Regulatory modes of rod outer segment membrane guanylate cyclase differ in catalytic efficiency and Ca 2+ -sensitivity Ji-Young Hwang 1, *, Christian Lange 1, †, Andreas Helten 1 , Doris Ho¨ ppner-Heitmann 1 , Teresa Duda 2 , Rameshwar K. Sharma 2 and Karl-Wilhelm Koch 1 1 Institut fu ¨ r Biologische Informationsverarbeitung 1, Forschungszentrum Ju ¨ lich, Ju ¨ lich, Germany; 2 The Unit of Regulatory and Molecular Biology, Departments of Cell Biology and Ophthalmology, NJMS & SOM, UMDNJ, Stratford, NJ, USA In rod phototransduction, cyclic GMP synthesis by mem- brane bound guanylate cyclase ROS-GC1 is under Ca 2+ - dependent negative feedback control mediated by guanylate cyclase-activating proteins, GCAP-1 and GCAP-2. The cellular concentration of GCAP-1 and GCAP-2 approxi- mately sums to the cellular concentration of a functional ROS-GC1 dimer. Both GCAPs increase the catalytic effi- ciency (k cat /K m ) of ROS-GC1. However, the presence of a myristoyl group in GCAP-1 has a strong impact on the regulation of ROS-GC1, this is in contrast to GCAP-2. Catalytic efficiency of ROS-GC1 increases 25-fold when it is reconstituted with myristoylated GCAP-1, but only by a factor of 3.4 with nonmyristoylated GCAP-1. In contrast to GCAP1, myristoylation of GCAP-2 has only a minor effect on k cat /K m . The increase with both myristoylated and non- myristoylated GCAP-2 is 10 to 13-fold. GCAPs also confer different Ca 2+ -sensitivities to ROS-GC1. Activation of the cyclase by GCAP-1 is half-maximal at 707 n M free [Ca 2+ ], while that by GCAP-2 is at 100 n M . The findings show that differences in catalytic efficiency and Ca 2+ -sensitivity of ROS-GC1 are conferred by GCAP-1 and GCAP-2. The results further indicate the concerted operation of two ÔGCAP modesÕ that would extend the dynamic range of cyclase regulation within the physiological range of free cytoplasmic Ca 2+ in photoreceptor cells. Keywords: phototransduction; guanylate cyclase; GCAP; myristoylation; k cat /K m . Photoexcitation of vertebrate photoreceptor cells leads to the hydrolysis of cyclic GMP (cGMP) and subsequent closure of the cyclic nucleotide-gated (CNG) channels in the plasma membrane. Restoration of the dark state of the photoreceptor cell requires the reopening of CNG-channels (reviewed in [1–3]). A critical step in this recovery process is synthesis of the second messenger, cGMP. Studies with vertebrate photoreceptor cells, constituting mainly rods, show that these cells express two types of a membrane bound guanylate cyclase termed ROS-GC1 and ROS-GC2 (alternatively used names are retGC1 and retGC2 and GC-E and GC-F; reviewed in [4,5]). ROS-GC1 has been purified directly from bovine and amphibian rod outer segments [6–9], and it is the only cyclase which has been cloned based on its amino acid sequence [8,10]. Human retinal diseases (LCA1 and CORD6) affect both rod and cone vision, but are only linked to the ROS-GC1 gene [11–17]. Knowledge about enzyme kinetic parameters of native photoreceptor guanylate cyclase are so far restricted to ROS-GC1. This is mainly because only ROS-GC1 has been purified from bovine retina and thus, probably, constitutes the main cyclase in bovine rod outer segment preparations. Reported K m -values for the substrate, GTP, range from 0.76–1.1 m M [6,7,9,18]. Turnover numbers (k cat ) of the purified enzyme range from 0.2–3.9 cGMPÆs )1 [6,9]. Small acidic Ca 2+ -binding proteins, called guanylate cyclase-activating proteins or GCAPs, regulate ROS-GC1. Three GCAP (GCAP-1, 2 and 3) isoforms have been cloned from retinal sources [19–23]. GCAP-1 and GCAP-2 are both expressed in rod and cone cells of different species as shown by immunocytochemistry [21,22,24,25]. Expression of GCAP-3 is more restricted; it is present in human cones, fish rods and cones, but not in mice photoreceptor cells [26]. Thus, GCAP-3 does not appear to be a general sensor of Ca 2+ -pulses linked with phototransduction. GCAP-1 and GCAP-2 contain one nonfunctional and three functional EF-hands. Through functional hands they detect changes in the intracellular Ca 2+ -concentration [Ca 2+ ] and modulate ROS-GC1. Dark adapted vertebrate photoreceptor cells have a cytoplasmic free [Ca 2+ ] of 500– 650 n M . This falls below 100 n M upon illumination [27–30]. GCAPs detect the fall and in their Ca 2+ -free form, activate ROS-GC1 [4,5,19–23]. The generated cyclic GMP replen- ishes the depleted pool and restores the channels in their open state. While there is wide agreement in the literature Correspondence to K-W. Koch, Institut fu ¨ r Biologische Information- sverarbeitung 1, Leo-Brandt-Strasse, Forschungszentrum Ju ¨ lich, D-52425 Ju ¨ lich, Germany. Fax: + 49 2461 614216, Tel.: + 49 2461 61-3255, E-mail: k.w.koch@fz-juelich.de Abbreviations: ROS, rod outer segments; ROS-GC1/GC2, photo- receptor membrane guanylate cyclases 1 or 2; GCAP-1/2, guanylate cyclase activating protein 1 or 2; NMT, N-terminal myristoyl trans- ferase; myr, myristoylated; nonmyr, nonmyristoylated; Rh, rhodopsin. Enzymes: guanylate cyclase (EC 4.6.1.2.) *Present address: Genetics & Molecular Biology Branch National Human Genome Research Institute National Institute of Health Bldg. 49, Rm 4A08, 49 Convent Drive, Bethesda, MD 20892–4442, USA. Present address: Instituto de Bioquı ´ mica Vegetal y Fotosı ´ ntesis, Centro de Investigaciones Isla de la Cartuja, Avda. Ame ´ rico Vespucio s/n, 41092 Sevilla, Spain. (Received 2 June 2003, accepted 28 July 2003) Eur. J. Biochem. 270, 3814–3821 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03770.x that these Ca 2+ -binding-proteins are powerful activators of the ROS-GCs [4,5,19–23], there is no unanimity on their specific expression in rods or cones and to which ROS-GC they are paired with. Some immunocytochemical studies show that GCAP-1 is the predominant form in cones and GCAP-2 in rods [31,32]; and recently the physiological role of GCAP-2 has been questioned, because the expression of GCAP-2 in transgenic GCAPs null mice did not fully restore the normal flash response of the wild-type [33]; and Howes et al. [34] in a subsequent study on transgenic mice showed, that the wild-type phenotype flash response could be rescued by the expression of GCAP-1 in the absence of GCAP-2. To better define the Ca 2+ -modulated ROS-GC trans- duction component in rod cells we addressed the following questions in the present work: how much of GCAP-1 and GCAP-2 is present in rod outer segments? Are there differences between GCAP-1 and GCAP-2 in their regula- tory properties and if so, what is the physiological signifi- cance of such a difference? The study demonstrates that GCAP1 and GCAP2 modulate the Ca 2+ signaling of ROS-GC1 in different ways. Experimental procedures Preparation of ROS Bovine ROS were prepared according to a standard protocol under dim-red light. This involved sucrose density centrifugation in the presence of a moderate salt concen- tration to minimize loss of cytoplasmic proteins [35]. ROS were at all time stored and handled in the dark. Rhodopsin (Rh) concentration was determined spectrophotometri- cally at 498 nm using a molar extinction coefficient of 40 000 M )1 Æcm )1 . Expression and purification of GCAPs GCAP-1 and GCAP-2 were expressed in E. coli and purified as described previously [36,37]. Myristoylated (Myr) forms of GCAP-1 and GCAP-2 were obtained by their coexpression with yeast N-myristoyl-transferase (NMT) (the plasmid pBB131 was kindly provided by J. Gordon, Washington University School of Medicine, St. Louis, MO, USA). As wild-type GCAP-1 does not contain a consensus site for yeast NMT, we used the mutant D 6 S- GCAP-1 that is functionally indistinguishable from myr GCAP-1 [38]. The degree of myristoylation was determined by HPLC. Heterologous expression of ROS-GC1 HEK293 tsA cells. HEK293 tsA cells were grown and transfected with ROS-GC1 in pcDNA3.1 expression vector by the calcium phosphate method [16]. The following modifications were applied: 22 h post-transfection cells were washed sequentially with NaCl/P i ,NaCl/P i plus EDTA and NaCl/P i and incubated in medium for 20–22 h. Cells were harvested by a short centrifugation step (200 g;5min; 4 °C), resuspended in NaCl/P i and centrifuged again. The pellet was resuspended in lysis buffer (10 m M Hepes/KOH pH 7.5, 1 m M dithiothreitol), sonicated and centrifuged to remove cellular debris (400 g;5min;4 °C). The supernatant was then centrifuged at 125 000 g for 15 min at 4 °Cto pellet the membranes. The membranes were resuspended in 10 m M Hepes/KOH pH 7.5, 250 m M KCl, 10 m M NaCl and 1 m M dithiothreitol. Protein concentration was deter- mined by the amido black method [39]. COS7 cells. COS7 cells were transfected with the wild-type recombinant ROS-GC1 cDNA in pcDNA3 expression vector by the calcium phosphate coprecipitation technique as described before [16]. Sixty hours after transfection, cells were washed twice with 50 m M Tris/HCl pH 7.5 buffer. They were then scraped into 2 mL cold buffer, homogen- ized, centrifuged for 15 min at 5000 g andwashedseveral times with the same buffer. Electrophoresis and Western blotting SDS/PAGE and Western blotting were performed accord- ing to previously defined protocols [16,22]. Chemiluminescence detection and quantitation of proteins Quantitative chemiluminescence analysis was performed on a Luminograph LB 980 (Berthold). Varying amounts (10–50 ng) of ROS-GC1 were electrophoresed with a sample of ROS or ROS membranes. After SDS/PAGE, proteins were blotted and probed with anti-ROS-GC1 Igs. Shortly before measurement, the blots were incubated with Western blotting ECL reagent 1 and 2 (Amersham) for 1 min and photon emission per second was detected at 480 nm. Varying amounts of ROS-GC1 were used for calibration and the amount of cyclase in ROS was obtained directly from the calibration curve. The predominant membrane guanylate cyclase in bovine ROS is ROS-GC1, the other isoform (ROS-GC2) that is expressed in photoreceptor cells consti- tutes less than 10% of the amount of ROS-GC1 (A. Helten and K W. Koch, unpublished observation). Amounts of GCAPs were obtained by a similar proce- dure but instead of a Luminograph we used a Kodak Image Station. Exact amounts of purified GCAP-1 or GCAP-2 for SDS/PAGE and Western blotting were determined from GCAP specific protein standard curves. These curves were created as follows: purified preparations of GCAP-1 and GCAP-2 were used to determine their molar extinction coefficients as described in [40]. The values were e 280 ¼ 28378 M )1 Æcm )1 for GCAP-1 and e 280 ¼ 37512 M )1 Æcm )1 for GCAP-2. Exact stock solutions of GCAP-1 and GCAP-2 were prepared using the molar absorbance coefficients and a calibration curve was made using the Coomassie Blue dye binding assay. GCAP solutions were adjusted every time by the use of these GCAP-1 or GCAP-2 specific calibration curves. Guanylate cyclase assay For reconstitution experiments, pure ROS were washed several times in low salt buffer and the membranes were prepared as described in [22]. GCAPs were reconstituted with washed ROS membranes and guanylate cyclase activity was determined as described in [22]. Determination Ó FEBS 2003 Differential Ca 2+ -modulation of guanylate cyclase (Eur. J. Biochem. 270) 3815 of guanylate cyclase activities in HEK and COS cell membranes was as described previously [16]. Incubation of ROS-GCs at a constant concentration of 2 m M EGTA and varying concentration of GCAPs (0–10 l M )were performed to obtain EC 50 values for the GCAP. Keeping the GCAP concentration constant while varying free [Ca 2+ ] yielded IC 50 values of Ca 2+ for each GCAP. Analysis of enzyme kinetics ROS-GC1 in washed ROS membranes was reconstituted with 2 l M of purified myr- or nonmyr-forms of GCAP-1 and GCAP-2. Guanylate cyclase activity was assayed at 2m M EGTA (£ 10 n M free [Ca 2+ ]) as a function of the substrate GTP. We used Mg 2+ as cofactor and kept the ratio of Mg 2+ to GTP at 5 : 1. Analysis of data was performed with ORIGIN 6.1 and SIGMA PLOT 4.2 software. A direct plot of activity vs. [GTP] gave in all cases a sigmoidal curve indicating cooperative substrate binding. In order to analyze data of a sigmoidal dependence in a linear Lineweaver–Burk plot [41], we determined a Hill coefficient n from a fit of the direct plot. Values of V max and K m were then determined from a plot of 1/V (reciprocal of guanylate cyclase activity) vs. 1/[S] n (reciprocal of the substrate concentration raised to the power of n). Preparation of Ca 2+ -buffer and determination of free Ca 2+ concentration Free [Ca 2+ ] concentrations of buffer solutions were calcu- lated using the program WEBMAX 2.0 (Stanford University, CA, USA). The free [Ca 2+ ] of guanylate cyclase assay buffer solutions was checked at 30 °CwithaCa 2+ -sensitive electrode (World Precision Instruments, Inc.) using calibra- tion standards in the range from 10 )8 to 10 )1 M free [Ca 2+ ]. Results The Ca 2+ -sensor proteins GCAP-1 and GCAP-2 are almost equally expressed in native bovine ROS membranes Molar concentrations of GCAPs have not been reported so far, although they are indispensable for a full quantitative description of phototransduction. We determined these values by the following procedure: 0.5–9 ng of purified GCAP-1 and 2–10 ng of GCAP-2 were loaded onto PAGE. ROS containing 5–25 lg of rhodopsin were loaded on the same gel. After electrophoresis, proteins were electrotrans- ferred to a blot membrane and probed with antibodies against GCAP-1 or GCAP-2. Bands were visualized by chemiluminescence. Chemiluminescence intensity was line- arly dependent on the amount of antigen and was used to quantify the amounts of GCAPs. An example is shown in Fig. 1A,B, where the amount of GCAP standards showed a linear increase in chemiluminescence intensity. The analysis of several Western blots revealed a ratio of 1 : 1200 ± 360 (N ¼ 12) of GCAP-1 to rhodopsin and 1 : 1100 ± 560 (N ¼ 5) of GCAP-2 to rhodopsin. About 25% of GCAP-1 is lost during the purification of ROS on a sucrose gradient, whereas none of GCAP-2 was lost (data not shown). Taking into account the loss of GCAP-1 the ratio of GCAP-1 to rhodopsin is 1 : 900. These values correspond to a cellular concentration of 3.3 l M GCAP-1 and 2.7 l M GCAP-2. Thus, both GCAPs are present in bovine rods in nearly equal concentrations. Catalytic efficiency of ROS-GC1 Previous determinations of the ratio of ROS-GC1 to rhodopsin in bovine ROS, based on the purification of ROS-GC1 from ROS preparations ranged from 1 : 104 [6] to 1 : 440 [9]. We re-examined the amount of ROS-GC1 by a different approach using the chemiluminescent densito- metric analysis of blot membranes as described above for the GCAPs. For example, ROS with 2 lg of rhodopsin Fig. 1. Quantitation of GCAPs in purified bovine ROS. Increasing amountsofpurifiedGCAP-1(A:a,0.5ng;b,1ng;c,2ng;d,3ng;e, 4ng;f,5ng)andGCAP-2(B:a,2ng;b,3ng;c,4ng;d,5ng;e,6ng; f, 7 ng; g, 8 ng; h, 9 ng) were used to create a calibration curve. The two GCAP-1 bands visible in the calibration row are nonmyristoylated and myristoylated GCAP-1. Protein samples were electrophoresed and blotted together with a sample of ROS containing 5 lg(A)or10lg (B) of rhodopsin. GCAPs in these samples were detected by specific antibodies and ECL. Band intensity (arbitrary units) was linear within the tested range, ROS with 5 lg rhodopsin contained 1.8 ng of GCAP- 1 corresponding to a ratio of 1 : 1600 GCAP-1 to rhodopsin, ROS with 15 lg rhodopsin contained 6.5 ng of GCAP-2 corresponding to a ratio of 1 : 1340 GCAP-2 to rhodopsin. 3816 J Y. Hwang et al. (Eur. J. Biochem. 270) Ó FEBS 2003 contained 28 ng of ROS-GC1 (not shown). This relates to a molar ratio of 1 : 200 ROS-GC1 to rhodopsin. Analyzing four different blots revealed a mean ratio of 1 : 260 ± 60, which is consistent with previous estimates [6,9]. The functional unit of ROS-GC1 is a dimer [16,42,43] that is correspondingly present in a ratio to rhodopsin of 1 : 520. The cellular concentration of the ROS-GC1 dimer therefore is 5.8 l M (cellular concentration of rhodopsin ¼ 3m M ). We used this value to calculate the turnover number of ROS- GC1: V max in washed ROS membranes was 3.0 nmol cGMPÆmin )1 per mg Rh, which is synthesized by 4.8 · 10 )2 nmol ROS-GC1 dimer (1 mg of Rh ¼ 25 nmol Rh; ratio of ROS-GC1 to Rh is 1 : 520). Thus, the resulting turnover number or k cat of ROS-GC1 in native membranes is 1.0 s )1 , which is very similar to previous determinations of the purified enzyme in detergent (0.2–1.3 s )1 ; [6]). From these numbers we derive the catalytic efficiency expressed as k cat / K m as 0.77 · 10 3 M )1 Æs )1 (K m ¼ 1.3 m M ,Table1).Further- more, our data show that the sum of the molar concentra- tions of GCAP-1 and GCAP-2 equals the molar concentration of one ROS-GC1 dimer. GCAP-1 and GCAP-2 influence the catalytic efficiency of ROS-GC1 in different ways Equal amounts of GCAPs in bovine rods provoke the question, why cells express equal levels of protein isoforms with similar properties. It is reasonable to suggest that they differ in regulatory features. We tested this assumption by testing two key aspects of ROS-GC1 regulation, change in catalytic efficiency and Ca 2+ -sensitivity. First, how does the interaction with GCAP-1 and GCAP- 2 influence k cat /K m of ROS-GC1? ROS-GC1 in washed ROS membranes was reconstituted with a constant amount of myr- and nonmyr-forms of GCAP-1 or GCAP-2. As the half-maximal activation (EC 50 )ofROS-GC1byGCAP-1 and GCAP-2 occurs well below 1 l M [37],wechosea saturating concentration of 2 l M for all experiments. The ROS-GC1 activity was measured as a function of [Mg- GTP] at a constant free [Ca 2+ ]of5n M . The catalytic parameters were then determined from a Lineweaver–Burk plot (Fig. 2A–D). The results are listed in Table 1. The values of k cat were calculated from the values of V max (expressed as nmol )1 Æ min )1 per mg Rh) assuming a ratio of ROS-GC1 dimer to Rh of 1 : 520 as described above. Myr- GCAP-1 increased the catalytic efficiency of ROS-GC1 about 25-fold (k cat ¼ 5.9 s )1 ; k cat /K m ¼ 19.5 · 10 3 M )1 Æs )1 ). This increase in catalytic efficiency was significantly less pronounced (3.4-fold) when GCAP-1 was not myristoyl- ated, the k cat /K m value was 2.6 · 10 3 M )1 Æs )1 . When the effect of GCAP-2 was assayed in the same manner the results were different. Both GCAP-2 forms, myristoylated and nonmyristoylated, increased V max and decreased K m to a similar degree, i.e., for GCAP-2, the influence of the Table 1. Catalytic parameters of GCAP-dependent activation of ROS-GC1 in washed ROS membranes. Assays were performed at low [Ca 2+ ](2 m M EGTA). V max (nmolÆmin )1 per mg Rh) K m (m M ) k cat (s )1 ) k cat /K m (10 3 M )1 Æs )1 ) n ROS-GC1 3.0 1.3 1.0 0.77 1.38 + myr D 6 S-GCAP-1 17.0 0.303 5.9 19.5 1.68 + nonmyr GCAP-1 10.1 1.335 3.5 2.6 1.6 + myr GCAP-2 13.7 0.478 4.8 10.0 1.84 + nonmyr GCAP-2 14.0 0.619 4.8 7.7 1.86 Fig. 2. Lineweaver-Burk plots of GCAP- dependent regulation of ROS-GC1. (A) ROS- GC1 reconstituted with 2 l M myristoylated D 6 S-GCAP-1.ReciprocalsubstrateGTP concentration 1/[S] is given in m M )1 .Plotwas linearized by using the apparent Hill coefficient n ¼ 1.68 (Table 1) to yield 1/[S] n . Reciprocal activity 1/V is expressed as nmol )1 Æmin per mg Rh. Kinetic analysis was performed in the same manner with 2 l M nonmyryistoylated GCAP-1 (B), 2 l M myris- toylated GCAP-2 (C) and 2 l M nonmyris- toylated GCAP-2 (D). Data points represent mean of triplicates ± SD. Details are given in (A) and in Table 1. Ó FEBS 2003 Differential Ca 2+ -modulation of guanylate cyclase (Eur. J. Biochem. 270) 3817 myristoyl group was not very pronounced. Myr-GCAP-2 increased the catalytic efficiency k cat /K m by 13-fold, while nonmyristoylated GCAP-2 increased k cat /K m by 10-fold. Thus, myristoylation plays a significant role in the ability of GCAP1, but not of GCAP2, to activate ROS-GC1. Ca 2+ -sensitivities of GCAP-1 and GCAP-2 Second, do GCAPs differ in their Ca 2+ -sensitive regulation of ROS-GC1? To answer this question, we performed experiments with ROS membranes at different free [Ca 2+ ] and 2 l M of D 6 S-GCAP-1 or GCAP-2. A typical result is shown in Fig. 3A,B. A striking difference was observed at the free [Ca 2+ ] at which activation of the cyclase is half- maximal. The IC 50 value obtained with GCAP-2 was significantly lower than that obtained with D 6 S-GCAP-1. The results were reproducible with different preparations of bovine ROS and GCAPs: the dose–response curve was alwaysshiftedtolower[Ca 2+ ]withGCAP-2.TheIC 50 value for Ca 2+ determined from two to three independent titration curves for GCAP-2 was 100 ± 32 n M and that for D 6 S-GCAP-1, was 707 ± 122 n M . The dose–response curves were, in both cases, cooperative with Hill coefficients of n ¼ 1.46 ± 0.11 and n ¼ 2.4 ± 0.0 for GCAP-1 and GCAP-2, respectively. These different Ca 2+ -sensitivities of D 6 S-GCAP-1 and GCAP-2 were independent of GCAP concentrations, as similar activation profiles were obtained 1 l M and 10 l M GCAPs. Next, we compared the Ca 2+ -dependency of the ROS- GC1 activity in whole ROS and washed ROS membranes in the presence of both D 6 S-GCAP-1 and GCAP-2 at a ratio of 1 : 1 (Fig. 3C) that corresponds to the cellular ratio in bovine rods. The IC 50 values were similar for both curves (138 n M for whole ROS and 284 n M for reconstituted membranes). Cooperativity was slightly higher in incuba- tions with native ROS. Interestingly, the significant inhibi- tion of ROS-GC1 by GCAP-2 at high [Ca 2+ ] (Fig. 3B) was not seen in whole ROS and was less pronounced in reconstituted membranes, which indicated a compensatory effect by D 6 S-GCAP-1 (or native GCAP-1 in whole ROS). Thus, reconstitution of ROS membranes with GCAPs at physiological concentrations can reproduce the activation profile of ROS-GC1 in native ROS. We conclude from these results that GCAP-1 and GCAP-2 confer different Ca 2+ -sensitivities to ROS-GC1 in bovine rods. We then proceeded to study systems of heterologously expressed ROS-GC1 and purified GCAPs to answer the question whether the reconstitution of the native activation profile requires components specific to native ROS mem- brane preparations or not. Thus, ROS-GC1 was hetero- logously expressed in HEK293 tsA cells, HEK293 cells and COS cells. Cell membranes were reconstituted with D 6 S- GCAP-1 and GCAP-2 and the EC 50 and IC 50 values were determined. EC 50 values were similar as described previ- ously [25,44] and maximal activity at saturating GCAP concentration was similar (data not shown). When we analyzed the activation profile as a function of free [Ca 2+ ], we obtained for ROS-GC1 that was heterologously expressed in HEK293 or HEK tsA cells, similar results as with native ROS-GC1: GCAP-1 and GCAP-2 differed in their Ca 2+ -sensitivities. Figure 4A shows a typical series of experiments with HEK tsA cells. The determined IC 50 values were 609 n M for GCAP-1 and 47 n M for GCAP-2. However, a difference in the IC 50 values was not observed when ROS-GC1 was expressed in COS cells (Fig. 4B). We observed an IC 50 around 100 n M for both GCAPs. It could be that one factor or component is present both in native Fig. 3. ROS-GC1 activity in bovine ROS membranes as a function of free [Ca 2+ ]. Membranes were incubated with a constant amount (2 l M )of(A)D 6 S-GCAP-1 (d)or(B)GCAP-2(s). Crosses (·)are control determinations without added GCAPs. These data were averaged from at least three experiments. The activity unit is nmol cGMPÆmin )1 per mg rhodopsin. The data were fitted by the modified Hill equation; V/V max ¼ –Z[Ca 2+ ] n /([Ca 2+ ] n + K n m )+1; V is the activity of ROS-GC1, V max is the maximal activity of ROS-GC1, n is the Hill cooperativity, K m corresponds to IC 50 of Ca 2+ -dependent ROS-GC1 activity, and Z is a constant taking into account that ROS-GC1 activity is not zero at high free [Ca 2+ ]. IC 50 ¼ 627 n M for D 6 S-GCAP-1 and IC 50 ¼ 123 n M for GCAP-2. (C) Regulation of ROS-GC1 in bovine ROS membranes by the combined action of GCAP-1 and GCAP-2. (m) Washed ROS membranes were reconsti- tuted with 1 l M D 6 S-GCAP-1 and GCAP-2 and the Ca 2+ -dependent activationprofileiscomparedwithdataobtainedwithanativeROS preparation (h). Washed ROS membranes did not exhibit any Ca 2+ - sensitive guanylate cyclase activity (·). 3818 J Y. Hwang et al. (Eur. J. Biochem. 270) Ó FEBS 2003 ROS and HEK cells yet it is missing in COS cells. The dependency on the membrane source for ROS-GC1 could also indicate that cyclase lacks some modification when it is expressed in COS cells for instance and this modification could be necessary to exert different actions of GCAP-1 or GCAP-2, but we have no experimental proof for this speculation. From the results obtained with native membranes, the native reconstituted membranes and with the reconstituted heterologous expression systems of HEK tsa and HEK 293 cells, we conclude that the differences in the Ca 2+ -sensitive regulation of ROS-GC1 by GCAP-1 and GCAP-2 are an intrinsic property of the ROS-GC1/GCAP complex. The effect does not particularly depend on the presence of ROS membranes and therefore does not require a specific component found exclusively in ROS. Discussion A key finding of this study is that the cellular concentration of GCAP-1 and GCAP-2 approximately sums to the cellular concentration of a functional ROS-GC1 dimer (about 6 l M ). As the apparent affinities of GCAPs for native ROS-GC1 are very similar [37], one molecule of GCAP-1 or GCAP-2 could assemble with one ROS-GC1 dimer and form a functional unit. This would lead to two different ROS-GC1 complexes, i.e., ROS-GC1/GCAP-1 and ROS-GC1/GCAP-2. Alternatively, one ROS-GC1 dimer could assemble with one GCAP-1 and one GCAP- 2 leaving a population of ROS-GC1 dimers uncomplexed with GCAPs. An experimental distinction between these possibilities is beyond the scope of this study and will require further examination. However, our results establish that GCAP-1 and GCAP-2 act on ROS-GC1 in different ways. This conclusion is further supported by the following two observations. First, the catalytic efficiency of ROS-GC1 depends on the myristoylation of its regulatory proteins GCAP-1 and GCAP-2. However, both proteins differ in this aspect, as the myristoyl group has a stronger impact in the case of GCAP- 1. So far, it is unclear by which mechanism the myristoyl group contributes to the interaction between both GCAPs and ROS-GC1. The fact that there is a differential impact, it implies that GCAP-1 and GCAP-2 have different target sites in ROS-GC1. Indeed, it has been shown that the target sites of GCAP-1 and GCAP-2 on ROS-GC1 do not overlap [44,45]. Furthermore, Hwang and Koch reported recently, that the myristoyl group controls the Ca 2+ -sensitivity of GCAP-1, but not that of GCAP-2 [36]. We emphasize, that our results add a new aspect to the complex regulatory features of ROS-GC1: the myristoyl group makes the GCAP-dependent increase in catalytic efficiency more powerful and this effect is more pronounced in the case of GCAP-1 than in the case of GCAP-2. A second important aspect of our results concerning the differential regulation of ROS-GC1 activity by GCAP-1 and GCAP-2 is that the two GCAPs are sensitive to different levels of free [Ca 2+ ]. The detection level of GCAP-2 is one order of magnitude lower than that of GCAP-1. However, at intermediate levels, both GCAPs can detect the changes in Ca 2+ intensity. Although the Ca 2+ IC 50 values of ROS-GC1 regulation reported in the literature vary over a large range [4], these differences have not been addressed as specific distinct properties of GCAP-1 and GCAP-2. On the contrary, it has been shown that the Ca 2+ -dependent activation profiles of GCAP-1 and GCAP-2 coincide (Fig. 8C of [21]). Only occasionally have differences between GCAP-1 and GCAP-2 activation profiles been noted but not explicitly discussed [36,46]. Therefore, we investigated this problem in a systematic manner by using different membrane systems as a source for ROS-GC1, i.e., native ROS membranes, HEK293tsA, HEK293 and COS cells. In addition we varied the concentration of GCAPs, when testing the activation profiles. Different Ca 2+ -sensitivities of GCAP-1 and GCAP-2 were observed with all mem- brane systems except COS cells. Thus, we conclude, that a differential Ca 2+ -sensitivity is an important intrinsic property of the ROS-GC1/GCAP complex. One reason, why different Ca 2+ -sensitivities for GCAPs have been overlooked previously, could be that they are still in a very narrow, yet physiological range of free [Ca 2+ ]. Therefore, it was vital for the interpretation of our results that we determined the free [Ca 2+ ] of our buffer solutions with a Ca 2+ -sensitive microelectrode. Taking into account that GCAP-1 and GCAP-2 target different sites in ROS-GC1, we hypothesize that regulation of ROS-GC1 is switched from a ÔGCAP-1 modeÕ to a ÔGCAP-2 modeÕ or vice versa. Shortly after illumination, Fig. 4. Activity of heterologously expressed ROS-GC1 that was reconstituted with GCAPs. (A) Activity of ROS-GC1 expressed in HEK293 tsA cells as a function of free [Ca 2+ ] at a constant amount of either D 6 S-GCAP-1 (d)orGCAP-2(s). IC 50 values were 609 n M (D 6 S-GCAP-1) and 47 n M (GCAP-2). (B) Activity of ROS-GC1 expresssedinCOScellsasafunctionoffree[Ca 2+ ] at a constant amount of either 4 l M D 6 S-GCAP-1 (d)or15l M GCAP-2 (s). IC 50 values were 109 n M (D 6 S-GCAP-1) and 111 n M (GCAP-2). Ó FEBS 2003 Differential Ca 2+ -modulation of guanylate cyclase (Eur. J. Biochem. 270) 3819 whenthefree[Ca 2+ ] begins to fall, GCAP-1 becomes active. When free [Ca 2 + ] further decreases, ROS-GC1 is switched to the ÔGCAP-2 modeÕ. As it is expected, that under constant background light the free [Ca 2+ ] reaches a new steady state value [29], these two modes could also operate at different background light intensities depending onthefree[Ca 2+ ]. Our results may help to explain the findings of a recent study on transgenic mice by Howes et al.[33].Theseauthors showed that expression of GCAP-2 in GCAP null mice led to a phenotype that is clearly different from the wild-type [34]. Flash responses from rods of transgenic mice differed from wild-type responses mainly in two aspects: the fast recovery shortly after the maximum response amplitude was missing and in some cases the flash response ended in a prominent undershoot. The first observation is consistent with a lack of GCAP-1 that would activate cyclase very early after a single flash or when the free [Ca 2+ ]hadjust begun to fall. The observed undershoot indicates a delayed activation of cyclase by GCAP-2. This is also consistent with our result, that GCAP-2 is activated at lower free [Ca 2+ ] than GCAP-1. The operating range of free [Ca 2+ ] for GCAP-2 will be reached later after flash illumination. On the other hand, expression of GCAP-1 can fully reverse the observed effect of the lack of GCAP-1 and GCAP-2. This observation can be explained by our results, that GCAP-1 activates ROS-GC1 over a larger range of free [Ca 2+ ], at higher free [Ca 2+ ] and therefore with an onset far earlier in the timeframe of a single-flash response than GCAP-2 does. GCAP-1 can therefore compensate the lack of GCAP-2. In summary, we have found that both GCAPs are necessary to restore the native [Ca 2+ ]-dependent activity profile of ROS-GC1. Their combined presence in a complex with the catalytic ROS-GC1 dimer may serve to extend the working range of the Ca 2+ -feedback on ROS-GC1 and therefore extend the dynamic range of cyclase regulation under varying light regimes. Acknowledgements This study was supported by the Deutsche Forschungsgemeinschaft (Ko948/5–3) and by USPHS awards EY10828, DC 005349 and HL58151. References 1. Pugh, E.N. Jr & Lamb, T.D. (2000) Phototransduction in verte- brate rods and cones: Molecular mechanisms of amplification, recovery and light adaptation. In Handbook of Biological Physics (Stavenga,D.G.,DeGrip,W.J.&Pugh,E.N.Jr,eds),pp.183– 255. Elsevier Science BV, Amsterdam, the Netherlands. 2. Burns, M.E. & Baylor, D.A. (2001) Activation, deactivation, and adaptation in vertebrate photoreceptor cells. Annu. Rev. Neurosci. 24, 779–805. 3. Kaupp, U.B. & Seifert, R. (2002) Cyclic nucleotide-gated ion channels. Physiol. Rev. 82, 769–824. 4. Pugh, E.N. Jr, Duda, T., Sitaramayya, A. & Sharma, R.K. (1997) Photoreceptor guanylate cyclases: a review. Biosci. Report 17, 429–473. 5. Koch, K W., Duda, T. & Sharma, R.K. (2002) Photoreceptor specific guanylate cyclases in vertebrate phototransduction. Mol. Cell. Biochem. 230, 97–106. 6. Koch, K W. (1991) Purification and identification of photo- receptor guanylate cyclase. J. Biol. Chem. 266, 8634–8637. 7. Hayashi, F. & Yamazaki, A. (1991) Polymorphism in purified guanylate cyclase from vertebrate rod photoreceptors. Proc. Natl Acad.Sci.USA88, 4746–4750. 8. Margulis, A., Goraczniak, R.M., Duda, T., Sharma, R.K. & Sitaramayya, A. (1993) Structural and biochemical identity of retinal rod outer segment membrane guanylate cyclase. Biochem. Biophys. Res. Commun. 194, 855–861. 9. Aparicio, J.G. & Applebury, M.L. (1995) The bovine photo- receptor outer segment guanylate cyclase: Purification, kinetic properties, and molecular size. Protein Expr. Purif. 6, 501–511. 10. Goracniak, R.M., Duda, T., Sitaramayya, A. & Sharma, R.K. (1994) Structural and functional characterization of the rod outer segment membrane guanylate cyclase. Biochem. J. 302, 455–461. 11. Perrault, I., Rozet, J.M., Calvas, P., Gerber, S., Camuzat, A., Dollfus, H., Chaˆ telin, S., Souied, E., Ghazi, I., Leowski, C., Bonnemaison, M., Le Paslier, D., Fre ´ zal, J., Dufier, J L., Pittler, S., Munnich, A. & Kaplan, J. (1996) Retinal-specific guanylate cyclase gene mutations in Leber’s congenital amaurosis. Nat. Genet. 14, 461–464. 12. Kelsell, R.E., Gregory-Evans, K., Payne, A.M., Perrault, I., Kaplan, J., Yang, R B., Garbers, D.L., Bird, A.C., Moore, A.T. & Hunt, D.M. (1998) Mutations in the retinal guanylate cyclase (RETGC-1) gene in dominant cone-rod dystrophy. Hum. Mol. Genet. 7, 1179–1184. 13.Duda,T.,Krishnan,A.,Venkataraman,V.,Lange,C.,Koch, K W. & Sharma, R.K. (1999) Mutations in the rod outer segment membrane guanylate cyclase in a cone-rod dystrophy cause defects in calcium signaling. Biochemistry 38, 13912–13919. 14. Duda, T., Venkataraman, V., Goraczniak, R., Lange, C., Koch, K W. & Sharma, R.K. (1999) Functional consequences of a rod outer segment membrane guanylate cyclase (ROS-GC1) gene mutation linked with Leber’s congenital amaurosis. Biochemistry 38, 509–515. 15. Tucker, C.L., Woodcock, S.C., Kelsell, R.E., Ramamurthy, V., Hunt, D.M. & Hurley, J.B. (1999) Biochemical analysis of a dimerization domain mutation in RetGC-1 associated with dominant cone-rod dystrophy. Proc. Natl Acad. Sci. USA 96, 9039–9044. 16. Duda, T., Venkataraman, V., Jankowska, A., Lange, C., Koch, K W. & Sharma, R.K. (2000) Impairment of the rod outer seg- ment membrane guanylate cyclase dimerization in A cone-rod dystrophy results in defective calcium signaling. Biochemistry 39, 12522–12533. 17. Gregory-Evans, K., Kelsell, R.E., Gregory-Evans, C.Y., Downes, S.M., Fitzke, F.W., Holder, G.E., Simunovic, M., Mollon, J.D., Taylor, R., Hunt, D.M., Bird, A.C. & Moore, A.T. (2000) Autosomal dominant cone-rod retinal dystrophy (CORD6) from heterozygous mutation of GUCY23D, which encodes retinal guanylate cyclase. Ophthalmology 107, 55–61. 18. Dizhoor, A.M., Lowe, D.G., Olshevskaya, E.V., Laura, R.P. & Hurley, J.B. (1994) The human photoreceptor membrane guanylyl cyclase, RetGC, is present in outer segments and is regulated by calcium and a soluble activator. Neuron 12, 1345–1352. 19. Palczewski, K., Subbaraya, I., Gorczyca, W.A., Helekar, B.S., Ruiz, C.C., Ohguro, H., Huang, J., Zhao, X., Crabb, J.W., Johnson, R.S., Walsh, K.A., Gray-Keller, M.P., Detwiler, P.B. & Baehr, W. (1994) Molecular cloning and characterization of reti- nal photoreceptor guanylyl cyclase-activating protein. Neuron 13, 395–404. 20. Dizhoor, A.M., Olshevskaya, E.V., Henzel, W.J., Wong, S.C., Stults, J.T., Ankoudinova, I. & Hurley, J.B. (1995) Cloning, sequencing, and expression of a 24-kDa Ca 2+ -binding protein activating photoreceptor guanylyl cyclase. J. Biol. Chem. 270, 25200–25206. 3820 J Y. Hwang et al. (Eur. J. Biochem. 270) Ó FEBS 2003 21. Gorczyca, W.A., Polans, A.S., Surgucheva, I.G., Subbaraya, I., Baehr, W. & Palczewski, K. (1995) Guanylyl cyclase activating protein. J. Biol. Chem. 270, 22029–22036. 22. Frins, S., Bo ¨ nigk, W., Mu ¨ ller, F., Kellner, R. & Koch, K W. (1996) Functional characterization of a guanylyl cyclase- activating protein from vertebrate rods. J. Biol. Chem. 271, 8022– 8027. 23. Haeseleer, F., Sokal, I., Li, N., Pettenati, M., Rao, N., Bronson, D., Wechter, R., Baehr, W. & Palczewski, K. (1999) Molecular characterization of a third member of the guanylyl cyclase-acti- vating protein subfamily. J. Biol. Chem. 274, 6526–6535. 24. Cuenca, N., Lopez, S., Howes, K. & Kolb, H. (1998) The locali- zation of guanylyl cyclase-activating proteins in the mammalian retina. Invest. Ophthalmol. Vis. Sci. 39, 1243–1250. 25. Otto-Bruc, A., Fariss, R.N., Haeseleer, F., Huang, J., Buc- zylko, J., Surgucheva, I., Baehr, W., Milam, A.H. & Palc- zewski, K. (1997) Localization of guanylate cyclase-activating protein 2 in mammalian retinas. Proc.NatlAcad.Sci.USA94, 4727–4732. 26. Imanishi, Y., Sokal, I., Sowa, M.E., Lichtarge, O., Wensel, T.G., Saperstein, D.A., Baehr, W. & Palczewski, K. (2002) Character- ization of retinal guanylate cyclase-activating protein 3 (GCAP3) from zebrafish to man. Eur. J. Neurosci. 15, 63–78. 27. Ratto, G.M., Payne, R., Owen, W.G. & Tsien, R.Y. (1988) The concentration of cytosolic free calcium in vertebrate rod outer segments measured with Fura-2. J. Neurosci. 8, 3240–3246. 28. Korenbrot, J.I. & Miller, D.L. (1989) Cytoplasmic free calcium concentrationindark-adaptedretinalrodoutersegments.Vision Res. 29, 939–948. 29. Gray-Keller, M.P. & Detwiler, P.B. (1994) The calcium feedback signal in the phototransduction cascade of vertebrate rods. Neuron 13, 849–861. 30. McCarthy, S.T., Younger, J.P. & Owen, W.G. (1996) Dynamic, spatially nonuniform calcium regulation in frog rods exposed to light. J. Neurophysiol. 76, 1991–2004. 31. Howes, K., Bronson, J.D., Dang, Y.L., Li, N., Zhang, K., Ruiz, C.,Helekar,B.,Lee,M.,Subbaraya,I.,Kolb,H.,Chen,J.& Baehr, W. (1998) Gene array and expression of mouse retina guanylate cyclase activating proteins 1 and 2., Invest. Ophthalmol. Vis. Sci. 39, 867–875. 32. Kachi, S., Nishizawa, Y., Olshevskaya, E., Yamazaki, A., Miyake, Y., Wakabayashi, T., Dizhoor, A. & Usukura, J. (1999) Detailed localization of photoreceptor guanylate cyclase activating protein- 1 and -2 in mammalian retinas using light and electron micro- scopy. Exp. Eye Res. 68, 465–473. 33. Howes, K.A., Pennesi, M.E., Sokal, I., Church-Kopish, J., Sch- midt, B., Margolis, D., Frederick, J.M., Rieke, F., Palczewski, K., Wu, S.M., Detwiler, P.B. & Baehr, W. (2002) GCAP1 rescues rod photoreceptor response in GCAP1/GCAP2 knockout mice. EMBO J. 21, 1545–1554. 34. Mendez,A.,Burns,M.E.,Sokal,I.,Dizhoor,A.M.,Baehr,W., Palczewski, K., Baylor, D.A. & Chen, J. (2001) Role of guanylate cyclase-activating proteins (GCAPs) in setting the flash sensitivity of rod photoreceptors. Proc.NatlAcad.Sci.USA98, 9948–9953. 35. Koch, K W., Lambrecht, H G., Haberecht, M., Redburn, D. & Schmidt, H.H.H.W. (1994) Functional coupling of a Ca 2+ /cal- modulin-dependent nitric oxide synthase and a soluble guanylyl cyclase in vertebrate photoreceptor cells. EMBO J. 13, 3312–3320. 36. Hwang, J Y. & Koch, K W. (2002) Calcium- and myristoyl- dependent properties of guanylate cyclase-activating protein-1 and protein-2. Biochemistry 41, 13021–13028. 37. Hwang, J Y. & Koch, K W. (2002) The myristoylation of the neuronal Ca 2+ -sensors guanylate cyclase-activating protein 1 and 2. Biochim. Biophys. Acta 1600, 111–117. 38. Krylov, D.M., Niemi, G.A., Dizhoor, A.M. & Hurley, J.B. (1999) Mapping sites in guanylyl cyclase activating protein-1 reguired for regulation of photoreceptor membrane guanylyl cyclases. J. Biol. Chem. 274, 10833–10839. 39. Schaffner, W. & Weissmann, C. (1973) A rapid, sensitive, and specific method for the determination of protein in dilute solution. Anal. Biochem. 56, 502–514. 40. Pace, C.N. & Schmid, F.X. (1997) How to determine the molar absorbance coefficient of a protein. In Protein Structure (Creighton, T.E., ed.), pp. 253–259. Oxford University Press, Oxford. 41. Segel, I.H. (1993) Enzyme Kinetics. John Wiley & Sons, Inc., New York. 42. Yu, H., Olshevskaya, E., Duda, T., Seno, K., Hayashi, F., Sharma, R.K.,Dizhoor,A.M.&Yamazaki,A.(1999)Activationofretinal guanylyl cyclase-1 by Ca 2+ -binding proteins involves its dimeri- zation. J. Biol. Chem. 274, 15547–15555. 43. Ramamurthy, V., Tucker, C., Wilkies, S.E., Daggett, V., Hunt, D.M. & Hurley, J.B. (2001) Interactions within the coiled-coil domain of RetGC-1 guanylyl cyclase are optimized for regulation rather than for high affinity. J. Biol. Chem. 276, 26218–26229. 44. Krishnan, A., Goraczniak, R.M., Duda, T. & Sharma, R.K. (1998) Third calcium-modulated rod outer segment membrane guanylate cyclase transduction mechanism. Mol. Cell. Biochem. 178, 251–259. 45. Lange,C., Duda, T., Beyermann,M., Sharma,R.K.&Koch, K W. (1999) Regions in vertebrate photoreceptor guanylyl cyclase ROS-GC1 involved in Ca 2+ -dependent regulation by guanylyl cyclase-activating protein GCAP-1. FEBS Lett. 460, 27–31. 46. Dizhoor, A.M., Boikov, S.G. & Olshevskaya, E.V. (1998) Con- stitutive activation of photoreceptor guanylate cyclase by Y99C mutant of GCAP-1. J. Biol. Chem. 273, 17311–17314. Ó FEBS 2003 Differential Ca 2+ -modulation of guanylate cyclase (Eur. J. Biochem. 270) 3821 . Regulatory modes of rod outer segment membrane guanylate cyclase differ in catalytic efficiency and Ca 2+ -sensitivity Ji-Young. calibration and the amount of cyclase in ROS was obtained directly from the calibration curve. The predominant membrane guanylate cyclase in bovine ROS is