Cysteine-rich protein 2, a novel substrate for cGMP kinase I in enteric neurons and intestinal smooth muscle.

Nitric oxide/cGMP/cGMP kinase I (cGKI) signaling causes relaxation of intestinal smooth muscle. In the gastrointestinal tract substrates of cGKI have not been identified yet. In the present study a protein interacting with cGKIbeta has been isolated from a rat intestinal cDNA library using the yeast two-hybrid system. The protein was identified as cysteine-rich protein 2 (CRP2), recently cloned from rat brain (Okano, I., Yamamoto, T., Kaji, A., Kimura, T., Mizuno, K., and Nakamura, T. (1993) FEBS Lett. 333, 51-55). Recombinant CRP2 is specifically phosphorylated by cGKs but not by cAMP kinase in vitro. Co-transfection of CRP2 and cGKIbeta into COS cells confirmed the phosphorylation of CRP2 in vivo. Cyclic GMP kinase I phosphorylated CRP2 at Ser-104, because the mutation to Ala completely prevented the in vivo phosphorylation. Immunohistochemical analysis using confocal laser scan microscopy showed a co-localization of CRP2 and cGKI in the inner part of the circular muscle layer, in the muscularis mucosae, and in specific neurons of the myenteric and submucosal plexus. The co-localization together with the specific phosphorylation of CRP2 by cGKI in vitro and in vivo suggests that CRP2 is a novel substrate of cGKI in neurons and smooth muscle of the small intestine.

Nitric oxide, a candidate of nonadrenergic, noncholinergic neurotransmission of the enteric nervous system (1, 2) has a variety of physiological functions in the gut (3,4). The inhibitory effect of NO on smooth muscle cells, either of the gut wall (5) or intestinal blood vessels (6), regulates gut motility and blood flow. Further, NO modulates neurosecretion from enteric nerve terminals (7). A main signaling pathway of NO in the gut is the activation of soluble guanylyl cyclase (8 -10), which results in an increase of intracellular cGMP. Cyclic GMP analogues elicit NO-like effects on smooth muscle (11) and isolated nerve terminals (7) causing smooth muscle relaxation and release of vasoactive intestinal polypeptide, respectively. A major target of cGMP in smooth muscle is the cGMP-dependent protein kinase I (cGKI) 1 (12)(13)(14). The two existing isoforms of cGKI, cGKI␣ and cGKI␤, show a tissue-specific distribution (15), with cGKI␤ being the predominant isoform in the small intestine (5). The importance of the NO/cGMP signaling in the gut has been demonstrated by gene ablation. Knock-out mice lacking a splice variant of the neuronal NO synthase (16) or cGKI (17) showed reduced neuronal inhibition of gastrointestinal smooth muscle (17,18), clear signs of pyloric stenosis (16,17) and severe disturbances of intestinal motility (17).
The molecular mechanisms of cGMP signaling distal to cGKI in intestinal smooth muscle and enteric neurons are not yet understood. In smooth muscle, activation of cGKI lowers cytosolic Ca 2ϩ concentrations (19) and hyperpolarizes the plasma membrane (20), effects that are compatible with the smooth muscle relaxing activity of the enzyme. Several targets for cGKI have been implicated in mediating these effects (for review see Ref. 21;Ref.12). It is assumed that at least part of the inhibitory responses mediated by cGKI depends on the phosphorylation of the IP 3 receptor (22) and on the stimulation of the large conductance K Ca channel (BK) (23,24). Furthermore, the relaxation of smooth muscle by NO/cGMP is also associated with an increase in the extent of phosphorylation of small proteins like heat shock related proteins that appear to be regulatory components of the actin based cytoskeleton or that seem to interact with intermediate filaments (25). A further substrate of cGKI involved in actin filament assembling and cell motility is the vasodilator-stimulated phosphoprotein associated with focal adhesions and areas of dynamic membrane activity (26). These proteins may be involved in relaxation and contractions of vascular smooth muscle by regulating cytoskeletal organization and cell shape. In contrast to vascular smooth muscle, little is known about potential targets of cGKI in the gastrointestinal tract, although the relaxing effect of cGKI on this type of smooth muscle is well described (5,27). In addition, the downstream effectors of cGKI in enteric neurons remains to be established. To identify targets for cGKI, we used the yeast two-hybrid system with cGKI␤ as a bait and a rat cDNA library from intestinal smooth muscle including the myenteric plexus for the interactor hunt. By applying this method we identified cysteine-rich protein 2 (CRP2) as a specific substrate. CRP2 is phosphorylated by and co-localized with cGKI. The function of CRP2 is not resolved yet, but related cysteine-rich proteins have been shown to be involved in cytoskeletal organization (28) and developmental processes (29,30).
Antibodies-A polyclonal antibody specific for cGKI␤ (AB-cGKI␤) (15) was used for Western blot analysis at a dilution of 1:500. A polyclonal antibody specific for cGKI (AB-cGKI-common) was raised against recombinant cGKI␣ in rabbits and purified on a cGKI␤ affinity column. AB-cGKI-common recognizes cGKI␣ and cGKI␤ with identical affinity as evaluated with pure cGKI isozymes. CRP2 was detected with the polyclonal antibody AB-CRP2 raised in rabbits against recombinant CRP2 expressed in bacteria. Sensitivity and specificity of AB-CRP2 was analyzed with CRP2 expressed in bacteria and COS cells. A commercially available, murine monoclonal antibody was used to detect the FLAG epitop of the FLAG/CRP2 fusion protein (Sigma).
Isolation of Poly(A) ϩ mRNA from Rat Intestinal Longitudinal Muscle Layer with Attached Myenteric Plexus (LM/MP)-LM/MP was isolated as described previously (5). Total RNA was extracted from LM/MP according to the manufacturer's protocol supplied with the RNA isolation kit from Stratagene (Heidelberg, Germany). Poly(A) ϩ mRNA was isolated from total RNA by affinity chromatography using oligo(dT)cellulose of the poly(A) ϩ mRNA isolation kit from Stratagene.
cDNA Library/Two-hybrid Screen-The cDNA library of LM/MP was constructed using the SuperScript TM Choice System (Life Technologies, Inc.). 5 g of mRNA and 80 ng of random hexamer primer were used for reverse transcription. For insertion of the synthesized cDNA into the two-hybrid expression plasmid pJG4 -5 (34), a XhoI site was included into the random hexamer primer, and EcoRI adapters were ligated to the cDNA. The cDNA was fused to the "acid loop" DNA activation domain of pJG4 -5, yielding pJG4 -5/cDNA. Recombinant plasmids were transformed in Escherichia coli Sure® (Stratagene) cells. 1.2 ϫ 10 6 independent clones carrying pJG4 -5/cDNA plasmids were propagated. 5 g of the pJG4 -5/cDNA plasmids were used for transformation of EGY48 yeast cells. Growth and maintenance of the yeast cells has been performed as described (34). The cDNA of cGKI␤ as a bait was inserted into the expression plasmid pEG202 (34) in frame with the DNA binding domain of LexA, yielding pEG202-LexA/cGKI␤. The expression of the hybrid protein was checked by immunoblot analysis with AB-cGKI␤. pEG202-LexA/cGKI␤ and the reporter plasmid pSH18-34 (34) were transformed by lithium acetate (35) in EGY48. pJG4 -5/cDNA was introduced into EGY48 in a second round of transformation. Interaction of cGKI␤ with a protein encoded by the cDNA library results in activation of the reporter genes. Positive clones were identified by their blue color on Xgal plates and their ability to grow on Leu Ϫ plates in the presence of galactose. 8.5 ϫ 10 5 independent yeast transformants were recovered on glucose/Ura Ϫ His Ϫ Trp Ϫ plates after 3 days at 30°C and harvested as described (31,34). After determination of the replating efficiency about 2 ϫ 10 7 galactose viable cells were screened on selection plates. cDNA of positive yeast colonies was isolated by polymerase chain reaction, classified by restriction mapping and sequenced. Specificity was validated by transforming the cDNA of positive clones together with pSH18-34 and pEG202-LexA/cGKI␤ or an unrelated bait into EGY48.
Bacterial Expression and in Vitro Phosphorylation of CRP2-CRP2 was expressed as fusion protein to glutathione S-transferase (GST) using pGEX-6P-1 (Amersham Pharmacia Biotech) and the proteasedeficient E. coli strain BL21. Expression and purification of the GST fusion protein was performed according to the manufacturer's protocol. Removal of GST from CRP2 was achieved by digestion with PreScission Protease (Amersham Pharmacia Biotech). Purity of CRP2 was analyzed by Coomassie Blue-stained SDS-polyacrylamide gels. CRP2 expressed in bacteria was phosphorylated in vitro by incubation at 30°C with cGKI␣, cGKI␤, cGKII, and cAK (20 -100 nM) in a final volume of 20 l 50 mM MES buffer, pH 6.9, containing 1 mM MgAc, 400 M EDTA, 25 mM NaCl, and 10 M dithiothreitol. Phosphotransferase activity of cGKs was stimulated with 10 M cGMP. The reaction was started by addition of 100 M [␥-32 P]ATP (2000 cpm/pmol ATP) and stopped by addition of Laemmli buffer (62.5 mM Tris, pH 6.8, 1.25% SDS, 5% ␤-mercaptoethanol, 0.01% bromphenol blue, 10% glycerol). Proteins were separated by SDS-PAGE and blotted to PVDF membranes. Incorporated radioactivity was visualized by autoradiography and phosphoimage analysis (BAS-1500, Fuji, Software TINA 2.0, Raytest, Straubenhardt, Germany) and calculated by counting the photoluminescence of 1 pmol of [␥-32 P]ATP spotted onto the PVDF membrane.
The putative phosphorylation site Ser-104 for cGKI was mutated to alanine by polymerase chain reaction, yielding pcDNA3-CRP2/S104A. Mutated CRP2 was sequenced to confirm the mutation. Phosphorylation of the mutant was analyzed in vivo as described above.
Immunological Localization of CRP2 in Transfected Cells and Intestinal Tissue-COS-7 cells transfected with pcDNA3-FLAG/CRP2 were fixed in Zamboni's solution 48 h after transfection. Rats were perfusion fixed with Zamboni's solution, and segments of the duodenum were immersed (4 h) in the same fixative. After cryoprotection in 20% phosphate-buffered sucrose, 12-m-thick cryostat sections were mounted on poly-L-lysine coated slides and air dried (1 h). Following a 5-min rinse in TBS (136 mM NaCl, 2.7 mM KCl, 25 mM Tris, pH 7.4), cover glasses with transfected cells as well as the duodenal sections were incubated (1 h) in TBS containing 1% bovine serum albumin, 5% normal goat serum, and 0.5% Triton X-100. All incubations were performed at room temperature. Antibodies employed were diluted in TBS containing 1% bovine serum albumin and 0.5% Triton X-100. Immunological detection of CRP2 was performed by incubating with AB-CRP2 (1:1000) overnight. Binding of AB-CRP2 was visualized with a goat anti-rabbit IgG antibody (1:400) tagged with indocarbocyanin (Cy3, Dianova, Hamburg, Germany). Transfected cells and sections were rinsed in TBS and coverslipped in TBS:glycerol 1:1, pH 8.6. Co-localization of CRP2 with cGKI was investigated by a sequential double immunostaining protocol. Initial experiments showed that AB-cGKI-common and AB-cGKI␤ revealed identical immunological reactions in intestinal sections of the rat. However, because of its higher sensitivity, AB-cGKI-common was preferred over AB-cGKI␤ for double staining experiments. AB-cGKIcommon was conjugated with digoxigenin using a commercially available kit (Roche Molecular Biochemicals). To prevent unspecific binding of the digoxigenized AB-cGKI-common, an additional blocking step (1 h) was introduced using 10% normal rabbit serum in TBS containing 1% bovine serum albumin and 0.5% Triton X-100 prior to overnight application of the antibody (1:750). After rinsing with TBS (4 ϫ 5 min), a fluorescein isothiocyanate-tagged sheep anti-digoxigenin antibody (1: 100) was applied for 1 h. The incubation was terminated by several rinses in TBS and coverslipping in TBS-glycerol. Controls included omission of the primary antibody or its replacement by normal rabbit serum or antibodies preabsorbed with the respective antigen. To further characterize CRP2 immunostaining, sections were co-incubated in separate trials with AB-CRP2 and a monoclonal mouse anti-vimentin antibody (Dako) that specifically stains glia cells in myenteric ganglia (36). Binding of the vimentin-specific antibody and AB-CRP2 was revealed using goat anti-mouse IgG antibody tagged with Cy3 (Dianova) and goat anti-rabbit IgG antibody labeled with carbocyanin (Cy2) (Amersham Pharmacia Biotech) (1:200), respectively. Sections were analyzed by confocal laser scanning microscopy (Bio-Rad MRC 1000 attached to a Nikon Diaphot 300). Fluorochromes were excited with 488 and 568 nm lines, respectively, by a Krypton-Argon laser. Single optical sections were taken with a 20ϫ objective lens (numerical aperture, 0.75) and various zoom factors. When controls and "full" incubations were compared, special care was taken to keep the pinhole and gain of the photomultiplier constant. Two channel scans were coded green and red, and merged images were documented using the software package Corel Photo Paint.

Identification of CRP2 as Interactor of cGKI␤ in Yeast-To
identify proteins of enteric neurons and intestinal smooth muscle cells interacting with cGKI␤, a two-hybrid screen was performed using cGKI␤ as a bait and a cDNA library of the longitudinal muscle with attached myenteric plexus of the rat small intestine. Prior to the screen several control experiments were carried out demonstrating the nuclear localization of the LexA-cGKI␤ hybrid bait, its binding to the lexA-operator of the lacZ and Leu reporter genes, and lack of activation of the reporters by the bait itself. The expression of the hybrid bait LexA-cGKI␤ was analyzed by immunoblotting with a cGKI␤specific antibody (15) showing a 100-kDa protein that comprises the 75-kDa cGKI␤ and the 24-kDa LexA (data not shown).
The interactor hunt with cGKI␤ as bait yielded the complete coding region of CRP2 as interactor of cGKI␤ (Fig. 1). The rat intestinal CRP2 consists of 208 amino acids with a corresponding molecular mass of 23 kDa. The protein consists of two LIM (LIN-11, IsC-1, Mec-3) domains each containing two paired zinc fingers with a two-amino acid linker. Each LIM domain of CRP2 is followed by a glycine-rich domain and a putative nuclear localization signal (FGPKG). An identical protein has previously been cloned from rat brain (37). Rat CRP2 is distantly related to other cysteine-rich proteins that are grouped into the family consisting of CRP1 (29), CRP2/SmLIM (38), CRP3/MLP (39), and that formed by CRIP, which is characterized by a unique LIM domain (40) (Fig. 1). Rat intestinal CRP2 shares several structural features with other CRPs, i.e. the LIM domain(s) followed by the glycine-rich repeat and the nuclear localization signal.
In Vitro Phosphorylation of Recombinant CRP2 Expressed in Bacteria-The specificity of the interaction observed in yeast cells was further evaluated by studying the potential role of CRP2 as a substrate for cGKI␤. Initial phosphorylation experiments have been performed with CRP2, which was expressed as GST fusion protein in bacteria and subsequently purified on glutathione-Sepharose. The GST tag was removed by site-specific cleavage with PreScission protease. Coomassie staining of the purified protein fraction demonstrated the purity of recombinant CRP2 (Fig. 2). The phosphorylation of CRP2 by cGKI␤ CRP2, a Novel Substrate of cGKI was time-dependent and stimulated by the addition of cGMP. cGKI␤ and cGKI␣ specifically phosphorylated recombinant CRP2 to a similar extent. In contrast, phosphorylation by cGKII in the presence of cGMP was about 7-fold less effective and comparable with the phosphorylation of CRP2 by cGKI␣ and cGKI␤ in the absence of cGMP. No phosphorylation was observed with the catalytic subunit of cAMP kinase. The kinetic constants of the phosphorylation reaction were 5.8 Ϯ 0.9 M and 83.7 Ϯ 9.5 nmol phosphate ϫ min Ϫ1 ϫ mg Ϫ1 cGKI␤ (n ϭ 4) for K m and V max , respectively (Fig. 3). A similar K m value has been reported for a peptide derived from the IP 3 receptor channel that is phosphorylated by cGKI in vitro (22). The V max value for CRP2 is 20-fold higher than that for a protein of similar molecular size, i.e. the 21-kDa protein Rap1B (41), which is an in vitro substrate for cGKI. These results demonstrate that CRP2 is specifically phosphorylated by cGK type I isozymes and may function as an in vivo substrate for cGKI. In fact, CRP2 contains a putative site for cGK-dependent phosphorylation (RKTS) in between the two LIM domains. In contrast to CRP2, such a consensus sequence for cGMP or cAMP kinases is not present in the other CRPs (Fig. 1).
In Vivo Phosphorylation of CRP2 by cGKI␤ at Ser-104 -To evaluate the potential role of CRP2 as an in vivo substrate for cGKI␤, COS cells were transfected with CRP2 and cGKI␤ simultaneously or with either protein alone. CRP2 and cGKI␤ were expressed to a similar extent under each condition (Fig.  4). Immunocytochemical analysis demonstrated a cytosolic and perinuclear localization of CRP2 (Fig. 4A). The nuclear membrane is intensively stained, although protein extracts from transfected COS cells suggest that CRP2 is almost completely present in the soluble fraction. Transfected cells were metabolically labeled with [ 33 P]ortho-phosphate and stimulated with 8-pCPT-cGMP. Immunoprecipitation with CRP2-specific antiserum demonstrated that CRP2 was only phosphorylated when cGKI␤ was co-expressed (Fig. 4, B and C). In contrast, no phosphate was incorporated when CRP2 was expressed alone, although comparable levels of CRP2 were precipitated under CRP2, a Novel Substrate of cGKI both conditions. This suggests that CRP2 was specifically phosphorylated by cGKI␤. Furthermore, COS cells do not contain endogenous CRP2 or any other unspecific proteins that superimpose the CRP2 signal, because no signal in the immunoblot and the autoradiogram was observed when immunoprecipitation was performed using COS cells expressing only cGKI␤. The specificity of the immunoprecipitation was evaluated by using preimmune serum, which did not precipitate CRP2 (data not shown).
CRP2 contains a single putative phosphorylation consensus site (RKTS) for cyclic nucleotide-regulated serine/threonine kinases. To examine whether phosphorylation of CRP2 by cGKI␤ occurs at this site, Ser-104 was mutated to Ala. COS cells were transfected either with wild type or mutant CRP2 in the presence or the absence of cGKI␤. The expression levels of cGKI␤, CRP2, and CRP2/S104A were comparable under each combination (Fig. 4C). In contrast to the wild type protein, CRP2 was not phosphorylated when Ser-104 was mutated to Ala (Fig. 4C). The in vivo phosphorylation of CRP2 by cGKI␤ in COS cells together with the interaction of both proteins in yeast cells supports the idea of a physiological role for CRP2 as a mediator of cGMP/cGK signaling in the intestine.
Co-localization of CRP2 and cGKI in Rat Intestine-A prerequisite for the interaction of CRP2 and cGKI␤ in the intestine is the co-localization of the two proteins in specific cell types. CRP2 was detected in smooth muscle cells of the muscularis mucosae and in the inner part of the circular muscle layer (Fig.  5, A and B), which is separated in the small intestine from the outer part of the circular muscle layer by the deep muscular plexus. CRP2 was further localized in neurons of the submucosus and myenteric plexus as well as in nerve processes projecting from the myenteric plexus into the longitudinal and circular muscle layers. The neuronal origin of CRP2 in myenteric ganglia was confirmed by double staining with the CRP2specific antibody and an anti-vimentin antibody. The intermediate filament vimentin is not expressed in neuronal cells and accounts for a marker of glial elements in the myenteric plexus region (36) (Fig. 5D). In the mucosa CRP2 is present in fibroblast like and smooth muscle cells running along with the longitudinal axis of the villi (Fig. 5A). Preabsorption of the CRP2-specific antiserum with bacterial expressed CRP2 completely suppressed the immunological staining (Fig. 5C).
cGKI was found in the muscularis mucosae and the longitudinal and circular muscle layer with the highest level in the inner part of the circular muscle layer. In addition, pronounced expression was also observed in a cell population of the deep muscular plexus and in another population surrounding myenteric ganglia (Fig. 6A). A weak expression of cGKI was further detected in specific neurons of the myenteric and submucosal plexus (Fig. 6A). Double staining analysis using both antibodies revealed co-localization of CRP2 and cGKI in smooth muscle cells of the inner part of the circular muscle layer, in the muscularis mucosae (Fig. 7A), and in distinct neurons of the myenteric and submucosal plexus (Fig. 7). These results make it conceivable that CRP2 serves as a substrate of cGKI in the intestine. (1 mCi/dish) as described under "Experimental Procedures." At the end of the incubation period, 100 M 8-pCPT-cGMP was added for 10 min to stimulate cGMP kinase activity. 200 g of soluble proteins were used for immunoprecipitation of CRP2 with AB-CRP2. Immunoprecipitated proteins were separated by SDS-PAGE and blotted to PVDF membranes, and phosphate incorporation into CRP2 was analyzed by phosphoimaging. Lower panel, CRP2 immunoreactivity after immunoprecipitation (IP) was visualized by staining of the blot membrane with the mouse monoclonal FLAG antibody. One representative experiment of three with similar results is shown. C, phosphorylation of CRP2 at Ser-104. Upper panel, Western blot analysis of soluble extracts (15 g) of COS cells transfected with wild type (WT) CRP2 or mutant (Mut) CRP2/S104A and cGKI␤ using AB-cGKI␤ and AB-CRP2. Lower panel, autoradiogram (ARG) of phosphorylated and immunoprecipitated CRP2. In vivo phosphorylation and immunoprecipitation was performed as described above.

DISCUSSION
In an approach to identify substrates of cGKI␤ in neurons and smooth muscle of the small intestine, we used the yeast two-hybrid system, thereby isolating CRP2 as a novel substrate for cGKI␤. Further analysis of the interaction revealed that CRP2 is specifically phosphorylated by cGKI in vitro and in vivo. The phosphorylation site Ser-104 for cGKI exhibits the consensus motive RKTS, which is located in between the two LIM domains. The co-localization of both proteins in specific neurons and smooth muscle of rat small intestine further strengthens the idea that CRP2 represents a cGKI substrate.
A specific site of co-localization of cGKI and CRP2 is the inner part of the circular muscle. cGKI expression in this layer is increased when compared with the outer circular muscle. Apart from cGKI, CRP2 is also highly abundant in this muscle segment. In the small intestine the inner part of the circular muscle is separated from the outer part by the deep muscular plexus. Hence, the inner circular layer may be strongly under the influence of NO diffusing from deep muscular plexus nitrergic neurons to the smooth muscle cells of this layer. Further, the inner circular layer is characterized by a rich extrinsic and intrinsic innervation. This is reflected by the high number of contacts between muscle cells and nerve endings (42). As a functional consequence the inner circular layer differs from that of the outer layer in excitation contraction coupling.
CRP2 and cGKI are also co-localized in the muscularis mucosae, which represents the boundary between the mucosa and the submucosa. The contractile potential of this muscle layer may influence intestinal function by facilitating the emptying of the crypt luminal content. Moreover, movement of the villi caused by the muscularis mucosae may modulate the unstirred layer adjacent to the absorptive epithelium (43). In addition to the inner circular muscle and the muscularis mucosae, CRP2 and cGKI are co-localized in specific neurons of the myenteric and submucosal plexus, although the chemical coding of these neurons is not yet resolved. NO synthase has recently been FIG. 5. CRP2 is expressed in neurons and smooth muscle of rat small intestine. A-C, immunological localization of CRP2 in the small intestine with AB-CRP2 by confocal laser scan microscopy. The binding of AB-CRP2 was visualized with a goat anti-rabbit IgG antibody tagged with Cy3. The specificity of the immunological reaction was analyzed by preabsorption of AB-CRP2 with CRP2 (10 g/ml) expressed in bacteria (C). CRP2 was present in neurons of the myenteric (MP) and submucosal plexus (SP) as well as in nerve bundles projecting in the longitudinal (LM) and circular (CM) muscle layers (B). The innermost part of the circular muscle layer (ICM), the muscularis mucosae (MM) (A and B), smooth muscle cells of the villi (SMC), and fibroblast-like cells (FLC) of the mucosa (A) were also positive for CRP2. D, the origin of the CRP2 staining in the nerve plexus was neuronal (second antibody labeled with Cy2; green). Glia cells surrounding CRP2 positive neurons were stained with an anti-vimentin antibody (second antibody labeled wit Cy3; red). The two immunological stains did not merge, thereby excluding the possibility that CRP2 is localized in glia cells.
FIG. 6. cGKI is expressed in neurons and smooth muscle of rat small intestine. A, immunological localization of cGKI in the small intestine with AB-cGKI-common conjugated with digoxigenin by confocal laser scan microscopy. The binding of AB-cGKI-common was visualized with a Cy3-tagged sheep anti-digoxigenin antibody. B, the specificity of the immunological reaction was analyzed by preabsorption of AB-cGKIcommon with 30 g/ml purified, recombinant cGKI␤. cGKI was expressed in the longitudinal (LM) and circular (CM) muscle layers; in the muscularis mucosae (MM) and in specific neurons of the myenteric plexus (MP, arrow). ICM, inner part of the circular muscle layer.
found to co-localize with cGKI in some neurons (5). However, it remains speculative whether the CRP2/cGKI containing neurons are positive for NO synthase. The co-localization in specific muscle layers and neurons is consistent with the idea that CRP2 and cGKI interact in vivo; however, the localization of the two proteins was not identical. CRP2 is more widely expressed in neurons, whereas cGKI has a broader distribution than CRP2 in smooth muscle.
The cellular alterations in response to the phosphorylation of CRP2 by cGKI are unclear as is the physiological role of CRP2 itself. Preliminary data from co-transfected COS cells suggest that CRP2 phosphorylated by cGKI has a tendency to become enriched perinuclearly. 2 However, further studies with intestinal cells have to be carried out to confirm the redistribution. CRP2 was recently cloned from rat brain (37). The protein is highly expressed in heart, lung, placenta, kidney, and to a lower extent in several brain regions, skeletal muscle, stomach, liver, intestine, testis, and spleen (37,44). The human homologue of CRP2 was localized on chromosome 14q32.3 (44,45), a hot spot of translocation in tumor development (46) noted in T-cell leukemia (47). Therefore, it has previously been hypothesized that human CRP2 is involved in growth and development and may play a role in the pathogenesis of the mentioned malignancy (45).
Further speculations on the functions of CRP2 refer to the roles of related proteins exhibiting 30 -45% homology. These proteins are CRP1 (29,48), CRP2/SmLIM (38,49), and CRP3/ MLP (39). There is growing evidence that CRPs mediate protein-protein interaction by their two LIM domains (50,51). Several CRPs are co-localized with zyxin and ␣-actinin along the actin stress fibers and at focal adhesion plaques (28,(52)(53)(54). Thus, the CRPs may serve as adapter molecules or as scaffolds for the coordinated, localized assembly of multimeric complexes like those present at focal adhesion plaques. Hence, CRPs might be involved in signaling pathways coupling extracellular signals via integrins and adhesion plaque proteins to changes in cytoskeletal architecture. Probably by acting as a scaffold protein to promote protein assembly along the actinbased cytoskeleton, CRP3/MLP affects myogenic differentiation (39). CRP3/MLP-deficient mice have soft hearts with disruption of the cardiomyocyte cytoarchitecture (30) and develop dilated cardiomyopathy after birth. Also other CRPs like CRP1 and CRIP are expressed ontogeny-dependently, supporting their roles in neuronal and intestinal development (29,40).
Based on the structural homology of CRP2 to CRP1, CRP2/ SmLIM, and CRP3/MLP, it is conceivable, although not proven, that rat intestinal CRP2 is also localized at cytoskeletal structures. There is accumulating evidence that phosphorylation of cytoskeletal proteins by cGKI regulates cytoskeletal organization. Apart from CRP2, cGKI phosphorylates the vasodilatorstimulated phosphoprotein (55), a protein that has been implicated to control the actin assembly by its ability to associate with profilin (56). In endothelial and aortic smooth muscle cells, disassembling of focal adhesions triggered by the antiadhesive extracellular matrix proteins thrombospondin and tenascin requires cGKI (57), suggesting that cGKI participates in weakening cell-matrix interactions and regulates cell locomotion. Taken together, these observations give reason to implicate that cGKI is localized to focal adhesion plaques and phosphorylates specific proteins like CRP2 and vasodilatorstimulated phosphoprotein, thereby regulating specific functions such as cell-matrix interactions, assembling of focal adhesion plaques, and cell motility. As a consequence of CRP2 phosphorylation in enteric neurons, cytoskeletal changes may alter vesicular transport and release of neurotransmitters. In conclusion, cGKI␤ phosphorylates CRP2, a protein that may exhibit multifunctional regulatory properties.
Acknowledgments-A considerable part of the work was carried out at the Institute of Pharmacology and Toxicology. We are grateful for 2 A. Huber, unpublished results. FIG. 7. Co-localization of CRP2 and cGKI. A-D, immunohistochemical double staining of duodenal segments with CRP2-and cGKI-specific antibodies. CRP2 was detected with a Cy3-labeled second antibody (red). AB-cGKI-common was tagged with digoxigenin. The binding of this antibody was visualized with a fluorescein isothiocyanate-labeled antidigoxigenin antibody (green). Co-localization of both proteins resulted in a yellow color and could be demonstrated in the inner part of the circular muscle layer (ICM) and in the muscularis mucosae (MM, A). In addition both proteins were expressed in special neurons of the myenteric (MP) and submucosal (SP) plexus.