Polyamines Transduce the Nongenomic, Androgen-Induced Calcium Sensitization in Intestinal Smooth Muscle

Abstract

Androgens regulate body development and differentiation through a variety of genotropic mechanisms, mostly in reproductive organs. In recent years a different scenario for sex hormone actions has emerged: the intestinal muscle. Thus, although estrogens relax intestinal muscle, androgens are powerful inducers of mechanical potentiation. This effect of androgens was intriguing because it is observed at physiological concentrations, is mediated by nongenomic mechanisms, and involves a phenomenon of calcium sensitization of contractile machinery by stimulating phosphorylation of 20 kDa myosin light chain by Rho-associated kinase. Here we have deciphered the molecular mechanisms underlying calcium sensitization and mechanical potentiation by androgens in male intestinal muscle as well as its tight relationship to polyamine metabolism. Thus, androgens stimulate polyamine synthesis, and the inhibition of polyamine synthesis abolishes androgen-induced calcium sensitization and 20 kDa myosin light chain phosphorylation. We demonstrate that the first molecular step in the induction of calcium sensitization is a nonconventional activation of the adaptor protein RhoA, triggered by a transglutaminase-catalyzed polyamination of RhoA, which is then targeted to the membrane to activate Rho-associated kinase. Altogether, these results demonstrate that the physiological levels of androgens, through the modulation of polyamine metabolism and posttanslational modification of RhoA, activate a new signal transduction pathway in the intestinal smooth muscle to induce calcium sensitization. Furthermore, apart from being one of the few physiologically relevant nongenomic effects of androgens, these results might underlie the well-known gender differences in intestinal transits, thus expanding the nature's inventory of sex hormones effects.

Androgens are primary regulators of cell growth and differentiation, spermatogenesis, skeletal development, maintenance of bone metabolism, and development of secondary sexual characteristics both in males and females. These actions are primarily considered to be genotropic, and the major mechanisms accounting for these actions are through activation of high-affinity intracellular receptors, which are members of the nuclear receptor family, through the modification of DNA transcription upon binding to specific androgen response elements in target gene promoters (1). In addition, in recent years, compelling evidence has demonstrated that androgens (and other steroid hormones) can also exert many actions by nongenomic (or alternative) mechanisms (2–4). Such nongenomic effects are characterized by a time frame not sufficiently long to allow gene transcription/translation and are often initiated by the interaction of androgen molecules with membrane and/or cytoplasmic target proteins, even in cells lacking androgen receptors and in target and nontarget tissues (5, 6). However, the molecular mechanisms by which androgens and their receptors transduce cellular signals to provoke these effects are considerably diverse, from activation of different signaling cascades (7–10) to the direct modulation of ion channels (11–13).

Perhaps the best-studied acute action of androgens is the modulation of calcium homeostasis. Indeed, androgens are able to modify the intracellular Ca²⁺ levels within seconds to minutes in different types of muscle cells such cardio myocytes (14), skeletal muscle cells (10), and vascular myocytes (11, 15). In vascular smooth muscle, androgens reduce vascular tone and induce vasorelaxation, both in vivo and in vitro, by directly negatively modulating L-type Ca²⁺ channel activity and/or positively regulate large conductance Ca²⁺-activated potassium channels, rather than by activating the nitric oxide generation at the endothelial lining (reviewed in Reference 15). However, a major concern of the physiological relevance of these acute effects of androgens is that, in most cases, conclusions were drawn based on experiments conducted at nonphysiological androgen levels (>1 μM).

In contraposition to the relaxing effects of androgens reported to date, we have recently demonstrated that androgens induce the potentiation of contractile activity in intestinal muscle secondary to the induction of calcium sensitization of contractile apparatus (16–18). These effects are observed at low physiological concentrations of androgens and are strictly dependent on the activation of intracellular androgen receptors and the reduction of T to dihydrotestosterone (DHT) (16–18). In addition, androgen-induced potentiation of muscle force development and calcium sensitization is nongenomic in nature, given that they can be observed in the presence of inhibitors of transcriptional and translational processes and originate shortly after exposure to androgens (16–18). Our experimental evidence also showed that calcium sensitization induced by androgens was due to modifications in the phosphorylation state of the 20-kDa myosin light chain (LC₂₀) upon activation of Rho kinase (ROCK) (16–18). Although the final stage in the induction of calcium sensitization in the ileum and colon is the increase in the LC₂₀ phosphorylation state, the signaling pathways activated by androgens differ between intestinal segments. Thus, although in the ileum activation of ROCK directly phosphorylates LC₂₀ (16), in the colon, a protein kinase C-mediated activation of CPI-17 (an endogenous inhibitor of the myosin light chain phosphatase) triggered by ROCK is responsible for the effect (18).

In the current study, we have aimed at deciphering the molecular players and mechanisms upstream in the signaling pathway leading to ROCK activation in this novel nongenomic action of androgens. We demonstrate the key element is the activation of the small GTPase RhoA, a member of the Rho family of signaling adaptors (19, 20), which is switched on by androgens in an unprecedented manner. We show a common mechanism whereby androgens activate polyamine synthesis, which act as second messengers to cause the polyamination of the small GTPase RhoA through a transglutaminase-catalyzed reaction. This nonconventional activation of RhoA provokes the prolonged stimulation of its downstream effector, ROCK, to alter the phosphorylation status of LC₂₀ and, eventually, to induce calcium sensitization and mechanical potentiation. Interestingly, this nonconventional mechanism of RhoA activation has been previously demonstrated to be responsible for the dermonecrotizing effects of toxins of the Bordetella genus (21, 22), pinpointing the opportunistic strategies of some pathogenic agents.

Materials and Methods

Materials

5α-Dihydrotestosterone, putrescine, monodansylcadaverine, nifedipine, ionomycin, and mouse monoclonal anti-β-actin antibody were obtained from Sigma. α-Difluoromethylornithine (DFMO) was purchased from Bachem. L-[1-¹⁴C]ornithine and [1,4(N)-³H]putrescine were from PerkinElmer España. The goat polyclonal antibody recognizing LC₂₀ and the mouse monoclonal antibody directed to RhoA were purchased from Santa Cruz Biotechnology, Inc. The rabbit polyclonal antibody against phospho-Ser¹⁹-LC₂₀ was from Cell Signaling Technology. The mouse monoclonal antibody directed to α₁-Na⁺/K⁺-ATPase was from Millipore Iberica. The antirabbit, antimouse, and antigoat horseradish peroxidase-linked antibodies, the Hybond-P poly(vinylidenedifluoride) membranes, the enhanced chemiluminescence kit, and the protease and phosphatase inhibitor cocktails were purchased from Roche Diagnostics. The recombinant RhoA control protein and G-LISA kit were obtained from Cytoskeleton Inc.

Animals and tissues

All procedures were performed in accordance with the European Community and the University of La Laguna Ethics Committee guidelines for the care of laboratory animals.

Distal colonic and ileal segments from male Swiss CD1 mice (average weight 28 g) were dissected and placed in aerated cold physiological salt solution (PSS), containing (in millimoles) NaCl, 126; KCl, 4.5; MgSO₄, 1.0; CaCl₂, 2.0; NaH₂PO₄, 0.56; Na₂HPO₄, 1.44; glucose 15.0, pH 7.4. Longitudinal strips (1.5 cm long) of intestinal tissues were mounted in water-jacketed organ baths and incubated in aerated PSS at 37°C. Tissues were equilibrated at a resting tension value of 0.5 g. Ca²⁺-free solutions were made by replacing all CaCl₂ from PSS and supplemented with 25 μM EGTA.

Isometric tension of isolated muscle strips was measured using force transducers and sampled, digitized and filtered as described by Díaz (23).

Muscle permeabilization

Experiments in calcium-permeabilized smooth muscles were performed following the methodology established previously by our group (16, 18). To obtain simultaneous recordings of control and experimental conditions within the same animal, colonic and ileal tissues were cut into two halves (∼ 0.4 cm long), equilibrated at a resting tension of 0.5 g, and incubated for 30 minutes in aerated PSS. Calcium permeabilization was carried out by incubating intestinal preparations for 5 minutes in the permeabilization solution containing the following: ionomycin, 25 μM; EGTA, 1.6 mM; NaCl, 126 mM; KCl, 4.5 mM; MgCl₂, 1.2 mM; Tris-HCl, 20 mM (pH 7.4); and glucose, 15 mM. Tissues were then exposed to nifedipine (5 μM) for an additional 2–3 minutes before the application of calcium pulses.

Isolation of smooth muscle microsomes

Ileal and colonic muscle strips were dissected and immersed in liquid nitrogen. Homogenization was carried out in lysis buffer (25 mM HEPES; 150 mM NaCl; 10 mM MgCl₂; 1 mM EDTA; 10% glycerol; 1 mM Na₃VO₄; 25 mM NaF; 1× protease inhibitor cocktail; 0.01% phosphatase inhibitor cocktail, pH 7.5) and ultracentrifuged at 100 000 × g for 60 minutes at 4°C. The pellet was resuspended in 100 μl Mg²⁺ lysis buffer (Upstate) and centrifuged at 14 000 × g for 10 minutes at 4°C. The supernatant, containing the microsomal fraction, was collected and stored frozen.

Ornithine decarboxylase activity assays

Ornithine decarboxylase activity was determined following the procedures previously described (16, 24). Specific ornithine decarboxylase activity was assessed by the rate of decarboxylation of L-[1-¹⁴C]-ornithine from tissues incubated in the presence of dimethylsulfoxide (DMSO) or DHT and in the presence or absence of the irreversible ornithine decarboxylase (ODC) inhibitor DFMO (10 mM).

Western blot analyses

For total protein extraction, male murine intestinal tissues were processed for SDS-PAGE as previously described (4, 16). Equal amounts of extracted protein were electrophoresed on 12.5% SDS-PAGE, transferred to Hybond-P membranes, and submitted to Western blotting. Alternatively, to optimize resolution in the low-molecular-weight region, protein extracts were run on SDS-PAGE with modifications in the pH of the running buffer (8.5 instead of 8.8) and in the acrylamide to bisacrylamide ratio (30:0.8 instead of 30:0.15). Membranes were first preincubated at room temperature for 1 hour in 5% blotting-grade blocker nonfat dry milk (BioRad Laboratories) in Tris-buffered saline with 0.1% Tween 20 (TBS-T) or, alternatively, in 5% BSA diluted in TBS-T for the incubation with antibodies against phosphoproteins. Specific bands were visualized by incubating with horseradish peroxidase-linked secondary antibodies followed by membranes treatment with the Amersham enhanced chemiluminescence kit (ECL).

Activation of RhoA

For the analyses of RhoA activation, colonic smooth muscle strips from the same animal were incubated in PSS at different times with either DHT (10 nM) or vehicle (DMSO, 0.1%). The amount of GTP-bound RhoA (activated RhoA) was inmunodetected in cellular extracts using a G-lisa RhoA activation assay. Constitutively active RhoA was used as positive control.

Cross-link immunoprecipitation kit and 2-dimensional gel electrophoresis

Tissue samples were homogenized in immunoprecipitation buffer (25 mM Tris-HCl pH 7.4; 150 mM NaCl; 1 mM EDTA; 1% Nonidet P-40; 5% glycerol; and Halt protease and phosphatase inhibitor cocktail). For RhoA immunoprecipitation, the Pierce cross-link immunoprecipitation kit (Thermo Scientific) was used, following the procedure described by the manufacturer. Briefly, the anti-RhoA monoclonal antibody was captured to protein A/G agarose resin and covalently immobilized to the support with disuccinmidylsuberate. Then the antibody resin captured in spin columns was incubated overnight at 4°C with protein sample extracts. After washing, proteins bound to immobilized anti-RhoA antibody were eluted with 60 μL elution buffer provided in the kit. Antigen purity free of antibody fragments was tested in SDS-PAGE and Western blotting, using the same anti-RhoA monoclonal antibody.

Eluted samples were resuspended in a ready Prep 2-D starter kit (Destreak reagent; Bio-Rad Laboratories), adding 0.75% 3–10 pH ampholines (Bio-Rad Laboratories). Samples were made up to 125 mL and loaded on immobilized 7-cm pH 3–10 gradient strips (Bio-Rad Laboratories). Rehydration was performed by 14 hours in-gel active rehydration at 20°C with a constant current of 50 mA/strip, followed by isoelectrofocusing, which was performed at 20°C at 20 000 V/h for 4 hours. Strips were equilibrated for 10 minutes in equilibration buffer I [6 M urea; 0.375 M Tris-HCl, pH 8.8; 2% sodium dodecyl sulfate; 20% glycerol; 2% (wt/vol) dithiothreitol], followed by 10 minutes in equilibration buffer II [6 M urea; 0.375 M Tris-HCl, pH 8.8; 2% sodium dodecyl sulfate; 20% glycerol; 2.5% (wt/vol) iodoacetamide]. Two-dimensional separation was performed on 12.5% SDS-PAGE. Proteins were then transferred to Hybond-P membranes and immunoblotted with anti-RhoA-specific antibody using Transblot Turbo (BioRad Laboratories) and processed for Western blotting.

Determination of RhoA polyamination

Ileal and colonic muscles were incubated in aerated PSS and metabolically labeled with 5 μCi of [1,4-³H(N)] putrescine dihydrochloride (1.0 mCi/mL) for 60 minutes in polyamine-free medium. Tissues were then incubated with vehicle (DMSO, 0.1%) or DHT (10 nM) for an additional 60 minutes in the presence of radioactive putrescine, and then total extracts and microsomal fractions were obtained. The polyamination of RhoA was detected by autoradiography and immunoblotting. Protein extracts were processed for either 1-dimensional electrophoresis on SDS-PAGE or cross-link immunoprecipitation for 2-dimensional gel electrophoresis. After the immunoblotting, membranes were washed in TBS-T and air dried. Radioactive bands and spots were visualized by autoradiography using Kodak XAR films, which were exposed for 15 days to capture the weak signal from ³H-labeled RhoA.

Relative quantification of mRNA levels by real-time quantitative RT-PCR

Total RNA was purified from 10- to 20-mg tissue samples, preserved in RNAlater (Ambion), using the Illustra RNA spin mini-RNA isolation kit (GE Healthcare), followed by deoxyribonuclease I digestion, acid phenol-chloroform extraction, and ethanol precipitation (25). cDNA samples were obtained with the Transcriptor first strand synthesis kit (Roche) using anchored oligo(deoxythymidine)₁₈ primers and 5 μg of total RNA as template. Real-time amplification reactions were performed using SYBR Green detection and run in triplicate on the LightCycler 480 platform (Roche). Additional information is provided in the Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org.

Statistical and mathematical analyses

Results are expressed as mean ± SEM. Differences between sample means were assessed by ANOVA or Kruskal-Wallis test followed by a Student-Newman-Keuls t test, post hoc Tukey honestly significant difference test, or Mann-Whitney U test where appropriate. Comparison between segments from the same animals was assessed by a paired t test or a Wilcoxon signed-rank test. Frequency analyses were performed using Fast Fourier transforms following the procedures reported by Díaz (26).

Results

Androgens and polyamines induce calcium sensitization in ileal and colonic smooth muscle cells

We have previously shown that application of physiological doses of T or DHT cause a considerable increase in the mechanical responses of colonic and ileal longitudinal muscles to extracellular stimuli, ie, CaCl₂ and carbachol (16, 18). Androgens also induce calcium sensitization in permeabilized smooth muscle preparations, which is secondary to the augmented phosphorylation of LC₂₀ at Ser¹⁹ relative to total LC₂₀ (16–18). For introductory purposes these effects were reproduced in the present study and shown in Figure 1A. Preexposure to DHT (10 nM for 90 min) elicited a significant increase in mechanical tension in ileal and colonic preparations under all calcium pulses, including at subthreshold extracellular calcium concentration (50 μM), which was not observed in vehicle-treated tissues (Figure 1A). These effects of DHT were accompanied by increased phospho-Ser¹⁹-LC₂₀ relative to total LC₂₀ both in ileal and colonic (but not duodenal) longitudinal muscles from the same animals (Figure 1B). In the ileum, these effects of androgens were mimicked by the diamine putrescine (16). Now we report that the calcium-sensitization effects of putrescine also occurs in colonic muscle (Figure 1C). Thus, incubation of colonic tissues with putrescine (500 μM for 90 min) increases not only the mechanical response to calcium pulses but also the ratio phospho-Ser¹⁹-LC₂₀ to total LC₂₀ to a similar extent than for ileum (Figure 1C), indicating that polyamines (at least the diamine putrescine) activate a calcium-sensitization process in longitudinal smooth muscle from both intestinal segments, therefore mimicking the effect of androgens.

Figure 1.

DHT and putrescine induce calcium sensitization in colonic and ileal longitudinal smooth muscle. A, Typical responses of ionomycin-permeabilized ileal and colonic longitudinal muscles preincubated with vehicle (0.1% DMSO, trace in black) or DHT (trace in gray) to extracellular calcium. Insets, Summaries of the results from another 8–10 experiments. **, *, P < .01 and P < .05 compared with vehicle, respectively. B, Representative immunoblot analyses for total and phospho-Ser¹⁹-LC₂₀ in colonic, ileal, and duodenal muscle extracts preexposed for 90 minutes to either vehicle (V; 0.1% DMSO) or DHT (10 nM). C, Upper panel, Typical responses of ionomycin-permeabilized colonic muscle preincubated with vehicle (V; trace in black) or putrescine (PUT; trace in gray) to extracellular calcium. Inset, A summary of the results from another seven experiments.*, P < .05 compared with V. Middle panel, Immunoblot analyses for total and phospho-Ser¹⁹-LC₂₀ in colonic and ileal muscle extracts, exposed for 90 minutes to either vehicle (V) or putrescine (PUT). α-Actin was used as a control of protein load. Lower panel, Densitometric values of phospho-LC₂₀ relative to total LC₂₀ and relative to the average of immunosignals obtained with vehicle-treated tissues (V). Four experiments were performed for each tissue and treatment. **, P < .01 compared with V.

Androgens stimulate polyamine synthesis

Critical enzymes involved in the generation of polyamines in mammals are ODC (EC 4.1.1.17) and S-adenosylmethionine decarboxylase (SAMDC; effective concentration 4.1.1.50). First, we have quantified the expression of genes coding for ODC and SAMDC in intestinal muscle, Odc1, and Amd, respectively (Supplemental Table 2). Our results using real-time quantitative RT-PCR indicate that both Odc1and Amd genes are expressed in the smooth muscle of ileum and colon, their expression levels of Odc1 being significantly higher in the ileum than in the colon. Compared with the kidney, used here as positive control, Odc1 and Amd expression levels in both intestinal tissues were approximately 10% and 125%, respectively (see Supplemental Table 2). We have previously shown that DHT increases ileal muscle ornithine decarboxylase activity, the rate-limiting enzyme in the generation of intracellular polyamines (16). Although significant, these effects were observed at supraphysiological concentrations of androgens (16). However, whether stimulation of ODC activity by DHT occurs in the intestinal muscle at the physiological dose used here and within such small time course was unknown. Therefore, we analyzed the change in ODC activity in both intestinal sections in response to DHT (10 nM for 60 min). Intestinal homogenates (colon and ileum) from the same animals were incubated with either vehicle or DHT and in the presence or absence of the ODC competitive inhibitor DFMO (27). The results illustrated in Figure 2 show that both intestinal muscles exhibited sizable DFMO-sensitive ODC activities under vehicle and DHT-treated conditions (P < .05 for both ileum and colon). Compared with controls, preincubation with DHT for 60 minutes significantly increased ODC activities in both intestinal segments.

Figure 2.

Androgens stimulate ODC activity in ileal and colonic smooth muscles. Muscle strips were preincubated with vehicle (V; 0.1% DMSO) or DHT (10 nM) for 60 minutes before ODC activity measurements. ODC activities in colonic and ileal preparations from the same animals were measured in response to DMSO and DHT and in the presence or absence of DFMO (10 mM). * #, P < .05 and P < .001 compared with vehicle and DHT+DFMO, respectively. Results correspond to data from four animals.

Blockade of polyamine synthesis prevents androgen-induced calcium sensitization, LC₂₀ phosphorylation, and contractile potentiation

We examined the effects of pharmacological inhibition of ODC activity on the stimulatory effect of androgens. The results depicted in Figure 3A show that preexposure to DFMO (500 μM for 60 min) completely abolished DHT-induced calcium sensitization in colonic longitudinal muscle. Similar results were observed in ileal segments (not shown). In agreement, preincubation with DFMO completely prevented the increase of phospho-Ser¹⁹-LC₂₀ induced by DHT in ileal and colonic tissues (Figure 3B). We also tested for the effects of polyamine biosynthesis inhibition on androgen-induced potentiation of intestinal contractile activity (Figure 3C). Preincubation of intestinal tissues with DFMO (500 μM for 60 min) before application of T (10 nM for 90 min) completely prevented the potentiation of contractile responses to extracellular calcium (2 mM) or carbachol (CCH; 1 μM) (Figure 3C, left lower panel). However, secondary incorporation of putrescine (500 μM for 90 min) to DFMO+DHT-pretreated tissues efficiently restored the response of intestinal tissues to CaCl₂ and carbachol, similarly to T alone (Figure 3C, right lower panel). In this case, we could observe a considerable increase in the response to putrescine (∼300% on peak phasic tension for both stimuli) compared with control tissues. Taken together, these experiments demonstrate that polyamine synthesis is an absolute requirement for the contractile potentiation induced by androgens.

Figure 3.

Inhibition of polyamine synthesis prevents androgen-induced calcium sensitization, LC₂₀ phosphorylation, and mechanical potentiation. A, Effects of DFMO on DHT-induced calcium sensitization in permeabilized colonic muscle. Inset, A summary of the results from another three experiments. ***, P < .005 compared with DHT. B, Western blot analyses for total and phosho-Ser¹⁹-LC₂₀ in muscle extracts. Bottom panel, Densitometric values of phospho-Ser¹⁹-LC₂₀/total LC₂₀ relative to vehicle-treated tissues. Four experiments were performed for each tissue. *,**, P < .05 and P < .05 compared with V; #, ##, P < .05 and P < .01 vs DHT. C, Upper panels, Effects of vehicle (DMSO; 0.1%, 90 min) and T (10 nM, 90 min) on consecutive stimulation with CaCl₂ and CCH in ileal segments. Left lower panels, Effects of DFMO on T-induced contractile potentiation. Right lower panel, Effects of incorporation of putrescine on the responses to CaCl₂ and CCH in (DFMO+T)-treated ileal muscle. Recordings from the same mice were taken at times 0 and 90 minutes after the addition of putrescine or vehicle. DFMO was added to the bath 45 minutes before the application of

Androgens and polyamines induce the translocation of RhoA to the plasma membrane

The hypothesis that polyamine biosynthesis underlies the stimulatory effect of androgens led us to explore whether androgens and polyamines modify the dynamics of RhoA compartmentalization. Using microsomal preparations from colonic and ileal tissues exposed to DHT (10 nM for 60 min), we could observe that androgen stimulation triggered the translocation of RhoA to the sarcolemma (Figure 4A). In addition, preexposure to DFMO (10 mM for 30 min) totally abolished the increase of the RhoA to α1-Na⁺K⁺-ATPase ratio elicited by DHT (Figure 4A), indicating that DHT-induced translocation of RhoA is dependent on the activation of the polyamine biosynthetic route. To further reinforce this point, we then proceeded to isolate intestinal microsomes from putrescine-incubated tissues (500 μM for 60 min). The results shown in Figure 4B demonstrate that putrescine incubation brings about a significant increase of plasmalemmal RhoA normalized to α1-Na⁺/K⁺-ATPase levels. These observations indicate the involvement of polyamine synthesis in the modulation of RhoA dynamics induced by androgens.

Figure 4.

Androgens and polyamines induce the translocation of RhoA to the plasma membrane. A, Left panel, Immunoblotting assay with anti-RhoA and plasma membrane marker (α1-Na⁺/K⁺-ATPase) in whole-tissue extracts (WTE) and microsomal fractions obtained from the colon and ileum. Tissues were incubated for 60 minutes with vehicle (V; 0.1% DMSO), DHT, or DFMO+DHT. DFMO-treated tissues were allowed to preincubate for 30 minutes before the addition of DHT. Right panel, Densitometric values of RhoA relative to α1-Na⁺/K⁺-ATPase expressed as the percentage of vehicle. #, *, P < .05 vs vehicle-treated tissues and DHT, respectively. B, Western blot analyses in whole-tissue extracts and microsomal fractions for RhoA in colonic and ileal muscle in response to putrescine incubation. Tissues were incubated with putrescine (PUT) before the preparation of extracts and microsomal fractions. Right panel, Densitometric analyses for RhoA bands relative to α1-Na⁺/K⁺-ATPase. ***, P < .005 vs V. Four assays were performed under each condition in all experiments. The cytosolic marker heat shock protein 90 (HSP90) was used as a control of microsome purity.

Androgens induce polyamination of RhoA

We next tested for potential postransductional modification of RhoA by polyamines. Previous studies on Bordetella bronchiseptica have shown that the toxin produced by these bacteria-causing lethal dermonecrosis and splenoatrophy catalyzes the polyamination of RhoA by cross-linking ubiquitous putrescine (and also spermidine and spermine) to Gln⁶³ (21, 22). It would be expected that if polyamines covalently bind to RhoA, a shift in the isoelectrofocusing of the protein might be observed. Therefore, we performed 2-dimensional electrophoresis of RhoA purified from vehicle- and DHT-treated ileal and colonic muscle extracts. Tissue protein extracts were processed for cross-link immunoprecipitation of RhoA protein. This method irreversibly binds the antibody to agarose beads, and enabled the purification of RhoA without contamination by antibody fragments and then allowing the enrichment of the protein for 2-dimensional gel electrophoresis. Then electrophoresed samples were immunoblotted and incubated with anti-RhoA antibody, observing the presence of a single spot at approximately 5.86–5.88 isoelectric point (pI) in DMSO-treated ileums (Figure 5A, upper panel), and two spots with pIs 5.88 and 6.24 in DHT-treated ileums (Figure 5A, middle panel). Similar results were obtained in the colon.

Figure 5.

Androgens induce the polyamination of RhoA in colonic and ileal muscles. A, Representative 2-dimensional immunoblotting analysis of RhoA isoforms in ileum. Cross-link immunoprecipitated tissue samples were run on 2-dimensional electrophoresis in pH 3–10 gradient strips, followed by second-dimensional separation on 12.5% SDS-PAGE. Muscle preparations from the same animal were treated with either vehicle (0.1% DMSO) or DHT (10 nM) for 60 minutes. Arrows indicate the calculated isoelectric points. Recombinant RhoA protein was used as a control. B, Western blot and ³H-putrescine (H³-PUT) autoradiography obtained in another experiment on ileal muscle processed as in A. Autoradiograms for vehicle and DHT membranes are shown in the middle panel. Calculated isoelectric points are indicated with arrows along with the single radioactive spot only in DHT-treated samples (full autoradiogram images are included in Supplemental Figure 2). C, left panel, Representative immunoblots of RhoA in microsomal fractions. Tissues were metabolically labeled with ³H-putrescine. Radioactive bands detected by autoradiography on the same membranes ([H³]RhoA), corresponding to polyaminated-RhoA, are shown in the lowest row. Right panel, Densitometric analyses for the 28-kDa band (polyaminated RhoA) relative to unmodified RhoA (30 kDa) and normalized to α1-Na⁺/K⁺-ATPase, in a set of four different experiments. ***, P < .005 vs vehicle (0.1% DMSO).

It is worth mentioning that in each of these experiments both treatments were performed in ileal and colonic segments from the same animals, thereby reducing interanimal variability. Furthermore, the empirical pI values (5.85–5.88) estimated for the more acidic spots in DMSO and DHT were nearly identical with the pI value of 5.89 for the recombinant RhoA protein run on parallel 2-dimensional gels as a control (Figure 5A, lower panel). In silico analysis of RhoA protein from Mus musculus (accession number NP_058082) indicates that the average pI for the 193-amino acid RhoA protein is 5.82, thus very similar to the values observed here. The more alkaline spot arising from DHT treatment indicate a net positive charge gain in RhoA by incorporation of a functional group such as a basic amines carried by polyamines. Indeed, our in silico analysis indicates that substitution of Gln⁶³ by Lys or Arg in RhoA, thereby incorporation of a single amino group, gives rise to a protein with average isoelectric point at 6.25, which fits neatly with our experimental value (6.24). This raised the hypothesis that RhoA is polyaminated upon androgen treatment. To confirm this possibility, we metabolically labeled intestinal tissues with ³H-putrescine before treatments with DMSO or DHT. Then muscle extracts were processed for cross-link immunoprecipitation and 2-dimensional gel electrophoresis as described before. After immunoblotting with specific anti-RhoA antibodies, a radioactive spot as a result of ³H-putrescine incubation was exclusively observed in DHT-treated samples (Figure 5B). This ³H-carrying spot isoelectrofocused at pI 6.31 likely corresponds to a polyaminated form of RhoA protein generated in response to DHT treatment.

We next tested whether these modifications of RhoA were reflected in the dynamic of translocation of RhoA in response to androgens. We adapted the electrophoresis conditions to obtain a fine resolution for low-molecular-weight protein separation. The results in Figure 5C (also in Figure 6, A and B) show that incubation with DHT (but not with DMSO) gives rise to the appearance of two closely spaced RhoA bands in ileal and colonic microsomal fractions, migrating at approximately 30 and 28 kDa in both tissues. These results encouraged us to ascertain whether the faster-moving band observed in Western blots in the microsomal fraction of intestinal muscle might correspond to the polyaminated RhoA spot observed in the 2-dimensional assays described above. We pretreated tissues with ³H-putrescine as indicated before and isolated microsomes from control and DHT-treated tissues. Microsomal extracts from ileal and colonic muscles were separated and visualized by Western blotting and autoradiography (Figure 5C, lowest band). The results demonstrated that radioactive putrescine was exclusively incorporated to the 28-kDa DHT-induced RhoA band and demonstrate that androgens elicit not only the polyamination of RhoA but also its translocation to the membrane fraction. The possibility that transglutamination occurs on Gln⁶³ of RhoA was also tested by proteomic approaches.

Figure 6.

Inhibition of TG2 prevents RhoA polyamination and calcium sensitization. A and B, Representative Western blots for RhoA in whole-tissue extracts (A) and microsomal fractions (B) from colonic and ileal muscles. α-Actin and α1-Na⁺/K⁺-ATPase were used as normalizing markers. Four experiments were performed for each tissue and experimental condition. **, ***, P < .01 and P < .005 vs vehicle-treated tissues, respectively; ###, P < .005 vs DHT. C and D, Effects of the specific TG2 inhibitor MDC on DHT-induced mechanical potentiation (C) and LC₂₀ phosphorylation (D) in permeabilized ileal muscle. Inset in C, Summary of the results of three experiments in the ileum. ***, P < .005 vs DHT. D, Densitometric values of phospho-LC₂₀ relative to total LC₂₀ calculated for vehicle (V), DHT, or MDC+DHT. At least three experiments were performed for each tissue and treatment. *, P < .05 compared with V; #, ##, P < .05 and P < .01 compared with DHT, respectively.

We also addressed the time course of RhoA activation (GTP bound RhoA) in response to DHT. Ileal and colonic segments from the same animal were exposed to either vehicle or DHT for different times from 0 to 60 minutes, and the levels of RhoA-GTP in cellular extracts were immediately determined. Remarkably, the results showed that DHT treatment was unable to induce any appreciable increase of cellular RhoA-GTP within the time course (0–60 min) of the experiment (Supplemental Figure 1). However, in vehicle-treated tissues, we have observed a slight although significant reduction of RhoA-GTP as measured at 20 minutes, rendering a significant difference compared with DHT-treated tissues. Because this difference was due to the reduction of RhoA-GTP rather than to an increase in the response to DHT, we concluded the hormone does not alter the rate of RhoA activation.

Inhibition of transglutaminase prevents RhoA polyamination and calcium sensitization

We further explored the biochemical mechanism for RhoA covalent polyamination. We speculated that an endogenous transglutaminase activity might be responsible for the transfer of polyamine groups to an acceptor Gln in RhoA in response to androgens. One of the transglutaminases ubiquitously expressed in intestinal tissue is tissue transglutaminase 2 (TG2) (28). We have explored the potential participation of TG2 in the response to androgens by using the specific inhibitor monodansylcadaverine (MDC). Ileal and colonic tissues were incubated with MDC (200 μM for 45 min) before application of DHT (10 nM for 90 min). The results showed that preincubation with MDC vastly reduced the amounts (and even provoke the disappearance) of the 28-kDa polyaminated RhoA band in whole extracts (Figure 6A) and prevented its translocation to the membrane fraction (Figure 6B). These results demonstrate the involvement of TG2 in the androgen-induced polyamination of RhoA.

To further demonstrate that TG2-mediated RhoA polyamination is essential for the induction of calcium sensitization, we assessed the effects of MDC on DHT-induced contractile response of ileal and colonic muscles to calcium and also on the phosphorylation status of LC₂₀. The results demonstrate that MDC completely abolished the potentiation of mechanical response to calcium of DHT-treated tissues (shown in Figure 6C for ileum). Paralleling this finding, inhibition of TG2 prevents the increase in phospho-Ser¹⁹-LC₂₀ relative to total LC₂₀ induced by DHT and reduces the amount of phospho-LC₂₀ to levels similar than in vehicle-treated tissues (Figure 6D).

Discussion

The naturally occurring polyamines putrescine, spermidine, and spermine are considered essential growth factors in eukaryotic cells. Physiological functions of polyamines include nucleic acid packaging and regulation of gene expression as well as modulation of cell proliferation, differentiation and development (29). High intracellular concentrations of these amines are found not only in tissues with high cell-turnover but also in quiescent tissues, such as muscle and neuronal tissues, suggesting that polyamines are important also for processes other than growth. Indeed, emerging evidence has also shown that polyamines can modulate second-messenger and signal transduction pathways (30) in addition to ion channels function (31). Here we demonstrate that polyamines are also capable of inducing the mechanical potentiation of intestinal muscle by activating a calcium sensitization phenomenon, which is triggered by the increase in the phosphorylation status of LC₂₀ at Ser¹⁹. Interestingly, the specific phosphorylation of LC₂₀ at this residue appears to be the common denominator of calcium sensitization processes in response to GTPγS and different agonists (32–34).

Importantly, the effects of polyamines observed here mimicked those of DHT, the active metabolite of T in target tissues, on distal ileal and colonic muscles (16–18). Indeed, we have previously shown that calcium sensitization and potentiation of contractility induced by androgens occurs at physiological concentrations of androgens (EC₅₀ values below 1 nM for both T and DHT) and within a short time course (16–18). Moreover, we have shown that these mechanical effects of androgens on intestinal muscle are strictly dependent on the activation of androgen receptors (which are expressed in ileal and colonic muscles as two receptor isoforms) but independent of transcriptional and translational processes (16–18), therefore involving nonconventional cellular signals.

Several studies performed in nonmuscle and muscle preparations have shown that androgens affect polyamine metabolism at different levels, from regulation of ODC and SAMDC gene expression (35–37) to direct stimulation of ODC enzymatic activity (24, 38, 39). Although less explored than for genomic regulation, it seems that direct regulation of ODC by androgens involves complex, cell-specific, posttranslational mechanisms initiated upon activation androgen receptors (24, 38, 39). Our present data show that not only intestinal muscle expresses the genetic machinery for polyamine synthesis but also that androgens increase ODC activity in the colon and ileum. Accordingly, the increase of ODC activity induced by androgens leads to augmented intracellular putrescine concentration from decarboxylation of L-ornithine, which would explain the similarity of putrescine and androgen responses and that secondary incorporation of putrescine to DFMO-treated tissues rescued the potentiation mechanism. Paralleling these observations, incubation with the ODC inhibitor abolished calcium sensitization and returned levels of phospho-Ser¹⁹-LC₂₀ to control values, hence confirming the hypothesis that androgens induce sensitization through the stimulation of polyamine synthesis.

Work from several laboratories, including ours, have implicated the monomeric GTPase protein RhoA in the induction of calcium sensitization in smooth muscle in response to GTPγS and agonists (33, 34, 40). It has been shown that agonists activate RhoA, allowing its interaction with the effector ROCK, leading to its activation. Subsequently activated ROCK increases phospho-LC₂₀ either by direct phosphorylation at the same site that is phosphorylated by myosin light-chain kinase or by inhibition of myosin light-chain phosphatase (32–34, 40, 41). Our recent studies demonstrate that the increased LC₂₀ phosphorylation caused by androgens depends on the intestinal segment (16–18). Hence, although in ileal muscle the augmented level of phospho-Ser¹⁹-LC₂₀ is rather due to direct phosphorylation by ROCK, in colonic tissues, ROCK activates the smooth muscle-specific protein CPI-17, an endogenous inhibitor of myosin light-chain phosphatase, favoring the phosphorylated state of LC₂₀. Despite these differences between segments, in both cases ROCK is located upstream in the signaling pathway, and inhibition of ROCK activity with Y27632 prevents LC₂₀ phosphorylation and halts calcium sensitization and potentiation of contractility in ileal and colonic muscles (16–18).

One critical step in the stimulation of ROCK activity by androgens is the upstream activation of RhoA. It is known that activation of RhoA involves GDP/GTP exchange reaction and its translocation to the plasma membrane (40, 42). In agreement, we could observe that androgens induce the rapid translocation of RhoA to the microsomal fraction and, in line with the above results; this effect is mimicked by putrescine but completely prevented by the inhibition of ODC activity with DFMO, providing a direct link between androgen stimulation and polyamine-induced activation of RhoA and ROCK.

The clear involvement of intracellular polyamines in response to androgens as well as the undetectable levels of conventional RhoA activation (RhoA-GTP) suggested some sort of direct interaction between RhoA and polyamines. Indeed, this hypothesis was corroborated in 2-dimensional immunoblotting experiments of RhoA immunoprecipitated from the ileum and colon, which demonstrated that androgen treatment gave rise to a second RhoA isoform, which incorporated the radioactivity from labeled putrescine, and likely corresponds to polyaminated RhoA. In addition, this newly formed isoform coexisted with the canonical unmodified 30-kDa RhoA band in ileal and colonic microsomal fractions upon DHT or polyamine treatments. Taken together, these results demonstrate a novel posttransductional regulation of RhoA through covalent cross-linking with polyamines and substantially differ from classical G protein-induced RhoA activation and signaling elicited by agonists.

In our study, we further attempted to identify the biochemical process responsible for RhoA polyamination. It has been shown that RhoA is an in vivo substrate of tissue TG2, and it has been demonstrated that transamidation of RhoA promotes the formation of stress fibers and focal adhesion complexes in cultured cells (43). Here, by using the pseudosubstrate inhibitor of TG2, MDC, we have demonstrated that a transglutaminase-catalyzed polyamine transfer is responsible for the nonconventional RhoA activation observed in response to DHT. The inhibitor prevents not only the appearance of polyaminated RhoA but also its translocation to the plasma membrane both in whole extracts and microsomal fractions. That this TG2-mediated transamidation reaction is essential for the induction of calcium sensitization triggered by androgens was corroborated by the dramatic depression of contractile force and the reduction of phospho-LC₂₀/total LC₂₀ in the presence of the TG2 inhibitor.

In summary, we demonstrate here that polyamines are the signaling transducers for androgen-induced calcium sensitization of intestinal longitudinal smooth muscle. At physiological concentrations, androgens activate androgen receptors to trigger a cascade of signaling events involving the stimulation of ornithine decarboxylase and a concomitant increase in intracellular polyamines. Polyamines act as molecular signals to nonconventionally activate RhoA through a TG2-catalyzed reaction, which produces polyaminated RhoA. This molecular form of RhoA translocates to the plasma membrane to activate ROCK and, eventually, to unbalance the phosphorylation state of LC₂₀ toward its phospho-Ser¹⁹ form, leading to calcium sensitization of the contractile apparatus. A cellular model depicting the interactions of the different elements leading to calcium sensitization is shown in Figure 7. The novelty of our results lies not only in the fact that intestinal tissues must be considered new targets for androgens but also because at physiological concentrations they activate a novel signaling pathway in which intracellular polyamines act as modulators of signal adaptors involved in endocrine regulation of intestinal smooth muscle. From the heuristic point of view, our results provide one of the few biologically relevant scenarios for the emerging field of nongenomic actions of steroid hormones.

Figure 7.

Signaling model depicting the mechanisms leading to calcium sensitization by androgens in ileal and colonic smooth muscle cells. For details, see Discussion. AR, androgen receptor; CaM, calmodulin; IP3, inositol 1,4,5 triphosphate; MLCK, myosin light-chain kinase; MLCP, myosin light-chain phosphatase.

From a pathophysiological point of view, our data might explain, at least in part, why intestinal transit times seem to be lower in females than in males, especially during the luteal phase of menstrual cycle or during pregnancy (44–46), a finding that may be accounted for by two reasons: 1) the low circulating androgen levels in women prevent calcium sensitization of colonic contractile machinery, and 2) estrogens tend to reduce colonic contractility, although in this case the mechanism appears to be associated with a reduction in muscle excitability, at least in the short time course (4, 47). Altogether these observations suggest an endocrine scenario in which the balance between circulating estrogens and androgens conditions the contractile behavior of the distal intestine. Furthermore, it is known that colorectal motility disorders like chronic constipation due to slow transit, one common and most difficult to treat subtype of constipation, are more common in female than in male patients (48). Thus, our present data showing the involvement of androgens and polyamines in the modulation of intestinal contractile activity might pave the way to develop novel strategies to treat severe types of chronic and severe idiopathic constipation.

Funding Statement

This work was supported by Research Grants SAF2007–66148-C02–02 and SAF2010–22114-C02–01/02 from the Spanish Ministry of Science and Innovation (Spain) and the European Regional Development Fund Programme.