RhoA-mediated Ca 2 (cid:1) Sensitization in Erectile Function*

A Rho-kinase inhibitor increases corpus cavernosum (CC) pressure in an in vivo rat model suggesting that Rho-mediated Ca 2 (cid:1) sensitization of CC smooth muscle maintains the flaccid (contracted) state. We directly demonstrate Ca 2 (cid:1) sensitization of permeabilized rabbit and human CC and identify a highly expressed molecular component of this pathway. Ca 2 (cid:1) sensitization of force induced by endothelin or GTP (cid:2) S was significantly greater in CC than in rabbit ileum smooth muscle and was accompanied by a 17-fold higher RhoA content. Pull-down assays with the RhoA binding domain of mDia showed the high RhoA content of CC to be available for activation by GTP (cid:2) S. Ca 2 (cid:1) sensitization induced by endothelin, phenylephrine, or GTP (cid:2) S was completely relaxed by the Rho kinase inhibitor Y-27632. Human and rabbit CC both express the phosphatase inhibitor CPI-17, the myosin phosphatase regulatory (MYPT-1) and catalytic (PP1 (cid:3) ) subunits, and two isoforms

Relaxation of the smooth muscle trabeculae of the corpus cavernosum (CC) 1 and of the helicine arteries leads to blood filling of the sinuses, occlusion of the venous outflow (reviewed in Ref. 1), and penile erection. Nitric oxide (NO), generated by both nerves and the endothelial cells that cover the trabeculae of the CC, through stimulation of soluble guanylate cyclase and the generation of cyclic GMP (for review see Ref. 2), is thought to play a dominant role in relaxation of smooth muscle in this tissue. The pathophysiological relevance of the cyclic GMPprotein kinase G (PKG) pathway is evidenced by the successful use of cyclic GMP phosphodiesterase inhibitors (PDE-5) such as sildenafil (Viagra) in the treatment of erectile dysfunction (for review see Ref. 3). Other signaling pathways involving vasoactive intestinal polypeptide/cAMP may also be operative in relaxation of the CC. Severe erectile dysfunction in cGMPdependent kinase 1-deficient mice, with normal cAMP signaling, also demonstrated the importance of PKG and the inability of the cAMP pathway to compensate for the absence of the cGMP signaling cascade in vivo (4,5).
The contracted (resting) state of the CC smooth muscle is considered to be mediated by release of norepinephrine, endothelin-1, neuropeptide Y, prostanoids, and angiotensin II (1, 6 -11). Inhibition of these agonists at the receptor level or in their downstream signaling pathways should, like NO stimulation, also lead to a decrease in myosin regulatory light chain (RLC 20 ) phosphorylation and consequent relaxation through decreased cytoplasmic [Ca 2ϩ ] and/or inhibition of Ca 2ϩ sensitization (for review see Refs. 12 and 13). Calcium sensitization is brought about by agonist activation of heterotrimeric G-protein-coupled receptors, the exchange of GTP for GDP on the small GTPase RhoA, its activation and dissociation from its partner Rho-GDI (guanine nucleotide dissociation inhibitor). GTP⅐RhoA activates Rho-kinase, which inhibits myosin light chain phosphatase, resulting in an increase in RLC 20 phosphorylation and force at constant [Ca 2ϩ ] (13)(14)(15)(16)(17). Ca 2ϩ sensitization has been shown to make a significant contribution to agonist-induced contraction under physiological conditions in vascular smooth muscle (12,18). A recent study of an in vivo rat model demonstrated that a Rho-kinase inhibitor caused a dramatic increase in CC pressure and erection independent of NO (19) while exerting a minimal effect on systemic arterial pressure. This finding suggests that the RhoA-Ca 2ϩ sensitization pathway is present and active in the "resting state" of CC smooth muscle. These authors proposed that antagonism of Rho-kinase introduces a potential alternate avenue for the treatment of erectile dysfunction. The Rho-kinase inhibitor Y-27632 (20,21) also relaxed intact phenylephrine-contracted cavernosal strips of smooth muscle in a dose-dependent fashion (22,23) similar to other smooth muscles (20,24).
In this study we directly demonstrate very potent Ca 2ϩ sensitization of force in permeabilized rabbit and human CC in which Ca 2ϩ concentrations are clamped and show that a component of this pathway, RhoA, is very highly expressed in CC compared with its content in other smooth muscles; the RhoA binding partner, guanine nucleotide dissociation inhibitor (GDI), is also present at high concentrations. Using a Rho binding domain pull-down assay, we also show that, during Ca 2ϩ sensitization with GTP␥S, a large fraction of the endogenous RhoA is in the active, GTP␥S-bound form, and that the Rho-kinase inhibitor inhibits calcium sensitization induced by GTP␥S, phenylephrine, and endothelin-1. Furthermore, the magnitude of relaxation of maximal GTP␥S-induced calciumsensitized force by Y-27632 is 2-to 3-fold greater than by 8-Br-cGMP in rabbit CC. We also find that CC expresses telokin, a target of PKG (25), myosin phosphatase regulatory MYPT-1 and catalytic PP1␦ subunits (12,13,17), both Rho-kinase isoforms ROK␣ and ROK␤ (26 -29), and the myosin phosphatase inhibitor protein, CPI-17 (30).

EXPERIMENTAL PROCEDURES
Tissue Preparation and Force Measurements-New Zealand rabbits (2-3 kg) were anesthetized with halothane and killed by exsanguination according to protocols approved by the Center for Comparative Medicine at the University of Virginia. The penis was excised, and the CC was dissected free from the tunica albuginea. Human CC specimens were obtained at the time of implantation of a penile implant from four patients, one diagnosed with vasculogenic erectile dysfunction due to atherosclerotic disease and the other three with type II diabetes mellitus. These three patients had failed to achieve an adequate erection for intercourse following administration of the PDE-5 inhibitor sildenafil and intracavernous prostaglandin E1, which raises cAMP. Small strips (0.15-0.2 mm wide and 2 mm long) were cut with razor knives from the regions where longitudinal bundles could be well resolved under a dissecting microscope. The ends of the muscle bundles were tied with monofilament silk to a force transducer and a stationary hook and mounted on a bubble plate for force measurements. After recording the contractile response to high potassium, the muscle was transferred to relaxing solution (G1) containing 1 mM EGTA and permeabilized with 0.1 mg/ml Staphylococcus aureus ␣-toxin. Maximal Ca 2ϩ -activated force was determined at pCa 4.5 (CaG). All solutions and the use of ␣-toxin and A23187 have been described previously (15,31,32).
mDia Pull-down Assays-Affinity precipitation of active Rho was performed using the Rho binding domain (RBD) of the fusion protein, mDia (33). pEGX-4T-1 mouse mDia RBD (amino acids 2-304) was introduced in Escherichia coli BL21(DE3) pTrx, and the GST fusion protein was expressed, conjugated to glutathione beads, and purified. Small strips of CC (0.15-0.2 mm wide and 2 mm long) were permeabilized with ␣-toxin for 2 h at room temperature, washed in G1, and contracted with pCa6.5. After 10 min, 10 M GTP␥S was added for a predetermined amount of time, and tissues were snap frozen and pulverized at liquid nitrogen temperature. The powder was rapidly homogenized in the homogenization buffer containing 50 mM Tris-HCl (pH 7.5) 100 mM NaCl, 1 mM EDTA, 5 mM MgCl 2 , 10% glycerol, 50 mM NaF, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, 0.5% Nonidet P-40 and a protease inhibitor mixture (P-8340, Sigma Chemical Co., St. Louis, MO). Homogenized samples were centrifuged at 14,000 rpm at 4°C for 10 min, and a small quantity of supernatant was saved to be analyzed later for total RhoA content. The remaining supernatant was incubated with 30 g of GST-RBD fusion protein conjugated with glutathione beads at 4°C for 2 h. The beads were washed twice with homogenization buffer and then resuspended in 48 l of non-reducing sample buffer with 2 l of 1 M dithiothreitol and heated to 100°C for 5min. Positive and negative control samples, into which GTP␥S or GDP was exchanged into the endogenous RhoA of rabbit tissues, were homogenized in homogenization buffer and centrifuged at 14,000 rpm at 4°C for 10 min, and the supernatant was collected. For nucleotide exchange, 10 mM EDTA was added to each sample, 1 mM GDP was added to the negative control and 50 M GTP␥S to the positive controls. Samples were incubated at 30°C for 15 min, and then the reaction was stopped by addition of 60 mM MgCl 2 . These controls samples were then incubated with glutathione-mDia beads and run in parallel with the tissue samples. After binding to mDia RBD, the samples were subjected to SDS-polyacrylamide gel electrophoresis on a 4 -20% gradient gel. Bound RhoA was detected by Western blot using a monoclonal antibody against RhoA (Santa Cruz Biotechnology).
Western Blot Analysis-Strips of rabbit and human CC, strips of rabbit femoral artery and sheets of the longitudinal muscle of rabbit ileum at rest or following stimulation were homogenized in ice-cold buffer (250 mM sucrose, 25 mM Tris, 5 mM MgCl 2 , 5 mM EDTA, 1 mM dithiothreitol, and protease inhibitor mixture, see above), pH 7.4. Tissue homogenates were submitted to SDS-PAGE with 8% polyacrylamide for MYPT-1, PP1␦, ROK␣, and ROK␤ and 12% for all other proteins followed by transfer to polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat dry milk and phosphatebuffered saline containing 0.05% Tween-20 (PBST) for 1 h and then incubated with primary antibody for 2 h at room temperature. The following dilutions of primary antibodies were used: anti-RhoA (monoclonal antibody SC-418, Santa Cruz Biotechnology, Santa Cruz, CA) at 1:10,000, rabbit polyclonal anti-telokin antibody prepared against recombinant full-length rabbit telokin at 1:1,000, anti-MLC 20 (monoclonal antibody, Sigma) at 1:200, anti-GDI (rabbit polyclonal antibody, Santa Cruz Biotechnology) at 1:5,000, anti-Rho-kinase ROK␣ (monoclonal antibody, Transduction Laboratories) at 1:1,000, anti-ROK␤ (mono-clonal antibody, Transduction Laboratories) at 1:500, anti-MYPT-1 (sheep polyclonal antibody (Upstate, Waltham, MA) at 1:1,000, anti-PP1␦ (Upstate) at 1:5,000, anti-CPI-17 (rabbit polyclonal) at 1:2,000 (34). The blots were washed in PBST, incubated in horseradish peroxidase-conjugated secondary antibodies to mouse, rabbit, or sheep for 1 h at room temperature and detected with enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech). The protein signals were quantitated by densitometry using an Epson Expression 1600 scanner (Long Beach, CA) and the IMAGE Gel Quantitation Program (National Institutes of Health). Protein was determined with the Bio-Rad Protein Assay kit. Quantitation of RhoA, GDI, telokin, and MLC 20 in tissues was carried out in the linear range of the antibody reactivity.
For electron microscopy, strips of CC smooth muscle were fixed in 2% glutaraldehyde with 0.1% tannic acid in cacodylate buffer followed by osmium tetroxide, stained with uranyl acetate en bloc, dehydrated, and embedded in Spurr's resin. Sections were viewed in a Phillips CM12 electron microscope.

RESULTS
The respective activities of the Rho-kinase inhibitor Y-27632 and of 8-Br-cGMP to relax permeabilized Ca 2ϩ -sensitized CC muscle strips were compared (Fig. 1). Muscles were maximally sensitized with GTP␥S or endothelin plus GTP with Ca 2ϩ concentrations clamped at pCa 6.9 -6.7. GTP␥S (10 M) and endothelin (100 nM) were highly effective in eliciting Ca 2ϩ sensitization. 10 M GTP␥S induced Ca 2ϩ sensitization as follows: 64% Ϯ 11.6 S.D. of maximal Ca-induced force (pCa 4.5) (n ϭ 42) in rabbit CC and 52% Ϯ 4.9 S.D. (n ϭ 7) in several strips from three human CC specimens. Endothelin induced Ca 2ϩ sensitization as follows: 35% Ϯ 5.  Fig. 3. 8-Br-cGMP was more effective in relaxing endothelin than GTP␥S-mediated Ca 2ϩ sensitization, 96% Ϯ 4.4 S.D. (n ϭ 7) in rabbit CC. In several strips of human CC obtained from three of the patients, Y-27632 completely relaxed GTP␥S as well as endothelin plus GTP-induced calcium sensitization (n ϭ 3) with a similar sensitivity as found in rabbit CC (Fig. 4); 50 M 8-Br-cGMP was more effective than in the rabbit CC (compare Figs. 1 and 4). As in the rabbit CC (Fig. 1C), even when 8-Br-cGMP induced a large reversal of sensitized force, subsequent additions of Y-27632 induced a further relaxation (Fig. 4,  A and D).
The concentrations of the small GTPase RhoA, its partner Rho⅐GDI, and the protein kinase G (PKG) target, telokin, were estimated from Western blots of tissues, and known amounts of purified standards of recombinant proteins were loaded in the linear range of the antibody reactivity (Fig. 5). Rabbit ileum smooth muscle is included for comparison with rabbit and human CC. Because the CC has an exceedingly high content of connective tissue relative to smooth muscle, to obtain comparable estimates of the concentrations of RhoA, Rho⅐GDI, and telokin in smooth muscle, the concentrations were normalized to tissue myosin content estimated from known concentrations of purified turkey gizzard myosin loaded on the same gel (Fig.  5D). The myosin in the standards and tissue specimens was detected with an antibody to MLC 20 . Comparison of myosin contents detected by Western blotting with a MLC 20 antibody (Fig. 5D) clearly shows the lower myosin content of equal protein loads in the cavernosum samples compared with rabbit ileum as well as the significantly higher RhoA content (Fig. 5A) and Rho⅐GDI (Fig. 5B) in cavernosum compared with ileum at equal protein loads, which is obvious even without correcting for the lower cellular content of the CC samples. The relative amounts of myosin, telokin, and RhoA in the CC and rabbit ileum, estimated by comparison with known concentrations of those proteins, are shown in Table I. The myosin content per microgram of total protein in the ileum is 3.4 times that in the rabbit CC. The RhoA content per microgram of total protein was 5.3 ng in the corpus compared with 1.1 ng in the ileum sample. Correcting for the large contribution of connective tissue to total protein in the CC by normalizing the RhoA measurements to the myosin content results in a 17-fold higher RhoA concentration in the rabbit CC than in rabbit ileum; the relative expression of telokin was not significantly different. A similar 16-fold greater RhoA content was measured in human CC samples compared with rabbit ileum (Fig. 5). Normalization of Rho⅐GDI to myosin content in the same sample (not shown) also gave a content that was ϳ17-fold higher in CC than in ileum smooth muscle.
The density of myosin filaments seen in electron micrographs of rabbit CC smooth muscle (Fig. 6) was typical of other smooth muscles such as ileum, showing normal arrays of myosin and actin filaments with a high actin to myosin ratio, justifying our above normalization of RhoA content to myosin content. Incidentally, a well-developed sarcoplasmic reticulum forming surface couplings with the plasma membrane as well as the presence of a basement membrane were also observed.
To determine whether the very high concentrations of CC RhoA were accessible to activation during Ca 2ϩ sensitization with GTP␥S, we used a pull-down assay for GTP⅐RhoA. The mDia Rho binding domain (RBD) conjugated with glutathione beads extracted a larger fraction of RhoA⅐GTP␥S per total RhoA in the homogenate from rabbit CC compared with rabbit ileum under identical assay conditions (Fig. 7). The RhoA remaining in the supernatant may have also included some GTP␥S⅐RhoA, because the extent of extraction depends on the amount of mDia-bound beads used in the assay, which, for technical reasons, we could not increase further. Positive and negative controls illustrate that the assay extracts GTP␥Sbound but not GDP-bound RhoA from tissue homogenates treated to load the given nucleotide onto the endogenous RhoA. These results indicate that the highly expressed RhoA in CC is correctly folded and available to be activated by Ca 2ϩ sensitization.
Other proteins, thought to participate in Ca 2ϩ sensitization signaling pathways such as Rho-kinase, myosin phosphatase regulatory subunit MYPT-1, PP1␦, and the myosin phosphatase inhibitor protein CPI-17 were also detected in human and rabbit CC (Figs. 8 and 9). ROK␣ content at identical protein loading was greater in both human and rabbit CC compared with ileum, even without correcting for the lower myosin content in the CC (Fig. 8A). The reverse relationship occurred with ROK␤ content. Rabbit femoral artery contained both isoforms. The absolute amounts of ROK␣ and ␤ could not be quantitated due to the lack of recombinant ROK proteins and the unknown differences in sensitivities of the two antibodies used for detection. The MYPT-1 and PP1␦ contents in both human and rabbit CC, scaling to myosin contents, were similar to the phasic rabbit ileum and greater than in the tonic femoral artery (Fig.  8B). CPI-17 content in the rabbit CC (Fig. 9A), scaled to myosin C, telokin in rabbit tissues. Lanes 1-3, rabbit ileum smooth muscle; lanes 4 -6, rabbit CC; lanes 7-9, recombinant mouse telokin (FLAGtagged telokin runs higher than tissue telokin). D, MLC 20 . Lanes 1 and 2, purified chicken gizzard MLC 20 ; lanes 3 and 4, human CC; lanes 5 and 6, rabbit CC; lanes 7 and 8, rabbit ileum smooth muscle. The MLC 20 samples are from the same tissue specimens as used for A and C. The paired MLC 20 Western blot from the same tissue, used for quantitating the amount of Rho⅐GDI in smooth muscle excluding connective tissue, is not shown but was similar to D. content (Fig. 9B), was much closer to the femoral artery than the ileum. In the three samples of human CC, CPI-17 Western blots resulted in a lower molecular weight protein than found in rabbit CC, ileum, or femoral artery. DISCUSSION The RhoA signaling pathway that, through activation of Rho-kinase, inhibits myosin phosphatase resulting in an increase in MLC 20 phosphorylation and force in smooth muscle has been implicated in the regulation of tone in CC smooth muscle in studies in the rat in vivo and on isolated intact strips of cavernosal tissue from rat, rabbit, and human (19,23,35). The NO-cGMP pathway also plays a significant role in CC regulation (for review see Ref. 3) and may act on intact tissue by modulating both cytoplasmic calcium levels and/or Ca 2ϩ sensitivity (36). The use of permeabilized CC allowed us to test directly the role of RhoA-mediated Ca 2ϩ sensitization, and its reversal by a Rho-kinase inhibitor with Ca 2ϩ concentrations clamped allowing changes in force to reflect Ca 2ϩ -independent mechanisms, such as modulation of myosin phosphatase activity. The magnitudes of Ca 2ϩ sensitization induced by GTP␥S were unusually large, ϳ64% (range 44 -91%) of maximal calcium-induced force in rabbit CC and ϳ52% in human CC, compared with 20 -45% observed in other smooth muscles (24,37). Phenylephrine-induced Ca 2ϩ sensitization was only 10 -20%, similar to that found in other smooth muscles. The differences are likely due to GTP␥S being able to fully activate the entire RhoA pathway, whereas activation by agonists may be limited by availability of receptors and their coupling to RhoA and RhoA guanine nucleotide exchange factors. Although the human sample size is small (and included because of its relevance, despite the difficulty in obtaining such specimens), its somewhat smaller magnitude of calcium sensitization, compared with rabbit CC, does not reflect a deficiency in contractile proteins, because maximal Ca 2ϩ -induced forces were similar to rabbit CC. It could reflect the rundown of sensitized force of the human tissues from an unknown cause, or it could be related to disease. Further support for a physiological role of Ca 2ϩ sensitization arises from the report that endothelin-1 induces an intracellular Ca 2ϩ transient, which peaks at 250 -400 nM before returning to near resting levels of 60 -80 nM in about 2 min (38). Yet in intact tissue endothelin induces a maintained contraction suggesting that, if [Ca 2ϩ ] i is low as in cultured cells, The relative amounts of myosin, telokin, and RhoA in the rabbit ileum and CC were estimated by comparison with known concentrations of these proteins. 1 g of total protein extract contained 19 or 65 ng of myosin in the CC and ileum, respectively. The relative expression of RhoA normalized to myosin contact was markedly higher (17-fold) in the CC compared to the ileum, whereas there was no significant difference in the relative expression of telokin. The mDia precipitated RhoA (RB), as a fraction of WE, was greater in the CC than in ileal tissues (A and B). Starting tissue sample sizes were slightly different for CC and ileum as indicated in the WE lanes. C and D, control assays showing that, when the endogenous RhoA in tissue homogenates was loaded with GTP␥S or GDP, mDia beads extracted GTP␥S but not GDP RhoA. this tonic phase of force maintenance is due to the activation of the Ca 2ϩ sensitization pathway. The observed magnitude of Ca 2ϩ sensitization may be related to the very high RhoA content of human and rabbit CC smooth muscle (Fig. 5), which is 16-to 17-fold greater, respectively, than in rabbit ileum smooth muscle in which the Ca 2ϩ sensitization signaling pathway has been well characterized (31), but it may also reflect inhibition of an initially high myosin phosphatase activity in this tissue. Rabbit portal vein, femoral, and bladder have similar RhoA contents as ileum. 2 The MYPT-1 and PP1␦ contents of rabbit and human CC, when scaled to myosin content, are similar to that of the phasic ileum rather than the tonic femoral artery, in keeping with the higher phosphatase activity in phasic compared with the tonic smooth muscles (39). Current evidence strongly suggests that Ca 2ϩ -sensitized force and MLC 20 phosphorylation reflect inhibition of myosin phosphatase. GTP⅐RhoA activates Rho-kinase (16), and both translocate to the membrane with a time course compatible with the onset of Ca 2ϩ sensitization (40). Active Rho-kinase either directly phosphorylates and inhibits the regulatory subunit of myosin phosphatase (MYPT-1) or acts indirectly by phosphorylating another kinase, which in turn modulates the phosphatase activity (41). Our finding of a 17-fold greater RhoA content in CC than in ileum smooth muscle suggests that it may contribute to the flaccid state of CC. Mills and colleagues (42) have speculated that penile erection occurs as a result of two distinct processes. First, NO-initiated relaxation is mediated by a cGMPdependent increase in PKG activity with subsequent fall in [Ca 2ϩ ] i , decreased MLC kinase activity, and MLC 20 phosphorylation. In addition to mediating NO-induced relaxation, PKG may inhibit or prevent activation of RhoA and indirectly relax CC through disinhibition of MYPT-1 (43). The molecular basis and mechanism(s) of cGMP-induced relaxation of sensitized force remain to be determined. It is also tempting to speculate that dysregulation of the RhoA pathway may occur with prolonged, non-sexually mediated erection termed priapism. In this case the treatments of choice are ␣ 1 -adrenergic agonists (44). Lastly, recent data suggest that the Rho/Rho-kinase pathway is activated in the CC following castration and contributes to erectile dysfunction due to loss of testosterone (45). Thus the Rho kinase pathway may be involved in the physiology of erection and alterations in signaling contribute to erectile dysfunction.
A landmark paper by Narumiya and colleagues (20) described a series of compounds that potently relaxed agonistinduced smooth muscle contraction through inhibition of Rhokinase. One of these compounds, Y-27632, has been widely used to characterize Rho/Rho-kinase signaling pathways in many cell types. The IC 50 value for Y-27632-induced inhibition of GTP␥S-induced Ca 2ϩ -sensitized force in rabbit CC was 0.9 M, somewhat lower than the IC 50 of 2.8 M found by Rees (23), using Y-27632 in intact rabbit CC contracted with electric field stimulation. Y-27632 completely relaxed the GTP␥S Ca 2ϩ -sensitized force in both rabbit and human CC, whereas supra maximal concentrations of 8-Br-cGMP were much less effective in the rabbit (Figs. 1 and 2). In the three human CC tissue samples, 8-Br-cGMP induced a greater inhibition of GTP␥Ssensitized force than in the rabbit. In keeping with observations in cGMP-dependent kinase I-deficient mice in which PKA could not compensate for the lack of PKG (4, 5), the human samples in the present study showed a robust response to 8-Br-cGMP despite the patients' poor response to intercavernous prostaglandin E1 that raises cAMP. However, they also did not respond to the PDE-5 inhibitor sildenafil, which should lead to increased cGMP, perhaps indicating altered PDE-5 activity as these isolated permeabilized CC samples responded well to 8-Br-cGMP. In support of a significant role for the RhoA/Rho-kinase contribution, Chitaley et al. (19) found major increases in CC pressure with injection of Y-27632 in vivo, with or without electrical stimulation, as well as in the presence of nitric-oxide synthase inhibitors or cGMP inhibitors. This is consistent with the capabilities of the NO pathway and the Rho-kinase pathway to function independently. Thus, one would expect that the erectile dysfunction and low reproductive ability of cGMP-dependent kinase I-deficient mice (4) should be alleviated by inhibition of Rho-kinase. However, there may also be cross-talk between these pathways (46 -49), and their balance may be altered in different types of erectile dysfunction.
Both Rho-kinase isoforms ROK␣ (ROCKII) and ROK␤ (p160ROCK) (reviewed in Ref. 50) were present in CC where their ratio differed from that in ileal muscle. The two isoforms have similar functional domains and have no known functional differences or different targets, and their K i values for Y-27632 are very similar (21,50). In the absence of recombinant ROK proteins, total ROK content and the ratios of ROK␣ and ROK␤ could not be accurately determined. A rough estimate of total ROK from Western blots (e.g. Fig. 8A), scaled to the different myosin contents, suggests a higher ROK content of CC than ileal muscle. Thus, the large Ca 2ϩ sensitization observed in CC compared with other smooth muscles may reflect the 17-fold greater RhoA content that would favor formation of the RhoA⅐GTP⅐ROK complex and may also be related to a high content of the RhoA effector, ROK.
Stimulation of intact or permeabilized smooth muscle with 8-Br-cGMP or forskolin leads to phosphorylation of telokin, a 17-kDa protein whose sequence is identical to the C terminus of myosin light chain kinase (51) and is expressed in high concentrations in some smooth muscles such as ileum (25,43). The telokin content of smooth muscle correlates with the relevant effect of 8-Br-cGMP (25,43,52). The concentration of telokin relative to myosin in CC was similar to that found in ileum smooth muscle, ϳ70 M. Telokin has been suggested to either directly or indirectly activate myosin phosphatase, and its abundance in CC raises the possibility that it is also a target of PKG in this tissue and may contribute to cGMP-induced relaxation of Ca 2ϩ sensitization.
The finding of the phosphatase inhibitor protein, CPI-17 (30), in CC smooth muscle raises the additional possibility that this signaling pathway, which leads to Ca 2ϩ sensitization largely through PKC phosphorylation and activation of CPI-17, inhibition of PP1␦, and increased MLC 20 phosphorylation (53), might also contribute to erectile physiology. Indeed, phosphorylation of CPI-17, which converts this protein into a potent inhibitor, is reduced by treatment with 10 M Y-27632 in femoral artery (54). The migration of human CPI-17 at a lower molecular weight on SDS gels compared with the rabbit protein may reflect another isoform as reported for human aorta, in which two isoforms were detected (55), a truncation of the protein or proteolysis, although the latter is unlikely, because in the same samples there was no proteolysis of MYPT-1, which is highly susceptible to proteolysis. It is unclear if it is related to the disease state of the samples.
Erectile dysfunction is prevalent in ϳ9.6% of the population or 30 million men in the United States between the ages of 40 and 70 years, with diabetic men having a more than 3-fold increased prevalence (52% is those aged 55-59 years) when compared with non-diabetic men (56,57). Depending on the severity and etiology of erectile dysfunction, 20 -60% of men with erectile dysfunction exhibit a suboptimal response to phosphodiesterase inhibitors such as sildenafil (58). Our present findings as well as previous in vivo studies (19) suggest that Ca 2ϩ sensitization makes a sizable contribution to the contracted (flaccid) state of CC smooth muscle. Our results show that Ca 2ϩ sensitization can be completely reversed by watersoluble, orally active Rho-kinase inhibitors, suggesting that they may provide another line of therapy for these patients, particularly to the group resistant to therapy with PDE-5 inhibitors.