Regulation of polyphosphoinositide synthesis in cardiac membranes

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 271, No. 1, May 15, pp. 21-32,1989

Regulation of Polyphosphoinositide EUGENE Department

QUIST,’ NIMMAN

Synthesis in Cardiac Membranes

SATUMTIRA,

AND

PATRICIA

of Pha,rmacology, Texas College of Osteopathic Medicine, University .%%I0Ca,mp Bowie Boulevard.

POWELL of North Texas,

Fort Worth, Tera.s 76107

Received August 22,1988, and in revised form January

16,1989

The relative distribution of phosphatidylinositol (PI) and phosphatidylinositol-4phosphate (PIP) kinase activities in enriched cardiac sarcolemma (SL), sarcoplasmic reticulum (SR), and mitochondrial fractions was investigated. PI and PIP kinase activities were assayed by measuring 32P incorporation into PIP and phosphatidylinositol4,5bisphosphate (PIP,) from endogenous and exogenous PI in the presence of [-y-32P]ATP. PI and PIP kinase activities were present in SL, SR, and mitochondrial fractions prepared from atria and ventricles although the highest activities were found in SL. A similar membrane distribution was found for PI kinase activity measured in the presence of detergent and exogenous PI. PI and PIP kinase activities were detectable in the cytosol providing exogenous PI and PIP and Triton X-100 were present. Further studies focused on characterizing the properties and regulation of PI and PIP kinase activities in ventricular SL. Alamethacin, a membrane permeablizing antibiotic, increased “P incorporation into PIP and PIP2 4-fold. PI and PIP kinase activities were M$+ dependent and plateaued within 15-20 min at 25°C. Exogenous PIP and PIP, (0.1 mM) had no effect on PIP and PIPe labeling in SL in the absence of Triton X-100 but inhibited PI kinase activity in the presence of exogenous PI and Triton X-100. Apparent K,(‘s of ATP for PI and PIP kinase were 133 and 57 PM, respectively. Neomycin increased PIP kinase activity 2- to 3-fold with minor effects on PI kinase activity. Calmidazolium and trifluoperazine activated PI kinase activity 5- to 20-fold and completely inhibited PIP kinase activity. Quercetin inhibited PIP kinase 66% without affecting PI kinase activity. NaF and guanosine 5’-O-(3-thiotriphosphate) had no effect on PI and PIP kinase activities, indicating that these enzymes were not modulated by G proteins. The probability that PIP and PIP, synthesis in cardiac sarcolemma is regulated by product inhibition and phospholipase C was discussed. G 1989Academic Prrss. Inc.

Certain muscarinic cholinergic receptor agonists stimulate turnover of the polyphosphoinositides, phosphatidylinositol4phosphate (PIP)‘and phosphatidylinositol

4,5-bisphosphate (PIPJ in cardiac tissue (l-4). The muscarinic agonist carbachol stimulates a transient decrease in 32P-labeled PIP and PIP:! within lo-15 s in canine atria1 slices (4) and isolated atria1 myocytes (5). Following this decrease, there is an immediate increase in 32Pincorporation

i To whom correspondence should be addressed. a Abbreviations used: DAG, diacylglycerol; EGTA, ethylene glycol bis(p-aminoethyl ether) N,N’-tetraacetic acid; GDP&S, guanosine 5’-0-(2-thiodiphosphate); GTPrS, guanosine 5’-0-(3-thiotriphosphate); GppNHp, 5’-guanylimidodiphosphate; Hepes, 4-(2hydroxyethyll-1-piperazineethanesulfonic acid; IP, myo-inositol l-phosphate; 1,3,5-IP,, myo-inositol 1,3,5-trisphosphate; 1,3,4-IP8, myo-inositol 1,3,4-tris-

phosphate; 1,3,4,5-IP,, myo-inositol 1,3,4,5-tetrakisphosphate; PA, phosphatidic acid; PI, phosphatidylinositol; PIP, phosphatidylinositol 1,4-b&phosphate; PIP,, phosphatidylinositol 1,4,5-trisphosphate; SL, sarcolemma; SR, sarcoplasmic reticulum.

21

0003-986X39 $3.00 Copyright All rights

Cc:1989 by Academic Press, Inc. of reproduction in any form reserved.

22

QUIST,

SATUMTIRA,

into phosphatidic acid (within 30 s) as diacylglycerol (DAG)-produced from PIP and PIP2 breakdown is phosphorylated by DAG kinase (1,4). Concomitant with these changes, carbachol increases myo-inositol bisphosphate (IP,) and myo-inositol trisphosphate (IP,) release within 10 s in chick heart cells with a delayed but prolonged (i.e., 30 min) increase in myo-inositol monophosphate (IP) (6). Mechanistically, muscarinic receptor agonists stimulate phosphoinositide phospholipase C which hydrolyzes PIP and PIP2 to diacylglycerol and to 1,4-IP2 and 1,4,5-IP3, respectively. Evidence to support a role of guanyl nucleotide regulatory (G) proteins in muscarinic receptor activation of phospholipase C in heart was reported by Jones et al. (6) and is consistent with studies on G protein regulation of phospholipase C in a number of other cells (7,8). Because muscarinic receptor activation produces only a transient decrease in PIP and PIP2 levels (4), PIP and PIP2 must be rapidly resynthesized from membrane PI to replace these lipids. At present, very little is known about the membrane location or the regulation of the PI and PIP kinases which synthesize PIP and PIP2 in heart. Evidence for the presence of PI and PIP kinase activities in rabbit ventricular sarcolemma (9) and sarcoplasmic reticulum (10) has been reported but little is known about the relative specific activities or properties of these kinases in heart membranes. Conceivably, PIP and PIPZ synthesis could be regulated by muscarinic receptor activation or by second messengers produced by activation of other receptors. In other tissues (ll-13), regulation of PI and PIP kinase activities by receptor and G proteins has been suggested. The information available on the properties and regulation of PI and PIP kinases in heart is limited and therefore this investigation was undertaken to determine the distribution of PI and PIP kinases in cardiac membranes and in particular to characterize the properties of these kinase activities in sarcolemmal membranes. The possibility that sarcolemmal PI and PIP kinase activities may be regulated by prod-

AND

POWELL

uct inhibition, G proteins, exogenous polyphosphoinositides, and inositol phosphates was also studied. MATERIALS

AND

METHODS

Muterials. [y-“‘P]ATP (600 mCi/mmol) was ohtained from ICN, Inc. (Irvine, CA). Neomycin sulfate, alamethacin, saponin, guanosine 5’-0-(2-thiodiphosphate), guanosine-5’-O-(3-thiotriphosphate), 5’-guanylimidodiphosphate, phosphatidylinositol (soybean), phosphatidylinositol 4-monophosphate (bovine brain), and phosphatidylinositol4,5-bisphosphate (bovine brain) were from Sigma Chemical Co. (St. Louis, MO). Calmodulin (bovine brain), calmidazolium, myo-inositol 1,4-diphosphate (tetraammonium salt), muo-inositol 1,4,5-triphosphate (trilithium salt), myo-inositol 1,3,4-triphosphate (tripotassium salt), and nayo-inositol 1,3,4,5-tetraphosphate (tetrapotassium salt) were purchased from Calbiochem (La Jolla, CA). All other reagents were of analytical or HPLC grade. SarcoPreparation of cardiac subcellular fractions. lemma (SL), sarcoplasmic reticulum (SR), and mitochondria were isolated from cardiac tissue by modifications of the procedures of Kidwai et al. (14) and others (15,16). Mongrel dogs of either sex were anesthetized with intravenous Surital and the hearts were quickly removed into cold isotonic saline and 0.1 mM EGTA. All subsequent steps were carried out at 5°C. The ventricles or atria were trimmed of fat and major blood vessels and 4-g pieces were transferred to 50ml centrifuge tubes. Tissue was covered with 8 ml of the saline solution and minced into approximately lmm pieces with scissors. The minced tissue was washed with the saline solution and resuspended with 20 ml of hypotonic homogenizing buffer (HB) containing 10 mM Hepes, pH 7.4, 2.0 mM MgClz, 0.5 mM dithiothreitol, and 0.1 mM EGTA. The tubes were left on ice for 10 min and centrifuged at 150s for 5 min. The tissue was then suspended with 18 ml of HB and homogenized four times for 10 s at a half-maximal setting with a Polytron leaving at least 30 s between each homogenization. The homogenates were filtered through 43’7-pm nylon mesh and 18 ml of homogenate was layered on three different discontinuous sucrose gradients to isolate enriched SL, SR, and mitochondria. Gradients consisted of 5 ml (top layer) and 6 ml (bottom layer) of (a) 8 and 22.5% sucrose, (b) 27 and 35% sucrose, and (c) 35 and 50% sucrose for isolation of SL, SR, and mitochondria, respectively. The sucrose gradient solutions also contained 10 mM Hepes, pH 7.4, and 0.1 mM EGTA. Tubes were centrifuged 60 min at 70,OOOgin a SW-27 swinging bucket rotor. The cytosolic phase was collected and dialyzed overnight at 5°C against 10 mM Tris-HCl, pH 7.6,0.1 mM EGTA, and 0.2 mM dithiothreitol. The

POLYPHOSPHOINOSITIDE fractions located at the sucrose interfaces were collected and resuspended with 40 ml of HB and centrifuged for 30 min at 70,OOOg.The pellets were resuspended with HB using a motor-driven Potter-Elvehjem homogenizer. Membrane and mitochondrial fractions were used immediately or frozen in liquid nitrogen and stored at -80°C. Marker enzyme assays. The relative purity of isolated SL, SR, and mitochondria was assessed by marker enzyme assays: Na+,K+-ATPase for SL, Ca’+,K’-ATPase for SR, and NaNa-sensitive ATPase for mitochondria by slight modifications of the procedures of Jones and co-workers (15, 17, 18). Na’,K’ATPase activity was measured as P, released from ATP in 25 mM Hepes, pH 7.4, 100 mM NaCl, 20 mM KCI, 5.0 mM MgClz, 5.0 mM NaN,, 0.1 mM EGTA, 40 pg saponin, 2.0 mM ATP, and was that activity inhibited by 0.1 mM ouabain. Ca’+,K+-ATPase activity was assayed in 25 mM imidazole-HCI, pH 7.0, 3.0 mM MgCle, 100 mM KCI, 2.0 mM NaN,, 2.0 mM ATP, and Ca”‘,K’-ATPase activity was the difference in activity in the presence of 100 @M CaCl, or 100 PM EGTA. NaN,-sensitive ATPase activity was determined in the same medium used for Na+,K+-ATPase activity in the presence and absence of 5.0 mM NaNa. Final volumes for all ATPase assays were 1.0 ml and the tubes contained 7-15 pg protein. Incubations were 30 min at 37°C. Protein was assayed according to Peterson (19). Phosphoinositide kinase assay. Membrane PIP and PIP, were labeled with 32P by incubation in standard medium (0.25 ml final volume) containing 5.0 mM M&l,, 10 mM Hepes, pH 7.4, 0.1 mM EGTA, 0.2 mM NaN,, 25 mM NaCl, 25 mM KCl, 12.5 wg of alamethatin, 0.25 mM [y-32P]ATP (5-10 PCi), and 18 Ng of membrane protein. In some studies, exogenous PI, PIP, and PIP, were included with or without 0.25%, Triton X-100 and other variations as indicated in the text or legends. Tubes were usually incubated at 25°C and reactions were stopped by the addition of 2.0 ml of chloroform:methanol:HC1(20:40:1) and held for 20 min on ice (20). Chloroform (0.75 ml) and 0.50 ml of water were added to each tube. After vortexing and centrifugation at 500s at 5°C the aqueous-methanol phase and the protein interface were removed by aspiration and the chloroform phase was then washed one time with 3.0 ml of 0.1 N HCI. One milliliter of the chloroform phase was dried under nitrogen and resuspended with 0.03 ml of chloroform:methanol:HCI (60: 3O:l) and 0.02-ml aliquots were spotted on K6 silica gel thin-layer chromatography plates (Whatman, 0.25 mm). Plates were developed for 15 cm in chloroform:methanol:H,O:NH,OH (450:350:75:25) and “Plabeled PIP and PIPa were detected by autoradiography overnight with XAR-5 X-ray film (Kodak). PIP (R, = 0.32) and PIP:! (Rf = 0.16) were identified using authentic standards as previously described (20). Ra-

23

SYNTHESIS

diolabeled lipids corresponding to PIP and PIP2 were scraped from the plates and counted in 8.0 ml of Ecolite. The standard deviation of triplicate points was usually less than 7% of the mean. RESULTS

Distribution

ofPI and PIP Kinase

Activities Sarcolemma, sarcoplasmic reticulum, and mitochondria were isolated by a relatively rapid and simplified procedure which entails homogenizing the tissue in hypotonic buffer and directly layering the filtered homogenate on sucrose gradients. In preliminary investigations, a number of sucrose gradient combinations were screened in order to optimize yield and purity of the fractions as assessed by marker enzyme assays. Consistent with other studies, SL (1617, Zl), SR (15,16,22), and mitochondria (16, 22) migrated at mean bouyant densities corresponding to approximately 20-22, 30-35, and 45-50% sucrose, respectively. A fraction consisting of a mixture of SL and SR migrated between SL and SR at a density of approximately 27% sucrose and was not used in these studies. These results indicate that the hypotonic conditions used here to homogenize the tissue did not affect the density of the membrane fractions. However, homogenization in hypotonic medium was found to increase the relative purity of the SL. For instance, the specific activity of Na’,K’-ATPase in SL fractions was threefold greater in tissue homogenized in hypotonic medium than in otherwise identical medium containing 0.25 M sucrose or mannitol. The difference was not due to differences in the permeability of SL vesicles to substrates since Na+,K+-ATPase activity was measured in the presence of saponin or alamethacin (see below). It is more likely that under hypotonic conditions the vesicles are leaky and do not trap contaminants as readily. Directly layering the filtered homogenate on the sucrose gradient may also reduce aggregation. The canine ventricular SL membranes which were used in most studies consistently had Na+,K’-ATPase (SL marker) and Ca2+,K+-

24

QUIST, SATUMTIRA,

AND POWELL

TABLE I DISTRIBUTION OF PI AND PIP KINASE ACTIVITIES IN ATRIAL AND VENTRICULAR ENRICHED SL, SR, AND MITOCHONDRIAL (MIT.) FRACTIONS UTILIZING ENDOGENOUS PHOSPHOLIPID SUBSTRATES’

Tissue fraction Ventricle SL SR Mit. Atria SL SR Mit.

Na+,K+-ATPase

Ca’+,K+-ATPase (pm01 P,/mg protein/h)

NaN,-ATPase

PI kinase PIP kinase (pm01 32P/mg/min)

51.2 8.1 2.5

4.9 14.2 3.1

15.6 25.3 160.0

103.2 35.8 14.7

21.0 10.2 3.0

32.4 6.8 3.1

2.8 7.1 2.4

11.2 31.7 171.0

81.7 27.7 10.1

14.7 8.1 1.9

’ Experimental results are representative of one of three separate experiments.

ATPase (SR marker) activities of 48-60 and 4.9 pmol Pi/mg/h, respectively (Table I). Corresponding activities in SR were 8.1 and 14.9 pmol Pi/mg/h and in homogenates 1.8 and 3.1 ymol P/mg/h. The yields of both SL and SR were approximately 30 mg membrane protein per 100 g tissue wet weight. PI and PIP kinase activities were distributed in SL, SR, and mitochondria albeit with large differences in specific activities in these fractions. PI and PIP kinase activities were approximately 3- and a-fold greater in SL than in SR from both atria1 and ventricular tissue (Table I). The specific activities were approximately seven times higher in SL than in mitochondria whereas Na+,K+-ATPase activity was approximately 5- and 25-fold greater in SL than in SR and mitochondria, respectively. Therefore, the large differences in the ratios of Na+,K+-ATPase versus PI and PIP kinase activities between SL and SR and mitochondria indicates that the PI and PIP kinase activities in SR and mitochondria cannot be explained solely by SL contamination. To determine if the relative differences in PI kinase activities in these different membrane fractions was due to actual differences in enzyme activity or to limitations in endogenous membrane PI substrate, PI kinase activity was measured in the presence of exogenous PI and Triton

X-100 (Table II). Results show that the distribution in PI kinase activity is similar in different ventricular membrane fractions if either endogenous or exogenous PI is used as a substrate. Note, however, that in the presence of 0.25 mM PI and 0.25% Triton X-100, PI kinase activity is approximately 2.5-fold greater than in controls using endogenous PI (Tables I and II). Triton X-100 completely destroyed PIP kinase activity in SL even if 0.1 mM PIP was added with or without 0.25 mM PI and a comparison of PIP kinase with exogenous substrate could not be done. In contrast, both PI and PIP kinase activities could be detected in dialyzed cytoplasm providing that exogenous PI or PIP2 and Triton X100 were included (i.e., no activities were

TABLE II PI AND PIP KINASE ACTIVITIES IN ISOLATED MEMBRANE FRACTIONS FROM VENTRICULAR TISSUE MEAS~JRED IN THE PRESENCE OF EXOGENOUS PI AND TRITON X-100 Tissue fraction

Triton X-100 (% w/v)

SL SR Mit.

0.25 0.25 0.25

PI (mM)

0.25 0.25 0.25

PI kinase PIP kinase (pm01 32P/mg/min) 253 158 36

0 0 0

POLYPHOSPHOINOSITIDE TABLE

25

SYNTHESIS III

EFFECT OF PIP AND PIP2 ON PI AND PIP KINASE ACTIVITIES IN SARCOLEMMAL MEMBRANES AND CYTOSOL Tissue fraction

Triton X-100 (% w/v)

PI (mM)

PIP, (mM)

PIP (mM)

SL SL SL 0.25 0.25 0.25 0.25

Cytosol Cytosol

0.25 0.25

Cytosol Cytosol

0.25 0.25

0.25 0.25 0.25 0.25

0.1

19.8 17.2 16.6

0.1 0.1

19.4 246.0 159.2 167.3 -

0.0 0.0 0.0 0.0 -

0.25 0.25

0.1 0.1 0.1

detectable in the absence of Triton X-100) (Table III). Cytosolic activities were approximately lo- to E-fold lower in cytosol than in SL membranes. Detergent is probably required for the kinases to interact with the exogenous substrates. The difference in detergent effects on membrane versus cytoplasmic PIP kinase activity possibly is a reflection of structural differences of these enzymes. In contrast to Varsanyi et al. (9), y2P incorporation into phosphatidic acid (PA) was negligible in SL (and SR and mitochondria) possibly because DAG kinase activity was removed during preparation or endogenous DAG substrate was low in these membrane fractions.

Properties of Sarcolem~rnal Kimse Activities

0.1 0.1

PI and PIP

0.1

3.67 2.23

0.13 1.37

0.05 0.14

1.20 0.32

the membrane vesicles permeable to ATP and ions. Na+,K’-ATPase activity was also stimulated four- to fivefold by alamethacin whereas Ca2’,K’-ATPase activity (located on the cytoplasmic or outer surface of SR) was unaffected by this antibiotic (not shown). Because PI kinase activity is located on the cytoplasmic aspect of plasma membranes (2735), the large activation or unmasking effect of alamethacin on PI and PIP kinase activities indicates that the SL used here consist predominantly of sealed

m 600 -5 a R 400 {

In initial studies, incubation of SL with Mg-[y-“‘P]ATP resulted in low levels of 32P incorporation into PIP and especially PIP,. Labeling was increased over fourfold by inclusion of the antibiotic alamethacin in the incubation medium (Fig. 1). Jones et CLZ. (17) previously reported that alamethatin similarly “increased” or unmasked latent Na+,K+-ATPase and adenylate cyclase activities in cardiac SL by rendering

PIP kinase

(pm01 32P/mg/min) 107.3 99.2 111.9

0.1

SL SL SL SL Cytosol

PI kinase

l ye

/m’mye-

a 200’ ,d’ 0 0.0

0 5 119 alamethacln/,ug

FIG. 1. Influence of alamethacin tion into PIP (H) and PIP2 (0) in lemma1 membranes. Membranes min at 25°C with other conditions terials and Methods).

1.0 protein

1.5

on 32P incorporaventricular sarcowere incubated 15 standard (see Ma-

26

QUIST, 800 r

n--=

SATUMTIRA,

-m-------m

700 z 600 2 500 4 In 400

/ ,=

Time

(min.1

FIG. 2. Time course of 32Pincorporation into PIP (m) and PIP2 (0) in ventricular sarcolemmal membranes at 25°C with other conditions standard (see Materials and Methods).

right-side-out vesicles. Alamethacin promoted maximal activation of PI and PIP kinase activities at a ratio of 0.66:1 (wt:wt of membrane protein) and therefore alamethacin was included at this ratio in most assays in this study. Saponin (40 pg/ ml) also activated PI kinase and PIP kinase activities in SL approximately fourand twofold, respectively (not shown). The disportionate activation of these kinase activities by saponin suggested that saponin partially inhibits PIP kinase activity in addition to increasing the permeability of the vesicles. In addition, alamethacin may be more useful than saponin as a permeablizing agent because this antibiotic does not uncouple p adrenergic or muscarinic receptor G protein effects on adenylate cyclase (17,23). In SL membranes, labeling of PIP and PIP2 with 32P was linear over a concentration of 5 to 30 pg SL/O.25 ml after a 5-min incubation at 25°C (not shown). 32P incorporation into PIP and PIP, was linear for 5-10 min but plateaued after 15 to 25 min at 25°C (Fig. 2). At 30 and 37”C, PIP and PIP2 labeling was more rapid and plateaued within 10 and 5 min, respectively (not shown). However, the maximal level of labeling of PIP and PIP2 was similar at all temperatures. Plateauing at 25°C was not due to ATP depletion because it occurred at the same time in the presence of up to 2 mM ATP (not shown). PIP2 labeling was always much less (i.e., 20-33%) than

AND

POWELL

PIP labeling in isolated SL. Addition of aliquots of dialyzed cytosol with the SL did not increase PIP2 labeling, indicating that removal of a cytosolic cofactor of PIP kinase or PIP kinase itself was not removed during membrane preparation. The effects of exogenous PI, PIP, and PIP2 on membrane and cytosolic PI and PIP kinase activities were examined to determine if depletion of exogenous substrate or product inhibition was primarily responsible for the plateauing effect. Exogenous PIP (0.1 mM) or PIP, was not able to significantly affect either PI or PIP kinase activities in SL membranes in the absence of Triton X100 (Table III). These data could be interpreted to mean that PIP and PIP2 have no effect on PI and PIP kinase activities or more likely that they were unable to interact with the membrane-bound kinase activities. Assuming that the addition of a detergent might increase lipid-kinase interactions, the effects of PIP and PIP2 on SL and cytosolic kinase activities was studied in the presence of Triton X-100. In the presence of 0.25 mM PI and 0.25% Triton X-100,0.1 mM PIP or PIP2 inhibited PI kinase activity in SL membranes 32%. Similarly in cytosol, 0.1 mM PIP inhibited PI kinase activity approximately 39% in the presence of 0.25 mM PI and 0.25% Triton X-100 (Table III). Cytosolic PIP kinase was inhibited 73% by PIP2 in the presence of exogenous PIP. These observations provide evidence for product inhibition of PI and PIP kinase. Other properties of PI and PIP kinase activities in SL were studied in attempts to gain further insight into the mechanisms regulating PIP and PIP2 synthesis. The apparent K,‘s for PI and PIP kinases for ATP were 133 and 57 ELM, respectively (Figs. 3A and 3B). Determinations were made after a 5-min incubation at 25°C to minimize ATP loss by ATPase activities. PI and PIP kinase activities were MgC12 dependent and half-maximal and maximal activation of both PI and PIP kinase occurred with 1 and 5 mM MgC12, respectively (Fig. 4). It is of interest that the MgC12 concentration dependence required for optimal activity greatly exceeds the ATP re-

POLYPHOSPHOINOSITIDE

27

SYNTHESIS

A 00 i

4o /./ 20. /Y ,!+y-‘-: 0.0

o-0-0 0.1

0.2

0.3 [ATP]

04

0.5

mtd

FIG. 3A. ATP concentration dependence of “P incorporation into PIP (B) and PIPa (0) at 25°C. Ventricular sarcotemmal membranes were incubated 5 min with other conditions standard (see Materials and Methods).

quirement. This may indicate that M$ has a additional role in increasing PIP and PIP:! synthesis other than combining with ATP as a substrate. PIP and PIP, Hydrolysis To determine if 32P incorporation into PIP and PIP2 is a manifestation of synthesis or turnover, the rate of breakdown of “‘P-labeled PIP and PIP2 was assessed in SL. First, SL were prelabeled for 20 min with 200 PM [T-~~P]ATP and breakdown was measured after the addition of 2.0 InM cold ATP to dilute the hot ATP. With further incubation at 25”C, labeled PIP decreased very slowly at a rate of 15 pmol

6

100

_

a

20. 0-B 0

mM

FIG. 4. M&l, concentration dependence of azP incorporation into PIP (W) and PIP2 (0) at 25°C. Ventricular sarcolemmal membranes were incubated 5 min and other conditions were standard (see Materials and Methods).

“‘P/mg/min and PIP, labeling increased 15% after 10 min incubation. It is likely that the decrease in labeled PIP is due to both phosphoesterase degradation and conversion of PIP to PIP2 by PIP kinase activity. Second, if prelabeled membranes were washed with 10 mM Hepes, pH 7.4, and 0.1 mM EGTA by centrifugation, resuspended, and then incubated 10 min at 25”C, PIP decreased at a rate 10 pmol 32P/ mg/min and loss of labeled PIP2 was negligible. Breakdown was the same in the presence of 1,5, or 10 mM MgC&. Both methods were in relative agreement, indicating that the level of phosphoesterase activity is low in washed and nonwashed SL and therefore 32P incorporation into PIP and PIPe is a reliable estimate of synthesis by PI and PIP kinase activities. Regulation of PI and PIP Kinase Activities

a

E

[M9W

100

200

300 400 v/s x 103

500

600

FIG. 3B. ATP concentration dependence of “P incorporation into PIP (w) and PIP2 (0) plotted according to Eadie.

An important aim of this investigation was to determine if PI and PIP kinase activities in cardiac SL are regulated by G proteins or second messenger. Cyclic AMP or GMP (lo-100 PM) did not have any effect on PIP and PIP;: labeling. A number of guanyl nucleotides and NaF were tested to determine if PI and PIP kinases were regulated through G protein interactions. GTP, GppNHp, GTPyS (lo-100 PM), and 10 mM NaF had no effect on the rate or maximal

28

QUIST,

SATUMTIRA.

labeling of PIP and PIPe. These guanyl nucleotides and NaF were tested under a variety of conditions (i.e., at 25,30, and 37°C in the presence of 1 and 5 mM MgCl, with incubations from 1 to 25 min). GDPpS (100 pM) also had no effect on PI and PIP kinase activities making it unlikely that endogenous GTP was masking the effects of the other guanyl nucleotides. Therefore it is unlikely that PIP and PIP kinase activities are coupled to G proteins or muscarinic receptors in cardiac SL membranes. In this SL preparation, 10 mM NaF activates adenylate cyclase and PIP, phospholipase C activity lo- and 2.5-fold, respectively, providing evidence that Gs and Gp proteins are coupled to these activities in the presence of alamethacin (E. Quist and P. Powell, unpublished observations). Muscarinic receptor stimulated polyphosphoinositide turnover is known to generate various inositol phosphates (6) which could potentially regulate PI and PIP kinase activities. However, 100 pM concentrations of muo-inositol 1,4-bismyo-inositol 1,4,5-trisphosphosphate, phate, myo-inositol1,3,4-trisphosphate, or myo-inositol1,3,4,5-tetrakisphosphate had no effect on PIP and PIP2 labeling. Therefore inositol phosphates are unlikely to provide feedback inhibition or activation of PI and PIP kinase activities in cardiac SL membranes. Effects of Miscellaneous Agents ox PI and PIP Kinase Activities In other tissues, neomycin has marked and variable effects on PIP and PIP2 turnover (20,25,26). In cardiac SL membranes, neomycin increased PIP2 labeling 2- to 2.5fold with half-maximal activation occurring at 0.05 mM neomycin (Fig. 5). The effects of neomycin on PIP labeling were minimal and low concentrations of neomycin slightly inhibited (11%) and higher concentrations activated (23%) PI kinase activity. Neomycin is a hexavalent cation which binds avidly to highly negatively charged PIP and PIP2 (25, 26). Neomycin could therefore act by neutralizing the negative charge on the inositol phosphates

AND

POWELL

100

0' 0.0

0.2

0.4 INeomycln]

0.6

0.8

1.0

mM

FIG. 5. Effect of neomycinon 32P incorporation into PIP (m) and PIP2 (0). Ventricular sarcolemmal membranes were incubated 15 min at 25°C and other conditions were standard (see Materials and Methods).

of PIP and PIP2 leading to decreased binding of these lipids to membrane proteins. By this mechanism neomycin and possibly other cations could increase PIP kinase activity by making more PIP available for phosphorylation; or it is intriguing to speculate that reduced binding of PIPe to PIP kinase itself may relieve product inhibition. The flavenoid, quercetin was an effective inhibitor of PIP kinase (not shown) but did not have any effect of PI kinase activity over this same concentration range. Previously, Cachet and Chambaz (34) reported that quercetin inhibited PIP kinase in rat brain membranes. The selectivity of quercetin for PIP kinase activity may allow this agent to be useful in the study of differential effects of altering PIP and PIP:! levels on membrane-bound enzymes or ion channels in isolated membranes. Calmidazolium, a calmodulin antagonist, was initially used here to investigate a possible role of calmodulin in regulating PIP and PIP, synthesis. However, calmidazolium (50 pM) increased PIP labeling 20-fold and simultaneously inhibited PIP2 labeling 100% (Table IV). These effects were evident in the absence of Ca2+ or in SL prewashed with 10 mM Hepes, pH 7.4, and 0.1 mM EGTA to remove endogenous calmodulin. The phenothiazine derivative trifluoperazine had qualitatively similar effects on PI and PIP kinase activities. These agents increased PIP synthesis as effec-

POLYPHOSPHOINOSITIDE TABLE

Condition None 10 FM calmidazolium 25 PM calmidazolium 50 PM calmidazolium 100 PM trifluoperazine

IV

PIP (pmol/mg/ 15 min) 245 335 1425 5400 3039

PIP, (pmol/mg/ 15 min) 50 66 59 0 0

’ Incubation was at 25°C and other conditions were standard except alamethacin was not included. Results are representative of one of three separate experiments. Standard deviations are less than 7% of the means of triplicate points.

tively in the absence of alamethacin and therefore these agents can make the membranes permeable to ATP. From these observations, it is likely that the effects of calmidazolium or trifluoperazine are related to increasing membrane fluidity rather than inhibiting calmodulin. Recently KIockner and Isenberg (37) reported that eaImidazolium and trifluoperazine reduced membrane currents across cardiac SL by a mechanism suggested to involve partitioning of these agents into the membrane lipids. The effects of calmidazolium and trifluoperazine on PI and PIP kinase activities resemble the actions of the nonionic detergent Triton X-100 on these activities in kidney (33) and erythrocyte (38) membranes. The actions of these amphipathic molecules on PI kinase could be explained by displacing bound PIP from PI kinase as a result of nonspecific conformational changes (i.e., removing product inhibition?) and/or by fluidizing the membrane and thereby increasing PI mobilization to the kinase. DISCUSSION

The objectives of this study were to determine the relative distribution and specific activities of the PI and PIP kinase ac-

SYNTHESIS

29

tivities in SL, SR, and mitochondria, characterize the PI and PIP kinase activities in cardiac SL vesicles, and determine if these activities are regulated by various second messengers and G protein interactions. For these studies, a method was developed here to rapidly and efficiently isolate SL, SR, and mitochondrial enriched fractions from cardiac atria or ventricular tissue. The relatively short period required for the separation of membrane fractions and the paucity of steps in the procedure may diminish damage of the membranes by endogenous proteases, phosphatases, or mechanical disruption yielding membrane derivatives more closely approximating their state in intact cells. Distribution Activities

and Properties

of PI and PIP

The functional importance of PI and PIP kinases in cardiac SL membranes is clear. Muscarinic receptor agonists stimulate a rapid turnover of polyphosphoinositides, presumably in cardiac sarcolemmal membranes, which involves breakdown and resynthesis of PIP and PIP2 (4, 6). In other tissues, receptor-stimulated turnover of polyphosphoinositides has been shown to occur in the plasma membrane (8,27). Previous workers have provided evidence for the existence of PI and PIP kinase activities in rabbit ventricular SL (9) and SR (10) but the specific or relative activities in these membrane preparations was not reported. The results of the present study confirm and extend the above observations and show that the specific activities of PI and PIP kinases are higher in SL membranes than in SR or mitochondria isolated from both atria and ventricles (Table I). A similar distribution of PI kinase activity in these membrane fractions was found using either endogenous or exogenous PI. This observation is consistent with studies in other types of ceils in which PI kinase activity has been shown to be highest in the plasma membranes of liver (28, 29), brain (30, 31), pituitary (32), and kidney (33). The distribution of PIP kinase activity in cells is more complex and high

30

QUIST,

SATUMTIRA.

activities have been reported in plasma membranes of kidney (33) and brain (31, 34) and soluble forms have been reported in brain (30,34) and rat pituitary GHB cells (32). In cardiac tissue, significant amounts of PI and PIP kinase activity are present in SR and mitochondria which indicates a widespread distribution of PI and PIP kinases throughout the heart cell membranes in agreement with studies performed with membranes from liver (2729), brain (31), and kidney (33). The functions of PIP and PIP2 at organelle sites other than SL are currently unknown but these lipids likely have structural and functional importance at these sites. In cardiac tissue, cytosolic PI and PIP kinase activities were detected at levels approximately 15 times lower than in SL membranes (Tables I-III). A considerable difference in the sensitivity of SL and cytosolic PIP kinase to Triton X-100 was noted and therefore membrane-bound and cytosolic PIP kinase may be structurally different. PI and PIP kinase activities were then more thoroughly investigated in cardiac SL because muscarinic receptors are located on these membranes and SL would be the most likely site for potential muscarinic receptor G protein regulation of PIP and PIP2 synthesis. This does not exclude the possibility that receptor-mediated changes in intracellular ion concentrations or the production of second messengers could also regulate PIP and PIP, synthesis at other organelle sites. The results of this investigation show that there are marked similarities between the properties of PI and PIP kinase in erythrocyte (20,36) and plasma membranes from other cell types (29-33). These activities are Mg2+ dependent and maximal activation of PIP and PIP, synthesis occurs at 5-10 mM MgC& which are concentrations lo- to 20fold higher than for the optimal ATP concentration requirements. This suggests to affect that Mg2+ may act allosterically these activities in virtually all membranes studied. Mg”+ and other cations may increase activity by binding to the phosphates on PIP and PIP2 and thus reducing

AND

POWELL

their negative charge. In this way, PIP and PIPa binding to proteins, including PI and PIP kinase, would be reduced and this interaction may provide an important clue as to the mechanism of product inhibition of PI and PIP kinase activities in a number of membranes (see below). In cardiac SL (Fig. 2) and in other types of plasma membranes (24, 30, 33,41), PIP and PIPe synthesis reaches a maximal level or plateaus after short time periods (i.e., 5 to 15 min). The explanation for this phenomenon has not been adequately explained; however, it is unlikely that depletion of membrane PI is the sole determinant of plateauing because SL membranes possess 10 to 30 nmol of PI/mg protein and after maximal synthesis of PIP has been achieved only 0.8 nmol or 2.5% of the PI would be utilized (Fig. 2). Product inhibition of PI and PIP kinase activities appears to be a plausible explaination which has been reported to operate in erythrocyte membranes (20, 24,39). In support of this concept, direct inhibition of PI and PIP kinase activities by exogenous PIP and PIP, has been reported in liver (29), bovine retina (41), and lymphocyte membranes (42). In the present study, exogenous PIP and PIP,, had no effect on PI and PIP kinase activities in SL unless a detergent was added (Table III). This observation suggests that an amphipathic agent is required in this system to allow the PIP and PIP2 to interact with the kinase activities. Product inhibition of both PI and PIP kinase in cardiac cytosol was observed and therefore product inhibition can occur without the requirement of a membrane environment. Product inhibition of PI and PIP kinase activities could then result from direct binding of PIP or PIP, with PI and PIP kinases through electrostatic interaction as well as hydrophobic attractions. For instance, interaction of the negative phosphate groups on PIP and PIP2 alone appear to be insufficient since exogenous PIP and PIP, were unable to inhibit PI and PIP kinase in the absence of detergent. Presumably, in the presence of detergent, PIP and PIP2 could interact with PI and PIP kinase through hydrophobic in-

POLYPHOSPHOINOSITIDE

teractions of the fatty acids and hydrophobic regions of the kinases as well as through electrostatic interactions. However, binding of PIP or PIP2 to kinases might be anticipated to be weaker in the presence of detergents because Triton X100 may compete for lipid fatty acid binding to proteins and the conformation of the solubilized protein might also be altered by the detergent. This could explain why product inhibition in the artificial detergent environment appears to be less complete than in the membrane (Table III) and (41, 42). There is direct and indirect evidence to suggest that product inhibition by PIP and PIP2 is highly dependent on a strong negative interaction of the phosphates with a positive charge on the kinases. First, PI itself does not produce inhibition of PIP or PIP kinase activities and second cations such as Mg”+ (1-5 mM) and neomycin (0.05 to 0.1 mM) increase PIP or PIP2 synthesis presumably by binding to inositol phosphates and reducing the negative charge. In erythrocyte membranes NaCl or KC1 (100-200 mM) have effects similar to lower concentrations of MgCI, (5-10 mM) on PIP and PIP:! synthesis and therefore it is probable that cations reduce product inhibition by reducing interaction of PIP and PIP2 with PI and PIP kinases. The idea of product inhibition of PI and PIP kinase is not universally accepted and Imai et al. (32) reported that mechanisms other than product inhibition may control PIP and PIP, synthesis in GH3 plasma membranes. Second Messenger an.d G Protein Regulation Because PIP and PIP2 synthesis in cardiac SL is rapid, it is apparent that PI and PIP kinases could be important regulatory enzymes modulating PIP and PIP2 turnover. These activities determine the concentrations of PIP and PIP, in the membrane and could indirectly modulate the rate of PIP and PIP, breakdown by limiting the supply of these substrates to phospholipase C. In the present study, the potential effects of a number of second mes-

31

SYNTHESIS

sengers and G protein activators on PI and PIP kinase in cardiac SL were examined. The results, however, were negative since cyclic AMP, cyclic GMP, or high concentrations of inositol phosphates (i.e., 1,4IPe, 1,3,4-IP3, 1,4,5-IP3, 1,3,4,5-IPJ had no effect on PI and PIP kinase activities. The possibility that PI or PIP kinase activities may be controlled by guanyl nucleotide regulatory protein was exhaustively examined here in cardiac SL but the results were also negative. In contrast to our study using cardiac SL, GTPyS (presumably acting on G proteins) has been reported to increase PIP kinase in human placental membranes (13) and epidermal growth factor (11) and glucagon (12) have been shown to increase PI kinase activity in A431 cells and hepatocytes, respectively. The lack of effect of G protein activators on PIP and PIP2 synthesis in cardiac SL suggests that muscarinic receptor-mediated increases in polyphosphoinositide synthesis occur secondarily to phosphoinositide phospholipase C-mediated breakdown. For instance, a decrease in membrane PIP and PIP2 from phospholipase C activation could activate PI and PIP kinase activities by releasing product inhibition. Collectively, studies on the regulation of PIP and PIP2 synthesis suggest that a diverse number of regulatory systems exist in various tissues or cells and the properties of PI and PIP kinases will have to be individually determined in each cell type to understand regulation of polyphosphoinositide synthesis. ACKNOWLEDGMENTS We thank Tung Tran for his excellent technical support. This research was supported by grants from the American Heart Association, Texas Affiliate, the American Osteopathic Association, and the NIH-HL. REFERENCES 1. QUIST, E. E. (1982) B&hem. Pharnmcol. 31,31303133. 2. BROWN, S. L., AND BROWN, J. H. (1983) Mol. Ph,armacol. 24,351-356. 3. BROWN, J. H., AND BROWN, S. L. (1984) Fed. PTOC. 43,2613-2617.

32

QUIST,

SATUMTIRA,

4. QUIST, E. E., AND SATUMTIRA, N. (1987) Biochem. Pharmacol. 36,499-505. 5. MATTERN, J., SATUMTIRA, N., AND QUIST, E. E. (1985) Phwmacologist 27,170. 6. JONES, L. G., GOLDSTEIN, E., AND BROWN, J. H. (1987) Circ. Res. 62,299-305. 7. COCKROFT, S., AND GOMPERTS, B. D. (1985) Nature (&w&n) 314,534-536. 8. BERRIDGE, M. J. (1987) Annu. Rev. B&hem. 56, 159-163. 9. VARSANYI, M., MESSER, M., BRANDT, N. R., AND HEILMEYER, L. M. G., JR. (1986) B&hem. Biophys. Res. Commun. 138,1395-1404. 10. ENYEDI, A., FARAGO, A., SARKADI, B., AND GARDOS, G. (1984) FEBS L&t. 176,235-238. 11. PIKE, L. G., AND EAKES, A. T. (1987) J. Biol. Chem. 262,1644-1651. 12. WHIPPS, D. E., ARMSTON, A. E., PRYOR, H. J., AND HALESTRAP, A. P. (1987) Biochem. J. 241,835845. 13. URUMOW, T., AND WIELAND, 0. H. (1986) FEBS Lett. 207,253-257. 14. KIDWAI, A. M., RADCLIFFE, M. A., DUCHON, G., AND DANIEL, E. E. (1971) Biochem. Biophys. Res. Commun. 45,901-910. 15. JONES, L. R., BESCH, H. R., JR., FLEMMING, J. W., MC~ONNAUGHEY, M. M., AND WATANABE, A. M. (1979) J. Biol. Chem. 254,530-539. 16. MORCOS, N. C., AND DRUMMOND, G. I. (1980) Biu chim. Biophys. Acta 598,27-39. 17. JONES, L. R., MADDOCK, S. W., AND BESCH, H. R., JR. (1980) J. Biol. Chem. 255,9971-9980. 18. WILLIAMS, L. T., AND JONES, L. R. (1983) J. Biol. Chem. 258,5344-5347. 19. PETERSON, G. L. (1977) Anal. Biochem. 83, 346356. 20. QUIST, E. E. (1982) Arch. B&hem. Biophys. 219, 58-64. 21. KUWAYAMA, H., AND KANAZAWA, T. (1982) J. Biothem. 91,1419-1426. 22. ST. LOUIS, P. J., AND SULAKHE, P. V. (19’76) Canad. J B&hem. 54,946-956.

AND

POWELL

23. FLEMMING, J. W., STRAWBRIDGE, R. A., AND WATANABE, A. M. (1987) J. Mol. Cell. Carcliol. 19, 47-61. 24. QUIST, E. E., AND BARKER, R. C. (1983) Arch Biothem. Biophys. 222,170-178. 25. ORSULAKOVA, A., STOCKHORST, E., AND SCHACT, J. (1976) J. Neurochem. 26,285-290. 26. LODHI, S., WEINER, N. D., AND SCHACHT, J. (1976) Biochim. Biophys. Actu 426,781-785. 27. SEYFRED, M. A., AND WELLS, W. W. (1984) J. Biol. Chem. 259,7666-7672. 28. SEYFRED, M. A., AND WELLS, W. W. (1984) J. Biol. Chem. 259,7659-7665. 29. MICHELL, R. H., HARWOOD, J. L., COLEMAN, R., AND HAWTHORNE, J. N. (1967) Biochim. Biophys. Acta 144,649-658. 30. KAI, M., AND HAWTHORNE, J. N. (1969) Ann. N. Y. Acad. Sci. 165,761-773. 31. STUBBS, E. B., JR., KELLEHER, J. A., AND SUN, G. Y. (1988) Biochim. Biophys. Acta 958, 247254. 32. IMAI, A., REBECCHI, M. J., AND GERSHENGORN, M. C. (1986) Biochem. J. 240,341-348. 33. HUGGINS, C. G., HURST, M. W., Tou, I. S., AND LEE, T. C. (1969) Ann. N. Y. Acad. Sci. 165,790-800. 34. COCHET, E., AND CHAMBAZ, E. M. (1986) Biochem. J. 237,25-31. 35. HARWOOD, J. L., AND HAWTHORNE, 3. N. (1969) Biochim. Biophys. Acta 171,75-88. 36. HOKIN, L. E., AND HOKIN, M. R. (1964) Biochim. Biophys. Acta 84,863-875. 37. KLOCKNER, U., AND ISENBERG, G. (1987) Amer. J. Physiol. 253, H1601-H1611. 38. BUCKLEY, J. T. (1977) Biochim. Biophys. A&z 498, 1-9. 39. BUCKLEY, J. T., AND HAWTHORNE, J. N. (1972) J. Biol. Chem. 247,7218-7223. 40. QUIST, E. E., AND POWELL, P. (1985) Lipids 20, 433-438. 41. VAN ROOIJEN, L. A. A., ROSSOWSKA, M., AND BAZAN, N. G. (1985) Biochem. Biophys. Res. Commun. 126,150-155. 42. O’SHEA, J. H., HARFORD, J. B., AND KLAUSNER, R. R. (1986) J Zmmunol. 137,971-976.