500)— .1 ”\X . \._ 0k . . X I I I l I L 4 6 8 IO 12 14 Frociion Figure 4b. [\3 (J1 T l T l l f l x C x/ LOOOP- o O—---O +ROS Q ~ xwx XWX ~ROS é . 0 x o \ 500l- \ . O .\ O . ' U '0 o o Xbez {) x\>< I l I L I I I 2 4 6 8 IO 12 14 Hudson Figure 4c. 26 DE 81 paper and develOped as described in Methods. After chromatography, papers were cut into 5 mm strips from which radioactivity was determined. UV bleaching for 5 min resulted in a shift in 32p migration from fraction 3 to fraction 5. Again, 0.15 mg/ml BSA did not alter this characteristic (Figure 4b). Inclusion of 0.75 mg/ml ROS however, apparently diminished 32R in fraction 5, perhaps by scattering and absorption of UV light thereby reducing photolysis. It also consistently resulted in a new peak of radioactivity in fractions 10 and 11 (Figure 4c). Migration on 0E 81 paper was also enployed as an indication of XrBZP~8-azido-GTP stability to various solvent treatments. Unphotolysed $432P-8-azido-GTP migration was essentially unaffected by incubation in diethyl ether or acetone for 1 hr at 4°; 0.1% SDS, 0.1% BME for 10 min at 90°; or 5 or 100 mM NaH2P04 pH 7.6, SmM MgClz for 2 hr at 4°. A portion of the label in samples treated with 6 M urea at 4° for 24 hr; water at 4° for 36 hr; 30% ethanol, 7.5% acetic acid for 1.5 hr at 25°; or 10% tri- chloroacetic acid (TCA), 10 mM Na2P207 for 1.5 hr at 4° exhibited a greater migration on 0E 81 paper than intact 8-azido-GTP. The renainder of the label migrated as intact 8—azido—GTP. Incubation in 10% TCA, 10 mM Na2P207 at 90° for 10 min resulted in nearly complete loss of label from a location on DE 81 corresponding to intact 8—azido-GTP to a location correSponding most closely to free inorganic phosphate. This instability of 8-azido-GTP in hot TCA proved useful in corroborating photoaffinity labeling as opposed to TCA-stable phosphorylation. These results also indicated the necessity of rapid and mild treatment of the label and labeled products. r 27 ROS Protein Labeling. ROS were incubated with'E-32P-8-azido- GTP as described in Methods, and the proteins were separated into supernatant and particulate fractions by centrifugation as described in Methods. These fractions were analysed by autoradiography of SOS gels for 32P incorporation. The Coomassie blue staining pattern and associated autoradiograms of pellet and supernatant fractions are shown in Figure 5. Pellet Protein Labeling. Rhodopsin monomer and multimers were the principal protein bands in the pellet, and also accounted for the majority of the label incorporation (Figure 5, lanes 1 and 2). Rhod0psin has been shown to be a substrate for phosphorylation by either ATP E 87 J or GTP E 88 J. Phosphorylation by rhodopsin kinase is activated through conversion of rhodopsin by light to an "active” substrate E 88 - 90 J. Reaction of rhodopsin with Zl3zP—8-azido- GTP did not require UV bleaching and the product survived treatment with hot TCA as well as dialysis against both 6 M urea or water at 40. o(-3ZP-GTP did not label rhodopsin, but X—32P-GTP did. Preboiling of the ROS or inclusion of an excess of EDTA prevented labeling. These results indicate that rhod0psin labeling arose from enzymatic phOSphorylation as opposed to a photoaffinity reaction. The extent of rhod0psin phosphorylation as a function of duration of bleaching by visible light, or by visible plus UV light is plotted in Figure 6. Reaction conditions are described in the Figure legend. 32P incorporation into rhodopsin required visible light. A 5 min incubation in the dark resulted in minimal label incorporation. Labeling was not enhanced by the addition of UV bleaching and at longer times may have been impaired by UV light. 28 fjgg:§_§. ROS Protein Labeling with X¥32P-8-Azido-GTP. ROS were — incubated with x-32p-8-azido-GTP (0.5 )JCl/nmol) and illuminated as described in Methods. The samples were separated by centrifugation into supernatant and pellet fractions. These were applied to SDS gels and electrOphoresis was performed as previously described [ 59 J. Lane 1: Coomassie blue stained ROS pellet proteins. Lane 2: Autoradiogram of lane 1. Lane 3: Coomassie blue stained ROS supernatant proteins. Lane 4: Autoradiogram of lane 3. Rhodopsin monomer (rholl, dimer (rhoz), PDE (aiand (3 subunits [ 57 J), GdF<4wm Figure 8. F gure 9. RhodOpsin Labeling as a Function of Nucleoside Triphosphate Concentration. ROS were incubated with K-BZD-S— azido—GTP (0.5 pCi/nmol) in the presence of varying concentrations of either ATP or GTP. Otherwise samples were treated and illuminated as described in Methods. ROS pellets w re applied to SOS gels, and subjected to electrOphoresis. Autoradiograms of the SOS gels were analysed as described in the legend to Figure 6. ()~-(): ATP concentration was varied. )(--)(: GTP concentration was varied. I! l/ \ M I! O.\w\.0 their O O 2 l ozjwm/j Hzmomma m>_._.<4m_m ” 400 200 I00 WES]. PM Figure 9. (A) ‘0 Table 1. Effect of Prebleaching on RhodOpsin Phosphorylation and Gm Labeling. Treatment rhodopsin Get 1. Prebleached 13.2 12.0 2. Prebleached + GTP 17.6 21.6 3. Prebleached, then GTP 20.9 18.4 4. Control + GTP 18.2 24.0 5. Control 27.1 22.5 ROS were incubated with K-32P~8-azido—GTP (0.5 pCi/nmol) in the absence of ATP. Samples were treated and illuminated as described in Methods with the following modifications. Line 1: ROS were prebleached with visible light in 100 mM NaH2P04 pH 7.6, 5 mM MgClZ for 5 min on ice prior to addition of 8—azido-GTP. Line 2: ROS were prebleached as in line 1 except that SOO’uM GTP was also present. Line 3: ROS were prebleached as in line 1, but SOOIuM GTP was added to the labeling reaction along with the 8-azido-GTP. Line 4: ROS were not prebleached, but SOOJuM GTP was included in the labeling reaction. Line 5: ROS were not prebleached nor were any nucleotides added other than 8—azido-GTP. Autoradiograms of SDS gels were analysed as described in the legend to Figure 6. The scanning densities of rhodopsin and G¢(labeling were analysed and automatically normalized separately. Values are expressed in relative percent labeling. 40 prebleached samples from one—half of control to the same level as unbleached ROS in the presence of GTP (Table 1, lines 2 and 3 vs 4). The effect was most pronounced if GTP was added to the assay after prebleaching. The effects of prebleaching on Ge(labeling will be discussed below. Supernatant Protein Labeling. Incubation of ROS with X432P- 8~azido~GTP under simultaneous UV and visible light as described in Methods, followed by centrifugation, yielded predominantly one labeled protein in the soluble fraction (Figure 5, lane 4). This protein corresponded in migration to the oisubunit of G protein (Figure 6, lane 4) [ S8, 59 J, with an estimated molecular weight from SOS gel migration of 40,000. Labeling of Gogwas stable to dialysis against 6 M urea and partially so when dialysed against water at 40. Neither «r32P-GTP nor YL3ZP-GTP introduced label into Gotunder identical conditions. The label appeared unstable to hot TCA treatment, although it was difficult to resuspend and solubilize the TCA pellets for $03 gel analysis. Consequently, not all the protein would enter the gel. Quantitation therefore depended upon radioactivity in a TCA pellet relative to radioactivity in a control pellet where 6132P—8razido—GTP was added with the TCA. By this comparison there was about 15,000 cpm of TCA stable radioactivity in the ROS pellet, but less than 100 cpm in the ROS supernatant. Finally, Gu_labeling required UV light (Figure 10). These results indicate that a specific photoaffinity reaction between X-32P-8~azido-GTP and Ba occurred. Another protein in the supernatant besides Gx_also exhibited a UV dependence of labeling (Figure 10, upper band). It can best be Figure 10. Visible and UV Light Dependence of GotLabeling. ROS were incubated for varying times under visible or visible plus UV light with X432P-8-azidoeGTP (0.5 pCanmol). Otherwise samples were treated as described in Methods. The supernatants were collected by centrifugation and subjected to SOS gel electrophoresis as described in Methods. An autoradiogram of the SOS gel is shown. Lanes 1 - 6: Visible plus UV light bleaching for O, 1, 3, 5, 7, and 10 min respectively. Lanes 7 - 12: Visible light bleaching only for O, 1, 3, 5, 7, and 10 min reSpectively. Lane 13: Incubation in darkness for 5 min. The locations of G¢(and Ge,are indicated. Figure 10. 42 23456789|OH|2|3 43 seen as a band with an approximate molecular weight 0. 65,000 in Figure 12, lane 9. he have not characterized its labeling, nor have we identified it. The possible indentity of this protein includes rhodOpsin kinase [ 90 J or the @-subunit of tubulin [ 81 J, both of which have reported molecular weights of about 55,000. As shown in Figure 11, photolabeling of GoLWBS linearly dependent on duration of UV + visible bleaching. Bleaching with visible light alone resulted in a low, time independent background labeling which may be due to a contamination of the supernatant by rhodopsin. ROS incubated in the dark for 5 min with 5-32P-a-azido~oip did not exhibit any labeling. As in the case of rhodopsin, G«;phctolabeling exhibited a linear dependence on‘5-32P-8—azidonfilP concentration up to at least 200 pM (Figure 7 and Figure 12, lanes 6 - 9). However, Ger photolabeling was almost inversely proportional to ROS concentration (Figure 8), so that at 92 pM rhodOpsin in the assay, virtually no Ga~ labeling was detected (see also Figure 12, lanes 1 ~ 5). This was probably due to scattering and absorbance of UV light by ROS proteins thereby decreasing the photolytic activation of X-32P-8-azido—GTP. When GTP concentration was varied in a competition experiment with 3—32P—8-azido-GTP, as shown in Figure 13, there was an increase in Go_._. :7" ‘ "twine Figure 16. 58 band (indicated as ‘x' in the Figure). Dilution in our 'S-ZOO' buffer (10 mM Tris—Cl pH 7.8, 10 mM BME, 1 mM EOTA, 1 mM Nah3, 0.1 mM PhSF, 501 mM NaCl) also inhibited G protein labeling (lanes 13 - 15). This probably reflects a dependence of Go x j 47: l i L ~00 -7 -e -5 -4 -3 I": a C rmtne lo. M, [GTP or Nag-GT8 DISCUSSION . . . L *9 . - we have demonstrated that X—JLP-8-a21do-GTP labels both rhodopsin, via phosphorylation, and G protein, via a photoaffinity reaction. Rhodopsin labeling was identified as phosphorylation by the following criteria. The reaction required visible light, but was independent of UV bleaching (Figure 6). The label was stable to hot TCA treatment and dialysis against both urea and water. GTP has been reported to be a good substrate for rhod0psin kinase with a Km of 200 pM [ 88 j. G protein labeling in the presence of the ROS was found to vary with incubation conditions. Goglabeling required UV bleaching (Figure 10) and appeared to be unstable to hot TCA treatment. Ga1was not labeled by eithercx-32P-GTP or X—3ZP-GTP. These results indicate a photoaffinity reaction mechanisn for Gaglabeling. 63 labeling in the presence of the ROS also required UV bleaching (Figure 14). A small amount of label seen in the absence of UV light may be due to rhodopsin, which has an SDS gel mobility similar to that of GE,(Figure 5). Rhodoosin in ROS membranes does not sediment as well under the hypotonic conditions employed for 63 labeling as it does under isotonic conditions [ 57, 58 ]. Stability of Ge labeling was not examined, nevertheless, the dependence of the reaction on UV light argues strongly for a photoaffinity mechanism. The conditions 64 65 which favor G“ or 65 labeling will be discussed below. Soluble purified G prOtein was also labeled by K—32P~8-azido- GTP. Both Gx_and 6% were labeled without UV treatment (Figure 16. lanes 4 and 6), although UV bleaching did enhance labeling in some cases. Unlabeled GTP inhibited the reaction, as did the inclusion of EDTA. The mechanism of this reaction is unclear. We are unable to eXplain the difference between our results and those of Takemoto at 31. E 80 3, who reported a UV dependent labeling of Ge under nearly identical conditions. It is possible that protein concentration, the only non-identical condition, may play a role. Takemoto_et_al. did not indicate the protein concentration in their reaction [ 80 ]. The UV independent labeling shown in Figure 16, lanes 4 and 6, would not have been detected in our ROS labeling experiments (Figure 12, lanes 1 - 5 for example). 100 pg G protein was used in the soluble labeling reactions. The extent of 32p incorporation for that amount of protein was comparable to labeling observed in the ROS (Figure 12, lanes 1 - 3). Yet there was only about l‘pg G protein contained in the ROS in those reactions [ 54, 88 ]. Thus UV independent labeling in the ROS, if it occurred to the same extent relative to soluble protein, would have been only 1% of the total label incorporated. Several variables, some in unexpected ways, were found to affect ’3 . . the amount of “2P incorporated into Gag —Azido-GTP Concentration. Even though it did not maximize 32P incorporation (Figure 7), ZO‘pM 8-azido-GTP was routinely used because G protein was completely eluted from bleached ROS manbranes (Figure 17) and PDE was near fully activated (Figure 18) at 66 that concentration. In addition, a low concentration should have minimized 32R incorporation into SOS gel bands from non-photoaffinity, non-GTP specific reactions. At 200 pN 8-azido-GTP, at least 8 SDS gel bands in the ROS supernatant were labeled, although UV dependence, and the identity of these bands, were not explored (Figure 12, lane 9). Limits to Labeling. In terms of label incorporation, the most productive condition occurs when UV light photoactivates 8-azido—GTP while it and the G protein are in a complex on the ROS membranes. Several factors then, can decrease labeling. One arises from any G protein/8-azido—GTP interactions that lack the proper orientation to react covalently upon UV photolysis of the azido group. UV bleaching of 8-azido-GTP away from the binding site effectively reduces labeling because the photolysed label still has a competitive affinity for the binding site, but cannot covalently attach. Access- ibility of the photolabel to UV light can also be limiting. This was illustrated in Figure 8, where increasing ROS concentration decreased the amount of Gaglabeled. This was probably due to UV light absorption and scattering by ROS, limiting UV availability to photoactivating reactions. Finally, instability of the label between reaction and analysis probably reduced the extent of labeling. We attempted to minimize this by use of a short centrifugation time and low temperatures, and by staining and destaining as briefly as possible. Stimulation of Labeling bv Nucleotides. An unexpected effect on Gorand rhodopsin labeling was observed with both ATP and GTP. Neither significantly inhibited labeling by 8-azido-GTP. On the contrary, 67 ATP activated rhodopsin phosphorylation by 8-azido-GTP 5-fold at 4 pM ATP (Figure 9). GTP at 80 pM activated G°(photolabeling by 2-fold (Figure 13). Recently, GTP was reported to enhance rhodopsin phosphorylation by ATP E 92 J. Analogous effects of ATP and GTP on 8-azido—GTP phosphorylation were apparently observed here. Effect of Prebleaching. Another set of observations on 8-azido- GTP labeling of ROS proteins may be explained by the same mechanism. A decrease in both rhodOpsin phosphorylation and Ga:photolabeling was obse ved by prebleaching. As prebleaching should have increased the number of rhodopsin phosphorylation sites E 89 J and perhaps membrane bound G protein E 60 ], we might have expected an increase in labeling. As shown in Table 1, just the opposite occurred. Furthermore, addition of 500 pM GTP, during or after prebleaching, largely reversed the effect of prebleaching for both rhodopsin phosphorylation and Gozphotolabeling. This suggests that prebleaching may deplete an endogenous pool of nucleoside triphosphates. That pool may enhance the reaction of 8-azido-GTP with both rhodopsin and G protein. Rhodopsin can be phosphorylated at up to 7 sites E 93 ]. While no direct evidence of c00perativity has been reported, a population of extensively phOSphorylated rhodopsin has been separated from largely unphosphorylated rhodOpsin E 90 J, suggestive of cooperativity. It is possible that 8~azido—GTP is a poor substrate for the initial phOSphorylation event, but adequate for subsequent phosphorylations. Prebleaching may then act through depletion of an endogenous pool of ATP, perhaps through Mg-ATPase E 94, 95 ]. SOOJuM GTP replenishes that pool, perhaps generating ATP through a 68 nucleoside diphosphate kinase reported to be in ROS E 96 3, thus allowing an initial phosphorylation, which can then lead to subsequent phosphorylations by 8-azido-GTP. in support of this idea, 4 pM ATP, which maximally stimulated phosphorylation in non-prebleached experiments (Figure 9), is close to the Km of ATP reported for rhodopsin kinase E 90 ]. Alternatively, there could be a regulatory site for ATP which stimulates phosphorylation by GTP (or 8-azido—GTP), analogous to GTP activation of phoSphorylation by ATP reported by Swarup and Barbers E 92 3 through a regulatory nucleotide binding site E 38 3. GTP Allosteric Site. Several anomalous results have been observed relevant to ODEphotolabeling by 8-azido-GTP. One is that binding did not saturate at higher 8-azido-GTP concentrations (Figure 7). Assuming that Ga_labeling is relevant to PDE activation, then the KO for 8-azido-GTP on Go(should be about lOluM (from Figure 18). Consequently binding should have saturated by IOO‘pM. A second anomaly appeared in the inability of GTP to inhibit labeling. In- stead, GTP increased Gcelabeling, reaching a peak at 80‘pM (Figure 13). Such effects might be explained by a non-GTP-specific interaction of 8-azido-GTP with G“. This explanation would require reconciliation with the data of Fung §£_§l- indicating G«;binding by GMPPNP E 68 J and the higher level of UV-dependent Go(labeling with 8-azido»GTP than labeling of other proteins in the ROS. Alternatively, there could be an allosteric binding site for GTP which increases Ga/GTP interactions. It is interesting to note that while we observed a half~maximal activation of POE at about 1 pM GTP (Figure 18), we previously found that maximal activation was achieved at 100 pH in a 69 broad curve of PDE activation by GTP E 54 J. This is consistent with GTP activation of Go(labeling by 8-azido—GTP and non-saturation behavior. Several other researchers have reported a high Km GTPase E 41, 48, 95, 97 3. It is possible that GTP at higher concentrations, on this high Km GTPase or elsewhere, may allosterically stimulate Gu/8-azido-GTP interaction. Inhibition by prebleaching would be consistent with this idea (Table 1), acting through depletion of the allosteric site. Added GTP might replenish that site and reverse the effect of prebleaching on Gozlabeling. This possible allosteric site is discussed further in Experimental Recommendations. Differential GTP Binding. Godchaux and Zimmerman described two distinct binding sites for GTP on G protein, one of which could also accept GDP E 64 J. Baehr_et_al., on the other hand, observed only 1 binding site for GMPPNP per G protein although they would not have detected a binding site with a KA greater than 104 M‘1 E 63 ]. The Km for binding of GMPPNP was reported to be less than 1 pM E 63 ]. The data presented here are consistent with either two distinct binding sites or a G protein/GTP complex at one site that can exist in more than one conformation. Under hypotonic conditions, 20 pM 8-azido—GTP eluted the bulk of G protein from bleached ROS (Figure 17), indicating that a substan- tial portion of the G protein was interacting with 8-azido—GTP. That interaction resulted in labeling of the GB subunit (Figure 15). Under isotonic conditions, nearly undetectable amounts of G protein were eluted by 8—azido—GTP (data not shown). when 6-32P-8-azido— GTP was used with unwashed ROS under isotonic conditions, what little detectable G protein was found in the supernatant was labeled on Go( 70 (Figure 5). This suggests that under isotonic conditions only a small portion of the G protein interacted with 8-azido-GTP and was consequently eluted. The actual process of photolabeling, as Opposed to binding, was apparently required to elute G protein under these conditions. Alternatively, G protein may have interacted in more than one way with 8-azido-GTP and the conformation which favored photolabeling also favored elution from the membranes. Unwashed ROS incubated under isotonic incubation conditions exhibited preferential photolabeling of Go(with 8-azido-GTP (Figure 16, lane 1). Labeling was virtually eliminated if the reaction was carried out under hypotonic conditions (Figure 15, lane 2). Prebleached and prewashed ROS, on the other hand, exhibited preferential labeling of Ge,under hypotonic conditions (Figure 15, lane 4), and if incubated with 8-azido—GTP under isotonic conditions, exhibited virtually no labeling (Figure 15, lane 3). PDE was absent from the prebleached, prewashed ROS (see Figure 17, lanes 1 and 2, showing POE appearance in the supernatant), due to hypotonic elution E 59 ]. POE may also have been released from the ROS, though still present in the medium, during hypotonic labeling of unwashed ROS. Thus, binding of POE to the membrane and/or G protein may have directly influenced whether Gx_became labeled. Even when both POE and G proteins were still on the membrane, the orientation of PDE may have influenced accessibility of G to GTP or its analogs. Either he absence of POE, or hypotonic conditions or both, permitted preferential 8—azido-GTP access to Gg. Fung 33121; found GHPPHP labeling of G“ using purified G protein reconstituted with bleached rhodopsin vesicles under hypertonic conditions E 68 J. PDE was not present. Takemoto et al. reported G5 71 photolabeling using purified, soluble G protein under the same solvent conditions E 80 3. While we could not reproduce Takemoto's results, we did observe Ge,photolabeling under a proscribed set of conditions. We have not yet pursued characterization of that labeling because it represented conditions which were presumably less physiological than those in which Gagwas photolabeled. It does represent, however, the conditions under which many researchers purify G protein E 58, 63, 68, 80 J, and is therefore relevant to much current research. Models of G Protein/GTP Interaction. At least two models of G protein/G—azido—GTP interaction are consistent with our results. The first entails two distinct binding sites, as suggested by Godchaux and Zimmerman E 64 3. One binding site is apparently on Gogand the other on Gg. Ga interaction with 8-azido-GTP may be favored by isotonic conditions with PDE present. The 66 binding site may interact favorably with 8—azido-GTP under conditions mimicked by hypotonic incubation in the absence of PDE. A second model involves only one binding site located near the interface between Ga and Gg. akemoto.gt~al. E 80 3 suggested this possibility to explain the difference between their results and those of Fung et_gl. E 68 ]. Salt concentration, presence of PDE, and membrane association of either POE or G would all affect the conformation around the binding site. This would then be reflected by differential subunit binding of GTP and its analogs. Our results do not resolve the identity or location of the GTP binding site(s) on G protein. Rather, our results indicate that the interaction of G protein with GTP might occur by more than one 72 mechanism leading to different subunit binding. How these mechanisms relate to the functions of G protein in the ROS, particularly PDE activation, awaits further characterization. Experimental Recommendations. Earlier in the discussion I described some anomalous behavior of G“ photolabeling in the ROS. The labeling reaction appeared to be non-saturating with respect to 8-azido-GTP concentration (Figure 7), and was not competitive with GTP (Figure 13‘. Rather, GTP increased labeling. Based on reports of a high Km GTPase from several laboratories E 41, 48, 95, 97 J, and the GTP concentration dependence of PDE activation E 54 ], I proposed that an allosteric GTP site may exist which could account for the anomalous behavior of Go(photolabeling. In order to test this hypothesis, the relationship between Gcc photolabeling and elution of photolabeled G«_fron the ROS must be further clarified. That is, both 8-azido—GTP and GTP enhanced the appearance of photolabeled Gacin the supernatant obtained by washing ROS. Does this enhancement reflect increased labeling of Ga;or increased elution of labeled product? This can be tested simply by addition of high concentrations of 8—azido-GTP or GTP after UV irradiation but before centrifugation. Alternatively, Go depleted ROS hypoton'c W 4' 11qu lCEDA82CEHUE58 l O-ObM NoCl (zillion 1 90,000): 9 60m CEAE pellet pooled DEAL-Z Dealt imchms 1 3200 do filtration 1 l l P W G profeln frocims FIG. 1. Flow chart of washing procedures for removal of soluble proteins from ROS membranes and of operations sub- sequently performed using 4 parallel solutions indicated under "ROS Washing" under "Materials and Methods." 7. -—.~ . "—1 m-.__h._ «-.....A—fi- v” _.~_ ._... 12503 persulfate and 0.375 pl of N,.\'..\".;\”-temrnethylethylenediamine per ml of gel. The negative electrode bufier contained 8 g of Trlzma base. 14.4 g of glycine. 1 g of SDS. and 1 ml of 2-mercaptoethsnol per liter at pH 8.7. The positive electrode buffer was a 1:10 dilution of the same solution. Samples contaiim'lg 5—30 .ug of protein were boiled for '2 min in 10% glycerol. 1% SDS. 1": 2-rnercsptoethanol. mixed with bromphenol blue dye. and subjected to electrophoresis for 3-5 h at a constant 130 V (about 25 mA) per 24-sample slab. Bands were visualized by staining with Coomamie brilliant blue R. .Vau't'e PAGE—Electrophoresis was performed in 8% acrylamide. 0.213% Na ‘~methylenebisacrylarnide gels in 0.37 M Ml. pH 8.7. in glass babes 1.0.5 x 12 cm). Negative electrode buffer contained 3 g of Trims base and 14.4 g of glycine per liter at pH 8.7. with the positive elecrrode buffer a 1:10 dilution of the same solution. Gels were subjected to pro-electrophoresis for 3 h at 2 mA/gel with 0.37 at Tris-Cl. pH 8.7. in the negative electrode chamber to remove an. reacted ammonium persult'ate and NJJ..V',.\"'-tetrametbylethylene~ dimme. Samples containing 10—50 ug of protein were subjected to elecuophoresls for 3—5 h at .3 mA/gel and stained with Coomassie blue. Sucrose Gradient Sedbnentotton—Following DEAR chromatog- raphy. protein mmples (100-150 ,ngl were diluted to 0.15 ml in either 15 or 250 mM NsCl. 10 mar Tris-Cl. pH 8.0. at 4 °C and were layered onto gradients which were 4.9 ml. 57-20% sucrose (w/v) in the same solution as the applied protein. Catalase (232.0%). sldolase (158.000). and bovine serum albumin (643.0(1)). all from Pharmscia. were applied to a separate gradient as standards. The grad-.ents were centrifuged in a Beckmsn SW 50.1 rotor at 40.000 rpm for 14 h at t ’C. Fractions of 0.25 ml were collected from the bottom of the gradient. analyzed by SDS PAGE. and assayed for phosphodiesterase activity. In other experiments. ‘2-mercaptoethanol or EDTA was added to the protein and gradient solutions to maintain the concentration of these reagents to which proteins had been exposed during DEAE chromatography. Analytical Ultrocenrrifugorz'on-Protein samples were dialyzed against 10 man TrioCl. pH 8.0. 250 mM NaCl, 1 ms! MgCh. 1 ms: NsNz. 20% glycerol. and centrifuged in the same bufier at 20 ’C in a Beckman model B ultracentnfuge at 36.000 rpm. Boundary velocities were measured by monitoring Am. Sedimentation coelficients were calculated according to Cher-Jeni“ (21). Catalan» (320(1)? -- 113) and aldolue (saw - 735) (2'2) were run as standards in the same buffer to venfy the Viscosity correction. Sedimentation coefficients and M. were estimated by comparison to those of the ‘2 standards. Amino Acid Analyst's—This was graciously performed by Alan J. Smith of the Biochemistry and Biophysics Department at the Uni< versity of California at Davis on a Dumm model D-SIJO instrument. Materials—Acrylamide. .V,.V'-rneth_\lenebisacn'lamide. and SDS were untamed from Bio-Rad. All other chemicals. except where indicated. were obtained from Sigma and were reagent grade. RESULTS DEAE-cet’lulose Purification of Soluble ROS Proteins Purified unbleached ROS were washed in solutions contain- ing either 2-mercaptoethanol or EDTA. both. or neither of these additions. in the concentrations given in the legend to Fig. 2. The figure presents the elutlon profiles for each of the 4 soluble protein supernatants obtained by washing. Fractions indicated by shaded areas contain ed almost exclusively P and G proteins. as revealed by SDS PA gels. Other proteins were found predominantly in earlier fractions. in Fig. 2a. the large absorbance observed in the first 14 fractions was for the mos: part not amociated with protein. as determined by the method of Lowry et at. (19). In each shaded peak. P and G were found in the same proportion (about 1:2. gtg) in all fractions. as determined by densitometry of SDS PA gels Moreover. this proportion appeared to be the same for all 4 treatments. C learly, P and G were not resolved by DEAE-cellulose. re. gardless of the treatments employed. Resolution by Gel Filtration DEAE-cellulose fractions from each of the shaded peaks in Fig. 2 were pooled. concentrated on Amicon filters. centrifuged at 90.000 X g. and applied to a Sephacryl S-‘ZOO column. The l .Z.e.|‘. . a... .A‘ I! '.<- “is? 4....- . 12504 centrifucation step was included to remove aggregated protein identified as largely P by SDS PAGE. Fig. .3 displays the S- 200 fractionation profiles for each of the 4 DEAE pools. together with SDS-PA gels of pooled peak fractions from So 3.. O U j "5 Noemi 0 t3. - -1.— \ O l.- L—‘A O MCI. M ‘ o NoCl, ti Pic. 2. DEAR-cellulose chromatography ofbypotonic amper- natant proteins. Samples were obtained in one of the t washing solutions described under ‘Materials and Methmb": (a) 10 mat 2- mercaptcethanoL 1 mat EDTA; lb) 10 out Z-mcrcsptoetnanol only: (c) 1 ma: EDTA only. 'dl neither Z-merceptoethanol nor EDTA. The some solutions were used to equilibrate and alone columns. but the elutmg solutions contuned 0—50') am NaCl applied as a linear gm. dien: lshown as a dotted lute with concentration scale on the right- hand ordinate). ROS from approximately 1500 retinas were washed to obtain the 200 mg U30 per column! applied to the columns. Three- ml fraCtions were collected. Absorbance at 280 nm is plotted usinst fraction number. Fractions indicated by shaded areas were pooled for PA gels and for further treatment. Dark Resolu:ion 200. The PA gel banding patterns observed confirm previous reports by Kuhn (Bl and by Baehr 9! al. (14>. Baehr et al. refer to the P protein as phosphodiesterue and to the G protein as 50K. which is close to the sum of its subunit M. values as judged on gels Kiihn l8) and Godchaux and Zim- merman ( 12: have ascribed GTPase and GTP-binding activity to a protein exhibiting the banding pattern of G. and Show:- awe et al. have called this protein G in the case of the frog (‘23). [n the following. we use the subscripts a, B. and y to desngnate the 3 subunits of both P and G proteins. in Fig. 30, the first peak contained exclusively P protein. as visualized by the high M. band on the gel. This band was resolvable into 2 bands with M, of 88.000 (Pu) and 85.000 (Pu) using gels prepared according to Laemmli [24). A third band (Psi . - 13.000l could not be differentiated from the dye front on the gel shown, although it was demonstrated on 15% PA gels. The second peak contained exclusively G protein bands of M. = 40.000 (G.,) and 35.000 (G,,l. A third band of .l . 3:: 10.000 (G.) again could not be differentiated from the dye front but appeared on 15‘?- gels. When 2~mcrcaptoethanol alone was present (Fig. 3b). only G was found in the late fractions. but a complement of G remained unresolved from P in the early fracnions. The minor peak at the left oi the early major peak was determined to contain largely aggregated P. Following sedimentation of ag- gregated P. both G and some P remained in the supernatant. In Fig. 3c both P and G were retained longer than by the other columns and eluted together in a broad peak of late fractions. In the early fractions. the G subunit with M. a 35.000 appeared to be enriched compared to the heavier G subunit and to P. In Fig. 3d the number of the peak fraCtion of unresolved P.G proteins was the same as that of pure P in Fig. 3a. but the peak was broad. and the amount of G relative to P and of the two heavy G polypeptides relative to each other varied across the peak. Pro. 3. Gel filtration of pooled peak fractions from DEAE chroma- tography. Pools obtained under each of the 4 conditions were concentrated. clar- I _ - l— « 0.4.. b . ilied by centnfucatton. and applied to a Sephncryl 3-200 column as described in text. SDS PA gels sliced from slabs ap- pear next [0 pooled fractions from which they were derived- in! 10 ms! S-mcrcup- ' toethanol. 1 mM EDTA: lb) 10 mil 2- 0.4 __C mercaptoethanol only. (cl 1 mm EDTA only: (0'3 neither 2~mercaptoethan01 nor ,- EDTA. In (bl the leading trac'ions of the E double Deal: at left contained preciptta- 5 0.37 ble inactive P and relatively little G. CO (\I <1: 02*- p 20 fraction -~m“vvmy o.’ -~- I. -.«v—’w—..-.,-‘ .m. .-..‘._.~.-". a. . - <- ‘ ‘. ‘ .. . '. ‘ . T T. 1 I ‘ i 1 Oran-d - - ~ g - P— “ 003’ c l" ‘ 4 'm: 8 al.—Ll : ' - ‘1 o.2~ ' ~ 0.5 ~ ' " ~ 40 60 2b 40 ’ froclicn ea--.c ._ -.- Dar/t Resolution These results indicate that only when 2-mercaptoethanol and EDTA were present (Fig. 30) was P completely resolved from G. In addition. pure G was obtained only when Z-mer- captoethanol was present (Fig. 3. a and b). However in Pig. 3, c and 0’, G was found in large excess over P in the later S~200 peak fractions. in some experiments 1 not shown) G could be obtained pure even in the absence of ‘Z-mercaptoethanol. In the presence of 1 my. EDTA. the relative purity of P was linearly related to the concentration of 2-rnercaptoethanol used up to the point of purity. This relationship was deter- mined by densrtornetry of stained gels (results not shown). Ten mM 2-merceptoethanol was usually sufficient to resolve all P Erorn G. Phosphodiesteraae Activity of Fractions of Various Degrees of Resolution Throughout the foregoing procedures. fractions were as- sayed for phosphodiesterase activity. All assays were done in the light, with GTP absent. Table I presents the activities of pooled fractions from each of the four experimental proce- dures involving presence or absence of 2-mercaptoethanol and EDTA. In each set of 5 rows. activities are presented for each Stage of the purification beginning with the isownically washed ROS (Fig. 1) to which the P prOtein was still bound. The pooled DEAE peak fractions were the same as those indicated by shaded areas in F 1g. 2. and the 5-200 pools were the same as indicated by PA gel insets in Fig. 3. TABL£ I Phoephodiesterase activity of fractions of various degrees of resolution ROS (900 mg of protein) from 1500 retinas were split into 4 aliquots treated With or without 10 mm 2-merceptoethanol or 1 min EDTA as described under “Materials and Methods" (see Fig. 1). Indicated fractions were assayed for phosphodiesterase activity as described under "Materials and Methods." Total activity recovered at each stage and change in specific activity (-fold purification) relative to the ROS are also given. The pooled DEAR pea}; fractions were those indicated by shaded areas in Fig. 2. The 5-200 fractions were those indicated by PA gel insets in Fig. 3. TPlane 110-1 7 l . P q- I 'FOld Fraction ll’roteinl mg?" ; “£35? :punfi- : I actiflty l -> 1mm“ : ; total/fl , mg imimmgHWI’mmE ROS 325 . 0.13 v 29." 1.0 (a) same” + EDTA . i l ; “I I Isotomcally washed ROS . 254 l 0.41 1 10;.“ 3.1 Hypotomc supernatant ( 59 f 0.36 ' $1.24 2.7 Pooled DEAE peak fractions 3 13 3 0.46 . 5.95 ' 3.5 3:200 early pealt lpure P) g 3.7 . 1.43 . $.29 ' 10.9 5.200 late peak (pure G) : 6.5 l 0.06 i 0.39 . as (1)) +354}: - EDTA 1 : l lsotonically washed ROS . as .‘ 0.35 I 33.2 g 2." Hypotonic supernatant ; 64 l 0.68 . 43.52 5.2 Pooled DEAE peak fractions 3 16 i 0.71 : 11.36 g 5.4 S-QOO early peak (P and G) , 8.0 . 0.2’.’ l 2.16 . 2.1 S-‘ZOO late peas .pure G) , 4.3 . 0.15 l 0.64 l 1.2 (c) —BME + ED‘TA l . l ‘ Isotonically washed ROS } 236 l 0.35 9 s9 53 3 2.9 Hyputonic supernatant , 48 1.30 l 62.40 ( 9.9 P00led DEAE peai't fractions 1 15 l 0.58 ' 13.20 1 6.7 3-200 peak 1? and G) l 6.3 1.67 10.52 1 12.7 5.200 trailing edge (P and G) , 3.2 5 (.70 i 5.44 ‘ 13.1 (d) ~BME — EDTA 1 l . Isotonicallywashed ROS ' 2‘24 9 0.48 ' 107.52 _ 3.7 Hypotonic supernatant l 46 l 1.49 ' 65 54 1 11.3 Pooled DEAF. peak fractions ; 13 ; 0.96 ; 12.4.3 ; 7.3 8-200 early peak (P and G) 1 2.9 f 4.40 ; 12.76 “ 13.5 5-200 late peril; 4? and G) l 4.2 I 0.86 l— 33.61 . 6.6 " Total protein. b BME. vaercaptoet'nanol. 12505 A marked elevation of total activity was observed in the isotonically washed ROS. regardless of whether 2—mercapto- ethanol or EDTA was present. It appears likely that the increase resulted from removal of inhibitory factors or alter- ation in regulatory interactions of the remaining proteins. These possibilities were not explored. The supernatants obtained by hypotonic washing of the ROS in the media shown in Table 1, fractions b. c. and d. exhibited increased phosphodiesterase specific activity rela- tive to the isotonically washed ROS. The greatest activity was found in the supernatant obtained in the absence of ‘2-mer- captoethanol and EDTA (Table 1. fraction d) While the lowest was found in their presence (Table 1, fraction a). SDS PA gels of the supernatants (not shown) were all similar. Since most of the hypotonic supernatant protein was P and G (in a ratio of approximately 1:2 on a mg/mg basis) relatively small in- creases in specific activity were brought about by DEAE and S-200 chromatography. Furthermore. recovery of protein and of activity from DEAE was consistently low 125-4092 protein and iii-30% activity). The highest activity (4.40 pmol/min/mg) exhibited in Table I was that of P still unresolved from G (fraCtion d. S-‘ZOO early peak). The lowest aCtivity (0.27) in any S-‘200 fraction con- taining P was observed in the unresolved P obtained with '2- mercaptoethanol alone (fraction b. early peak). This was also the fraction containing a significant component of aggregated P (Fig. 36. leading edge of the early peak). The activity of pure P obtained by use of ‘2-mercaptoethanol and EDTA (fraction 2:. early peak) was intermediate between these ex- tremes. as was that of the unresolved P,G fractions obtained with EDTA alone (fraction c. peak). In other preparatiom using the same reagents phosphodiesterase activuy of the pure P was as high as 4.8 pmol/min/mg, and actiVity of P still unresolved from G (without ‘2-mercaptoethanol or EDTA) was as high as 17. These results demonstrate that pure P was not as active catalytlcally as when unresolved from G. The most aetive form of P was found with an excess of G present and with the two heavy G subunits in roughly equal propor. dons (d in Fig. 3 and Table l). The least active form of P contained P in excess of G. and a significant amount of the P had aggregated (Fig. 3 and Table I. b). In the case of the P forms mth intermediate activity, one had no C (Fig. 3 and Table I, a). and the Other was noticeably deficient in the heaViest G subunit. Thus while the presence of G was impor- tant to catalytic activity of P, the state of the G protein was evidently important as well. When we purified P using light and dithiOthreitol according to the procedure of Kuhn 116). we observed activity of 2.7 iunol/min/‘mg in pure P immediately after its removal from the ROS. This activity was comparable to that of P purified by our dark resolution method. However. Baehr er al. ('14) reported activity of 15 ‘amoi/min/mg in their purified P. We have found activities that high only in the presence of G. Aetivity of our pure P after readdition of G and P,G-depleted ROS membranes was typically 15-23 .crnol/rnin/rng of P pro- tein (11). Differences in G Fractions The phosphodit-Sterase activity found in G-containing frac- tions was virtually zero for pure G. In addition to tne different gel patterns (Fig. 3) and apparent effects on actitity of P (Table I) already described different pure G [motions had different effeCts on activity of pure P when G and P were mixed together. Table 11 illustrates this pomt. The pure G obtained With Q-rnercaptoethanol plus EDTA inhibited the catalytic activity of both pure and unresolt ed P. while pure G obtained in the absence of :2-mercaptoethanol and EDTA -ua..a_. .".'. ' I '_ fi. ' V . - ' .l . l. ’ I, ‘ ‘ . . . a .l . ‘ . . u ‘ l. 4 ‘ V I It. 11‘“, ’ I ’ v s '_ ‘..‘-, I. - " '. ' ‘ ' . I > 1 \- .... .uuvm'm‘w53‘J. m— ‘-Iu‘-" v.—-.l...-.41 . "m; . .., ’. ).-». n -'.' 5. "I ._ a._- ._ r (xv—f1:- ...‘. 41-.» I: 1. with: A ~‘Jr _ _ ,. . .‘Q,-»’ 'a . '. . . _ I. ' ' 1 I . . ' i I V . i. ...l i..:.- ...-auburn-.- .. . , _',, " "' ‘ . .Q— ‘_' |I I - “I‘M‘t’firerau w- iv..- .» .-- or)". . i ,. ,_ .. 1:4-'42 «‘f I“ L?- «n—gv ' F3" ‘ b .’. E; «81‘ «to “I 52"" (32?. II ‘-. .4“ “3&5 12506 (from is. different preparation than that shown in Fig. 30" did not. We show elsewhere that these two kinds of G proteins were identical on SDS PA gels. but not on native gels (11). Furthermore. the pure G which inhibited catalytic activity of P in Table ll promoted reconstitution of full activity of P in the presence of GTP and bleached ROS membranes previ- ously depleted of soluble proteins (11). Taken together. these observations argue that at least some form of G regulates the activity of P by physically interacting with it. Neither of the above G proteins exhibited bound nucleotide detectable by Ans/Am ratios, although deteccability of nucleotide by this method is not quantitative. In any case the physiological interaCtion may still require GTP. Stability of P Protein Pure P protein obtained With 2-mercaptoethanol and ED’I‘A (early peak. Table In and Fig. 3a) subsequently ex‘ hibited aggregation and lost its phosphodiesterase activity upon storage at ~20 ‘C. The two impure P fractions obtained without ‘2-mercapthoethanol {Table l and Fig. 3. c and d) appeared Stable in storage. The impure partly aggregated P obtained With 2-mercaptoethanol l Table lb and Fig. 3b) lost some of its activity in storage but had the lowesr specific activity of any of the gel filtration fractions to begin with. These observations led us to compare stability of activity with the tendency to aggregate as reflected by the presence of multiple P bands on native PA gels. At least through the stage of its elution from DEAE«celluiose. impure P obtained with 2-mercaptoethanol and EDTA appeared to have retained catalytic Stability and had shown no evidence of aggregation. Its SDS gel (Fig. 40’) displayed both P and the heavy subunits of G. Its native gel (Fig. 40) showed a single low mobility band and a diffiise nonuniform high mobility band. Its activity was 1.38 ,umol/miL/rng. When this unresolved P,G fraction was subjected to S-‘ZOO gel filtration, retaining '2-mercaptoethanol and EDTA. the resulting early peak contained pure P which exhibited a single high M. band on SDS gel (Fig. 46’) but multiple low mobility bands on native gel (Fig. 4b). Activity was only 0.67. The late peak from the same column contained pure G, as confirmed by the SDS gel (Fig. 4e). Its native gel contained a diffuse high mobility band. U, instead of applying the unresolved P.G DEAE pool directly to gel filtration, it was first dialymed to remove EDTA and add Mg“, the multiple bending of pure P seen on native gel wu noticeably decreased (Fig. 4c; compare With Fig. 46). and its activity was increased T-fold. to 4.80. If. in addition. ‘2.- mercaptoethanol was removed prior to gel filtration. P could no longer be resolved from G. but its activity increased still further, and multiple banding was no longer detectable on the native gel of the unresolved proteins (Fig. 4d). Even more significant. we believe. was the absence of any evidence of the difiuse high mobility bands seen for pure C (Fig. 4e) or for G still unresolved from P in the DEAE pool (Fig. 4a). Though TABLE ll Phosphodiesterase activity of P and 0 fractions separately and mutual together after separation Pure G protein derived in the presence l+—-l and absence (--) of both 2.mercaptoethanol and EDTA was added to pure P or P unre- solved from G. Phosphodjesterasc activity is expressed as cpm/assay. About 3 ug of the P component and 7 ug of G were present in the assays. The P and PG proteins used here were not from the same preparation. No additions ~~G --C- No additions .35 39 ‘4'? (pure) 840 496 856 —-—P.G (unresolved) £318 252 506 MW . . mwmm-amm~-'~vv-‘ Dark Resolution an - . - ... - ‘- " - O a O a ‘ l‘ - i '1' ll ~' . «t I ’ 3 o- h u- “' " o a' c c‘ d d' e e‘ LBE 0.67 4 SO 1690 (OJ Fm. 4. Stability of P protein. Phosphcdiesterase activities (pmol/min/mgi appear beneath the PA gels. Tracks o and a’ are native (a) and SDS (6’) PA gels of pooled DEAE fractions obtained in the presence of 2-mercaptoethanol and EDTA (as in Fig. 2a, shaded area). This pool was then split into 3 aliquots; one aliquot was applied to Sephacryl S-200 equilibrated in 2-mercaptoethanol and EDTA. The early peak from the column (obtained as in Fig. 3a and Table l (a), S—QOQ early peek) appears in tracks b (native) and b’ ' (SDS). The late peak from S-200 (correponding to Fig. 30 and Table Ila). 5-200 late peak) appear: in tracks 2 (native) and e' (SDS). A second aliquot was dialysed to remove EDTA and to add 1 mM Mg“ before 3200 chromatography. The early peak which resulted appears in tracks c lmtivei and c' (SDS). while the late peak was identical With that shown in e and e’. The third aliquot was diolyzed to remove both ‘2-mercaptoethanol and EDTA and to add l mu Mg“ before S- 200 chromatography. The early peak which resulted appears in tracks d and cf. The late peak had a native gel similar to a and an SDS gel simiLar to a' but with less P. Its actiVity was 0.6 mol/min/mg. ROS from approximately 1200 retinas were washed to obtain the protein used in this experiment. G was unmistakably in evidence in this fraction (Fig. 4d'i, it was clearly tightly coupled to P. These results emphasized that stable activity of P required prevention of aggregation. This was accomplished either by maintaining the interaction of P with G or to some extent by including Mg“ in the medium during isolation of P from G. Once multiple bending of P had occurred. we tried by a variety of procedures to reverse it, with no success. We also tried, by removing 2- mercaptoethanol afterwards. to name maximal activity to the pure P obtained in the presence of added Mg“ (Fig. 4, c and c’). Activity increased only 29%. Thus mere removal of 2- mercaptoethanol could noc offset removal of G. Sucrose Density Gradients To determine whether the resolution of P from C after FAB chromatography was uniquely dependent upon gel filtration, we undertook their resolution by sedimentation velocity on sucrose gradients. Pooled DEAE pealt tractions from the experiments described in Fig. '2 were applied to the gradients. The analysis of gradient fractions by SDS PAGE appears in Fig. 5. In the aliquots prepared with 2-mercapto- ethanol and EDTA (Fig. Sal. complete resolution oi P and G was obtained on the sucrose density gradient. With 2vmercapo methanol only (Fig. 5b) resolution was nearly complete. With EDTA only (Fig. 5c) resolution was less complete than with Lmerceptoethanol. and in the absence of both reagents (Fig. 501’) resolution was poor. This outcome roughly paralleled the results or" gel fil ration (Fig. 3). However. separation appeared to have been better. and unlike the gel filtration. most of the G procein was completely resolved after any of the four treatments. The peak fraction number was the same after all four treatments for both P and G proteins. which was not the case in gel filtration (compare Figs. 3 and 5). inclusion of high, or low salt concentration. B-mercapcoethanoi. or EDTA in the gradient solutions to maintain the concentrations to which pr0teins had been exposed during DEAE chromatography had no effect on the results. """“Vi~Anr-~w.»-. -. -. -—-"‘-‘. ...- i . for w“ ;-' ~41: A._.>_1§_.;,';‘“3.. er 35?? 'L“ 2'1, a \' . r- " . Ifii»f”‘_ . A, L ‘ ‘4‘“ Tt‘: 55.33.54‘ if??? I: .4' 1 ‘ '. . t" at? E."§fi+- Trot” v. 1’ -' 5 was. 1“ 3‘6 .. 4 ' 1‘ a. \ Ami 1‘" (A. n \ 9" . . ...: 3....) n u I I. n I ‘ . I, ' V ”...-“1’.“ “Am”... ‘.-'I_‘ “'A‘.. 21 . 4--) '. u ' . 1 ._... -w-—.———r< ...-.. s. 5. z‘ ‘. .._~v -. . r ,1 k ' , . Dark Resolution 1-..- 9'071201415 90H 011121.124 Pic. 5. Resolution of P from G by sucrose gradient sedimen- tation following DEAE chromatography. Proteins were obtained by DEAE chromatography in (a) 10 ms! Z-mertaptoethanol and 1 am EDTA: Iby I-merccptoethanol only“. Ic') EDTA only: 1a": neither 2—mercaptoethanol nor EDTA. Sedimentation was performed as de- scribed under “Materials and Methods." 0.25mi fractions were ana. lyzed by SDS PAC E. Sucrose gradient fraction number: are indicated beneath the gels. Lower numbers represent greater sedimentation velocities The tractions shown in the figure were taken from gradients containing 250 mu NaCl but neither fl-mercnptoethanol nor EDTA. Molecular Weight Estimates (a) Sucrose Density Gradient Cenm'fugation—This pro- n’ded an eatimate of M, based upon comparisons with stan- dard proteins. The .11, of pure P was approximately 185,000 (Fig. 60). judged in relation to aldolase (158.030), cataiase (232,000). and bovine serum albumin (66.000). G was esti- mated to have a M, of 35.000. Unresolved P.G (Fig. 5d) had a M, equal to that of pure P. Using the same method, Saab: 91 al. 4141 estimated the M, of P to be 170.000 and that of G to be 80,000. (bl Gel Filtration—Pooled DEAE peak fractions similar to the shaded areas in Fig. ‘2. a and b. and marker proteins were separately applied to a Sephadex G400 column. The P protein eluted prior to catalase (232.000) but after blue dettran which defined tne void volume (Fig. 65!. By extrapolation from the 4 points of the standard curve. P was estimated to have a M, of 250.000. G was estimated to have a .11, of 50,000. The P prorein peak eluted from G-200 after 2-mercaptoethanol and EDTA treatment in this instance contained a small 640%) amount of G. The peak obtained after '2-rnercaptoethanol treatment alone contained a much greater amount of G but produced an identical M, estimate. (c; SDS PA Gels— Using Sigma .11, standards and gels made according to Fairbanks et al. (20) our e5timote of M, for each of the two large P subunits was 85.000. since we seldom saw separation of the 85,000 subunits on these gels. They did separate into two bands of equally intense staining on Laern~ mli gels 1'24}. Estimated M. for the separated subunits was 88.000 \P..) and 85.000 (Pm. The third subunit of P was visualized on 15% PA gels made according to Fairbanks et al. ('20) and had an estimated M, of 13.000 (P , l. Thus the estimate for M, of P derived from PA gels is 183,000. For the G subunits. our estimates were 40.000 (G.,). 35.000 (Ga). and $10,000 111,). providing an estimated M, of 85,000 for G. (d) Analytical L'lrracenrrz',"ugazion—Pure P and P still unresolved from C but exhibiting a single band on native gel (see Fig. 4d) exhibited very sirnilar sedimentation profiles on analytical ultracentrifugation. The 330,... values of borh. estic mated by reference to catalase and aidolase in the same buffer, were 8.2. indicating a .V, of 165,000 if the protein was spherical. 12507 I V Y T a 150?- P ‘4 1 m l 4r th .m J I 3 ml 6 a r- l 4 mm. . a .a \i 22: fraction in l cordon P am» .1 mating» é . 1 cytochrome c {U 13 57 2.5 v.1v° Pro. a M, estimates for pure P and pure G. a. estimate by sedimentation velocity on sucrose gradients. LogmM. is plotted against fracuou number. which is proportional to velocity. Unresolved P.G exhibited the same velocity as pure P. b. M. estimate by gel filtration on Sephadex G620). Logic. M, of pure P and pure G is calculated by reference to protein standards (Pharmacist as a function of elution volume 1V.» divided by voad volume (V0). Unresolved P.G exhibited the same elution volume as pure P. TABLE 111 Summary of molecular weight determinations on pure P. pure G, and PG proteins P P10 G M, by gel filtration 250.000 250.000 50.000 M, by sedimentation veloc- 155.000 185.000 35.000 ity M, by SDS PAGE 58.00) 38.000 35,000 85.000 40.000 40.000 35.000 35.000 13.000 13.000 5 10.000 510.000 sea... from analytical ultra- 8.2 8.2 centnfugation M, calculated trom 320.. 165.000 16.300) Unresolved P.G had a minor component (13% of total absor- bance) ot’ slower migrating material. There was only 3% of such material in the case of pure P. However. the bulk of the G apparently comigrared with P. The P "G ratio for the unresolved P.G examined here was “2:1 (ugpgi. estimated from plots of Staining density on SDS PA gel versus micrograms of pure P or G applied. The specific absorption of P at 280 nm was approximately 0.425 nil-cm"i -mg", while that of G was 0.927. This implies that more than 50% of the observed A23. in the analytical ultracentrit‘uge was attributable to G. Thus the 13% slower monng material. if indeed it was G, represented at most a quarter of the measured amount of We performed these centrifugations in 20% glycerol. because without glycerol. pure P was unstable even at liquid N: temperature. The analytical ultracentrifugation profile ot‘the deStabilized P (without glycerol) was polydisperee. With a .. y‘"m-,“’fim.—‘w> mwh.‘figvv-wafifi~m——M“W . .. . . .10“. . hon-1 .‘ D ‘ . . ... .d ., .i . . .-_ I. I" I n I.“ .. - .. .~ . .. ~. .-o., “...-yrzurhua-g. r_~.,c.,..‘~.- .' v.--" . . ‘7'. H.“ . 2508 Tut: l‘v" Ammo acid composum of P and G proteins Analyses were normalized to 10 pg of protein per sample. excluding tryptcp han Which coes not survive the and hydrolysis used Duplicate analyses were averaged :range/L‘ and then converted to nanomolcs of amino aczd. Analysis does not differentiate between side chain acid and amide forms. P 0 Amino acid 1mol :nngez 2 nrnel trance/2 C ylteine 1.53 (M 2..) .1 Aspamc acid: 9.02 0.5 11.09 0 4 Threonine 4.14 0.1 5.13 t) Senna 4. 71 0.1 8.59 0.1 Praline 3.32 0.3 1.70 O (32ch acnd 10.82 0.4 8.86 0.1 Glycine 5.9) 0.4 6.99 0.5 Alanine 5.43 04‘. 5.73 0.2 Va‘une 4.82 0.1 3.4-3 0.7 Methionine 0.23 0.1 0.20 02 lsoleucine 3.39 0 3.91 0.1 Leucme 7.32 0.! 7.16 0 Tyrosine 1.82 0.1 1.42 0.2 Phenylalanine 3.92 0.1 3. ll 0 Histidine 1.97 0.2 1.30 0.1 Lysine 5.14 0.! 4.76 . 0.1 Arg'inine 3.71 0.1 4.40 0.1 considerable amount ot'aggregated material which sedimented very rapidly to the bottom of the cell. Table III summarizes the M, estimates by the four preceding methods. Carbohydrate and Lipid Determmations No evidence of carbohydrate content in either P or G was found when SDS PA gels were Stained using the procedure of Racusen (25). Oil Red 0 and Sudan Black S stains for lipid content were also negative. Amino Acid Analyst's Table IV presents the amino acid composition of resolved P and G prOteins. There appear to be sufficient differences in the profiles of these proteins to minimize the possibility that P and G are closely homologous proceins. DISCUSSION Purified P and G proteins derived trom darkcadapted ROS appear to be essentially identical with the major prOteins which Baehr et al. (N) and Kuhn (16) have reported. We have confirmed Kuhn's method of resolving P and G using light and GTP. in the process of dark resolution of the proteins. a considerable amount of new information about their properties and relationship has been obtained. The purified P was not stable. We believe a major part of this instability resulted from a tendency to aggegate when sepa- rated from the G protein. Consistent with this interpretation are the multiple banding of the inactive form of P on native gels and its appearance, verified by SDS gels, in an early peak ninning at the void volume on gel filtration isee Fig. 3b). This peak was diminished by prior centrifugation, and the pellet contained largely inactive P protein. The separation of G horn P under the action of high 2-mercaptoethanol and Withdrawal of Mg“ appears to have been a crucial event initiating aggre- gation. However, the mere presence of G clearly did nor prevent significant aggregation of P in all cases. Nor only did different forms of pure 6 exert different effects on the activity of P. but the relative amounts of the two heaviest G subunits present also appear to have been important to the activity of P. Among the possible alterations in G responsible for desta- bilization of P, one (aggregation) was nOt observed. The """"-‘N*“e-.~m.,.g...--‘ . ,. Dark Resolution destabilizing effects on P attributable to removal of G could be partially prevented by Hg" or to a smaller extent by removal of Q-mercaptoethanol before final separation of P from G. Deatabiliza tion of P appeared irreversible once it had occurred. However. if it had not already occurred by the time P was resolved from G, subsequent destabilization could be prevented by storage in l mu MgCh, 50% glycerol w, v). 250 mM NaCl. 10 mM Tris~CL pH 7.8. Because G appeared in most instances to protect P from aggregation and loss of catalytic activity, we infer that physical interaction with G protein is essential to the stable catalytic activity of P. This inference is reinforced by observations in the accompanying report on reconstitution of light and GTP- dependent catalytic activity from P, G. and soluble protein- depleted ROS membranes ill). The manner and Stoichiome- try in which G and P physically interact is clearly of great interest. Although substantial amounts of G could easily be resolved from P Without special measures (by DEAE chromatography followed by sucrose gradient sedimentation. Fig. 5) we have demonstrated that in the absence of 2-mercaptoethanol and EDTA. P could not be resolved from some amount or form of G by an anion exchange resin (Fig. 2‘), gel filtration (Fig. 3), native PA gels (Fig. 4). sedimentation velocity (Fig. 5). or analytical ultracentrifugation (Fig. 6). In addition. the two proteins were not resolved by GDP-Sepharose and Blue Seph- arose" Adventitious copurification of two separate proteins through all these conditions would require a high degree of similarity. which is not corroborated by the ammo acid anal. )ees l'l‘able IV). We conclude, therefore, that P and G c0purify as a multisubunit complex. Regardless of the stoichiometry in the P'G complex. there is apparently an excess of G in the ROS ill). The M, estimates summarized in Table [II demonstrated three important features of the apparent complex. One was that gel filtration prowded a higher M, escimate than the other methods. This may have resulted from an asymmetric conformation (nonunity axial ratio) of both the P:G complex and pure P or from an unusual interaction With the gel. Among possible causes of asymmetry neither carbohydrate nor phos- pholipid residues would appear plausible. because neither was found in cytochemical staining of gels. A second feature was that the M, of the complex fell in the range of 165,000-250000. Preliminary sedimentation equilib- rium studies3 of P and P26 suggesr more than one species of both proteins. despite apparent homogeneity on SDS and native gels, and as judged by other physmal measures. This may reflect aggregation and may also explain loss of activity observed under some conditions. Perhaps the most striking feature was that the M, of re- solved ? and of the P:G complex was virtually identical using all methods. This necessitates careful attention to the stoichi— ometry between the subunits of P and G in the P:G complex. The lower M, estimates would be consistent with the arrange- ments, P,,G..a, or PJ.G..5.., while the higher estimate would argue for either P,..(G..a.)2. P;,(G.,g,)2, or P.u.G..fi.. Additional possxbilities must also be considered. More than one form of the complex could exist in equilibrium. complicating the de- termination ot' stoichiometry. AnOther possibility is that the minor Iyl subunits of both G and P may make larger contri- butions than we have allowed in the foregomg. Their impor- tance cannoc be neglected. Since the present work was com- pleted. Hurley and Stryer have reported that P. (M, = 13.000) is an inhibitor of catalytic activity and proposed that a func- J R. E. Kohnken. D. M. Eadie, A. Revzin. and D. G. McConnell. unpublished observations. -~rn-~.--.—— 7.7-, - ,.\, ‘_ M ._ 7».n-er..-. ‘.- 7 ' . , ‘ ‘ . - ' . r.’ ' . ' ~ ‘ ' ' a L01“. Jo‘s-ido- A.L..A“ A hub—2...: -.~ . .‘ - “. o l N . Judie-\- - .H . . ,. . . Ion-.N‘—-I u. ' . .‘ _ ' ‘ _ "r. -. ...: l.¢.a"-Jt:l-lr-l' 0 Dark Resolution 1‘509 :ion of G is to remove it from the other P subunits (26). They have also proposed that G. specifically binds GTP and is responsible for the initial amplification of the effects of light (15). Consistent with this proposal is the high activity ex. hibited by the P20 complex in which G., participated dispro- portionately compared to C; (Fig. 40"). The present study demonstrates that when removed from ROS membranes in darkness. the catalytic protein P forms a multisubunit complex with a regulatory protein G. leaving a considerable amount of G protein in excess of the complex. Although the sull’hydryl reagent and cation chelator used in the study permitted resolution of P from G after their removal from the membrane. the relationship of P to G or of either protein to the membrane before their removal in darkness is unknown. In the absence of the membrane. G appears to be essmtial to the Stable catalytic activity of P. The relevance of this regulatory interaCtion to the native state is supported by the observation. made elseWhere. that G is required for recon- stitution of light and GTP-dependent phosphodieSteras-e ac- tivity when purified P and G proteins are added to ROS membranes depleted of soluble proteins (11). Acknowledgments—We gratefully acknowledge critical reading of earlier drafts of the manuscnot by Christine Collins. James H. Asher. Jr., W. W. Wells. J. L Fairley. and C. H. Suelter. REFERENCES 1. London. (3.. Salomon. Y.. Lin. M. C., Harwood. J. P.. Schramm. 31.. Woifl'. J., and Rod'tiell. M. H974) Proc. Natl. Acad. Sci. U. S. A. 71. 3087-13030 2. Perkins. J. P.. Johnson. G. I... and Harden. T. K. H978) Adv. Cyclic Nucleotide Res. 9. 19~J2 3. Ross. E. M.. llaga. T.. Howlett. A. C., Schwmmeier. J., Schleil'er. L S . and Gilman. A. G. (1978) Adv. Cylic Nucleotide Res. 9. 53—68 4 Bitensky, M. W.. Wheeler. G. L... Aloni. 3.. Vetury, 8.. and Memo. ‘1'. ll9’78i Adv. C yclic Nucleotide Res. 9. 553472 5. Robinson. P. P... Kawamura. 8.. Abramson. 8.. and Bownds, M. D. (1930) J. Gen. Physwl. 76. till-8&5 6. Miki. INC. Keirns. J. J.. Marcus. F. P... Freeman. J., and Bitensky. M. W. Il9’3i Proc. Natl. Acad. Sci. L'. S. A. 70. 3820-3824 1'. Liebman. P. A., and Pugh. E. N., Jr. ”980) Nature 287. 734—736 . Kill'm. H. (1980) Nature 283. 587-589 . Godchaux. W.. [It and Zimmerman. W. F. (1979) Exp. Eye Res. 28. 48$500 10. Mild. N.. Bamban. J. H., Keirns. J. J., Boyce. J. J., and Bitensky, M. W. (197.5% J. Biol. Chem. 250. 6320—6327 11. Kohnken. R E.- Eadie. D. M” and Mc Council. D. G. (1931) J. 3wl. Chem. 256. 12510-12516 12. Godchau. W. III. and Zimmerman. W. F. (1979i J. Biol. Chem. 254. 7874-7884 13. Robinson. W. 8.. and Hagins, W. A. (1979} Nature 280. 398-400. 14. Baehr, W.. Devlin. M. J. and Applebury. M. L (1979) J. Biol. Chem. 254. 11669-11677 15. Fun. 3. K-K. Hurley. J. 8.. and Stryer. L. (1981; Proc. .VatL Acad. SCI. U. S. A 73. 152-156 16. Kuhn. H. [1960) Neurochem. Int. 1, 2159—285 17. McConnell. D. G. (1965) J. Cell Biol 27. 459-473 18. McConnell, D. G., Dangler, C. A... Eadie. D. 51.. and Litman B. J. (1981) J. Biol. Chem. 255. 4913-4918 19. Low. 0. H.. Rosobrough. N. J., Farr. A. I... And Randall. R. J. (1951) J. Biol. Chem. 193. 265—275 20. Fairbanks. G.. Ste-ck. T. l... and Wallach. D. F. H. (1971) Bio- chemistry 10. “2606-2617 21. Chen'enlte. C. H. (1969) .4 Manual ofMethoda for theAnalyttcal Ultracentrifuge. Spinco Division of Beclunan Instruments. inc. Stanford Industrial Park. Palo Alto. CA 94304 22. Handbook of Biochemistry (Sober. H. A., ed) (1968). Chemical Rubber Publishmg C0,. Cleveland. Ohio. p. C-11 '23. Shinozawa. T., Uchida. 8.. Martin. E. Cafiso, D.. Hubbell. W" and Bitensky, M. W. (1980! Proc. Natl. Acad. Sci. U. S. A. 77. 1408-1411 24. Laemrnli. U. K. (1970) Nature 227. 880-685 25. Racusen. D. (1979) Anal. Biochem. 99. 474-476 26. Hurley. J. 8.. and Stryer. L. (1981) Biophys. J. 33. 203: IT-PM- C2) 100'! . '7‘ .‘0—._. ..-..na —.- Mr_~'_‘~...- . "‘r‘.-"'_ ”flu-pow - ~‘fi’_,.-—--—17.—y—~ \- I . " . ¢ . , '- 3.-.A...~.’.¢d A... \ - , . 4 ‘ l - ‘nilhv l .n {E}. -. f ‘ . ., l. ‘.‘ . . . a. .- 1—I-‘kh.. §'.vun4_‘ ... Tn; Menu. 04 Slatoacu Cums-r37 Vol. 256. No. .3 issue ofCicembu l'). pp. llbul—t'lllé. lg! Pt-Ateu U! L'.S.A The Light-activated GTP-dependent Cyclic GMP Phosphodiesterase Complex of Bovine Retinal Rod Outer Segments RECONSTITUTION FROM CATALYTIC AND REGULATORY PROTEINS IN THE PRESENCE or MEMBRANES DEPLETED OF SOLUBLE PROTEINS‘ (Received for publication. April 20. 1981) Russell E. Kohnken. Deborah M. Eadie. and David G. McConnelli From the Deparanent ofBiochemzstry. Michigan State University. East Lansing. Michigan 48824 Reconstitution of GTP- and light-dependent cGMP phosphodlesterase activity in bovine rod outer segment (ROS) fragments depleted of soluble proteins required two distinct soluble components. one (F) catalytic and the other (C) regulatory. Three forms of G were tested. All promoted increased GTPase activity of soluble pro- tein-depleted ROS membranes (DROS), but only a form isolated in the presence of 2-mercsptoethanol or EDTA was effective in reconstituting phosphodiesterase ac- tivity. Reconstituted membranes had about 65% of the orig- inal activity. but the degree of acrivation by light and GTP was the same. 10- to 100-fold. Maximal activation of the reconstituted syStem, as in the case of the native ROS. occurred in the range of 20-200 um GTP. Activa- tion was proportional to the amount of G protein, rel- ative to P. up to at least a lO-fold excess by mass (20- fold molar excess assuming the molecular weight of P to be about twice that of G), to the amount of membrane protein up to 20-fold excess by mass relative to P, and to the amount of P protein up to 12.5 pg per 50 pg of ROS membrane protein. Pretreatment of DROS in light or dark with concan- avalin A bad no significant effect on reconstitution. However. inhibition of reconstitution by 48% or more was observed after 8 h of incubation of DROS with a 400-fold molar excess (with respect to rhodopsin) ofN- ethylmaleimide. Trypsinization of DROS doubled its ability to promote reconstitution. suggesting a trypsin- sensitive component in ROS membranes which is in- volved in pbosphodiesterase activation. Although the form of the reconstituted P:G complex may not be the same as in vivo. these results support the concept that P and G interact natively as a complex. with G exerting regulatory control on P subject to modulation by light. GTP. and the ROS membrane. In the bovine retinal rod photoreceptor, cyclic GMP metab- olism is externally regulated by light t1-3). The light-activated cGMP phosphodiesterase of the rod outer segment is appar- ently bound under physiological conditions to the F108' mem- ‘ This work was supported by Unzted States Public Health Service Grant EY-1574 to DGM. by funds from the College of Osteopathic Medicine at Michigan State University. and private sources. The cosns of publicatzon of this uticle were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 13 US.C. Section 1734 solely to indicate this fact. t T) uhom reprint requests =hould be addressed. ‘Tbe abbreviations used are. HOS. retinal rod outer segments: DROS. ROS membranes depleted of soluble proteins: Con A. concan. branes and requires GTP t‘or acrivation However, it is easily removed from the membranes by hypotonic washing (4, 5) and can be restored to the membranes in the presence of isotonic salts (6). Separated from the membranes, it is no longer light sensitive and is not activated by GTP (6, 7). In a companion paper (8) we have reported that if the enzyme was solubu‘ized in darkness it “as found to consist of a catalytic (P) protein complexed with a regulatory or stabi- lizing protein ‘0). in concert with R08 membranes G also displayed GTPase activity (9). When removed in darkness. P could be resolved from C using 2-mercaptoethsnol and EDTA, but the pure P exhibited a tendency to aggregate and catalytic instability resulting from the dissocmtion of P from G. These findings implied that the G protein was essential to stable catalytic activity of P (8“. The present report extends this inference by showing that G is absolutely required for reconstitution of the light-acti- vated GTP~dependent phosphodiesterase activity in P,G-de- pleted ROS membranes. Quantitation of the reconstituted activity and preliminary characterization of its soluble and membrane components are also reported. EXPERIMENTAL PROCEDURES Preparalicn of ROS and DROS—Cattle eyes were obtained from Murco. Inc. of Plainwell. MI. and ROS were prepared on sucrose gradients as described previously :8). DROS were prepared by twice washing ROS in dim red light at 4 °C with 10 volumes of 0-25 M sucrose. 10 mm Tris-Cl. pH 5.0. 150 mat KCl With (a) 10 ms! 2- mercaptoethanol and 1 mM EDTA. or (bl no additions. ROS were sedimented from these media by :entritugauon at 30.000 x g {or 30 min. The resulting pellets lline ‘3 in Table ll were resuspended in 0.25 M sucrose. 10 mM Tris-CL pH 8.0. containing the same additions. centrifuged at 93.000 x g for 90 min. and this procedure was repeated. The pellets (line 3 in Table l; were resuspended in the watching media. Some were scored direc:ly at ~20 °C in the dark. while others were bleached on ice betore storage. Preparation of P. P:G. and G—Soluble components were prepared in darkness as described in the accompanying report (6) or in the light as described by Kuhn (10). PhosphodicsteraseActivity—This was essay/ed as oescnbed in the .iccotnpanvmg paper {8). Phosphodiestemse reconstitution sassy: entailed mixing of P or PrG complex. G. GTP. and DROS at .30 ’C for 5 min or less prior to initiation of reaction by addition of cGMP. Pabacrylamide Gel Electrophoresu—Tbis was performed usinz sodium dodecyl sulfate and nondenaturing acrylamide gels as de scribed in the accompanying paper 18). GTP Binding—These assays were performed essentially as de- scribed oy Godchaux and Zimmerman «9). Up to 100 Au_ of protein were assayed at 0 ’C in 100 ms! NaCl. 1 mm MgClg. 15 mat 2- mercaptoethanol. 20 mm Tr's»Cl. pH 7.5. with ‘20 py. [So‘HlflTP aval'm A; NEH. .‘V—ethylmaleimide; EGTA. ethylene glycol bisui- aminoethyl ether)N.Ni\"..\".:etraacetic acid; SDS. sodium dcdecyl sulfate; PA. poly-acrylamide: GE. gel electrophoresis e‘ '~‘ , o‘ _ .' i. if. ‘ l . ' . v - . _ . .‘ .1». l4: ' lIohaL—stu3“. J - 12510 ’5‘ '-("'L'77_~"‘ »*-r~Q’:.~-‘*\fifi p. WWW?“‘—-‘m.~ 3"“. \ -y—uw’u ... .. .‘ .. _ -~,v-_w-- w, ‘.-.‘.1.‘-- .7- . . .. . t, ,‘ -.*--__ -....._. ...-.. ,7 - , ' v i i ’ i ‘ 1“; ,u‘, odds-L.» . h—l - . . . ..-.‘a‘..'-m'-~" l I b ‘. I ... u _-‘i e ‘ ...-AAI-‘V .' .V “ b . - s ‘ . ‘ - _. . n‘. l.:-IL .‘U. N.- 1‘ 4;.31- .-. ...i 1404‘ ".a -~ \h- . ..- 17.4 .A-o .. . o D- " ' - «vaginal. ...‘J. A.k¢‘-|, ton-“i.- . Phosphodiesterase Reconstitution 12511 60,000 cpm/ hatch in a iOOoul volume. After a Sonia incubation. assay suspenmons were filtered on Millipore 0.45 pm type HA filters. These were rinsed with 20 mM Tris-Cl. pH 7.3. 1 mar MgCh. Radioactivity returned on the filter was used to calculate the GTP bound. Back- ground values were determined in may: in which a large excess of unlabeled GTP had been added prior to reaction. All samples had been previously stored at -20 ’C. GTPnse Activity—This was assayed largely as described by God- chaux and Zimmerman (91 Up to 100 pg of prorein in 200 pl was incubated for 5 min at 30 °C in '20 mM Tris—Cl. pH 7.5. 100 mint NaCl. 15 mm '2-mercaptcethanol. 1 mM MgCl: containing 100 list GTP. The reaction WM terminated by boding {or 2 min. Samples were then centrifuged to pellet protem. and a 20-;i1 aliquot of the supernatant was analyzed by high pressure liquid chromatography. GTP and GDP were separated on a Whatman Partisil SAX column with a running buffer of 280 mu NaPO.. pH 6.0. 3 mac 31301.60? was deteczed and quantitated by absorbance at 254 nm. Concnnavalm .«l Treatment—DROS, either unbleached or pre- bleached on ice in room light. were incubated for 30 min at 20 °C in 10 um Tris-CL pH 7.8. 0.1 that 5111012. 150 mu NeCl. and molar ratios of Cor. A.,/rhodopsin up to 12. Some assays also contained 100 mM a- methylmannostde to inhibit Con A binding. All samples were then bleached 5 min before addition of other reconstitution components. and phosphodiesterase aciiv'ty was assayed as described above. .N'-et/':_vlmaleim1de Treatment—DROS. either unbleached or pro bleached. were incubated either for 15 min or 8 h at 20 ’C in 10 mM Tris-Cl. pH 7.8. 10% glycerol. 1 IBM N353. and concentrations oi'NEIM indicated in the legend to Fig. 3. Reactions were terminated by addition of a 5—fold excess of dithiothreitol over NEM. All samples were bleached for 5 min before addition of other reconstitution components. and phosphodiesterase activity was assayed as described above. Trypsm Treatment—Prebleached or unbleached ROS or DROS (80 ,ug) were incubated for 0.5. 3. or 10 min at '20 ‘C in a reaction volume of 167 ml containing 20 pit/mi of try-pain. 10 mm ”Ins-Cl. pH 7.8. 1 ms: MgCLc 10% glycerol. The reaction was terminated or prevented by addit.on of a S-t‘old excess of soybean trypsm inhibitor. All samples were then bleached for 5 min before addition of other reconstitution components. and phosphodieSterase aCLiVIty was as sayea as described above. Rhoa'opsm Conant—This was measured in 1% Emulphogene. 0.1 M Nugou. pH 7.0 Ill). using the molar extinction coefficient 40.600 at 498 um (1:2). Protein-This was determined according to 1.0va er al. 1131. Materials—Electrophoresis reagents were obtained from Bio- Rad. [S-‘HlGTP mus from New Brcland Nuclear, and all other materials ireagent gradel were from Sigma. RESULTS The underljn'ng strategy in these procedures was to recon- stitute light plus GTP-requiring cGMP phosphodiesrerase actiVity by adding membrane-free soluble protein fractions to membmnes depleted of soluble proteins. varying the compo- nents of the soluble fractions. Since full acrivation of ROS phosphodiesrerase requires bmh light and GTP. it has become common pracrice to prebleach the ROS before assay 46. 143. Therefore. light activation was away-ed using prebleached membranes in the presence or absence of GTP. It has also been shown that when soluble proteins are hypotonically removed from the ROS in darkness. they can be reacted in built to the membranes in darkness. using isotonic salts. to achieve full reconstitution of phosphodiesterase and GTPase activities (15). The object of the experiments which follow was. therefore. to (1th the soluble fraction into those pro- tans which were essential to the reconstituted activity and those which were not. Comparison of Phosphodiesterase Reconstitution Using Resolved and Unresolved Components—The first row in Table 1 presents. the specific activity of unwashed ROS With and Without GTP. Row 2 gives the actitity of the ROS after isotonic washing. This procedure did not remove the major soluble proreins P and G. as evidenced by SDS PA geLs and by phosphodiesterase assays. but it might have removed a -r . _ . -r.""“ .fl”.. *s'r'rt' P‘CW‘W, .e—nr—v-c. wan,” 2......- «up-.. TABLE 1 Reconstitution ofphosphoa'iea‘eemse activity with resolved and unresolved soluble components Assays were performed as described under "Experimental Proce- dures" using prebleached DROS and with or Without GTP. Both total acrinty 4cpm/assayi and specific acrivity (moi/min/tnm are included in the table because tne observed effects are more difficult to follow With either column alone. Total Sample and my description 91?: ~67? cor? 4m» -01“? ”.V in; cpm,’aamy Mir-l: min 1mg" 1. ROS 50 240 3166 0.15 1.98 ‘2. ROS iso washed 54 +40 2519 0.16 0.97 3. DROS ldepletedl To 364 490 0.124 0.143 4. DEAE pool - BME" - 10 1088 4.33 "..59 1.03 EDT A (P20 + G) 5. DROS «.70 mg) l- DEAE 80 1464 1915 0.436 0.570 pool - BME - EDTA (10 mg) 5. S-200 P Pool —- BME - 3 680 638 6.48 6.08 EDTA lP:Gl 7. DROS (70 ygi ... 3-200 73 578 769 0.159 0.259 P Pool — BME - EDTA (3 pg) 8. DEAE pool + BME + 13 209 232 0.383 0.425 EDTA (P + Go 9. DROS (7011.5) + DEAE 83 ‘240 4214 0.069 1.22 pool + BME e EDTA 413 pgl ‘ 10. S-200 P Pool + BME + 4 1406 1.148 8.37 8.02 EDTA (P1 11. DROS (70.17;; + S-‘ZOO 74 588 342 0.139 0.271 P pool é BME «- EDTA 44 ug) Total protein. ’ BME. 2-mercaptoethanol. light or GTP-dependent activator of phosphodiesterase.‘ As shown in row 3. the DROS exhibited less than 10": of the GTP-dependent phosphodiesterase activity of the ROS. Most of the P prOtein (from DEAEocellulose in the absence of 2-mercaptoethanol and EDTA (8)) added in the experiment shown in row 4 was still complexed with G. as evidenced by subsequent sucrose velocity sedimentation and gel filtration (8|. its activity was stable but was inhibited by GTP When asayed tn the absence of DROS. When added to the DROS (row 5) its activity was increased only 31‘} by GTP. The unresolved P:G complex derived from S-200 gel filtration of the protein without 2-mercaptoethanol and EDTA was also stable but largely unresponsive to GTP, whether alone 1 row 8) or with the DROS lrow 71. in contrast with the previous observations. the DEAE pool with ‘2-mercaptoethanol and EDTA (row 8) contained physi- cally distinct P and G proteins. as evidenced by subsequent resolution by sucrose velocity sedimentation or gel filtration (8). The specific activity of the pool in the absence of GTP was about 7-t’old lower than that of the corresponding pool derived without 2-mercaptoethanol and EDTA lrcw «H and about Z-fold lower when in the presence of GTP. However. when this pool was added to the depleted pellet ”'0“! 9) activrty increased almost 1.5-fold in the presence of GTP. so that the final activity was twice as high as that in row 5 and was comparable to the activrty of the undepleted ROS «rows 1 and 2). These observations implied that whereas G stabilized the catalytic activity of P when complexed with it. it did not promote reconstitution and may even have prevented it. On the other hand. when G was separable from P but still tn : lsotonic washing did not consistently decrease phosphodiestemse activity cf ROS. in some cases. i: was tncreased (9). ~— wgp‘p;grm.“ v05; — -.7-\_-‘—§_ 3..-.._.q ”-v-v—o -.q- yen-‘Vf r1»"“_' - an- -e-‘.-—~—--.'-—"' ‘ 12512 Phosphodiesterase Reconstitution solution with it. the specific activity of P was markedly lower. but reconstitution was dramatic. The resolved P protein derived from gel filtration with 2- mercaptoethanol and EDTA present (row 10) was more active than most such fractions we have observed and had no: yet destabilized into multiple bands on native gels (8t. Yet despite its high acuvity it failed (row 1'0 to promOte the reconstitution manifested by the tool in row 9 and was actually inhibited by DROS. This comparison underscored the requirement for G. which was present in the systems of rows 5 anti 9 but absent from those of rows 10 and 11. it did not ans“ er the question. however. as to why G fostered reconstitution in row 9 but was either nonparticipating or inhibitory in rows 5 and 7. To pursue the relationship of P to G further. we conducted the following experiments. Reconstitution uztr': Resolved Components—Fig. 1 shows SDS and native PA gels of components relevant to reconsti- tution experiments which follow. The G protein in track 5 had the same SDS staining pattern as that in track 1. However. the native staining patterns were not the same (tracks 6 and 2: see legend to Fig. 1‘2. The P proteinused in these studies (track 4) had the same native gel pattern as the P:G complex (track 8) but a different SDS gel pattern (compare tracks 3 and 7}. Table ll presents the results of an experiment with resolved components only. The essentials of the experiment have been reproduced repeatedly. The activity of resolved P protein was relatively insensitive to GTP but was inhibited by either DROS or G protein, whether CT? was present or not. in the absence of GTP. the separate inhibitory effeczs of DROS and G were additive. G procem also inhibited the residual activity of DROS. unless GTP was present. However, when all com- ponents were present. GTP caused a lO-fcld activation. in Fig. 2 the relative contributions of each component to phosphodiestemse reconStitution are examined. Optimizing these contributions produced activation of up to lOOd'old relative to that of the combined components without GTP. 4 '3 t- (a L.‘ L“: " ’J t i it =1 a 00 2 ' 5 5 323456789 Pro. 1. SDS and native PA gels of? and G frncrlons. Electro- phoresis and isolation were performed as described in :he accompa- nying Jaoer (St. Tracks l and ‘2 are SDS (1) and native (2| geLs of G protein resolved by Sephacryl 3-200 filtration in the presence of 2- meroaptoethanol and EDTA (5). Trucks .3 and 4 are SDS (3* and native (4) gels of P pretein resolved from the same G as in tracks 1 and 2. Tracks 5 throuzh 6 are gels of corresponding proteins prepared in the absence of 2-mercaotouthanol and EDTA: (53 SDS gel of G (the '2 faint bands in the middle of t. e gel are a nonprotein artifact); (6) native gel of G. When this protein was combined With the prorein in track '2. the faint upper band in 2 was noticeaoly darkened: (fl SDS gel of PG complex. emibiting relative depletion of G. (3"; (8t native gel of the P13 complex in track 7; (9) SDS gel of DROS. exhibiting largely r’nodoosm monomer (lower band) and dimer. The top of the gel 15 histrle in all '9 tracks. The thin band at the bottom m l. 5. 1nd 9 15 the dye tront. With whlcn (E. conugratea ‘n l and 5 (5). TABLE II Reconstitution ofphosphodresterase activity with the reached components (tracks In!) and DROS (Duck 9) identified tn F lg. I Assays contained 4 p3 of P. 20 pg of G. 60 pg of prebleached DROS. and 0.5 mu GTP where indicated. Activity is expressed in rpm: assay. All DROS were prebleached. When unbleached DROS were used. the activities obtained were comparable to those below except in the last column, where the activities were close to those seen under DROS alone. l P idioms woaos Gei‘ROSijGlPOGit-DROS —GTP7_t 0:35. tel 341 46 i 496 i 191 *GTPlfsag); 333; 541 .515 5 406 ; 20x4 Maximal activation of the reconstituted system. as in the case of the original ROS, occurred with a broad peak in the range of 20-200 it‘d GTP (Fig. 2a). (Using a pH assay for cGMP hydrolysis. ‘t'ee and Liebman (16) reported half-maximal ac. tivation of phosphodiesterase by subrnicromolar concentra- tions of GTP. with maximal activation occurring in a broad range from 0.3-100 lint.) L'smg G protein resolved With 2. mercaptoethanol and EDTA (Fig. 1. tracks I and 2) activation was proportional to the amount of G procein. relative to P, up to at least a 10-fold excess by truss (20-fold molar excess assuming molecular weight of P to be about twice that of G; Fig. 22), top curt'e). However. G protein resolved without 2- :nercaptoethanol and EDTA (Fig. 1. tracks 5 and 6‘: failed to promote reconstitution (Fig. 2b. bottom curt'ei. When we resolved P from C using Kuhn's procedure (which includes 1 mat dithiothreitol (10)). both P and G prornored full reconsti- tution. reinforcing the belief that they were the same as P and G resolved with 2-mercaptoethanol and EDTA (3). Activation of the reconstituted complex was proportional to the amount of DROS membrane protein up to a 20—fold excess by mass relative to P (Fig. 2c) and to the amount of P protein up to 12.5 pg per 50 pg of membrane protein (Fig. 2:1). in the presence of 100 um GTP. 5 pg of P. 25515: of G. and .30 pg of DROS. specific activity of fully reconstituted P averaged 1.36 :t 0.19 (B.E., N u 8) limol/min/mg of membrane plus soluble protein (68% of the activity of the untreated ROS) or 20.7 t 1.9 umol/min/mg of P protein only. Activity of pure P alone was on the average low but highly variable (3.38 2: 1.82 limol/ min/mg; X a 4). Reconstitution with PC but not Inactive P—Resolved P procein was earlier observed to be unstable conformationally (as judged by appearance of multiple bands on native gels) and catalyticaliy, unless protected by glycerol (.3). Destabilized inactive P net only failed to promOte reconstitution but ap- parently inhibited the residual acmity of other components. For example, in the experimental format of Table II it was found that stable P added to DROS resulted in activity greater than that o!‘ DROS alone. while destabilized P resulted in less activity. When all components including stable P were mixed together in Table 11. activity was about 5-fold greater than that of the DROS alone. Yet when destabilized P was used instead of Stable P. the activity was half that of DROS alone. On the other hand the PzG complex (Fig. 1. tracks 7 and 3) was active in reconsritution. although addition of resolved G was still required (results not shown). This explains the failure of RC: fractions described in Table i to promote reconstitution in the presence of DROS membranes. despite the presence of a form of G. In rows 6 and 7 of Table 7.. the G protein complexed with P apparently afforded protecnon for its cat- alytic activity but could not promote reconstitution. Whether this was due to an alteration in form of G or to insufficient quantity was not determined. However. it is clear that had G resolved with Z-mercaptoetbanol and EDTA (Fig. 1. tracks 1 and 2) been added. reconstitution would have occurred. in ...N .. I. A ... 4.. ,. . . ...u: ,. ‘R a”? . . 51‘qu I . 1 I U ,‘ . a» uni! T. .3 u. .2. .A.. . L (Tl-5‘... J «n.-I t v ' _ ' r ' > ' . an. "n.5,; (r, u. ..'. with“); W_~. . '~‘»'I'v‘ ‘1 7r Phosphodz'esterase Reconstitution r f T ‘ a 2"”. '1 ii so» « o .2 § :4 4 o a: ’ ‘L V T— Y j c J .7: l“ :2: 3.. ° i 8 _ 3oz» J r a r . do 150 :50 259 DROS protein, in FIG. 2. Contributions of resolved components to phospho- diateme activation. P and G were resolved in the presence of 2- mercaptoethsnol and EDTA i8). In a (bottom curve only). b. and c. assays were performed with and uithout P prom-m. The component on the abscissa was varied. To demonstrate the ezi'ect of that com- ponent on activation of P. the difference between phosphodicsterese actiVity with and without P was plotted. Activity a: zero concentration of the abscissa component was defined as 1, and all further sctiwties were normalized to that value. The average of three experiment: 'u rows 4 and 5. G procein was present in excess of the PtG complex. However. this species of G, obtained by hypotonic washing in the absence of ‘2-mercaptoethauol and EDTA, not only did not promote reconstitution but in the presence of GTP and absence of membranes may have inhibited activity. GT Pose and GTP Binding—GTP-binding activity of the ROS was not removed by isotonic washing in the presence of 2-mercaptoethanol and EDTA. but 60% of the activity was removed by hypotonic washing in the presence of 2-mercap- toethanol and EDTA. .\' either isotonic nor hypOtonic super- natants bound GTP, and the depleted activity was not re- stored When either supernatant was recombined With the isotonically washed or protein-depleted pellet. respectively. This raised the question of whether the G protein, a major. Still unresolved component in the hypotonic supernatant. had no residual GTP~binding capacity or had lost it in storage. Godchaux and Zimmerman observed that in the absence of 2- mercaptoethanol accivity was lost within '2. h at 0 °C (9). GTPase acrivity of the ROS was reduced 17% by isotonic washing with 2~mercaptoethmol and EDTA present and only 5% more by hypoconic washing with 2-mercaptoethanol and EDTA present. Both supernatant fractions contained low but measurable GTPase activity. W hen isomnic supernatant and pellet were recombined the resulting GTPase was the sum of their separate activmes while that of the recombined hypo- tonic supernatant and DROS exceeded their sum by 69%. Table III presents GTPase reconstitution data for various soluble fractions upon their recombination with the DROS. Only the resolved P fraction failed to promOte measurable reconstitution of GT Pase. All the other fractions contained G in some amount. and tne resolved G ( row 5) in particular gave the largest increase in reconstituted acuvity. The G protein in the P20 complex was apparently responsible for the GTPase acrivi'ty of that fraction. Attempted Blocking of Phosphodicsterase Reconstitution by Modifying DROS-The most heterogeneous component in “ .v. (. \ «v, we -.. _‘ wry ‘fwv 2 _‘ W“ ‘P'T'F: -- . - .II . . ' v , 12513 1 T I I b ,4. - ‘3 5:0. .. 0 .2 s . ° . a: 6 ”010.99 4 P U V ‘7 V 3 d 1 3: :30“ 1 8 2 2‘1- . '- 0 g [all at 2:5 ‘ 73 4 12:: P promo. #9 platted. In the standard experiment. 4 pg of P. ll) pg of G. 50 pg of DROS. and 0.4 mu GTP were included in the 0.5-ml assay volume except where a component was the independent variable. a. GTP was varied from 0-1 am. M. ROS: H. reconstitution with all components. o. G protein was varied from 0—40 ,ug; G protein resolved With (C--—C) and without (H) 2-mercaptoethanol plus EDTA. c. DROS was vaned from 0-200 pg. d. P protein was varied from 0- 12.5 pig. The plotted value is the difference in phosphodieeterase accivity with and without 0.4 mm GTP. Tau: Ill GTPase activity of the ROS and derivative fractions Washing solutions and procedures are described under “Experi- mental Procedures." as are procedures for determining GTPase activ- ity. Expected values in parentheses are sums of separate activities. " denotes inclusion of ‘2-mercaptoethanol and EDTA and " denotes their onussion. during washing and subsequent DEM-3 and S-200 chromatography (8). GTP”. in pmol of GDP/min/ my Component identity \ctivity Recomtituud ' wi bl ch ‘1” ubags ed 1. ROS 1093 2. "DROS 723 3. "Hypo Super 59 (782) 1322 4. "P (Fig. 1. tracks 3 and 4) 202 (9‘25) 666 5. "G (tracks 1 and 2) 68 (789i 2769 6. “P20 (tracks 7 and 8) 66 (739) 1524 1'. "G (tracks 5 and 97 234 (957) 1414 the recomtitution assay is the ROS membrane system from which most soluble proteins have been depleted. The mem- branes consist predominantly of discs or disc vesicles. which are comprised for the most part of rhodopsin and phospholip- ids (17. 18). SDS PA gels (Fig. 1) indicate other minor con- stituents. as would be expected since the DROS contain residual phosphodiesterase. GTPase. and GTP-binding activ- ities (Tables 11 and III). As a preliminary step in exploring whether reconstitution required binding of P and G proteins to rhodopsin. experiments were conducted attempting to alter possible binding sites on DROS. Concanat'alin A—This has been shown to bind to glycoside chains on the exposed intradiscal surface of rhodopsin (19, 20). On the assumptions that (a) P or G bind to or near these chains and lb) :1 significant proportion of the intradiscal surfaces of the discs are turned inside out by freeze-thawing 0 -.—0'—.—-~4.-'--;~-v7 ., 7-- r. .... —-.-—~ -.~v-- -..--- -— . — 1 ‘il I IJ I‘ll III-ll; lull‘lnt I.II l : |.IFH . iiill . y. _ . .1. o .- . u .“a If? 1.1... .. n.2,... . r '13 _ I ‘1»...- ... ', .‘. “In. H; l . . . ' ... nus...:..;-“J.i..4£.o-.-..u_'.-.v.t.. T Y n e— -—- +--~+-—- -... -.....L- - .....-4..- .... “r““'1'\‘t1"‘*‘ -* * PDE eating a «Hm/can _.-_,,.'P__....._ li:“‘“”r'"‘”“ . L- _.-__ -L... .. Pic. 3. Effect ofN-ethylrnaleimide on reconstitution. DROS were incubated With different concentrations of NEH as described under "Experimental Proceouree." Activity is expressed in epm.'asaa_~.- x 10". Treated DROS. P. G. and CT? were included in the assay. DROS were incubated as fellows: '}--Q. 15 mm in the light using prebleached DROS; 0—-—-O. 15 mm in darkness: [IL—Cl. 3 h in the light using pTEbIEJChed DROS: H 8 h in darkness. Incubations were terminated With excess dithiothreitol. On the extra ordinate a: left are plotted the activities of reconstituted DROS which were treated ‘Jfllh dithiothreitol alone for the same incubation times and light conditions. ‘3. 15 min in :he light using prebleached BEDS 0. 15 min in damned-1;: 5 h in the light using prebleached DROS; il 5 h in darkness T ABLE IV Effect of trypstmzation ' of DROS on Us contribution to phosphodi'esrerase reconstitution Dark or pre bleached mumbers in parentheses) ROS or DROS were incubated in buffer With or mthout trypsin as described under “-Ex perimental Procedures." All assays were performed in the light in the presence or absence of 0.2 mm GTP. 80 ,ug of ROS or DROS. 20,13 of G. and 5 pg of P. Except in rows 9 and 10, activity is expressed in micromoles of GMP formed per min per mg of total protein. In rows 9 and 10 specrfic actin'ty of P in the assay was calcuiated by subtract- ing total activmes of appropriate assays conducted with and without P and then dividing by nulligmn'is of P. For example. in row 7 the activity of 0.14 umol/rrun/mg corresponded to a total actinty of 0.14 umol/min'rng it 0.105 mg - 0.0147 ,umol/rmn. In row 5. hora! activity was calculated to be 0.01 X 0.100 - 0.031 umol/rnin. The difference due to P was 10,014? - owinooos - 2.7-1 umol/min/mg of P. The discrepancy between 2.74 and the 2.90 value in row 9 results from rounding errors in this enunple. Numbers are averages of duplicate experiments. v o“.'—-.,..——-.—.’ *GT? +GTP ‘.. ROS 04310.52) 319313.93) '2. R05 + T“ 20.86 mam 24.47 123.301 3. DROS 0.03 (0.02) 0.23 (0.25) 4. DROS + T 0.58 (0.60) 05610.81) 5. DROS -r G 0.01 10.02: 032610.291 6. DROS + T + G )191016) 0.31 -0.30) 7. DROS + G + P 0.14 40.13; 3.691335» 8. DROS + T + G + P 0.251027: 5.651454) 9. P «7.51 29012.33) 76.86 {68.6.31 10. Pr '6-6) 1.5-I 1.2.55: 1391019939) 11. P 5.1-1 5.14 12. G 002 0.02 “ Trypsin oreincunated With soybean trypsin inhibitor had no effect or; phUSphO¢fiSIEfL~'-2. ” T. trypsm. G. G protein; P. P procein. and/or hypotonic swelling. one would predict that Con A would prevent or diminish reconstitution. Regardless of whether DROS were prebleached or kept in darkness. Con A in concentrations up to a 1‘3-fold molar excesxa relative to rhodopsin had no significant effect on reconstitution. This _. H-,-'._ _..~_ ..n. "-'--.‘-~‘-’w"p"M-”“' .-- —.I -W—._ m,-—.".¢ q .-m-’—~-‘ a - ~ Phosphodiestcrase Reconstitution suggests some other site of interaction between DROS and the soluble components than the Con A-binding site of rho- dopsin. N-eth) lmaleilnide—Tliis reacts with sulfhydryl groups of proteins. Rhodopsin sulfnydryl exposure is increased by bleaching 121-23). although the precise Stoichiometry is still at issue. When we incubated unbleached DROS for 15 min at ‘25 ’C. using up to a 400-t'old excess (1 mat) of NEM relative to rhodopsin. there was no appreciable efiect on reconstitu- tion. However. when the incubation time was increased to 8 h there were inhibitory effects (Fig. 3). At the highest concen- trauon of NEM. reconstitution with unbleached DROS was inhibited 48%, compared to 25% with prebleached. In another experiment lnot shown! inhibition was even greater (about 50%). but the bleached DROS exhibited more inhibition than the unbleached. Incubation of DROS with a {CO-fold excess of dithiothreitol for 15 min had no appreciable effect on reconstitution. but again. an 8-h incubation markedly in- hibited reconstitution with both unbleached and prebleacheo DROS. Incubation for 8 h with neither NEM nor dithiothreo itol had no effect on reconstituted phosphodiestc-rase activity tsee zero concentration points on internal ordinate of Fig. 3). The inference to be made from these experiments is that sult‘hydryls may be involved in interaction of the soluble components with the membrane but probably not those which are exposed by Light. Tip-pain Treatm exit—This has been shown to increase phos- phodiesterase activity of the ROS ('24) and of the soluble P protein (25). We sought to determine the impact of trypsini- zation of the DROS on its contribution to phosphodiesterase reconstitution. In Table IV the data in rows 2 and 4 relative to 1 and 3 reveal that trypsin increased activity of both ROS and DROS. In general. this increase was not afi'ected by prebleaching the membranes. GTP had little effect on the activity of trypsinized membranes. The untrypsinized DROS (row 3: exhibited 6"? of the activity of the ROS (row I). The trypsinized DROS (row 4) exhibited less than 3% of the activity of the trypsinized ROS (row 2). Addition of G protein (row 5 relative to 3) did not have a large effect on the activity of DROS. whether CT? was present or not. After trypsinization of DROS. addition of G appears to have inhibited aetivity l row 6 relanve to 4). both in the presence and absence of GTP. Prebleaching of the DROS had no appreciable effect. Rows 7 and 8 represent the standard reconstitucion system. using untrypsinized and crypsinized DROS, respectively. Again. prebleaching had no appreciable effect. GTP activation was on the order of 20-fold. Trypsinization approximately doubled activity in the absence of GTP and induced about 50‘} increase of actwtty in the presence of GTP. In rows 9 and 10, the activation attributable to P was not noticeably dependent upon prebleaching of the DROS but was markedly GTP dependent. However. the GTP depend- ence of P reconStituted with the trypsmized DROS was double that with the untrypsmized DROS c'55«fold versus '26-fold activation by GTP). Ca“'-EGTA-Ca"‘ and EGTA were examined for their effect on phosphodiesterase activity of the ROS, pure P. and the reconstituted system. Neither Ca" nor EGTA alone at concentrations up to 1 in.“ affected phosphodieSterase activ- ity. However. an effect whicri we attribute to a Ca""EGT.X complex was observed. reSulting in almost complete inhibition of phosphodieSte-rase activity in reconstitution. in the soluble P protein. and in the ROS. This effec'. is under investigation. DISCL‘SSIO!‘ Reconstitution of light-dependent phosphodiesterase acnv- ~4"‘~"‘\"‘.$ “'M”.‘Qwu M‘VW . . l ’..- I, . V V v . ‘ , ' ~.| ‘. _ l. I «.41.». .’ul-A....~ i’- or...“ .v . . u. ‘1 ' i u ' - . ... “A “A‘s... Phosphodies (erase Reconstitution it); has been shown to require catalytic (P) and regulatory (G) proteins, GTP. and ROS membranes. in the absence of any single component. the remaining components continued to interact, though characterisrically in an inhibitory manner. as exhibited in Table II About 659': of the activity of the untreated ROS could be rec0vered upon reconstitution. but this represented more than 100% of the activity of the isotonically washed ROS. The discrepancv between the reconstituted DROS and the un- treated ROS may have resulted from removal of an uniden- tified regulator in the isotonic wash or from the presence of nonfunctional forms of P or G which when used in reconsti- tution could still bind each other or the membrane in an inhibitory manner. This may explain why our purified P (8) demonstrated lower activi " than that reported by Baehr et cal (25). Since the DROS were not entirely free of residual P or G proteins. the question arises whether the observed inhibition of phosphodiesterase activity by DROS when added to P was due to residual G or to some other residual component of the DROS. We estimate that there was at most 5 pg and probably less than 1 ug of G in the DROS. as compared with ‘20 ug of pure G added in the three righthand columns of Table II- Yet addition of DROS to P resulted in a greater inhibition than addition of G to P. We. therefore. consider it unlikely that inhibition of activity by DROS was due merely to residual G. A second question which focuses on the properties of the DROS is the reason for increased GTP dependence of the reconstituted DROS following trypsinization (Table 1V). Once again. the effect may be due to residual P or G in the DROS. but the possibility of some ether unique trypsin-enhanced component of the DROS must be considered. L'nequivocal answers to either of these questions probably require recon- stitution of DROS from lipids and purified membrane pro- terns. Another uncertainty is the physiological ratios of P and G to each other and to rhodopsin. Baehr et al. (25) estimated the ratio of P to rhodopsin in aim to be between U40 and 1/ 170 on a moi/moi basis. in Fig. 2c, we showed that the requirement for DROS approached an asymptote at about 100 .ug of DROS prorein per 4 pg of P protein. Assuming that 85 pg of this represents rhodopsin. that the M, of rhodopsin is 39,000. and that the M. of P is 185,000, the calculated ratio of P to rhodopsin was 1 to 100. midrange between the limits proposed by Baehr et ai. (‘25). The physiological P to G ratio is more difficult to determine. Based on removable protein recovered from hypOtonic washes. there was approximately twice as much G pretein as P protein estimated by SDS PAGE staining densities (data not shown}. implying a molar ratio of 4 G to l P. Reconstituted activity was proportional to added G up to a 20-fold molar excess of G over P. Interpretations include the possibility that the ROS phosphodiesterase operates at submatimal concen- trations of G, that the proportion of G involved in activation of P depends upon the amount of rhodopsin bleached, or that the isolated G protein included a significant proportion which had lost the capacity to promote reconstitution. While these alternatives are neither exhaustive nor mutually exclusive. the third one is consistent With our observation of apparent mul- tiple forms of G. 12515 of the phosphodiesteraae activity of the intact ROS. Of particular interest is the function of the G protein. While it apparently regulates catalytic activity of P. the manner in which it accomplishes this has not yet been clarified. The five principal functions which have been ascribed to G are GTP binding. GTP-G DP exchange. GTPase activity. GDP binding. and phosphodiesterase regulation (4. 9, 26—30). Again. these are not necessarily distinCt nor exhaustive. In addition, the functions may well be separable as proposed by Somers and Shichi (25) in the case of GTPase and GTP binding. Shino- zawa et al. (27) have reported a G protein in frog ROS which binds GTP and activates phosphodiesterase but is distinct fi’om an H protein which is required for their G to demonstrate GTPase activity. Robinson and Hagins reported differentia- tion of GTPase activity from both phosphodiesterase regula- tion and GDP binding (‘29, 30). In all cases that we know of, GTPase activity has been reported for the soluble protein only in the presence of the ROS membrane, which was true of our own observations (Table III). Our data also suggest se- parability of GTPase and GTP binding, since we were unable to detect GTP binding in reconstituted DROS which dis- played GTPase activity. In addition. only one of the three G forms that we observed promoted reconstitution of both GTPase and phosphodiesterase activities. Separability of ac- tivities ascribed to G may be related to different conformations of the protein or to variations in the interactions of its 3 subunits (8). Although it has been amply demonstrated that GTP hy- drolysis is not required for activation of phosphodiesterase (16, 37). the precise mec..anism by which GTP facilitates phosphodiesterase activation remains to be explained. Liebo man and Pugh have proposed that the phosphodiesterase catalytic protein (they call it D) is activated by bleached rhodopsin and then binds GTP and releases rhodopsin (31). Deacrivation occurs in two steps according to their scheme. The bleached rhodopsin is phosphorylated by ATP, and GTPase releases bound GTP from the catalytic protein. In contrast. Fun; et al. propose that the initial activation step entails binding of GTP by the regulatory protein (they call it “transducin") when catalyzed by bleached rhodopsin (32). In their scheme the GTP-regulator complex then activates phos- phodiesterase by removal of a small inhibitory polypeptide from the remaining subunits of the catalytic protein (33). We do not see these models as incompatible. since the multi subunit PzG complex described in the present and accompa- nying (8) reports may allow multiple modes of interaction between the subunits of P and of G and between elements of the P20 complex and the ROS membrane. We believe that analysis of these different modes requires resolution of the subunits of both P and G and characterization of their struc- ture and interaction. Acknowledgments—We gratefully acknowledge the assistance of Kathleen Clark in collection of the data for the top curve in Fig. 20 and critical reading of earlier drafts of the manuscript by Christine CoUins. Arnold Revzin, James H. Asher, Jr.. J. L. Farley. and C. H. Sueiter. REFERENCES . Woodmfi', M. L. Bownds. D.. Green. S. H. Morris. J. l... and Shedlovsky, A. (1977': J. Gen. Physiol. 69. 667-679 . Fletcher. R. T.. and Chader. G. J. (1976) Biochem. Biophys. Res. pa '9 Despite these uncertainties the reconstituted system closely resembled the untreated ROS in irnportant respects. Both 3- had essentially absolute requirements for P. G, GTP, and light (not shown; see caption to Table II). The optimal concentra- tion range for GTP (Fig. 2a) and the degree of activation (10— Commun. 64. 1297-1302 Gondis. C., Urban. P. E., and Mandel. P. (1977) Exp. Eye Res. 24. l7l-l77 4. Kilhn. H. (1980) Nature 283. 587-589 . Godchaux. W.. “I. and Zimmerman. W. F. (1979) Exp. Eye Res. ..icg§7_l-:,-W~I" "w wi" .,« V—vn, ~—-<~“§._— “...-- . . . - -‘ . . , . . . . , L’h. ..'.‘.;- $qu :- 5.1“. I 9L... ..- ,. I I...' -. . lKu-‘Q\- J b “a. I . I . .WL. I'LL“ I / . M '-- ..a‘ mn-“.—wm -v< 12516 Mild. N., Buaban. J. 34.. Keims. J. J., Boyce. J. J., and Bntensk'r. M. W. (1975) J. Biol. Chem. 250. 3320-6327 . Kohnken. R. E., Eadie. D. M. Rewin. A., and McConnell. D. G. (1981) 256. 12501-12509 . Godchaux. W . II}. and Zimmerman. W. F. {1979) J. Biol Chem. 254. 787+- 7884 . Kilian. H. (1980! .‘v'eurochem. Int. 1. 269.235 . McConnell. D. G.. Dangle-r. D. .3... Eadie. D. M. and Litman. B. J. 11931) J. BioL Chem. 356. 4913-4918 1 Wald. (3.. and Brown. P. K. (1953) J. Gen. Physiol. 37. ESQ—2m 1 . Lcwry. O. H., Rosebrough. N. J.. Farr. A. L. and Randal]. R J. 119513 J. Biol. Chem. 193. 265—275 V . Wheaten G. .... and Bitensicy. M. W. I1977) Proc. Natl. Acad. SCL 1". S. A. 74. 4238-4242 . Bitensky, M. W.. .‘vlflu. N. Keims. J. J., Ke'u'ns. Mu Buaban. J.. Freeman. J.. Wheel-er. M. A.. Lacy. J.. and Marcus. F. R- (1975) Adv. Cyclzc Nuclroude Res. 5. 213-240 Yee. R.. and Liebman. P. A. (1973) J. Baal. Chem. 254. 6902—8909 . Daemen. F. J. M. (1973) Biochim. Biophys. Acta 300. 255486 15. Neilsen. N C., Fleischer. S.. and McConnell. D. G. (1970) ....--~ ... ,_.-,. _.... ~. Bloc/tun. Biophys. Acta 211. 10-19 ' . Molday. R. 8.. and Molday. L L. 41979) J. Biol. Chem. 2.“. 4653- 4660 ' . Adams. A. J., Tanaka. 51.. and Shichi, H. (1975) Exp. Eye Res. 27. 535-605 _ ~-,—-~-..>- -— Phosphoa'iesterase Reconstitution . Fung. B. K--K.. and Hubbell. W. I... .1978) chhemisrn' 17. 4396- 4410 Rem C. M.. deGn'p, W. J., Daemen. F. J. 3.1.. and Boating. S. L (1978) .Biochun. Biophyx Acta 537, 145-152 McDowell. J. H., Mu. M. T., Griffith. K. D.. and Hugnve. P. A. (1979} Visa»: Res. 19. 1143-1145 . Bicensky. M. W. Wheeler. C. L. Alan. B. Vetury. 8.. and Mama. '1'. (1978) Adv. Cyclic Nucleottde Res. 9. 553-572! . Baehr, \\'., Devlin. M. J.. and Applebury. M. L. 11979) J. Biol. Chem. 254. 11669-11677 . Fung, B. K.-K.. and Stryer. L. (1980) Pmc. Natl. Acad. Sci. U. S. A. 77. W233 .. Shinouwa. T.. Uchida. S.. Martin. E., Cafiso. D.. Hubbell. W.. and Bitensky, M. i1930) Proc. Natl. Acad. Sci. D. S A. 77, 1408—1“! ' “ . Semen. R I... and Shichi, H. (1979) Biochem. Biophys. Res. Commun. 89, 479-480 . Robinson. W. E., and Hagins, W. A. (1979) Photochem. Photo'oml. 29. 693 . Robmson. W. E., and Hagins, W. A. «1979) Nature 280. 398-400 . Liebman. P. A.. and Pugh. E. N. Jr. 41980) Nature 287. 734-736 . Fung. B. K.-K.. Hurley. J. B.. and Stryer. L. (IBEX) Proc. Natl. Acad. Sci. U. S. A. 78. 152-156 . Hurley. J. 8.. and Stryer. L. (1981: Bwphys. J. 33. 208.1 (T-PMo C2) #1:};‘4: '5} _ r7 ‘ ‘J ‘1‘" 1.. ‘3' ...—\ ‘ ‘ «rt-Q.‘, T‘kri) I _~‘ A: LV‘ -.f' «2?:- ‘. \\ . r' ”i - wait ‘( (I; lowing“: . u-sx‘ggw ‘ p ' “1}.;(:~!-r - {2):- ‘ . v. . .': _ '. CquoJ-.. . .'..’ 'n‘ t ' t u -l t . "I ' z_.‘ ‘1 . ... . .. ‘. W.--‘«- . .- °.‘ _‘ 2 ' - . - '.' ' ‘. ". . _ u’. '| n - l " .' . ' . . I . -.' ,. s ' ' I" 'h - I I .V " n 1'... ’ .. . . . . . .-Mpo.~r..l' A-‘\~L .o ... . .. H4 ." .u. l . It .115 l .1 . . . ‘ ‘ and». Angus; .--~—, 'V' “‘|\‘.—4\-v"‘_ ' r 'm. wer-w~oo.~'ua—ar fim w‘w..a1< -. “I Janus). or Blctomcu. anumfin Val. 256. No. 23. in.“ at 10. pp. 12517-1233. 1961 Prom-d an U. Sol. Calmodulin in Bovine Rod Outer Segments“ (Received for publication. April 20. 1981) Russell E. Kohnkeni, James G. Chafouleas§. Deborah M. Eadielz. Anthony R. Means§. and David G. Mc Connellfl From the tDeparrment of Biochemistry, Michigan State University, East Lansing, Michigan 48824 and the §Department of Cell Biology, Baylor College of Medicine. Houston. Texas 77030 Calmodulin (CaM‘) has been determined to be present in bovine retinal rod outer segments (ROS) at a level of about 450 ng of Cal»! per mg of ROS protein by radioim- munoassay (RlA). This was about half the level de- teCted in a membranous retinal inner segment fraction. The ROS membranes contained two populations of high affinity binding sites for l"‘l-labeled Cu}! (K. = 0.2 x 10" M and =3.0 x 10”” M) with approximately equal semi-abilities. The ROS and two other membranous retinal fractions contained comparable numbers of ”‘1- CuM-blnding sites per mg of membrane protein. The binding of u“l—CaM to ROS was time and temperature dependent. Binding was also Ca” dependent and CaM specific since bomologs such as troponin C or par-val- bumin failed to displace 125loCch from ROS. The number ofCaM-binding sites on ROS membranes could account for 64% of the total CaM content of ROS determined by RIA. CnM does not stimulate the ROS phosphodiesterase which requires light plus GTP for activation. and the ratio of Cu.“ to pbospbodlesterase is very low compared to known Cam-regulated phos- phodiesterases. It seems unlikely. therefore. that the phospbodiesternse represents a CaMubinding protein present in the soluble fraction of ROS. Similarly rho- dopsin appears to be an unlikely candidate for a CaM- binding protein associated with the ROS. since it was calculated that there was 1100 rhodopsin molecules per CaM-binding site. Despite these arguments against its binding to rho- dopsin or to phospbodlesterase. CaM may play an im- portant role in the regulation of Ca"—sensitive proc- esses in the ROS. Understanding of that role requires identification of the membrane and soluble proteins that serve as specific Cali-binding proteins. The molecular events required for visual excitation are not well understood- It has been calculated that between the absorption ofthe initial photon and the electrical events which follow. an amplification on the order of lOt-t'old mus: occur in the photoreceptor cell. Most of this amplification apparently occurs in the outer segment (1). 1n the case of rod outer segments there is no observable continuity between the rho- dopsin-containing discs in the interior of the ROS‘ and the ‘ This research was supported by United States Public Health Sen'ice Gram EY 1574. by funds from the Micnigan State Universxty College of Osteopathic Medicine, and from private sources. The costs of publication of this article were defrayed in part by the payment of page cnarges. This article must therefore be hereby marked “adver- tisement" in accordance with 18 L'.S.C. Secrion 1734 solely to indicate this fact. i: "To whom reprint requests should be addressed. ' The abbreviations used are: ROS. retinal rod outer segment: CaM. calmodulin. EGTA. ethylene glycol bislB-ummoetnyl ether)- plasmalenuna. On the basis of electrophysiologicsl observa- tions (1-3). Hagins (4) proposed that Ca", released from the discs by light. travels rapidly through the cytoplasm to plug Na' channels in the plasmalemma. A great many experiments designed to test the calcium theory have met with mixed results l for reviews see Refs. 5 and 6). An alternative proposal implicating cGMP in transduction has evolved. since in the vertebrate ROS. cGMP levels rapidly decline in the light (7-10). Activation of the phosphodiWerase responsible for the hydrolysis of cGMP requires both GTP and contact with bleached rhodopsin (ll. 12). Although GTP hydrolysis is not required for phosphodiesrerase activation by light, the ROS also contain a GTPase which appears to be light activated (11, 13). Thus both cGMP and GTP have been proposed as critical agents in the amplification sequence be- tween ROS discs and the plasmalemma {12, 14-20). though the proposed linkage is neither as direct nor as explicit as in the case of Ca" (4). An obvious accommodation between the two theories could be achieved if it could be shown that the light-activated phosphodiesterase of the ROS was under regulatory control by calmodufin as has been demonstrated for brain phospho- diesterase (21). Liu and colleagues have reported that bovine ROS contained an accivator of brain phosphodiesterase (22) and that in dogs with inherited rod cone :lysplasm the actio vator appeared to be absent (‘23). In the present report we demonstrate the presence of CaM in purified ROS by radioim- munoassay but show that this Ca“-binding protein does not appear to regulate the COM? phosphodiesterase. However. specific binding sites for CaM are present on membranes prepared from purified ROS. The nature and Specd'icity of these binding sites have been investigated. MATERIALS AND METHODS ROS Isolation Fresh cattle eyes were obtained from Murco. Inc. of Plainwcll. Michigan and were enucleated While facial muscles were still twitch- ing. ROS were isolated essentially as described earlier (‘2-1). using continuous or discontinuous sucrose gradients and centrifugation st 100.010 x g. For the preparations described in this report. however. 0.15 5: KCl was present in all solutions. and ROS were separated from retinas by shaking in the medium as opposed to homogenization. All operations were camed out in dim red light and at 4 ’C unless otherwise indicated. Rhodopsm was determined as described by McConnell er al. 1‘35). The purity ratio of the ROS Lam/Am.) was 2.65 in 1% Emuiphocene. 0.1 M .‘ngOl-l. pH 7.0. Protein was deter- mined by the method of Lown' er al. (‘26). in addition to the ROS. other retinal frucuons were prepared and examined as indicated in Tsole l. .\'.N..\"..\'-tetraacetic acid; RIA. radioimmunoassay: BSA. bovine serum albumin: HE. hypotonic medium containing ESTA; DROS. ROS membranes depleted of soluble proteins. 12517 ." ».-.- ‘9‘ ...--- 77,.» “..--,v'. . ,M—- n. .....- ,.~'...-,.--I:no~ ‘m. “._ ,?-—,-—9 fl .‘_.',— 7- 'n .. .o . O s v p ' n'l‘ , .81. "oh. n-a. . 1.-..nrdf—Q—v- - v. 7"“. 12518 808 Washing ROS obtained from the sucrose gradient were. diluted and tiger ously shaken in ‘20 volumes of one of the following solutions: K. 10 mm Tris-Cl. pH 7.5. 0.15 M KCl: KS. 10 am Tris-Cl. 0.15 M KCl. 1 mm ethylene glycol biso’l-aminoethyl ether)NuV'uV"..V-tetruacetic acid; HE. 10 mu Tris-Cl. 1 mar EGTA; or H. 10 mM Tris-Cl only. The suspensions were centrifuged at 90.1120 x g for 90 min. Supernatant fractions were lyophilired and resuspended in H10 to the volume of the original ROS aliquot. Pellets were bleached and resuspended in the washing medium by homogenization at the original volumes. CaM content in resulting moles was determined by radioimmunoassay. Pellets from HE winning were aLso used for “’l-CaM-bindlng studies following dilution to 4 nag/ml of ROS protein with eirher id) 10 mu TraoCl. pH 7.5. 0.15 M K 1. l nut MgClz. and 0.5 ms: CaCl; for kinetic and equilibrium assays: or (b) 10 ms! Tris-Cl. pl-l 7.5. 0.15 \| KC). 1 ms: EGTA. and 5 pig/ml of bovine serum albunun for binding assays. Light Microscopy ROS or washed pellem were examined Without dilution on a Leitz Orthopiun large field microscope using a halogen light source and phase contrast. Micrograph were taken a: 80-123 tunes magnification using Kodak TX-135 black and White film (ASA 400). Preparation of Fractions for R IA Membranous fractions were (a) sonicated for 30 s at a setting of 50 high on a Brownwill Biosonik 1V sonifier and brought to 0.1% Triton 31-100. 6 mm berets. pH 8.4. and 1 mM EGTA; (b) resonicatet‘. in the same medium Both soluble and membranous samples were then diluted mm the medium to about 6L0 .191 ml of protein concentra:ion. heated at 90 ’C for 5 min. rapidly cwled. and centrifuged at 10,000 x g {or 30 mm. Supernatant solutions were then subjected to RIA by the method of Chal'ouleas et al. (‘27). Rat testis C334 was purified according to Dedmun er a! 125) and iodinated according to Chafoulcas ct al. (‘37). The “"ll-CaM contained 1.5 mol of ‘nlfimol 01 protein and retained mater than 95% bloactinty as determined by stimulation of bovine brain phosphcdiestersse. AnuoCaM was purified from sheep serum by affiruty chromatography (28). Association Assays HIE-washed ROS (0.25 mtg/ml) were incubated at ‘25 “C an a solution containing 3.67 an '“l-CaM "50.000 cpm/ug). 10 mM Tris-CL pH 7.5. 150 mm KCl. 1 nut Mng. 0.5 mM CaCln. (Additional assays were carried out in CaM/Tris~C1/KC1 as above. no Mg“. 5 mtg/ml of BSA. 1 mm EGTA. and 1 mm CaCli (15 11.x free Ca“) with identical results.) The reaction was initiated by addition of ‘wl-CaM. After indicated time intervals. 50-ul aliquots were withdrawn and filtered on OAS-um Millipore filters that hat: been soaked in assay buffer plus 5 mc/ml 0! BSA. The filters were dried by suction. rinsed With assay buffer. retitled. and counted. Control assays with 100mm excess of unlabeled CaM were similarly treated. Radioactivity bound to filters in the control assays did not vary with time and was subtracted from the experimental counts to give specifically bound CaM. Dissociation .‘Lways HES-washed ROS at 0.25 mg/ml were incubated for 35 min in the association assay medium to bring CaM binding to equilibrium. A 100-l'old excess of unlabeled CaM was added. and 50-;11 aliquots were removed at tuned intervals thereafter and treated as in the assocmtion assays. Control assays were analogous to those performed for associ- ation assays Equilibrium Bindrr g Assays HIE-washed ROS at 0.2 :ng/rni were incubated at 25 ’C ‘J'l the association assay medium containing concentrations of "ul-CaM (20.000 cpm/ n2) between 0.2 and 15 nM in a 100-111 volume. At 25 min assays were filtered as before and counted. Control assays were performed as above but in IOO-fold excess of unlabeled Calvl at each ”ii-Ca?“ concentration. Cation Dependence (a) Subsntution-HE-washed ROS were diluted to 4 Int/ml with 0.15 M KCl. 10 mm Tris-Cl. pH 7.3. 1 mM EGTA. 5 mg/ml of BSA. Twenty .13 of ROS. 3 11M ‘7‘l-Ca31 (30.000 cpm/ng), and Ca". Mr", .Vln'l', or 20” were then mixed in .1 total volume of 100 ill. With KC‘. Tris-Cl. EGTA, and BSA as above. After 25 min at 25 ”C assay - v--.« “hm! ~—~" "W‘. cru'w ~.V'V'--'. r1- Wr- .~~-4-- .. - ROS Calmoduiin . ‘ tures were filtered and radioactivity determined. Nonspecific binding was approximared as before lb) Competition—Mg“ was examined for its ability to compete with Ca“ in promoting "“I-CaM binding. A free Ca“ concentration (0.3 um) was chosen which yielded 50"; maximal binding of ”"l-CaM. Mg“ was added to various concentrations. and the assay was con- ducted as for substitution. Protein Competition HE-wrshed ROS at 0.2 mg/ml were incubated at '25 °C in 100 rd of the cation dependence medium which contained in addition to 3 nM ““l-CaM (30,000 cpm/uz) and 1.5 it.“ free Ca“. varying concentrations of troponin C. pmalbumin. or unlabeled C 3M. After 25 mm assays were filtered. and radioactivity was determined as before. Temperature Dependence Association and disociation assays were performed as described above at ‘2 ’C and at 24 “C. Phosphodiesterase Assay Phosphocfiesterase may!» were performed in the light. sith or Without GTP, using ROS. soluble protein components. and ROS membranes depleted of soluble proteins and prebleached before assay. The basic assay procedure and the assay of phc-Spbodiesterase recon- stituted from components are described elsewhere (‘29. 30). Chemicals Reagent grade chemicals were obtained from Sims except as noted. Pan-albumin and troponin C were gifts of Dr. James 0. Porter. RESULTS Radioimmun oassa y ofRetz'ncl Fractions—Table 1 presents the measured CaM content of the purified ROS and of other retinal fractions recovered from the sucrose gradient em- ployed to purify the crude ROS. Fifty-six per cent of the total protein and 60‘5- of the tatal CaM in the crude ROS were recovered in the gradient pellet. which contained melanin granules from the pigment epithelium underneath mitochon- dria and other organelles derived in large part from inner segments Organelles from internal retinal layers are mini- mized by preparing ROS by shaking as opposed to homoge- nization (31). The next highest recovery (9% of the protein. TABLE l Ca.” content or'ROS preparation and derived fractions Crude ROS prepared horn 300 retinas were layered on dlSCOX'lan- uous 1.10-1.15 g/ml sucrose demity gradients in a Becki-nan SW 25.1 rotor (24). Freedom 2-7 are layers or bands from those gradients after ccntrir'ugation for 1.5 h at 100.000 x g I24). The ROS were washed as described under "Materials and Methods." and the result- mg pellets and supernatant fluids were assayed for CaM by RlA as described under "Materials and Methods." .“5' CAM; Fraction Protein CaM mg pro- (em ’7‘! M 1. Crude R05 1454 1284 0.88 '2. 1.10 g,’:nl layer 133 101 0.76 3. R05 162 7 0.45 4. intermediate band 1 109 50 0.46 5. intermediate band ‘2 9‘. 49 0.5-4 6. Intermediate band 3 69 54 0.73 7. Gradient pellet 334 776 0.93 8. ROS hypotonic supernatant 3 2 5.3 1.68 9. R05 hypotonic pellet 12.1 9.1 0.75 10. R05 isotonic supernatant 2.0 7.2 3.60 11. ROS isotonic pellet 12.7 6.1 0.48 12. ROS hypotonic v- EGTA superna- 3.2 2.8 0 58 1.1111. 13. ROS hypotonic + EGTA pellet 11.5 1 b 0.15 14. ROS isotonic + EGTA supernatant 2.0 6.7 3.35 1'. ROS isotonic * EGTA pellet 13.1 ‘ ' 0.1'3 .v-r-"p.’_.‘-rwor’-—~..--_v,-—.flr,—— .r‘~-—— ,7. . - - -~r u..- . . I . . ALL—l.a.oAL—-—’J ...-... A“; — l I. 1"1‘ l . l, ' .._‘44J-.!¢-.4.~‘.-\‘ «Hall. ROS Calmoa'ulin 8% of the CaM) was found in the clear soluble protein remain- ing at the top of the gradient. Eleven per cent of the crude ROS protein but only 6% of the CaM. was recovered in the ROS. The remaining fractions on the gradient (fractions 4-6) were mixtures of ROS and inner segment membranes. In these fractions there was a direct correlation between the content of inner segment membranes and CaM (gig/mg ot' prorein). The ROS contained 0 45 pg of CaM per mg of protein. Washing the ROS with hypotonic solution released about 20"? of the total protein but a larger proportion of the Cam. Of the protein remaining in the ROS it has been reported that about 85% is rhodopsin (32-&4'i. Using this figure and the measured CaM content. the molar ratio of rhodopsin to CaM in the ROS can be calculated to be about 700 to 1. Depending upon conditions, between 37 and 30% of re. covered CaM could be removed from the ROS by washing. with the highest proportion being removed by isotonic KCl. 1 mat EGTA. Light micrographs of the ROS and pellets derived by washing the ROS in two of the media are shown in Fig. l. The ROS consisted predominantly of rod-shaped structures. 12519 Washing in isotonic KCl converted most ROS to ball-shaped forms probably by breaking the plasmalemma. Hypotonic washing left clumps of amorphous material which can be seen at the electron microscopic level to consist primarily of dis- rupted discs (35). EGTA accentuated the effects of washing in both cases. Assocmtion Studies of 1”I-CaM Binding by Washed ROS—ROS washed hypotonically in the presence of i met EGTA (Hvaashed ROS) were incubated with ‘fiI-labeled CaM for various times up to 1‘30 min. At the end ot‘incu'oation samples were filtered and the filters counted for retained ”51. The raw data. after conversion ot‘cpm/rnl to ma CaM. appear in Fig. 2. Equilibrium binding was observed by '20 min. .5. double reciprocal plot of these data appears as an inset in Fig. ‘2 and demonstrates that the approach to equilibrium was essentially first order. The association rate constant (A2.) was determined to be 0.0436 : 0.0044 (5.0.) nM" min", based on the average of 3 experiments (36). In these experiments [303.]. the initial concentration of unoccupied CaM-binding sites on the ROS, was estimated by RIA to be 5.0 mt. There- _—“i 7 _ fore. the initial mf-CaM concentration. [nil-C3310]. was ex- - i 0 ~ if; perimentally set at 3.67 mat. From these same data. an estimate ‘ ‘ ' ' " ' of 2.89 x 10’” M was determined for the equilibrium dissocia- tion constant or binding constant K4. with 1.9 x 10'9 at Cal“ bound at equilibrium Dissociation Studies-After incubation of labeled CaM and ROS had been allowed to reach equilibrium. a IOO-fold excess of unlabeled CaM was added. and aliquots were filtered at various times thereafter. in 4 separate experiments a plot of In CaM bound [CaMb] versus time appeared nonlinear (Fig. 3). and this was confirmed by an analysis of variance indicating a significant departure from linearity (' p < 0.01). Thus a single homogeneous population of binding sites did not adequately describe the dissociation data. However, an adequate fit was obtained by a function which assumed ‘2 separate classes of binding sites (36). For 5.67 nM CaM and 8 nM initially unoc- cupied ROS binding sites the maximal CaM bound was 1.087 on at the first site and 0.976 m at the second. For 3.67 mt Ca.“ and 5 nM initially unoccupied sites, the maximal CaM bound was 0.768 nM and 0.867 nit. respectively. The standard deviation was calculated at 15%. The dissociation rate con- stants derived for the two sites were 0.120 min" and 0.0084 min" with Standard deviations of about 209%. These values . i ' . a ., -..ot‘n.‘ \ ...___L.._.-‘. ___4.__i ~< l- _ ...- _ .._L ... v. D r- A‘L-'| . on, ‘ ‘, j! u .35 p __l ... - ..a.’ ‘ 3 1 be». J A3: -_ Half" . 1 vf-¢« 33—406 .1"; {43; (fl l d ,' 4 i .‘ .. : I -IJ~:‘A.I-..-'tt. } . ' ‘ V . '~ '. 5 6 6 if“... (mail a: s ' ' I " . Jr (b ,r’fl'oluc .. "ffif’ut’fi' o -- . ' ‘ «‘v . a 4 .14. I‘éb“ ,3.- 4‘ l FIG. .. Assoc-much of Ca.“ With HES-washed R03. Ending ‘-4‘ '5“ 4‘“ " I ‘ ‘ was agave-d as described under "Materials and Methods." Data are ‘ FIG. 1. Light micrographs of fresh (1:. x 20m. isotomcany averages (:S.E.) from .3 experiments. Inset. assomation data plotted - ' l : washed (K. x 300). and hypotonically washed (H. x 300) as double rc-Ciprocals. The line was fitted by least squares linear 7 — “ ROS. ROS and washed pellets were dispersed in (".25 M sucrose-0.01 recreation. The y intercept is the reciprocal of mammal [C.L‘vi} bound - '7 M Tris-Cl. pH .' 5. prior to examination under the microscope. at equilibrium. [RUSJ was 5 n5! and [Cab-H was .367 nm. :9 are; 34 $1}, «m‘ h‘ «- ., . A)" . I . g- 9.15" ‘ ‘ i . ,. . g; ‘ -4 1‘1 ,v« r A _ A"'\ A 1,. ’ . . ' 3" . 4" .-‘r . A ‘ L , -rmo¢nqv-rxv~t‘vvm—. Fl"! \"<.afl\-»'.‘\'l.' 'm*W'-"r 1" ' "~ r» --- ~ ~-‘ -- -~"-“' 4‘ ' " ' “ ' r" ' ' “"’ .- u . - . -- . r — - . ’4 4(5’: ‘Xi "1* .' 9 ...-“q. p- . 1.4:: 31‘ 447., 'A ’x mgr-3' . A. -.-;f:." » 12520 ..- - _.-.L___-_._._.._.l-_ --.. .. -u lbw-VFW time (rm) FIG. 3. Dissociation of ”‘l-CaM from HIE-washed ROS. Dis. sociation data were obtained as described under "Materials and Methods." The curve was generated using an equation for dissociation of one ligand from '2 distinct sites l 36!. Four experiments. only one of which is shown. gave similar results. r . l I l l 3!- L.H.. _-‘..-_ --_.__..-____.._.._L-____J a l ' ’ 3 EloMblnM 5 4 FIG. 4. Scatchard plot oi'CaM binding. Three experiments were performed as described under equilibrium binding assays in “Mate rials and Methods." Each set of data points was analyzed by least squares linear regression at [CaMhl 30.4 ms The slopes and x intercepts were averaged to yield K4 and [R080]. imply that the first type of binding site released CaM more rapidly than the other. despite the fact that bOth sites appar- ently bound comparable amounts of CaM. K4, estimated at 2.89 x 10" M from the association data. was also estimated from the ratio of the forward (k!) and reverse UL.) rate constants. Using k-l — 0.120 ruin" for the first of the '2 binding sites above and It; = 0.436 not“ min" from the association Studies. it'd was calculated to be 2.75 x 10” M, in good agreement With the K4 estimated from the axiomation data. For the second binding site. it“ was 0.0084 min" and K4 was calculated to be 0.19 x l0"’ M. Equilibrium Binding Studies—Equilibrium binding was studied using a fixed ROS concentration in the presence of varying concentrations of labeled CaM. The difference be- tween bound and total Cab-I equaled free CaM. A plot of bound over free CaM versus bound CAM is shown in Fig. 4. for [0151:] greater than 0.4 nM. The data have been fitted by a straight line whicn interCepts the x axis at 4.17 t 0.45 nM. This is. the calculated concentration of unoccupied Cam-bino- ing sites per 0.2 mg/ml of protein. It'd, calculated from the negative slope of the line. is (3.2 t 0.5) X 10” M. which is higher than the values obtained from asocn'ition and disso~ ciation data. No evidence of a concave upward trend appears at low concentration of CaM. as would be expected if a second binding site was present (36!. Binding at concentrations below 0.5 not was highly variable because the total bound radioao ROS Catmodulin tivity was less than twice background. Therefore. the only estimate of Kd for the possible second binding site is 0.19 x 10’5 M. which was obtained from the dissociation data. Cat'cium Dependence—Fig. 5 describes the Ca" depend- ence of CaM binding to the soluble protein-depleted ROS. The concentration of Ca“ required to promote half-maximal binding of CaM to ROS was about 0.3 not. and saturation was achieved by 10'5 M. These values are similar to the Ca" concentration required for the stimulation of phosphodiester- ass (28) or myosin light chain kinase (37) at similar (33le concentrations. Calcium Specificity—The ion specificity of CaM binding to ROS was investigated by adding increasing amounts of com. petitor to a solution containing a Ca“ concentration (0.3 psi) a- ' “’“i. -3... 5 a” l 5 l i 3‘ 4 albi- l :q- J g I .29 4: ‘ -: J —5 Ion Eta“). M Pro. 5. Ca” dependence of CLM binding. Free [Ca”} in the presence of l mu EGTA was varied using Kc..mn - 4.4 x 10‘ tot" (38). Data points are the average: (28.2.) of 5 experiments. H. in Tris-Cl; x-——x. in 100 nut 2-(N—morp'nolinoleth- anaulfonate (U. 3. Biochemical). pH 5.8. I met NaNJ. l mu EGTA. 10 rug/ml of BSA. Y t ‘1’ t Y' T l 7 ml- ' ! v :1 root- l fl ‘ l i“? 1 5 E l 253.- .4 1 ; E‘Gf 1 20- ' ' + z ”n“ h 1; . ...," -..—r/l 1 ii «3— 6 .. s -5 a -> be {ME’EM FM. 6. Cation dependence ot‘CaM binding. Free {.‘-ie"] in the presence of l mu EGTA was varied using K‘dn-zb‘l‘). - 56.4 x 10" it" KZn-EGT\ ‘ 142 X 10' M". and Kwnzn I 7.3 NI". These values were derived by a memo: similar to that of Santa: 081. Kain-r. - 9.55 x 10"“ M " in mo but 4.4 x 10° 31" ‘n "Fm-Cl. pH 7.5. at 25 ‘C. The. ratio between these two values was assumed to be constant for HI) and Tris-Cl. regardless of the identity of the cation. Thus Kan mm in H30 is JCS x U)“ M". When oivxded by the ratio. it gives a value of 142 x 10" M" in Tris-Cl. Assays were performed as described under “.‘vlateiials and Methods" for Ca” dependence. 10097: maximal binding was defined as dust observed in the presence of Ca". C—D Mn". 0—4. Zn"; x K. Mg“; - - -, C3“ nfrom Fig. 5). Data points for Mn" and Zn" are means : 8.8. for 3 or more experiments. ..wa- . u . l ' h ”turf.” " .. \f‘ . - . _-' o.’.-—-- ...-u no )- .~ . .r: .-1 . ‘ .- I ~ ’. "“I ' ‘ 'Ir" ’.: a Ali-14.4..”— l I . _ v ‘2. , . . o4 «luéwri'd‘uLi ’. ""', W13“??? "fswvmwrcw‘f-‘Tlrr': W‘s-r v-m-r ...—”~- ~ .1" ~— ‘ ROS C almoa’uiin which yielded 50‘? maximal binding (Fig. 5). Mg‘“ does not compete effectively with Cag‘ for EGTA and. therefore, will not significantly alter the free Ca” concentration (33). It was found that free Mg“ concentrations up to 1 mM (3 x lOJ'fold excess over free Cai’) neither altered CaM binding to the ROS nor substituted for Ca" in the bidding reaction. Because both Zn“ and Mn“ compete for EGTA (38). free 0112' con- centrations cannot be accurately determined. Therefore. in both cases these cations were substituted for Caz’. retaining EGTA in the assay. As seen in Fig. 6. both Zn“ and Mn“ promoted Cal»! binding. For 50% maximal CaM binding. only 50 nM Mn” or Zni‘ was required. as compared with 300 mm C a“. The same maximal binding was achieved with Ca“ and '. in“. Binding in the presence of Zn“ was less effective. and concentrations of this metal above 100 pm were apparently inhibitory. Cad-I Specificity-Gabi is structurally related to other Ca“— binding proteins (39. 40). and for this reason we introduced parvalbumin. troponin C. and unlabeled CaM into the binding assay to determine to what extent they would displace labeled CaM from ROS membranes. As revealed in Fig. 7. authentic CaM was able to displace 50% of tracer at a molar ratio of 3: l. Troponin C did show an effect, but the concentration l?‘ i i i ' 5 i ‘00 MM. I'M FIG: 7. CaM specificity of binding. Binding of ml-CaM (3 rm) was assayed at equilibrium with the indicated amounts of total CaM, troponin C. or par-Jalbumin. Error bars on the CaM curve indicate SE. for 4 experiments. H. CaM: C—-O. troponin C; x—x. parvalbumm. TABLE II Effects of calmodulin and Co” on phosphodtesterase density of unwashed ROS. soluble phosphodies(erase (P). and a reconstituted system. containing DROS and soluble proteins (P and G (30)) 5 pg of C334. 1 mM Caz’. 1 mM EGTA. and 400 ux GTP were added at indicated to an assay volume of 0.5 ml. Phosphodiesterase was assayed as described in Refs. 29 and 30. Actmty Is expressed in nanomoles ot' GMP/assay. In our Michigan State Unlverszty labora- tory. this CaM caused a 3-fold activation of commercml (Sigma) activator-depleted beef heart CAMP phosphodieStt-rase. and activa« tion was independently demonstrated With respect to activator-de- pleted rat brain CAMP phosphodiestemse (‘27). nos — ROS + - . DROS t P * G GTP GTP Somble P W No additions 223 366 34 9 159 +CaM 22" 354‘ 41 ii 136 +CaCl-g 21 418 :31 11 138 -EGTA ‘31 371 1 T 124 +CaM 4‘ EGTA 31“ 365“ 54) 13 147. ~Ca.‘d - CaCl; 3‘2 505 14 18 144 " Values are averages obtained in '2 other experiments and normal- ized to the data in the rest of the table. 'o-\'- v —‘£l“ -: - ”v! -—a .v—rWVH—fi .’«-5«§"‘f‘,- u --r 12521 required was 500 times greater than that for authentic CaM. This difference is similar to that observed for Stimulation of phosphodiesterase (41) and in competitive binding to anti- CaM (27). Pan-albumin was without effect at any concentra- tion tested. Temperature Dependence—The Caz‘dependent binding of ‘t‘I-CaM to ROS was temperature dependent. Rates of asso- ciation and dissociation were approximately 3 times faster at 24 °C than at 2 °C. However, equilibrium binding reached the same level at both temperatures. Absence ofPhosphodz'ester-ase Regulation—Table II shows data from experiments designed to determine whether ROS phosphodiesterase aetivity was affected by CaM. CaM was added in the absence or presence of Ca“ to preparations of the ROS. the soluble phosphodiesterase isolated from ROS membranes and a reconstituted system consisting of the solo uble phosphodiesterase tOgether with membranes depleted of soluble proteins (30). The phosphodiesterase activities present in both the ROS and the reconstituted system exhibit require. ments for light and GTP. Neither CaM alone nor CaM in the presence of added CBCi-g resulted in substantial activation of phosphodiesterase. Since EGTA failed to attenuate the phos- phodiesterase activity measured in any of the preparations it seems unlikely that the ROS phosphodiesterase is Ca" de- pendent. DISCUSSION The data presented in this report demonstrate the presence of CaM in the unextracted ROS. CaM does not seem to regulate the GTP plus light-activated phosphodiesterase. However, high afi'mity membranous binding sites for m’l-CaM were demonstrated to be present in the ROS. Three estimates were made of the dimociation equilibrium (binding) constant K4. The values obtained were 2.39 x 10" M from association data. 2.75 x 10" it (first site) and 0.19 x 10“ M from dissociation data, and 3.9 X 10" M from the slope of the Scatchard plot (42). Dissociation studies implied the existence of a second high afl'mity binding site with K.) lower by an order of magnitude than that of the first. Despite the facr that this second site bound about as much CaM as the first. it released it much more slowly. Whereas the identity of the CaM-binding component is unknown, it does not appear likely to be rhodopsin. CaM- binding sites were eatimated to be 4.17 nM at 0.20 mg/ml concentration of the PIE-washed ROS. Assuming 85% of the membrane pretein to be rhodopsin (M, =- 39,000 (32-34)). there would be only one CaM-binding site per 1100 rhodop- sins. Therefore. unless only random rhodopsin molecules bind CaM. it is unlikely that this protein consritutes a new CaM- binding protein. Moreover. since rhodopsin comprises such a large proportion of the membrane protein, the location of the CaM-binding sites must be relatively Circumscribed. Both the ROS plasmalemma and the intrinsic membrane protein local- ized at disc incisures and margins by Papermaster et at. (43) are possible binding sites of CaM. Neither lipid binding of CaM nor binding by a minor membranous contaminant which copurifies with the ROS can be rigorously excluded. However. in another experiment (data n0t shown) in which equilibrium binding was analyzed by the method used in Fig. 4 (42). two pertinent observations were made. First. at large CaM concentrations. the ratio of bound C aM/free CaM approached zero (0.04 at 70 mm). This argues against a significant nonsaturable binding of (38.34 through a hydrophobic interaction With lipids. Secondly. two potential contaminating membrane fractions. both HE washed. were examined for their CAM binding as compared directly to the ROS in the same experiment. One fraction contained mito- chondria and other inner segment organelles, and the other Q -~--—-‘o:——-‘ m-—‘v~-o.,' -—-‘w—.~.~- .. .. We “—7. w-_~_vv— AIL ~4: .. H 31... ~ .' L4 Auhuuul. - I .1‘ . "'4'. If .15...“ . ‘ . ' .r... Iii .. .10. 12522 contained largely retinal microsomes. For both fractions, the values of K. and the estimated binding capacities were com- parable to those of the ROS. Had either fraction been respon- sible for the binding exhibited in the case of the ROS, the binding capacity of the possible contaminant would have been ‘4 ‘0 R08 Caimodulm . Mild. N., Keir-m. J. J., Marcus. F. R. Freeman. J.. and Bitensxy. M. W. {1973) Proc. Natl. Acad. Scr. L'. S. A. 70. 3620-3824 . Mild. N., Bareban. J. M.. Keirne. J. J., Boyce. J. J., and Bitemky. M. W. (1975) J. Biol. Chem. 250. 6320—5327 . Manthorpe. M, and McConnell. D. G. {1975) Biochim. Biophys. .4 cm 403, 438-44' 5 expected to be much higher than that of the ROS. since the 10, wmdmg, M. I... Bownda. 1),, Green. 5_ a" Morrisey, J, L. and purity ratios (Am/Anni observed for the ROS preparations Shedlovsky. A. (1977) J Gen. Physiol. 69. 667-679 exclude the possibility of large contamination by either free. 11- Wirifeée; 0141.4 $331“!- 341- L (1977’ Proc- Natl- Acad. Sci. non. a ',- -- _ We calculated .. man .. ......m. 1::nesa111122-r.:.1w5.12:.191:... binding 3"” ‘1 10031) ”um5 t0 “)0_’h°‘3<>P=um Per 591““? 14. Liebman. P. A- and Pugh. 2:. .\'.. Jr. (1979) Viscon. Res. 19. 375- CaM calculated earlier. Such determinations are subject to 33.0 considerable variability from preparation to preparation. 15. Liebman. P. A., and Pugh. E. N., Jr. 11980) Nature 287, 734-736 Neverthelete. on the basis of the present estimates. there 35- 31:?2994; BM B.M..andLolley. R- “41973? V310" Res. appear to be about 50% more CaM molecules in 305. than ,_ ‘ , '. ,‘ ' , . ‘ there are membranous binding sites. This excess of CaM “' "g?”rf'm‘ L..anJ Bownds, D‘ (19:9) J’GM’P’USwe‘ ””29” “105“”193 “mild be cWWW“?!e “‘1‘ the Presence 0" “he? 13. Nicol. G. 1).. and Miller. w. H. «1978) Proc. Natl. Acad. Sci. U. CaM-binding preteins in the soluble fraction of ROS. The s_ ,1, 73, 521742-20 lizht plus GTP-requiring cGMP phosphodiesterase seemed to 19. Miller, W. H.. and Nicol, G. D. (1979) .V'ahrre 280. 634-66 represent a likely candidate since many other phosphodies- 20- Wig?- wfld Stryer. L- (1930’ PfOC- Natl- Acad- Sti- U. 3. terasa require CaM for activity (44). However. neither Ca" 6 " “ . l“, l , _ _ ‘ , a _ 9 a ...... :1 “1...”? In addition the phospnodiesterase seems to be present in large 525‘ 1&1.” excese relative to Ca“. There is one catalytic (P) protein 5. L3... 3" p. Krishna. c,“ Aguirre, G., and Chader, Q .1_ (1979) molecule of the ROS phosphodiesterase complex for each 100 Nature 280. 62—64 rhodopsin molecules «am. Together with the calculated esri- 24. Britairithorge.1 M and McConnell. D. G. (1974i J. Biol Chem. 249. mate of one Cal»! per 700 rhodopsins. it seems probable that 4'26“ 1° 1| 1 , . only 1 CaM molecule is present per 7 phosphodiesterase 25‘ P'1¢?°nMILDfG"Dmler'L' Aih‘ad‘f' D" ”"3“ banana-J. , _ . . _ . . . . “981) J. Biol. Chem. 258. 49lJ—49le 3103“” “S‘t'c’atfd “m“ "‘9 F103- Tlu-‘f ”“09 cow-“em 25. mm. o. H., Rosehrough. N. J.. Fan: A. I... and Randall. P. J. with observations on 15% polyacrylamzde gels of soluble r1951, .13.“ Chm. 193.265-2275 proteins derived from HE wasmne of the 303- Staining 27. Chnt'ouleas. J. o. Dedznan. J. a. Munjaal. RP- and Meal-15.14.. densities of the low molecular weight subunits (P. =3 13.000: R (1979) J. Bwl. Chem. 254. 1026246267 G, 5 10.000) of the phosphodiesterase complex (‘29) markedly 38- Dedmm J- EL 1,3025?- J- D- “Ck-wring L- “hm“: J- D.. and exceed that of a 17.000.daiton band wnich is presumably CaM. . Mem‘ A' R‘ ‘19' 'l J‘ 8‘“ Chem ‘5'" 84154422 . , . .... 29. hohnken. R. 3.. Eadie, D. 51.. Revun. A. and McConnell. D. G. Yet tn cells wnereCaM regulates phosphodiesterase, the Ca - (1981) J. Biol. Chem. 256. 1250242509 bmdmz worm 18 present In large excess relative to the so. Kolmken. n a. Eadie. D. M. and McConnell. n. G. .1931) J. enzyme (44). Nevertheless. despite these arguments against BtoL Chem. 256. 12510—12516 its implication in phosphodiesterase regulation, CaM may Still 31. Mcgonnell. D. 0.. Ozga. G. W.. and Seize. D. A (1968) Biochim. play an important. role in the regulation of Cab-sensitive 32 HEwphflgcfgzg-hJIt-QB v 8' L 235 114 ' ‘r -v ettzman. .\ I- .aureiew l0 ' . processes in the RQb.To dace l: duierentenrymes have been 33- Robinson. W. E., Gordon-Walker. A., and Bownds, D. ”972 reported to be CaM dependent (45). The man before us is to Vature New 3w L 235 1‘. 2 identify and charaCterize the CaM-binding proteins present in 34. Diana. 3 J. M. De G'np' w_ J" anu Jansen. P. A. A. ”972) R03 Eiochtm. Biophys. Acta 271. 419-428 35. McConnell. D. G. l1965) J. Cell Biol. ‘27. 453-473 REFERE-'CES 36. Rod'oartl. D. (1973) in Receptors for Reproductive Hormones 1. Hagms. W. A., Penn. R D.. and ‘(oshikami S. tlQTO‘ Biophys. J. (O'Malley, 3- W.. and M9198. A- R., 8d!” PP. 359.335- Plenum. 10.380412 New York 2. Hagins, w. A., Robin...“ w, E., and Yoshikami. s. {1975, Energy :37. Yam. x- Yazawa. M. Kakiucbi. 3.. Ohshinm. M.. and Uenishi. K. Transduction m Biological Systems. Ciba Symposium 31. l69- (1978) J: 3‘“ Chem. 2533 13354349 159 38.. Bartfai. T. :1979) Adv. Cyclic Nucleotide Res. 10. “219-242 3. Hagins. w- A.. and ‘i'oshikami. 5_ 1975) Ann. _\,t_ y_ Acad. Sci. 39. Watterson. D. M. Harrelaon. W. E., Jr., Keller. P. .‘A..Sh3rief. F.. 264. 314—325. and Vanaman. T. C. (1976‘! J. Biol. Chem. 251. 4501-4513 . Stevens. F. C., Walsh. H., Ho. H. C., Teo. T. 3.. and Wang. J. H. 4. Hams. W. A. 11972) Anna. Reta Biophys. Bioeng. 1. 131-158 - _ _ 5. Hubbell. W. 7.... and Bownds. D. M. t1979l Anna. Reta Neurosci. ”9"“ J- BiOL Chem. Q51. “954000 -- ' . , 2. 17-34 41. Dedman. J. R., Potter. 9. D.. and Means. A. R. (19”) u. 810.. 6. Kaup. L'. 3., Schnetknmp. P. P. 1.1.. and Junge, w. (1979) in Chm 352-“3-4-“0 . . .-. Detection and .lleaaurcment of Cgicmm 1,; C9113 (‘Aghiey' C. 42. Scatchard. G. \19192Arln. .V. Y. Acad. SC: 51. 050-012. C. and Campbell a K" eds) Dp. 237-303. Elsevier. AmSterdam 43. Pupemeater. D. 5.. Schneider. B. G., Zorn. M. A., and Kraehen- ' ' ‘ buhl. J. P. (1979) .1. Cei! BIOL 73. 415...:5 44. Cheung. W. Y. 6980) Science 207. 19-‘2.’ "R E- Kohnken l G- Chafouleas. D. M. Eadie. A. R. Means. and 45. Means. A. R., Tash. J. 3.. and Chaiouieas. J. G. .1932) Plusxoi. D. G. McConnell. unpublished observations. Rev. 62. in press -P— mun—7*. —_ ..v. V—V-v7 ‘ A. -1 . .HV.-‘— fi-w... ~(,'_.,- -.. ,P‘ -..-‘-"—~.-‘ ".--, . -. .11...- -._‘... ”—Awr‘fq.‘uqvc‘.-rrrmm_ 5‘3"! . ‘ \ u‘ AAAAA TTTTTTTTTTTTTT ”in\fixi‘mmmmmmuswimnmmuwufifs 3 3333333333333