Smooth muscles permeabilized with deter-. Physiol ..... quartz cuvette with 1 ml of Krebs solution containing one of ...... We thank Susan U. Mooers and Srinivasa.
Metabolic characteristics of cx-toxin-permeabilized
smooth
A
muscle
AND THOMAS M. BUTLER LAURA TRINKLE-MULCAHY, MARION J. SIEGMAN, Department of Physiology, Jefferson Medical College, Philadelphia, Pennsylvania 19107 Trinkle-Mulcahy, Laura, Marion J. Siegman, and Thomas M. Butler. Metabolic characteristics of a-toxinpermeabilized smooth muscle. Am. J. Physiol. 266 (CeZZ Physiol. 35): C1673-C1683, 1994.-Rabbit portal veins were permeabilized using StaphyZococcus aureus a-toxin, and adenosinetriphosphatase (ATPase) was measured as the formation of [“H]ADP, [“HIAMP, and [“Hladenosine from [3H]ATP in the solution bathing the muscle. The resting ATPase (1.96 2 0.15 mM/min, n = 13) is - 5-10 times higher than that measured in Triton X-100-permeabilized muscles (0.28 2 0.01 mM/min, IZ = 4), with nucleotide accumulating as ADP, AMP, and adenosine. The ATPase activity is also seen when the intact muscle is incubated in a Krebs solution containing 1 mM MgATP (2.76 2 0.10 mM/min, n = 73). This suggests that it is due primarily to an ecto-ATPase. The ectoenzyme is capable of hydrolyzing both ATP and ADP, and in both cases there is a higher rate at 3 than at 1 mM nucleotide. The high resting ATPase compromises the control of nucleotide concentrations within the permeabilized tissue even in the presence of an ATP-regenerating system consisting of phosphocreatine (PCr, 35mM) and creatine kinase (1 mg/ml). Treatment of the intact muscle with the ectonucleotidase inhibitor 4,4’-diisothiocyanatostilbene-2,2’-disulfonic acid (DIDS) followed by a-toxin permeabilization and inclusion of sodium azide in subsequent solutions reduces the ecto-ATPase by - 70%. Addition of PCr and creatine kinase then results in the maintenance of high IATP] and low [ADPJ in the muscle, and importantly, there are no significant changes in [ATP], [ADP], [adenosine/AMP], or the ADP-to-ATP ratio upon activation of the muscle in pCa 4.5. In general, the force output in high Ca2+ increased as the metabolic profile of the muscle improved. When ATPase was measured as the appearance of 132P]Pi from [32P]PCr and [y-;j2P]ATP, the cx-toxin-permeabilized muscle subjected to the above treatment showed only - 30% higher total ATPase under activated conditions compared with the freeze-glycerinated Triton-treated portal vein. The suprabasal ATPase is similar in both preparations. We conclude that the reduction of the basal ATPase by the DIDS-azide treatment permits both rigorous control of nucleotide contents and accurate measurement of ATPase activity in cx-toxin-permeabilized smooth muscle. ecto-adenosinetriphosphatase; 4 ,4’-diisothiocyanatostil 2,2’-disulfonic acid; sodium azide
bene-
of the properties and regulation of smooth muscle contractile proteins has been greatly facilitated by the use of various cell permeabilization techniques, which allow controlled manipulation of the cytosolic contents as a method for the study of intracellular processes. Chemical permeabilization techniques range from the use of detergents, which remove some or all cellular membranes while leaving the contractile proteins organized in their filamentous matrix, to bacterial toxins, which form small pores in the plasma membrane, allowing passage of ions and small molecules while retaining intracellular proteins and receptorINVESTIGATION
0363-6143/94
$3.00
Copyright
~3 1994
linked signal transduction pathways (for review, see Refs. 27,29). Smooth muscles permeabilized with detergents such as Triton X-100 and saponin retain Ca2+ sensitivity, with a force output nearly equal to that in the intact muscle (5,2S>. Triton X-100 removes both the plasma membrane and all intracellular membranes, leaving a simplified system in which regulation can be studied at the molecular level without interference from internal Ca2+ stores and endogenous second messengers. Saponin complexes with cholesterol in membranes to form large pores and thus selectively permeabilizes the cholesterol-rich plasma membrane, leaving internal Ca2+ stores intact. This system is useful for studying the effects of second messengers on release of Ca2+ by the sarcoplasmic reticulum. Recently, many investigators have turned to permeabilization techniques that leave the plasma membrane relatively intact. The maintenance of both intracellular and membrane-associated regulatory systems in these preparations allows the testing of a wider variety of physiological processes that potentially modulate the regulation of smooth muscle contraction. P-Escin, a saponin ester, is one such compound. It selectively permeabilizes the plasma membrane in the same way that saponin does but creates smaller pores and leaves receptor coupling intact (15). This system retains high molecular mass compounds but permits diffusion of such lower molecular mass compounds as calmodulin (17 kDa) and h eparin (5 kDa). Another compound, Staphylococcus aureus a-toxin, provides even more limited diffusion. This bacterial toxin inserts into the plasma membrane of susceptible cells and forms very stable nonselective pores of l- to 2-nm effective diameter. These pores permit compounds of a molecular mass of < 1 kDa to diffuse while retaining larger proteins (1). This preparation therefore allows controlled manipulation of the intracellular ionic composition and introduction of nucleotides into the cytosol while preventing the loss of intrinsic regulatory proteins. Several laboratories have used cx-toxin-permeabilized smooth muscle preparations to study the change in the Ca2+ sensitivity of force in response to agonist stimulation (4, 11, 15, 23) and the effect of nucleotides on cross-bridge kinetics under various conditions (24). In this study, we determined the metabolic properties of the cx-toxin-permeabilized smooth muscle. The results show that there is a substantial ecto-adenosinetriphosphatase (ATPase) activity associated with the smooth muscle plasma membrane that is not removed by toxin poration and is highly resistant to modification by inhibitors. Unless precautions are taken to reduce this high resting ATPase activity, both the maintenance of intracellular nucleotide concentrations and force output are compromised; this necessarily complicates the interthe American
Physiological
Society
Cl673
Cl674
METABOLISM
IN a-TOXIN-PERMEABILIZED
pretation of mechanical and energetic studies performed in the a-toxin-permeabilized smooth muscle. Therefore, although the cx-toxin-permeabilized muscle has the advantages of intact receptor coupling and retention of regulatory proteins, its metabolic state represents a serious disadvantage. For this reason, we have developed a protocol that reduces the ecto-ATPase activity of the muscle to the extent that an ATP-regenerating system is able to maintain muscle nucleotide contents. The procedure involves combined treatment with the ectonucleotidase inhibitors 4,4’-diisothiocyanatostilbene2,2’-disulfonic acid (DIDS) and sodium azide. This treatment improves the metabolic profile of the preparation, enhancing the contractile response and enabling experiments to be performed in the ar-toxin-permeabilized smooth muscle under conditions that require rigorous control of muscle nucleotide contents. METHODS
Solutions NormaZ Krebs solution. The normal Krebs solution contained the following (in mM): 118 NaCl, 4.7 KCl, 1.2 KH2PO4, 1.2 MgS04, 25 NaHCOi3, 1.9 CaC12, and 11 glucose. The solution was gassed with a mixture of 95% O&j% CO2 for a pH of 7.4. Krebs solution used for guinea pig portal vein was the same except that it contained 2.5 mM CaClz. Ca2+-free Krebs solution was similar except that no CaC12 was added, and it contained 2 mM ethylene glycol-bis( P-aminoethyl ether)N,N,N’,N’-tetraacetic acid (EGTA). For measurement of ectoATPase activity, 1 mM ATP and an additional 1 mM MgS04 were added to these Krebs solutions. The Krebs solution used for muscle dissection and storage contained 125 U/ml penicillin G and 0.3 mglml streptomycin sulfate. NormaZ relaxing and activating solutions. The normal relaxing solution for the permeabilized muscles contained 30 mM piperazine-N,N’-bis(2-ethanesulfonic acid), 2 mM EGTA, 1 mM MgATP, 20 mM phosphocreatine (PCr), 3 mM free Mg2+, 0.5 mM leupeptin, 1 mM dithiothreitol, and 1 mg/ml creatine kinase. 1,6-Diaminohexane-N,N,N’,N’-tetraacetic acid was added to bring the ionic strength to 200 mM, and the pH was adjusted to 6.8. In the activating solutions, the total EGTA was kept at 2 mM, and pCa was adjusted by the addition of Ca-EGTA. In some relaxing and activating solutions the PCr concentration was varied as noted in the text. Experiments were performed at 20°C. Other solutions. The solutions that lacked a regenerating system for ATP contained 0.2 mM Pl,P”-di(adenosine-5’) pentaphosphate, a myokinase inhibitor, and had no added PCr or creatine kinase. Rigor solutions contained neither ATP nor PCr. In the solution containing adenosine 5’-O-(y-thiotriphosphate) (ATPyS), the total EGTA was 20 mM, and the pCa was adjusted to 4.5 by the addition of Ca-EGTA. No ATP, PCr, or creatine kinase was added to this solution. All solutions were filtered before use (0.2-pm Acrodisc filter, Gelman Sciences). Materials. 12,8-sH]ATP (27 Ci/mmol) and D-[U-14C]mannito1 (32 mCi/mmol) were purchased from ICN Biomedicals. [y-““P]ATP (10 Ci/mmol) and [2,8-3H]ADP (33 Ci/mmol) were purchased from New England Nuclear. ATPyS was purchased from Boehringer Mannheim. With the exception of suramin (Miles), all of the inhibitors were purchased from Sigma Chemical.
SMOOTH
MUSCLE
MuscZe Preparation Portal veins were removed from female New Zealand White rabbits that had been anesthetized with pentobarbital sodium (Beuthanasia) and rapidly exsanguinated. The vein was cut open longitudinally, dissected free of adventitia, and divided longitudinally into muscle strips. These could be used the same day or stored overnight at 4°C for use the following day with no change in mechanical or enzymatic properties. The strips, averaging 6-8 mm in length, 2-4 mm in width, and 60-100 pm in thickness, were held as flat sheets at their in vivo length with one end attached to an isometric force transducer (Kistler Morse DSK-6) and the other end attached to a micrometer for adjustment of length. They were bathed in a Plexiglas chamber containing flowing oxygenated Krebs bicarbonate solution at 20°C. Force was recorded on a Brush 2200 recorder (Gould). A steady-state contraction was measured in response to a maximal dose (0.1 mM) of the aladrenergic agonist phenylephrine (PE), and the muscle was relaxed fully in Ca 2+-free Krebs solution. The muscle was then permeabilized by incubating it for 40 min in normal relaxing solution containing 2,500-5,000 U/ml S. aureus a-toxin (GIBCO). To test force in response to maximal thiophosphorylation of the myosin light chains at the end of an experimental protocol, the permeabilized muscles were incubated in rigor solution and then treated with 1 mM ATPyS (pCa 4.5) for 10 min. After several washes in rigor solution, the muscle was transferred to relaxing solution (1 mM ATP, 0 Ca2+ >, and force output was measured. In another set of experiments, muscle strips were permeabilized with 1% Triton X-100 alone in a rigor solution for 30 min or by a modification of the freeze-glycerination method described by Haeberle et al. (6). This modified method is described in detail in a previous publication (3) and involves brief treatment of the freeze-glycerinated strips with 1% Triton X-100 before use. When the experimental protocol required only chemical measurements, the muscle was mounted as a flat sheet of the same dimensions between stainless steel clamps attached to a plastic holder, which allowed it to be easily transferred from one solution to another. The solutions varied from 600 to 800 ~1 in volume, and they were continuously stirred in all cases. Measurement PermeabiZixed
of Steady-State Muscles
Resting
ATPase
Rates in
To measure resting ATPase, permeabilized muscles were transferred from rigor solution to a relaxing solution containing 1 mM ATP (5 l&/ml L3H]ATP). The solution had no added PCr or creatine kinase, and the myokinase inhibitor P1,P5di(adenosine-5’) pentaphosphate was present. Aliquots of the solution were taken at 0 and 10 min, added to cold 0.5 N HC104 which contained carriers for radioactively labeled compounds, neutralized with KOH, and buffered to pH 7.4 with 0.2 M triethanolamine. The nucleotides were then separated by liquid chromatography on Extract Clean disposable NH2 columns (Alltech Associates). Several columns were mounted in a vacuum manifold so that multiple samples could be processed simultaneously. After a wash of the column with 50 mM NH4H2P04 to remove any contaminants, the sample was loaded in this low-salt buffer. The nucleotides were eluted in nine 2-ml washes: adenosine/AMP with three washes of 50 mM NH4H2P04, ADP with three washes of 250 mM NH4H2P04-0.25 N HCl, and ATP with three washes of 2 N HCl. The eluant was mixed with 14 ml of scintillation cocktail (Ecolite, ICN Biomedicals) and subjected to liquid scintillation counting in a Beckman model LS 5000TA spectrometer. Representative samples were also subjected to high-perfor-
METABOLISM
IN a-TOXIN-PERMEABILIZED
mance liquid chromatography (HPLC) to compare the efficiency of the separation. The disposable column method, using the protocol described above, provides very good separation of the nucleotides. At the end of the experiment, the total volume of the muscle was determined by comparing the tritium present in the blotted muscle to the tritium in a known volume of the solution. ATPase was calculated as the increase in L3H]ADP, L3H]AMP and [3H]adenosine in the solution over time. This is equal to the decrease in [“H]ATP in the solution, because each nucleotide is taken as a fraction of the total tritium in the solution. The volume of the muscle ( - 1.6 ~1) is quite small relative to the volume of solution (700 pl), and therefore the total ATP in the solution (which starts at 1 mM) decreased by only 4.2 * 0.4% (n = 13) over the course of an ATPase measurement under conditions with the highest ATPase. Measurement of Steady-State in Intact Muscles
Ecto-ATPase
Rates
A similar experimental protocol was used to quantitate ecto-ATPase activity in intact muscles by measuring the appearance of [“H]ADP, [3H]AMP, and [3H]adenosine over time while the muscle was bathed in normal Krebs solution containing 1 mM ATP (5 @i/ml [3H]ATP). Ecto-adenosinediphosphatase (ADPase) activity was measured by incubating muscles in normal Krebs solution containing 1 mM ADP (5 kCi/ml L3H]ADP> and by following the appearance of [3H]AMP and [“Hladenosine over time. When nucleotide concentration was varied in these solutions over a millimolar range, MgS04 was added in equal concentration. No additional MgS04 was added when ATPase and ADPase activity was measured at micromolar concentrations of nucleotide. A general protocol was adopted to test ecto-ATPase inhibitors in this preparation. The ATPase was measured in the intact muscle, which was then treated with the inhibitor in normal Krebs solution. The inhibitor was then washed out completely in fresh Krebs solution, and the ATPase was measured again. To account for variations in ecto-ATPase activity among individual muscles, results were reported as a percentage decrease (or increase) of the ecto-ATPase for each muscle after treatment with the inhibitor. This protocol varied slightly according to the solubility and activation requirements of the particular compound. Treatment with the ATP analogues 2’- and 3’-0-(4-benzoylbenzoyl) adenosine 5’-triphosphate ATP (BzATP) and &azidoadenosine 5’-triphosphate (8-azido-ATP) involved suspending intact muscles, mounted as described above on plastic holders, in a quartz cuvette with 1 ml of Krebs solution containing one of the compounds. After a lo-min incubation of the muscle in this solution, the compound was photoactivated for 10 min by exposure to ultraviolet irradiation (350 nm for BzATP, 254 nm for 8-azido-ATP) in a model F-2000 fluorescence spectrophotometer (Hitachi Instruments). Inhibitors were initially surveyed in two muscles, with more than one inhibitor tested on muscle strips from the same animal. If an inhibitor showed > 10% inhibition of the ectoATPase, additional experiments were performed on muscles from other animals. When ecto-ATPase was measured in intact muscles that were not subsequently permeabilized, volume was estimated based on muscle area. The least-squares regression of muscle volume (measured in permeabilized muscles by total tritium content) vs. area was y = 0.093x + 0.202 (n = 25), where y is muscle volume (in ~1) andx is muscle area (in mm2>. The average muscle volume was 1.6 t 0.2 ~1.
Determination Concentration
SMOOTH
Cl675
MUSCLE
of Relationship Between Substrate and Ectonucleotidase Activity
ATPase and ADPase rates were measured in intact muscle as described above over a range of nucleotide concentrations from 5 PM to 3 mM. These measurements were paired in the same muscle, with ectonucleotidase first determined for a 1 mM concentration of nucleotide. After the nucleotide was washed out in Krebs solution, another concentration was tested. The rate of degradation at each nucleotide concentration in the same muscle was reported relative to the rate at 1 mM nucleotide. Determination
of Radiolabeled
Nucleotide
Content in Muscle
Permeabilized muscles that were previously in rigor solution were incubated for 10 min in either relaxing or activating solution containing 1 mM ATP (40 l&i [3H]ATP/ml). If PCr was present in the solution, creatine kinase was also added (1 mg/ml). Muscles were blotted to minimize the amount of adhering solution, frozen in liquid nitrogen, and pulverized in the presence of 0.5 N HC104 containing 1 mM unlabeled ATP, ADP, and AMP. After centrifugation, the supernatant was recovered, neutralized with KOH, and buffered to pH 7.4 with 0.2 M triethanolamine. The nucleotides were separated by HPLC on an Econosil NH2 column (Alltech Associates), using a mobile phase of 50-500 mM NH4H2P04, 10 mM disodium EDTA (Na2EDTA) at a flow rate of 2 ml/min. Fractions were collected, mixed with scintillant and subjected to liquid scintillation counting. The time at which each radiolabeled nucleotide eluted from the column was determined by monitoring the appearance of L3H]ATP and [“H]ADP standards and by control experiments, which showed apyrase-induced formation of 13H]AMP from 13H]ATP and subsequent formation of L3H]adenosine by the addition of 5’-nucleotidase. With this procedure it was not always possible to separate adenosine and AMP; therefore these are usually reported as a combined value, designated as adenosine/AMP. This method does not give the absolute concentration; rather each nucleotide is reported as a fraction of the sum of [3H]adenosine, [3H]AMP, [3H]ADP, and [3H]ATP in the muscle. To determine the actual concentration of each radiolabeled nucleotide in the muscle, a similar protocol was used, with the addition of 1 mM mannitol (8 FCi D-[14C]mannitol/ml) to the 3H labeling solutions as a volume marker. In this series of experiments, muscles were first exposed to a labeled activating solution for 5 min to ensure that all of the ADP on myosin was exchanged and then either transferred to a labeled relaxing solution or left in the labeled activating solution for an additional 3 min. After HPLC separation of mannitol, adenosine/AMP, ADP, and ATP, fractions were mixed with scintillant and counted in a spectrometer with appropriate channel settings for 3H and 14C. To determine the concentrations of radiolabeled ATP, ADP, and adenosine/AMP in the muscle, the ratio of disintegrations per minute (dpm) (nucleotide) to dpm (mannitol) was compared with the ratio of dpm (ATP) to dpm (mannitol) determined in similar procedures performed on the initial incubation medium in which the ATP concentration was fixed at 1 mM. Measurement
of Suprabasal
ATPase
Activity
of Muscle
Measurement of the suprabasal ATPase of the muscle upon activation in high Ca2+ was performed in the presence of PCr and creatine kinase. PCr is accessible to the interior of the muscle cells and in the presence of the normal intracellular creatine kinase activity would act to keep [ATP] high and [ADP] low in the muscle. Both relaxing and activating solutions contained 1 mM ATP, 2 mM PCr, and 1 mg/ml creatine
Cl676
METABOLISM
IN (x-TOXIN-PERMEABILIZED
kinase. [y-3zP]ATP was added to the solution and allowed to equilibrate until the PCr and the y-phosphate of ATP had the same specific activity of 32P. This equilibration takes - 2-4 h and can be monitored by sampling the solution periodically, separating Pi, PCr, and ATP by liquid chromatography, and subjecting samples to Cerenkov counting. ATPase activity was then measured by incubating the muscle in this solution and monitoring the appearance of [32P]Pi into the solution. Because the specific activity of 32P was the same in PCr and the y-phosphate of ATP, any [32PJPi formed reflected a change in PCr and/or ATP content. Under resting conditions, all of the [32P]Pi formed was from PCr, with no measurable change in ATP. Under activated conditions, 95% of the [32P]Pi formed was due to PCr breakdown, and the remaining 5% reflected a small decrease in ATP. This method therefore allows measurement of ATPase while [ATP] is kept high. Suprabasal ATPase was calculated as the ATPase under activated conditions minus that measured previously under resting conditions in the same muscle.
Statistics All data are reported
as means t SE.
RESULTS
Steady-State Resting ATPase in Intact and Permeabilized Portal Vein The resting ATPase of the ar-toxin-permeabilized portal vein was 1.96 t 0.15 mM/min (see Fig. l), with products accumulating as both ADP (47%) and adenosine/AMP (53%). This is quite high compared with the resting ATPase measured in portal vein permeabilized with Triton X-100 alone, as shown in Fig. 1. The fact that the detergent-treated muscles have a much lower ATPase suggests that the enzyme or enzymes responsible for the high resting ATPase in ol-toxin-permeabilized muscles are membrane associated. Treatment with the mitochondrial FOF1-ATPase inhibitors, oligomycin (1 PM, n = 2) and aurovertin B (1 mM, n = Z), or ouabain, a specific inhibitor of the membrane Na+-K+ pump (1 mM, n = 3), each resulted in a < 10% change in 3.0 1
I
2.5 2.0
T
1.5
1
1.0
0.5 0.0 2
Intact a-toxin
Toxin/ azide
DIDS/
toxin/ azide
Triton
Fig. 1. Ecto-ATPase activity of intact rabbit portal vein and ATPase after various permeabilizing procedures. ATPase is shown for intact muscles (n = 73) and for muscles under resting conditions when permeabilized with a-toxin with no addition of inhibitors (a-toxin, IZ = 13), with 10 mM sodium azide (toxin-azide, n = 6), and with DIDS-toxin-azide protocol described in text (n = 6). Also shown is ATPase after treatment with 1% Triton X-100 alone for 30 min (n = 4). Data are means + SE.
SMOOTH
MUSCLE
the ATPase under resting conditions. It is thus unlikely that these intracellular processes are responsible for the high ATPase activity. To test the possibility that the ATPase is due to an ATP-hydrolyzing enzyme associated with the external side of the plasma membrane (ecto-ATPase; for review, see Ref. 25), ATPase was measured in intact muscles bathed in Krebs solution containing 1 mM MgATP. A substantial rate of ATP degradation was observed (2.76 t 0.10 mM/min, see Fig. 1). Control experiments showed no net breakdown of ATP in the Krebs solution over time when no muscle was added (data not shown). These results confirm the presence of an ecto-ATPase and suggest that it could account for a substantial fraction of the high ATPase rate at rest in a-toxinpermeabilized muscles. Further characterization of the ectoenzyme revealed that it is not stimulated by Ca2+, since it functions at about the same rate in Ca2+-free Krebs solution (3.48 t 0.42 mM/min, n = 4). There is no decrease in the enzyme activity over time or with disruption of the endothelium by blotting the muscle with filter paper (data not shown). In addition to the ATPase activity, there is an ADPase activity that, in paired experiments at 1 mM substrate concentration, is 61 t 4% (n = 3) of the ATPase activity of the muscle. The possibility that the ectonucleotidase activity is related to the dense purinergic innervation of the rabbit portal vein was tested by measuring ecto-ATPase activity in the guinea pig portal vein, which reportedly shows less purinergic innervation (2). However, the ecto-ATPase activity is even higher (6.25 t 0.58 mM/min, n = 10) in the guinea pig portal vein. Relationship Between Nucleotide Concentration and Ectonucleotidase Activity Figure 2 shows ecto-ATPase and ecto-ADPase activities as a function of substrate concentration. Activities are shown relative to that at 1 mM nucleotide. Halfmaximal ATPase and ADPase activities both occur at > 0.5 mM nucleotide, and in both cases there is an increase in rate at 3 mM compared with 1 mM nucleotide. Inhibition
of Ecto-ATPase
A systematic study was initiated using various ATPase inhibitors in an attempt to characterize and reduce the high ectonucleotidase activity, and the results are summarized in Table 1. The various inhibitors were tested in muscles from both the same animal and from different animals. Although no specific inhibitor has been reported for any of the known ecto-ATPases associated with mammalian cells, there are some compounds that have been reasonably successful at decreasing enzyme activity. To take advantage of the enzyme’s position on the external side of the plasma membrane, intact muscles in some cases were treated with these inhibitors and then washed before permeabilization with a-toxin. As shown in Table 1, the ectonucleotidase is impervious to treatment with strong sulfhydryl alkylating
METABOLISM
0.0
0.5
1.0
1.5
2.0
IN a-TOXIN-PERMEABILIZED
2.5
3.0
IAW ww
B
1.0
Cl677
MUSCLE
very effective inhibitor of the ecto-ATPase, it adversely affects the mechanical response of the muscle. Moderate success was found with two inhibitors of ectonucleotidases: suramin (831) and sodium azide (14, 33). Suramin is a nonpenetrating trypanocidal agent that inhibits a variety of ecto-ATPases. However, it inhibits in a reversible fashion and must be present throughout the entire experiment. Although suramin is reasonably effective in this system in combination with azide (see Table l), the use of suramin is complicated by our preliminary finding that it inhibits creatine kinase activity (data not shown) and that it is a nonspecific inhibitor, possibly interfering with the action of certain G proteins and protein kinase C (18). Sodium azide also inhibits reversibly, and it must therefore be present throughout the experimental protocol. Although the inhibitory effect of azide on the ecto-ATPase is not very large (38%, Table l), its usefulness lies in the fact that it strongly inhibits the ADP hydrolyzing component of the enzyme(s). After azide treatment and a-toxin permeabilization, 88 t 4% of the degraded ATP accumulates as ADP, whereas only 12 t 4% is broken down to adenosine! AMP (n = 4). Although ATP is still hydrolyzed at a Table 1. Effect of various inhibitors activity of rabbit portal vein
0.5
0.0
SMOOTH
0.5
1.0
1.5
2.0
2.5
3.0
WPI e-w Fig. 2. ATPase and ADPase rates in intact rabbit portal vein as a function of nucleotide concentration ([nucleotide 1). ATPase (A) and ADPase (B) rates are reported relative to those at 1 mM nucleotide measured in same muscle. Lines show Michaelis-Menten equation in form V = V,,([nucleotidel/[nucleotidel + K,), with K, of 1.6 and 0.8 mM and V,, of 2.8 and 1.8 for ATP and ADP, respectively. Each symbol, except that at 1 mM nucleotide, represents 1 determination.
agents, such as IV-ethylmaleimide, diazotized sulfanilic acid, and p-chloromercuribenzoic acid. Covalent affinity labels of ATP binding sites offered greater specificity as ATPase inhibitors, but little success was found with the compounds Y-p-fluorosulfonylbenzoyl adenosine, 2’,3’dialdehyde ATP, and 8azido-ATP. BzATP was moderately successful at inhibiting the ecto-ATPase, but the photoactivation protocol makes it cumbersome to use on a routine basis. Compounds that irreversibly modify amino acids, such as the nonpolar cross-linking agent l,&difluoro-2,4-dinitrobenzene or 1-cyclohexyl-3-(2morpholinoethyl)carbodiimide (CMCD), were able to reduce the activity of the enzyme. Unfortunately, these compounds are probably permeant and too nonspecific to use in a preparation in which one wants to maintain the integrity of other membrane components. For example, after treatment with 100 mM CMCD for 30 min at pH 10, only - 10% of the original force in response to PE was generated in the intact muscle, and similar results were obtained for a pCa 4.5 contraction in the of permeabilized muscle (n = 2). Thiophosphorylation the light chains only returned force to - 30% of the maximal PE contraction. Thus, although CMCD is a
on the ecto-ATPase % Inhibition Ecto-ATPase
Ectonucleotidase
inhibitors
Sodium azide ( 10 mM) Suramin (200 PM) Suramin (200 PM)-azide (10 mM) Pl, P5-di(adenosine-5) pentaphosphate (0.25 mM) DIDS (5 mM) DIDS (5 mM)-azide (10 mM) Sulfydryl
3825 29’” 50+2
< lo* 38+12 5822
aZkyZating
agents
Diazotized sulfanilic acid (5 mM) N-ethylmaleimide (10 mM) p-Chloromercuribenzoic acid (100 PM) Protein
5’-p-Fluorosulfonylbenzoyl 2’,3’-Dialdehyde ATP 8-Azido-ATP (1 mM) BzATP (1.5 mM) BzATP (1.5 mM)-azide
< lo* -8tl2
< 10:‘:
cross-linkers
DFDNB (1 mM) DFDNB (1 mM)-azide (10 mM) 1-Ethyl-3-(3-dimethylaminopropyl) imide (100 mM) l-Cyclohexyl-3-(2-morpholinoethyl)carbodiimide (100 mM) ATP
of
39* 56"
carbodi< lo* 66+7
afinity
labels
adenosine (0.5 mM)
-429 < lo* < lo* 22+4 53+3
(10 mM) ATPase
(1 mM)
inhibitors
4cw-Phorbol 12,13-didecanoate (100 PM) Cibacron blue dextran (100 PM)
< lo* < 10*
Values are means t SE; n = 4-11 except where noted. In general, the latter are compounds that either showed < 10%) inhibition in 2 muscles or gave marginal inhibition and were then tested in combination with azide. Treatment times ranged from 15 to 60 min. DFDNB, 1-5-difluoro-2,4-dinitrobenzene; BzATP, 2’- and 3’-0-(4-benzoylbenzoyl) ATP. *n = 2.
Cl678
METABOLISM
IN CY-TOXIN-PERMEABILIZED
relatively fast rate, it accumulates primarily as ADP, which should be accessible to ATP-regenerating systems. One of the best inhibitors of the ecto-ATPase that we have tested is DIDS (see Table l), a compound that is most commonly used to inhibit anion transport but has also been reported to be an ectonucleotidase inhibitor (13). Treatment with 5 mM DIDS for 30 min in normal Krebs solution with 2% dimethyl sulfoxide before permeabilization, followed by a-toxin permeabilization with 10 mM sodium azide in all permeabilized muscle solutions, inhibits the ecto-ATPase by - 70% (see Fig. 1). Steady-State Muscle Nucleotide Contents Figure 3 shows ATP, ADP, and adenosine/AMP contents in a-toxin-permeabilized muscles under different conditions. Each metabolite is shown as a fraction of the total nucleotide content. In a relaxing solution without an ATP-regenerating system, the ATP in the muscle is depleted to roughly one-third of its initial concentration in the solution, with the remainder accumulating equally as ADP and adenosine/AMP in the muscle. Addition of an ATP-regenerating system (35 mM PCr and 1 mg/ml creatine kinase) results in a lower ADP content, but there is still a large accumulation of adenosine and AMP at the expense of the ATP content. After activation with pCa 4.5, ADP increases and ATP decreases further. It is obvious that the high resting ATPase of an ar-toxinpermeabilized smooth muscle severely compromises the maintenance of nucleotide concentrations in the tissue, even in the presence of an ATP-regenerating system. Treatment with sodium azide results in a higher ATP and lower adenosine/AMP content in the resting muscle
lo0 L
V.V
Rest Rest Act. Rest Act. Rest Act. -PCr +PCr + PCr + PCr Azide DIDShzide
Fig. 3. Adenine nucleotide distribution in or-toxin-permeabilized muscles under different conditions. Fraction of total tritium in adenosine/AMP (hatched bars), ADP (open bars), and ATP (solid bars) is shown for muscles incubated in 1 mM ATP containing C3H]ATP. Muscles were under resting conditions with no added Ca2+ (Rest) or activated with pCa 4.5 (Act). +PCr, PCr (35 mM) and creatine kinase (1 mg/ml) were present in incubating solution. Muscle contents are shown for muscles permeabilized with a-toxin with no added inhibitors, with 10 mM sodium azide, and with DIDS-azide protocol described in text. Data are means + SE; n = 7-10.
SMOOTH
MUSCLE
0.30 % c l H
0.25 0.20
3
0.15 3
0.10 m
0.05
w
L
A I
L
0.00 0.00
0.05
0.10
0.15
0.20
0.25
0.30
Fraction of Label in Adenosine Fig. 4. Relationship between fraction of label present in AMP vs. that present in adenosine under various conditions. Fraction of label present in AMP is plotted vs. that present in adenosine for cx-toxinpermeabilized muscles incubated in 1 mM ATP containing L3H]ATP, with PCr (35 mM) and creatine kinase (1 mg/ml). Muscles were either under resting conditions with no added Ca2+ (0), activated with pCa 4.5 (n), treated with 10 mM azide under resting conditions (W, or treated with 10 mM azide and activated with pCa 4.5 (A). Each symbol represents 1 muscle.
in the presence of PCr and creatine kinase (see Fig. 3). This is consistent with the fact that azide inhibits the overall ATPase activity as well as the ADPase component of the enzyme(s), causing the degraded ATP- to accumulate primarily as ADP, which would be accessible to the ATP-regenerating system. However, when the muscle is activated with high Ca2+, there is a significant decrease in the fraction of total adenine nucleotide appearing as ATP and an increase in ADP and adenosine/ AMP. More information concerning the effect of azide on nucleotide metabolism in the cx-toxin-permeabilized muscle can be obtained by considering those muscles in which it was possible to separate and quantitate both adenosine and AMP. The results for resting and activated muscles in the presence and absence of azide are shown in Fig. 4. In the absence of azide, there was an accumulation of both AMP and adenosine in the muscle. There is a significant (r = 0.80, P < 0.01) tendency for muscles that have higher adenosine to have lower AMP. This may result from the presence of variable amounts of 5’-nucleotidase, which if high would convert a larger fraction of the AMP that is formed to adenosine. In the presence of azide, there was a large decrease in both AMP and adenosine contents with an almost total suppression of AMP accumulation in the muscle. These data are consistent with the idea that azide inhibits an ecto-ADPase. A very sensitive indicator of the metabolic state of a tissue is the ratio of ADP to ATP. Ideally, this ratio should be quite low and should not change u.pon activation of the muscle if nucleotide contents are adequately controlled. Figure 5 shows the ADP-to-ATP ratios for the muscles whose nucleotide contents are shown for different conditions in Fig. 3. As expected, if there is no ATP-regenerating system present, the ADP-to-ATP ratio is very high even under resting conditions. Inclusion
METABOLISM
IN a-TOXIN-PERMEABILIZED
SMOOTH
Cl679
MUSCLE
and [adenosine/AMP] were - 0.1 and 0.015 mM, respectively, and did not change with activation. We conclude that treatment with DIDS and azide and the use of PCr and creatine kinase result in excellent control of nucleotide contents so that any ATP that is broken down is almost immediately resynthesized. Contractile Response of Intact and Permeabilized Smooth Muscle 01.
Rest Rest Act. Rest Act. Rest Act. - PCr +PCr +PCr + PCr Azide DIDS/azide Fig. 5. Ratio of ADP to ATP in a-toxin-permeabilized muscles under different conditions. Fraction of ADP compared with ATP is shown for muscles incubated in 1 mM ATP containing l”H]ATP. Data used to calculate this ratio were same as that shown in Fig. 3. Muscles were either under resting conditions with no added Ca2+ (Rest) or activated with pCa 4.5 (Act). +PCr, PCr (35 mM) and creatine kinase (1 mg/ml) were present in incubating solution. Ratio of ADP to ATP is shown for muscles permeabilized with a-toxin with no added inhibitors, with 10 mM sodium azide, and with DIDS-azide protocol described in text. Data are means * SE; n = 7-10.
of PCr and creatine kinase in the bathing medium results in a low ADP-to-ATP ratio at rest, but there is a significant increase when the muscle is activated in pCa 4.5. The presence of azide does not significantly change the ADP-to-ATP ratio at rest, and there is still an increase in the ratio upon activation of the muscle. There is, however, a significantly (P < 0.02) lower ADP-to-ATP ratio when the muscle is activated in the presence of azide than in its absence. These data show that inclusion of an ATP-regenerating system, even in the presence of azide, is not sufficient to prevent alterations in nucleotide contents upon activation of the cx-toxin-permeabilized muscle. We therefore measured the nucleotide contents in muscles that were subjected to the DIDS-azide protocol. This inhibitory treatment results in a resting ATPase that is less than one-half that of muscles treated with azide alone (see Fig. 1). Figure 3 shows that under these conditions the ATP content of the muscle is high while ADP and adenosi ‘AMP are quite low. Importantly, there i.s no change in the distribution of nucleotides when the muscle is activated. Also, Fig. 5 shows that there is no significant change in the ADP-to-ATP ratio when the muscle is activated after the DIDS-azide protocol. Because the combination of DIDS and azide appeared to give the best results with respect to lowering ATPase activity and preventing any change in the distribution of nucleotides upon activation, we performed further experiments to determine the actual concentrations of the various nucleotides under resting and activated conditions. This design included the use of labeled mannitol as a volume marker against which the nucleotides could be compared. The results are shown in Fig. 6. The concentration of ATP in both resting and activated muscles is very close to 1 mM, which is the concentration of ATP in the solution bathing the muscle. [ADP]
ne/
Figure 7A shows a typical force response of the intact portal vein to a supramaximal dose of PE, and the response to pCa 4.5 after a-toxin permeabilization. The force response after permeabilization was about half the maximal force induced by PE before permeabilization (see Fig. 8 for summary). Thiophosphorylation of the myosin light chains by incubation of the permeabilized muscle in ATPyS (pCa 4.5) led to a subsequent force response in a low-Ca2+ solution similar to that of the intact muscle (ratio = 0.97 t 0.07, n = 6). It is therefore not likely that the low force in response to high Ca2+ after a-toxin treatment resulted from partial permeabilization or damage to the force-generating capability of the contractile proteins. To investigate the possibility that the force output was limited by the metabolic state of the muscle, we determined the force-generating capacity of the portal vein in the presence of inhibitors of the ecto-ATPase. Figure 8 shows that treatment with azide or DIDS individually or in combination increased the isometric force output of the muscle compared with that observed with a-toxin treatment alone. Figure 7B shows the force response of a muscle subjected to the DIDSazide protocol. The force output of the intact muscle is not affected by DIDS treatment (ratio of PE-induced force before DIDS compared with after DIDS = 1.06 * 0.03, n = II), and the force response of the permeabilized muscle in pCa 4.5 was similar to that in the intact muscle (see also Fig. 8). In addition, the force output after thiophosphorylation of the myosin light chains was similar to the intact (ratio = 1.15 t 0.04, n = 4). A change in the Ca2+ sensitivity of force in a-toxinIpermeabilized smooth muscles has been previously obn g
m 5
b
0.016 L 0.012 -
% ” 4 .H
0.008 -
z5 3
0.004 f-
Ado/AMP
0.10
10 .
0.08
08.
0.06
06.
0.04
04.
0.02
. 02 00 . :
0.00 ADP
ATP
Fig. 6. Nucleotide concentrations of DIDS-azide-treated cx-toxinpermeabilized muscles incubated in 1 mM ATP and 35 mM PCr. Concentrations of adenosine/AMP (Ado/AMP), ADP, and ATP are expressed as mM (note change in scale for different nucleotides) for resting conditions with no added Ca2+ (open bars) and activated conditions in pCa 4.5 (hatched bars). Data are means t SE; n = 6 and 5, respectively.
Cl680
METABOLISM
IN a-TOXIN-PERMEABILIZED
5mNL 5 min
A
GTPyS -
pCa 4.5
pCa 6
B
SMOOTH
21) and a decrease in myosin light chain phosphatase activity (12). In this study, the force response to submaximal Ca2+ (pCa 6.0) with and without GTPyS was determined for muscles in the presence and absence of the ectonucleotidase inhibitors. Figure 7 shows typical force traces for two of these conditions, and the data are summarized in Fig. 8. GTPyS increased the Ca2+ sensitivity in all treatment protocols, but there are differences in the fraction of maximum force obtained at pCa 6.0 and the increase in force caused by incubation in GTP$. For example, in the permeabilized muscle with no inhibitors present, the force in pCa 6.0 is very small, and GTPyS increases force to somewhat less than half of that obtained in pCa 4.5. On the other hand, after the DIDS-azide protocol, force in pCa 6.0 is -40% of maximum and increases to 100% upon treatment with GTP+. Suprabasal
PE
PE
pCa 4.5
pCa 6
Fig. 7. Contractile response of rabbit portal vein before and after permeabilization with a-toxin. All muscles were initially contracted in 0.1 mM PE and relaxed in Ca 2+-free Krebs solution. A: muscle permeabilized with a-toxin (4,200 U/ml, 40 min) in relaxing solution (0 Ca) and contracted with pCa 4.5. After with no added Cazt relaxation in a 0 Ca solution, muscle was contracted with pCa 6, and 10 PM GTPyS was then added. B: muscle treated for 30 min with 5 mM DIDS in normal Krebs solution containing 2% dimethyl sulfoxide and then contracted a 2nd time with PE. Muscle was then permeabilized with a-toxin under same conditions described above; 10 mM sodium azide was present in all subsequent solutions. Contractile response of muscle to pCa 4.5 and 6.0 with subsequent addition of 10 PM GTPyS is also shown.
served by others (4, 11, 15, 23). It appears mediated by a G protein and can be induced addition of GTPyS, a nonhydrolyzable analogue The increase in Ca”+ sensitivity is associated increase in myosin light chain phosphorylation 12.
to be by the of GTP. with an (10, 16,
Activity
We have determined the ATPase activity of the muscle in the presence of PCr by measuring [32P]Pi formation from 32P labeled PCr and ATP. Figure 9 shows the ATPase under resting and activated conditions in muscles permeabilized by either freeze-glycerinationTriton treatment or by the DIDS-azide a-toxin procedure. The ATPase under both resting and activated conditions is somewhat higher in the a-toxin-treated muscle, whereas the suprabasal ATPase is very similar for both permeabilization procedures. The total ATPase under activated conditions is only - 30% higher in the a-toxin-permeabilized muscle compared with the freezeglycerinated Triton-treated muscle. These data show that the DIDS-azide a-toxin procedure described results in a preparation in which ATPase measurements can be made with accuracy and statistical variability comparable to the freeze-glycerinated Triton-treated muscle. DISCUSSION
This study was undertaken to determine the metabolic characteristics of the cx-toxin-permeabilized smooth
a 2.5 t-2 k 2.0
2
0.2 00.
ATPase
L
zi 1.0 z Lri w 08a 3 0.6 E ‘2 0.4 c
$
MUSCLE
Toxin
Azide
DIDS
DIDS/Azide
Fig. 8. Contractile response of