CCTeta, a novel soluble guanylyl cyclase-interacting protein.

Nitric oxide (NO) transduces most of its biological effects through activation of the heterodimeric enzyme, soluble guanylyl cyclase (sGC). Activation of sGC results in the production of cGMP from GTP. In this paper, we demonstrate a novel protein interaction between CCT (chaperonin containing t-complex polypeptide) subunit eta and the alpha1beta1 isoform of sGC. CCTeta was found to interact with the beta1 subunit of sGC via a yeast-two-hybrid screen. This interaction was then confirmed in vitro with a co-immunoprecipitation from mouse brain. The interaction between these two proteins was further supported by a co-localization of the proteins within rat brain. Using the yeast two-hybrid system, CCTeta was found to bind to the N-terminal portion of sGC. In vitro assays with purified CCTeta and Sf9 lysate expressing sGC resulted in a 30-50% inhibition of diethylamine diazeniumdiolate-NO-stimulated sGC activity. The same assays were then performed using BAY41-2272, an NO-independent allosteric sGC activator, and CCTeta had no effect on this activity. Furthermore, CCTeta had no effect on basal or sodium nitroprusside-stimulated alphabeta(Cys-105) sGC, a constitutively active mutant that only lacks the heme group. The N-terminal 94 amino acids of CCTeta seem to be critical for the mediation of this inhibition. Lastly, a 45% inhibition of sGC activity by CCTeta was seen in vivo in BE2 cells stably transfected with CCTeta and treated with sodium nitroprusside. These data suggest that CCTeta binds to sGC and, in cooperation with some other factor, inhibits its activity by modifying the binding of NO to the heme group or the subsequent conformational changes.

cGMP from GTP (1)(2)(3). Cyclic GMP is then involved in the activation of a variety of effectors such as cyclic nucleotidegated channels, protein kinases, and phosphodiesterases (4).
There has been evidence suggesting that there are in vivo mechanisms other than nitric oxide that regulate sGC activity. Bellamy et al. (5) show that deactivation of sGC on the removal of NO occurred 25-fold faster in intact cerebellar cells than the fastest estimate for purified sGC. These data suggest that there is probably another protein or group of proteins that regulate sGC activity, because it occurs in the order of seconds.
In addition to the indirect evidence of protein regulation of sGC provided by Bellamy et al. (5), PSD95 (post-synaptic density protein 95) (6) and heat shock protein 90 (HSP90) (7) have both been shown to interact with isoforms of soluble guanylyl cyclase. Both PSD95 and HSP90 were shown to localize different isoforms of sGC to the membrane. Additionally, HSP90 was shown to enhance the response of sGC to NO. In both of these cases, the interaction with sGC is modulated by interaction with a nitric-oxide synthase isoform. We hypothesized that a protein existed that modified sGC activity, independent of any interaction with nitric-oxide synthase isoforms.
To find this regulator of sGC activity, we performed a yeast two-hybrid screening with the ␤ 1 subunit of sGC as bait against a human brain cDNA library. The screening revealed CCT (chaperonin containing t-complex polypeptide) subunit to be an interacting protein. Immunoprecipitation and immunohistochemical studies suggest that such interaction occurs in vivo. This interaction inhibited NO-stimulated sGC activity both in vitro and in vivo, whereas basal activity or activation by allosteric regulator was not affected by CCT. CCT had no effect on the constitutively active ␣␤ Cys-105 sGC mutant that lacks a heme group. These studies suggest that CCT, in cooperation with some other factor, can mediate a novel type of NO-dependent inhibition of sGC.

MATERIALS AND METHODS
All of the materials were supplied by Sigma unless otherwise specified. Yeast Two-hybrid Screen-The full-length cDNA of the ␤ 1 subunit of sGC was cloned in-frame with the GAL4 DNA binding domain of the pGBKT7 vector between the EcoRI and SalI sites (pGB␤). This vector was transfected into AH109 yeast cells (Clontech). A human brain cDNA library cloned in-frame with the GAL4 activation domain of the pACT2 vector was transfected into yeast containing the pGB␤ vector. The resulting transfected yeast were then plated on quadruple dropout agar (-Trp/-Leu/-His/-Ade) with 5 mM 3-aminotriazole. The yeast were then allowed to grow for 3 weeks at 30°C at which time the positive clones that were greater than 2 mm in diameter were picked and grown in Leu-deficient medium overnight. The yeast were then lysed, and the pACT2 vectors containing the target proteins (pACT-prot) were amplified in TOP10 cells (Invitrogen). The pACT-prot were then transfected into AH109 containing pGB␤ and replated on quadruple dropout agar. 〉-Galactosidase filter lift assays (Clontech) were then performed on these clones to try and reduce the number of false positives. The positive clones from this screening were then transfected into AH109 and AH109 containing pGBKT7 as controls to see whether any of these positive clones could activate the reporter gene in AH109 by itself or in conjunction with the binding domain alone. This step further reduced the number of false positives.
␣-Galactosidase Assay-The plasmids with full-length and truncated ␤ 1 subunit (pGB␤, pGB␤⌬ 366 -619 , pGB␤⌬ 266 -619 , and pGB␤⌬ 100 -619 ) were transfected into AH109 containing pACT-CCT and grown in selective liquid culture(-Trp/-Leu) overnight. The yeast were then centrifuged at 10,000 ϫ g for 2 min, and 8 l of the conditioned growth medium were mixed with 24 l of assay buffer (33 mM p-nitrophenyl-␣-D-galactoside and 333 mM sodium acetate). This mixture was then incubated in a water bath of 30°C for 30 min, and the reaction was stopped by the addition of 960 l of stop solution (100 mM sodium carbonate). The optical density of each of these samples was then measured at 410 nm, and ␣-galactosidase activity was estimated. One unit of ␣-galactosidase is defined as the amount of enzyme that hydrolyzes 1 mol of p-nitrophenyl-␣-D-galactoside to p-nitrophenol and Dgalactose in 1 min at 30°C in acetate buffer at pH 4.5, as described in the Clontech yeast two-hybrid protocol.
Cloning, Expression, and Purification of CCT with Histidine Tag-Full-length CCT was amplified out of brain cDNA library (Clontech) using the 5Ј primer, 5Ј-TTAGAATTCCACATGATGCCCACACCAGTT-ATCCTATTGAAA-3Ј, and 3Ј primer, 5Ј-TGGCGCCGGCGTCAGTGGG-GGCGGCCACG-3Ј, and TaqDNA polymerase (Invitrogen). The resultant 1.5-kb PCR product contained the full-length CCT and was cloned into a TOPO2.1 vector. The wild type CCT gene alone codes for a protein with a molecular mass of a 45-kDa protein. The CCT ORF was then cut with EcoRI and ligated into pET 28a in-frame with an Nterminal His 6 tag. This His 6 -tagged CCT has a molecular mass of ϳ75 kDa. The CCT-pET 28a construct was then transformed into BL21 DE3 pLysS cells and selected on kanamycin (50 g/ml) LB plates. A colony from the plate was picked and grown overnight in culture. 10 ml of culture was then taken and diluted to an A 600 of 0.02 in 300 ml of fresh LB. After the A 600 reached 0.4, 1.5 mM isopropyl-1-thio-␤-D-galactopyranoside was added. The culture was then shaken for another 9 h, after which time the bacteria were harvested by centrifugation at 10,000 ϫ g for 15 min. The supernatant was discarded, and the cells were resuspended in hypotonic phosphate-buffered saline (10 mM sodium phosphate, 30 mM sodium chloride, pH 7.4) with 20 mM Hepes, 100 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, and 200 g/ml lysozyme and left on ice for 30 min. Following sonication, the lysate was centrifuged at 15,000 ϫ g for 45 min. The supernatant was then passed through a 5-ml packed column of nickel-tagged agarose (Novagen). The column was then washed with 12 column volumes of phosphate-buffered saline (50 mM sodium phosphate, 150 mM sodium chloride, pH 7.4) followed by 10 column volumes of 100 mM imidazole in PBS and eluted with 3 column volumes of 250 mM imidazole in PBS. This elution was then concentrated with an Amicon Y m -30 column and used for GTP conversion assays. The control ␤-galactosidase protein was purified in a similar manner.
Western Blots-Protein samples were resolved on 7.5% SDS-polyacrylamide gels and transferred to methanol-activated polyvinylidene difluoride membranes. After blocking the membranes, the following antibodies were used for detection: rabbit polyclonal soluble guanylyl cyclase antibodies (Cayman) at a 1:5,000 dilution; clone B4 of mouse monoclonal soluble guanylyl cyclase antibodies (8); mouse monoclonal His 6 antibody (Invitrogen) at a 1:3,000 dilution, and rabbit polyclonal CCT antibodies (a gift from Dr. Willison at the Institute of Cancer Research, Chester Beatty Laboratories, London, United Kingdom) characterized in detail previously (9) at a 1:10,000 dilution. Secondary horseradish peroxidase-labeled anti-mouse IgG and anti-rabbit IgG antibodies (Sigma) along with ECLϩ Western blotting detection reagents (Amersham Biosciences) were used for visualization of the bands.
Immunohistochemistry and Confocal Microscopy-Adult rats were terminally anesthetized with chloroform and perfused transcardially with 500 ml of ice-cold PBS followed by 500 ml of 4% paraformaldehyde in PBS. Brains were removed, post-fixed in 4% paraformaldehyde in PBS for 3-4 h, and immersed in 20% sucrose overnight. Serial sections of 100 m were cut in the coronal plane using an HM505E Microm cryostat and collected in PBS. Sections were blocked and permeabilized for 1 h in PBS containing 5% goat serum and 0.5% Triton X-100. Primary antibodies (1:300 for B4 mouse monoclonal sGC antibodies and 1:1000 for CCT rabbit polyclonal antibody) were diluted in the same solution and incubated with sections of rat brain overnight at 4°C.
Control sections were not incubated with any primary antibody. All of the sections were then washed three times for 10 min in PBS and incubated with goat anti-rabbit Alexa-568-coupled antibodies and goat anti-mouse Alexa-488-coupled antibodies at a dilution of 1:500 for 3 h at 4°C. The sections were washed three times for 10 min in PBS and then mounted.
Confocal microscopy was done on a Zeiss LSM-410. Sections through the tissue were imaged at 1 m. Eight sections were recorded per tissue slice. Brightness and contrast of images were adjusted in Adobe Photoshop (Adobe Systems).
Preparation of Lysates for Immunoprecipitation and CCT Expression-The mouse organs used for determination of CCT expression were homogenized in 4 ml of PBS with 0.5 mM phenylmethylsulfonyl fluoride, 0.2 mM benzamidine, and 1 M benzamidine. The whole cell lysate protein concentration was determined via Bradford assay. The mouse brains used for co-immunoprecipitation were prepared in the same buffer and then following homogenization, the lysate was centrifuged at 2,900 ϫ g for 20 min at 4°C. The supernatant was then taken and centrifuged for 45 min at 29,000 ϫ g and 4°C.
Co-immunoprecipitation-20 l of a 50% protein G-Sepharose beads (Amersham Biosciences) slurry was added to control and experimental tubes. 20 l of ascite serum with rat monoclonal CCT antibody (Stressgen) or rat preimmune serum were added to each of the two tubes and allowed to bind to the beads for 15 min at room temperature. 30 mM Phe-Arg and Phe-Leu dipeptides were added to the beads to block nonspecific binding sites and incubated for another 15 min at room temperature. The mixture was spun down for 5 s, and the supernatant was discarded. 940 l of PBS was then added to each tube along with 780 g of mouse brain lysate prepared as described above and incubated with constant mixing at 4°C overnight. The tubes were then spun down at 10,000 ϫ g and washed with 500 l of PBS five times. Immunoprecipitated proteins were then eluted with 50 l of Laemmli buffer. 20 l of the eluted samples and 5 l of lysate were loaded and resolved by SDS-PAGE and subjected to immunoblot analysis as described above.
Human sGC Expression and Purification-Human recombinant sGC expression and lysate preparation were done using the SF9 baculovirus system previously described (10).
GTP Conversion Assay-Purified CCT (3.25 g) and either sGCexpressing Sf9 lysate (5-8 g) or ␣␤ Cys-105 mutant sGC-expressing Sf9 lysate (5-8 g) or purified sGC (1.2 g) were incubated together in 20 l of sample buffer (2 mg/ml albumin, 50 mM triethanolamine, and 100 M EGTA) for 10 min at room temperature. Following the incubation, 40 l of reaction buffer (125 mM triethanolamine, 250 M EGTA, 2.5 mM IBMX, 2.5 mg/ml albumin, 2.5 mM cGMP, 0.125 mg/ml creatine kinase, 12.5 mM creatine phosphate, and 7.5 mM magnesium chloride) were added. To the mixture of sample buffer and incubation buffer, 40 l of substrate buffer containing various concentrations of DEA-NO (as indicated) or 100 M SNP or 2 M BAY and 200 M GTP/0.01 Ci of [␣-32 P]GTP were added. This mixture was then incubated in a circulating water bath at 37°C for 15 min. The reaction was stopped by the addition of 1 ml of 100 mM zinc acetate and 1 ml of 120 mM sodium carbonate. The tubes were then centrifuged at 1,000 ϫ g for 5 min, and the supernatant was loaded on to 1.5-ml aluminum columns. Synthesized cGMP was eluted with 10 ml of 100 mM Tris, pH 7.5, and then quantified on ␤ scintillation counter.
Cell Culture and Stable Transfection-FLAG-tagged CCT was amplified using the 5Ј primer, 5Ј-TAAGCTAGCCACATGATGGAT-TACAAGGATGACGACGATAAGCCCACACCAGTTATCCTATTGAAA-3Ј, and 3Ј primer, 5Ј-TGGCGCCGGCGTCAGTGGGGGCGGCCACG-3Ј, and cloned into the mammalian expression vector pMGH2 (Invivogen). BE2 cells (ATCC number CRL-2268) were plated at a density of 5 ϫ 10 5 cells/cm 2 in cell culture medium (50% Dulbecco's modified Eagle's medium, 50% Ham's F-12, 10% fetal bovine serum, 2.5 mM L-glutamine, 1.5 g/liter sodium bicarbonate, 0.5 mM sodium pyruvate, and 0.05 mM non-essential amino acids). Four 100-mm plates of 80% confluent BE2 cells were transfected using LipofectAMINE-plus (Invitrogen) according to the manufacturer's instructions. Two were transfected with pMG-CCT plasmid, and two were transfected with pMGH2 vector alone. To select the stable transfected clones, media with 400 g/ml hygromycin B (Calbiochem) were added to each of the plates every 4 days for 3 weeks. Stable colonies were then picked, transferred to fresh plates, and grown to confluency with 300 g/ml hygromycin B.
RNA Isolation and Reverse Transcriptase Reaction-RNA was isolated from confluent plates of BE2 cells stably transfected with a pMG-CCT construct and pMGH2 vector using RNeasy kit (Qiagen) according to manufacturers. The expression of FLAG-tagged CCT was tested by a reverse transcriptase reaction using the following primers: 5Ј-CA-CCCACCTCAGCATCTTGGGAT-3Ј (CCT-specific) and 5Ј-ATGATGG-ATTACAAGGATGACGACG-3Ј (FLAG-specific).
cGMP Tritium Radioimmunoassay of Stable Clones-The BE2 stables expressing CCT and control cells were split at 50% confluence and replaced with medium that did not contain hygromycin B. The stables were at 80 -90% confluence when the assay was conducted. The medium described by the ATCC for the BE2 cells was taken off the cells and replaced with Dulbecco's modified Eagle's medium (Sigma) with 2 mM IBMX for 1 h at 37°C with 5% CO 2 . 1 mM SNP was then added to the medium, and cells were incubated for another 15 min at 37°C with 5% CO 2 . This medium was removed, and the cells were washed twice with sterile PBS with 2.5 mM IBMX. The cells were trypsinized and harvested. The cells were centrifuged at 1000 ϫ g for 5 min, and the supernatant was discarded. The cells were resuspended in 125 mM triethanolamine, 250 M EGTA, and 2.5 mM IBMX. The cells were then sonicated and centrifuged at 12,000 ϫ g for 15 min. 80 l of supernatant, as determined by the Bradford assay, of the control and CCTexpressing cells were used for cGMP radioimmunoassay. The cGMP tritium radioimmunoassay (Amersham-TRK500) was done according the manufacturer's instructions.

RESULTS
Yeat Two-hybrid Screening-To find the putative protein(s) that interacts with human sGC, we performed a yeast twohybrid screen using human brain yeast two-hybrid cDNA library (Clontech). Twelve million cDNAs were screened against full-length human ␤ 1 protein, and 334 of these clones grew on selective medium. 12 of these clones were positive for the Xgalactosidase filter lift assay. Sequence analysis revealed two open reading frames of 1315 bases and 1350 bases in length, both of which were CCT. The specificity of the interaction was then confirmed by transfecting the activation domain plasmid containing the CCT gene into yeast with a plasmid encoding just the Gal4 binding domain without bait and also into yeast with no plasmid. Both of these transfections did not grow on selective media, indicating that the interaction is specific in yeast (data not shown).
Because CCT is a novel interacting partner of ␣ 1 ␤ 1 sGC identified via a yeast-two-hybrid screen, we next wanted to determine the expression pattern of CCT versus the ubiquitously expressed ␣ 1 ␤ 1 sGC. We found that CCT was abundant in many tissues. CCT was not expressed at all in the small intestine, but it was expressed in the testis, heart, brain, kidneys, and lungs (Fig. 1).
sGC Co-immunoprecipitates with CCT-After the expression pattern of CCT was determined, we tested the interaction between sGC and CCT by co-immunoprecipitation. Immunoprecipitation of CCT protein from the C57BL6 mouse brain revealed the co-precipitation of sGC. On the contrary, when immunoprecipitation was performed with the preimmune serum, neither CCT nor sGC was precipitated, confirming the specificity of sGC/CCT precipitation (Fig. 2). Based on the standard curves obtained from the Western blotting of purified CCT and sGC, we calculated that ϳ6% of the total CCT was precipitated together with ϳ2% of the ␤ subunit of sGC (data not shown).  Co-localization of sGC and CCT-After confirming the in vivo interaction between sGC and CCT by immunoprecipitation, we next decided to visualize the co-localization of CCT and sGC in rat brain by immunohistochemistry and confocal microscopy (Fig. 3). As seen in the image from all of the magnifications (Fig. 3, C, F, and I) CCT and sGC proteins do co-localize in the rat hippocampus along the dentate gyrus. From the highest magnification, it appears that sGC has cytosolic and perinuclear localization, whereas CCT localizes mainly in the cytosolic compartment (Fig. 3H). Such subcellular localization suggests that only the cytosolic fraction of sGC enzyme could efficiently interact with CCT protein.
Region of ␤ 1 Interacting with CCT-We next determined the region of the ␤ 1 subunit of sGC that interacts with CCT. As seen in Fig. 4, the first 100 amino acids of ␤ 1 subunit are not sufficient to interact with CCT protein. However, the truncated ␤ 1 subunit carrying 266 residues interacts with CCT with the same efficiency as the truncated ␤ 1 subunit with 366 residues. The highest strength of interaction was observed when full-length ␤ 1 subunit was tested. Thus, these data indi-cate that there are at least two regions in the ␤ 1 subunit, one between the residues 101 and 266 and another between residues 367 and 619, that are interacting with CCT protein.
CCT Inhibits NO-stimulated sGC Activity-To determine whether or not CCT has an effect on sGC activity, we expressed a full-length CCT tagged with hexahistidine at the N terminus in BL21 bacterial strain. Hexahistidinetagged ␤-galactosidase was also expressed in the same strain and used as negative control. The results shown in Fig. 5 indicate that ␤-galactosidase and buffer have a similar effect on DEA-NO-stimulated cGMP formation activity of sGC, whereas CCT inhibits this cGMP formation by 30 -50% in the 10 -100 M range of DEA-NO (Fig. 5). Furthermore, CCT seems to inhibit sGC-expressing Sf9 lysate by decreasing the V max of NO stimulated-sGC and not by affecting the EC 50 of sGC for NO. Note that there is no difference in cGMP formation activities among buffer, ␤-galactosidase, and CCT under basal conditions. When BAY41-2272 was used to stimulate the sGC-expressing Sf9 lysate, no difference in cGMP formation was seen between controls and CCT (data not shown). The fact that CCT inhibited NO-stimulated sGC, a heme-dependent agonist, but did not inhibit BAY-stimulated sGC, an NO-independent allosteric activator, or any of the basal activities of sGC implies that the mechanism of CCT inhibition may work through a modification of the NO-stimulated heme group or the consequent conformational changes that occur after NO binds to the heme group.
This mechanism of CCT inhibition of NO-stimulated sGC was further supported by the effect seen on ␣␤ Cys-105 sGC. This mutant lacks the heme moiety because of the substitution of His-105 residue coordinating heme iron and is insensitive to NO stimulation (11). However, it was shown that ␣␤ Cys-105 sGC has a constitutively increased activity, even in the absence of any stimulators, suggesting that the mutant sGC is already in an activated conformation (11). These properties of the mutant enzyme are also summarized in the Fig. 6A. Preincubation of the mutant enzyme with CCT has no effect on cGMP-forming activity of basal or NO-treated ␣␤ Cys-105 sGC similar to the control buffer or ␤-galactosidase protein (Fig. 6B). However, preincubation of wild type enzyme with CCT inhibit the SNPstimulated activity versus buffer and ␤-galactosidase controls. This lends further evidence to the hypothesis that CCT inhibits sGC activity through modification of an NO-activated heme group or a modification of the conformational changes induced by nitrosyl heme.
N-terminal Portion of CCT Is Crucial for sGC Inhibition-Further characterization of the inhibitory effect of CCT on sGC was done via deletional analysis. The original 1,350-bp coding region of CCT, retrieved from yeast two-hybrid screen, was subcloned into the pGEX5X-3 vector to express the GST fusion product, GST-CCT frag , in bacteria. This GST-fused fragment of CCT lacked 94 N-terminal residues. We compared the effects of this truncated CCT frag with the full-length CCT on the activity of SNP-stimulated sGC (Fig. 7). Bacterial lysateexpressing GST-CCT frag product decreased sGC activity by only 20% in comparison with the lysate-expressing control GST protein. However, the full-length CCT inhibited the activity of sGC-expressing Sf9 lysate by 68% (Fig. 7).
CCT Alone Requires Other Factors to Inhibit NOstimulated sGC-We next compared the effects of purified CCT on pure sGC. Interestingly, we found that purified CCT had no inhibitory effect on pure sGC unless Sf9 lysate was provided (Fig. 8). Thus, it appears that CCT plays a role of the adaptor protein, which recruits additional factors necessary for sGC inhibition.
In Vivo Inhibition of sGC-After showing the inhibitory ef- fect of CCT on sGC in crude lysates, we wanted to see whether this inhibition occurred in intact cell as well. Full-length CCT with an N-terminal FLAG was cloned into a mammalian expression vector, pMGH2. This pMG-CCT construct, along with the vector alone, was then transfected into BE2 cells, a neuroblastoma line that endogenously expresses sGC (11). The stable lines cells were selected on hygromycin for 3 weeks. Several Hyg R colonies were expanded, and the mRNA was purified. The expression of FLAG-tagged CCT in these stable lines was confirmed by reverse transcriptase-PCR specific only for flagged-CCT (Fig. 9A). Western blot analysis indicated that the expression level of sGC enzyme was not affected in these lines (Fig. 9B). We also performed Western blots to check the expression level of sGC in the stable line expressing CCT versus the control line with vector alone and found the expression to be approximately equal (Fig. 9B).
After establishing successful expression of CCT in BE2, in vivo cGMP accumulation was measured via radioimmunoassay to determine the effect of CCT on sGC in vivo. Both the control cell line and the CCT-expressing cell line (BE2-CCT) were treated with IBMX or 100 M SNP and IBMX for 15 min at 37°C with 5% CO 2 . The cells were then lysed, and cGMP level was estimated via cGMP radioimmunoassay. The results, shown in Fig. 9C, illustrate that CCT does inhibit sGC in vivo upon activation but has no affect on the basal activity of sGC. This concurs with the in vitro data presented in Fig. 5, which   FIG. 4. Defining the region of interaction between the ␤ 1 subunit of sGC and CCT. Progressive C-terminal deletions of the sGC ␤ 1 subunit were cloned in-frame to the GAL4 binding domain of the bait vector pGBKT7 of the yeast twohybrid system. The full-length CCT was cloned in-frame to the GAL4 activation domain of the target pACT vector of the yeast two-hybrid system. Both bait and target vectors were transfected into the AH109 strain of yeast, and the strength of the interaction was measured in vivo by the ␣-galactosidase assay. Interaction of p53 and the small T-antigen was used as positive control for in vivo interaction, whereas interaction of sGC ␤ 1 subunit and actin was used as negative control. Data are presented as the mean Ϯ S.D. of two independent experiments performed in tetraplicates. also indicated that CCT inhibits SNP-stimulated sGC but not the basal activity of sGC. Furthermore, the extent of observed inhibition is very similar in both cases. DISCUSSION To date, there have been no reports of a systematic search for proteins interacting with sGC, let alone affecting its activity.
The yeast two-hybrid screening described in this report identified the subunit of CCT as a novel protein directly interacting with sGC. The association of CCT with sGC was demonstrated not only through interaction in yeast but also through immunoprecipitation (Fig. 2) and colocalization in the hippocampal area of the brain (Fig. 3).
CCT is a eukaryotic cytosolic protein involved in the folding of other proteins. It is composed of two rings, one on top of the other, forming a cylinder in which the environment is amenable to protein folding. Each of the rings is composed of eight different 60-kDa subunits: ␣, ␤, ␦, ⑀, ␥, , , and (9). Actin and tubulin have been identified so far as two major substrates of CCT (13).
The CCT complex has the same function as the GroEs⅐GroEL complex in that they are both formed by two oligomeric rings to create a closed compartment in which protein folding can occur (14). In light of this fact, the reason for having eight different subunits to do the job of protein folding where two seem to suffice indicates that each of the subunits may have individual functions within the CCT complex or that these subunits may provide some specificity to the function of CCT complex. There is precedence for subunits of the CCT complex to have specific functions. CCT␣ has been detected in microtubule structures, and when antibodies to CCT␣ are added in vitro, microtubule polymerization is inhibited (15). CCT has been shown to be important for yeast morphogenesis (16). Interestingly, the subunits of CCT do not always display the same cellular localization, e.g. CCT has been specifically detected in the soluble fraction of Tetrahymena homogenate, whereas CCT␣ and CCT⑀ were detected only in the particulate fraction associated with microtubules (17,18). This finding suggests that the subunits of the CCT complex do not necessarily always form an oligomer or that this oligomer is not always composed of all eight subunits.
Numerous GST pull-down assays were attempted with GSTtagged CCT and purified sGC. However, sGC had a high affinity not only to GST alone but also to glutathione column used for these purposes and was consistently detected in control samples (data not shown). The only conditions that decreased the sGC background to acceptable levels (500 mM NaCl and 0.5% Nonidet P-40) were prohibitive of any protein interactions.
In addition to evidence of association between CCT and sGC in brain homogenates, we have also shown in vivo evidence for their association through co-localization in rat brains (Fig. 3). Interestingly, although sGC and CCT displayed a superimposable localization in rat hippocampus, both proteins showed only partial overlapping on a subcellular level. The immunohistochemical staining presented here (Fig. 3G) corrob- FIG. 6. CCT inhibits the NO-stimulated wild type sGC but not the constitutively active heme-deficient mutant, ␣␤ Cys-105 sGC. A, ␣␤ Cys-105 sGC has a high constitutive activity not affected by nitric oxide. B, lysates of Sf9 cells expressing wild type or ␣␤ Cys-105 sGC were preincubated with equal amount of ␤-galactosidase (␤-gal), CCT, or control buffer before NO-stimulated activity was determined. Data from a representative experiment performed in triplicate are presented as the mean Ϯ S.D. Three other experiments showed similar results.

FIG. 7. N-terminal portion of CCT is crucial for its inhibitory action.
Bacterial lysates expressing control GST, CCT enzyme lacking 94 N-terminal residues fused to GST (GST-CCT frag ), or full-length CCT were added to sGC-expressing Sf9 lysate, and SNP-stimulated activity of sGC was determined as described under "Materials and Methods." orates well with previously described cytosolic localization of CCT (17). Whereas CCT was fully cytosolic, sGC showed cytosolic and perinuclear localization. Such distribution suggests that most probably the cytosolic sGC is regulated by CCT. Although traditionally regarded as cytosolic protein, a number of studies also report that sGC can be associated with particulate fractions. Schmidt and co-workers (12) suggest a calcium-dependent sequestration of sGC to the membrane in platelets and endothelial cells. Koesling and colleagues (6) show sGC activity in the membrane fraction of the brain homogenates, which they attributed to the association between the ␣ 2 ␤ 1 isoform of sGC and PSD95, a protein localized at the plasma membrane.
From the deletional analysis, it appears that the association between sGC and CCT requires the N-terminal 94 residues of CCT (Fig. 7) as well as the sGC regions between the residues FIG. 8. CCT requires crude lysate to inhibit sGC. Equal amounts of purified CCT were added to purified wild type sGC or Sf9 lysates containing the same amount of wild type sGC. The activity was measured in the presence of 100 M SNP as described under "Materials and Methods." 100% activity for purified sGC is defined as 18 Ϯ 1.5 mol cGMP/ mg⅐min, and 100% activity for sGC-containing Sf9 lysate is defined as 22 Ϯ 1.8 nmol cGMP/mg⅐min. Data from three independent measurements performed in duplicates are presented.

FIG. 9. Effect of CCT overexpression on sGC in vivo.
A, a 300-bp amplicon specific for the FLAG-tagged CCT is detected in a reverse transcriptase-PCR reaction from the CCT-FLAG-transfected stable line but not from the line transfected with control vector. B, expression of sGC is equal in BE2 line transfected with pMG-CCT plasmid or control pMG vector. C, SNP-dependent increase in cGMP levels in BE2 cells is decreased in cells transfected with pMG-CCT plasmid but not the control pMG vector. 101 and 266 and between residues 367 and 619 (Fig. 4). Further studies will be required to determine the structural aspects of this interaction.
The association between CCT and sGC, along with some other factor (Fig. 8), resulted in a 30 -50% decrease in NOstimulated activity of sGC in vitro (Figs. 5 and 7), which correlates well with the 45% reduction of SNP-induced cGMP accumulation in BE2-CCT cells stably expressing CCT (Fig.  9). CCT does not affect the affinity of sGC heme to NO and does not act as an NO scavenger, because the EC 50 for DEA-NO (Fig. 5) and SNP (not shown) is not changed in the presence of CCT. The fact that CCT has no effect on activity of sGC stimulated by BAY41-2272, an NO-independent allosteric activator of sGC, or a heme-independent constitutively active mutant, ␣␤ Cys-105 (Fig. 6B), or the basal activity of sGC implies that CCT-mediated effects requires the nitrosyl heme or, most likely, the conformational changes in the sGC heme binding domain induced by NO.
From these data, one may conclude that CCT-mediated regulation is the next step in fine-tuning the response of sGC to NO. There have been no reports of any endogenous cellular factors that inhibit sGC. Therefore, the only regulation of cGMP forming activity of sGC would be the local concentration of NO, which is controlled mainly by the upstream NO producing enzyme, nitric-oxide synthase. Not one mechanism of negative regulation has been identified for sGC until now. HSP90 was shown to enhance SNP-stimulated sGC activity in cell culture and in vivo indirectly by adding the HSP90 inhibitor geldanamycin (7). When geldanamycin was added, SNP-stimulated activity was inhibited. Assuming that geldanamycin is relatively specific in its inhibition of HSP90, these results may indicate a role for HSP90 in sGC regulation (7).
CCT-facilitated inhibition of NO-stimulated sGC activity is the first example of an endogenous inhibition that affects only stimulated sGC. Interestingly, despite a direct interaction between sGC and CCT, the inhibition of NO-stimulated sGC is observed only after the addition of a crude lysate, which most probably means that there is some other signal or factor required to turn on the inhibition of sGC. Whether this signal or factor(s) is mediated by another subunit of CCT and whether NO (or cGMP) is necessary to recruit this additional factors requires further study.
Because sGC is involved in such a variety of physiological signals ranging from immunomodulation to vasodilatation to the inhibition of primary hemostasis, the regulation of sGC activity should likewise be complex. Interaction of sGC with CCT and the resulted inhibition described in this paper provide a novel aspect of this regulation.