- Email: [email protected]
Ca2+-stimulated phospholipid phosphoesterase activities in rabbit erythrocyte membranes
ARCHIVESOF BIOCHEMISTRYAND BIOPHYSICS Vol. 236, No. 1, January, pp. 140-149, 1985
Ca2+-Stimulated Phospholipid Phosphoesterase in Rabbit Erythrocyte Membranes EUGENE Department
of Pharnwxology,
Texas CoUege
Activities
QUIST
of Osteopathic Medicine, Fort Worth, Texas 76107
Received May 11, 1984, and in revised form August 21, 1984
The properties of the enzymes involved in Ca 2+-stimulated breakdown of phosphatidylinositol 4’-phosphate (PIP), phosphatidylinositol 4’,5’-bisphosphate (PIP2), and phosphatidic acid (PA) in rabbit erythrocyte ghosts were studied. At 25”C, 1 to 180 /LM Ca2+ rapidly stimulated the breakdown of PIP and PIPe, and maximal breakdown occurred within 10 minutes at all Ca2+ concentrations. The rate and the total amount of breakdown of PA, PIP, and PIP2 increased with Ca2+ concentration. MgCl, inhibited the rate of Ca2+-stimulated breakdown of PIP and PIP2 at Ca*’ concentrations less than 10 PM, but did not have any appreciable effects at higher Ca2+ concentrations. MgCl, also protected against Ca 2+-stimulated breakdown of PA. In the presence and absence of 5 mM MgC12, Ca2+ stimulated half-maximal breakdown of PIP and PIP2 at 2-3 ~.LMunder hypotonic and isotonic conditions. In the presence of 5 mM MgC12, Ca’+stimulated breakdown of PIP and PIP2 was associated with the release of Pi and inositol bisphosphate. In the absence of MgC12, Ca2+ stimulated the release of 32Plabeled Pi, inositol bisphosphate, and inositol trisphosphate from labeled PIP, PIPz, and PA. Ca2+ increased phosphatidylinositol content and decreased PIP and PIP2 content in these membranes. The results of this investigation suggest that Ca2+ stimulates the breakdown of polyphosphoinositides by stimulating polyphosphoinositide phosphomonoesterase and phosphodiesterase activities in rabbit erythrocyte ghosts. These activities were activated by less than 3 PM Ca2+ in the presence of MgCl, under hypotonic or isotonic conditions. These Ca2+-stimulated polyphosphoinositide phosphoesterase activities could therefore be active under physiological conditions in normal rabbit erythrocytes. o 19% Academic PRSS, IW.
In mammalian erythrocytes, phosphatidylinositol 4’-phosphate (PIP),l phosphatidylinositol 4’,5’-bisphosphate (PIP,), and phosphatidic acid (PA) are synthesized from membrane phosphatidylinositol and diacylglycerol, respectively, by Mgz+dependent kinase activities (l-4). Phosphoesterase activities are also present which hydrolyze these phospholipids, and together these enzymes account for the ’ Abbreviations used: EGTA, ethylene glycol bis(baminoethyl ether)-N,W-tetraacetic acid, PIP, phosphatidylinositol 4’-phosphate; PIP2, phosphatidylinositol 4’,5’-bisphosphate; IP2. inositol bisphosphate; IP,, inositol trisphosphate; PA, phosphatidic acid. 0003-9861/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.
rapid turnover of PA, PIP, and PIP2 in erythrocyte membranes (5-7). Few studies have been undertaken to study the phospholipid phosphoesterase activities; however, Ca2+ has been shown to stimulate the breakdown of PIP and PIP2 by a Mgz+independent phosphodiesterase activity (5, 6). This activity hydrolyzes PIP to diacylglycerol and inositol bisphosphate (IPz), and PIP2 to diacylglycerol and inositol trisphosphate (IPB) (5, 6). Although a number of workers have suggested that PIP and PIP, may also be broken down by a Ca2+-stimulated phosphomonoesterase activity (7-g), direct evidence for the existence of this enzyme activity in ery140
RABBIT
ERYTHROCYTE
MEMBRANE
thocyte membranes is lacking. Presently there are conflicts in the literature regarding Ca2+ affinity of polyphosphosphoinositide phosphosphodiesterase activity (5, 6, 10, 20). In rabbit (5, 13) and human (5) erythrocyte membranes, Ca2+ has been shown to stimulate polyphosphoinositide breakdown at l-10 PM Ca” under hypotonic conditions. Ca2+ concentrations as low as 10 PM in the presence of A23187 have also been reported to stimulate the breakdown of 32P-labeled PIP and PIP2 in intact erythrocytes (10, 11). In contrast, Downes and Michell (20) have reported that, under isotonic conditions in the presence of physiological MgC&, greater than 100 PM Ca2+ is required to stimulate polyphosphoinositide breakdown. Therefore, according to this report (20), activation of Ca2+ polyphosphoinositide phosphodiesterase would never occur in healthy human erythrocytes because intracellular Cast concentrations would never be expected to increase above 10 PM. However, if this enzyme were activated at ~10 PM Ca’+, it could possibly regulate erythrocyte function under various conditions in the circulation. In this study, the properties of Ca2+-dependent PA, PIP, and PIP phosphoesterase activities in freshly prepared rabbit erythrocyte ghosts were examined under a number of conditions. These studies provide evidence that rabbit erythrocyte membranes possess Ca2+ polyphosphoinositide phosphomonoesterase activity in addition to polyphosphoinositide phosphodiesterase activity. This study also shows that these activities are half-maximally stimulated by 2-3 I.LM Ca2+ under isotonic conditions in the presence of MgC12. MATERIALS
AND
METHODS
Materials. [y-JLp]ATP (25 Ci/mmol) and qi (carrier free) was purchased from ICN (Irvine, Calif.). Reference phospholipids, L-a-phosphatidylinositol (98% from soybean), L-a-phosphatidylinositol 4’,5’-diphosphate (98% from bovine brain), L-a-phosphatidylinositol 4’-monophosphate (98% from bovine brain), and Dowex-1 (1 X 8-400) were purchased from Sigma Chemical Company. Silica gel 60 thin-layer plates (0.25 mm) were purchased from E. Merck. CHC& and CH,OH were HPLC grade. Methylamine (40%) was purchased from MCB. All other chemicals were reagent grade.
PHOSPHOLIPID
PHOSPHOESTERASES
141
Preparation of erythrocyte membranes. Fresh heparinized blood was drawn from New Zealand white rabbits. Erythrocytes were washed free of plasma, and white cells in isotonic saline (4) and ghosts were prepared by washing packed erythrocytes three to four times with 10 vol of 20 mM Tris-HCI, pH 7.6, at 5°C (4). Ghosts were used within 1 h of preparation. Phosphmylatim studies To phosphorylate PA, PIP, and PIP2 with 32P, 3 ml ghosts was incubated for 20 min at 25”C, in a final volume of 7.5 ml, in 25 mM imidazole-HCI, pH 7.0,5 mM MgClz, 1 mM [r-J2PjATP, and 1 mM EGTA. After incubation, the ghosts were washed twice with 25 ml 10 mM Tris-HCl, pH 7.6, at 5°C and 20,OOOg.The washed ghosts were resuspended to 4.5 ml in the washing medium, and 0.25ml aliquots were incubated in a final volume of 0.5 ml with a final composition of 25 mM imidazole, pH 7.0, 1 mM EGTA, and usually 1 mM ATP. MgClz and CaCI, were varied in this medium, and the free Ca ion concentration was calculated using a CaEGTA stability constant of lO”-~ (14, 15). In some experiments 80 mM NaCl and 30 mM KC1 were also included to increase the ionic strength to 150 mM. The tubes were incubated at 25”C, and the reaction was stopped with 1.0 ml 10% trichloroacetic acid. The tubes were centrifuged at 2500 rpm at 5”C, and 1.0 ml of the supernatant was collected for further determination of radiolabeled water-soluble products (see below). The pellets were further washed with 3.0 ml HzO, and 2.0 ml CHCla:CHaOH:HCI (20:40:1) was added to each tube, and the tubes were left 20 min at 5°C. After this time, 0.75 ml CHCIB and Hz0 were consecutively added to the tubes, and the tubes were centrifuged at 25009 for 10 min. The upper phase and protein interface were removed, and a lml aliquot of the CHCls phase was dried under Nz and resuspended to 40 111with CHCI,:CH,OH:HCl (6:3:0.1). Twenty-microliter aliquots were spotted on silica gel 60 plates, and lo-p1 aliquots were analyzed for total lipid phosphorous (16). Thin-layer plates were developed in CHC1a:CH30H:Hz0:30% NH, (25:35:7.2:2.5), which completely separates radiolabeled PA, PIP, and PIP2 as previously described (4). Labeled phospholipids were detected by autoradiography, scraped, and counted in 8.0 ml Tritisol (4). Measurement of radiolabeled, water-soluble pm&As. The =P-labeled Pi, IPz, an IPa released from phospholipids in the presence of various concentrations of CaClz and MgClz were determined by anionexchange chromatography (6, 17). The supernatant obtained after stopping the reaction with trichloracetic acid was neutralized with NaOH, and 12 ml Hz0 was then added to each tube. Ten-milliliter aliquots were applied to l-ml Dowex-1 (formate form) columns, and radiolabeled Pi, IPz. and IPB were eluted with 10 ml 0.1 M formic acid:0.2M ammonium formate, 0.1 M formic acid:0.4 M ammonium formate, and 0.1 M formic acid:1 M ammonium for-
142
EUGENE
mate, respectively. One-milliliter aliquots of each fraction were counted in 10 ml Tritisol. Labeled PIP and PIP* were prepared and separated by phosphorylating ghosts and by thin-layer chromatography methods described above. IP, and IPI standards of unknown positional isomerism were prepared from 3aP-labeled PIP and PIP2 by hydrolysis for 7 min in 5 N HCl at 100°C (25). After removal of the organic products and unhydrolyzed polyphosphoinositides by extraction with chloroform (25), the aqueous phases, containing primarily IPz or IP, and other minor water-soluble products, were lyophilized. The identities of labeled IPz and IP3 obtained by anionexchange chromotagraphy or by hydrolysis of PIP and PIP2 were also confirmed by determining their relative mobilities on Whatman No. 1 chromatography paper developed in n-propranol:concentrated ammonia:water (5:4:1) as previously described by others (5, 26). “*Pi (carrier free) was also used as a standard. Lktermination
of phphoinositide
content
in ghosts.
Ghosts were incubated in the identical medium used above to label the phospholipids with 32P in a final volume of 5 ml ghosts (l-2 ml) were added to each tube. The reaction was stopped with 7.5 ml 5% trichloroacetic acid, and the tubes were centrifuged at 25009 for 10 min. The pellet was washed with 7.5 ml Hz0 at 30009 for 15 min. Lipids were extracted from the pellets with 7.5 ml CHC13:CHIOH:HCl (20:40:1) containing 0.01% butylated hydroxytoluene for 20 min at 5°C. CHC& (2.5 ml) and 1.5 ml HaO) were consecutively added, and the tubes were centrifuged for 10 min at 2500g. The HzO:CH,OH phase was discarded, and the CHCla phase was washed three times with 3 ml 0.1 N HCl. Three milliliters of the chloroform phase was dried under nitrogen, and the lipids were resuspended with 100 pl CHCl,: CH30H:HCl (63:O.l). For total phospholipid determination, lo-p1 aliquots were analyzed for phosphorus (16). To separate PI, 80-~1 aliquots were spotted over 1.5 cm on silica gel 60 thin-layer plates. These plates were developed in CHC13:CH,0H:CHaNH2 (65:35:10), which routinely separates phosphatidylinositol (Rf 0.25) from phosphatylserine and phosphatidic acid (H0.35) as previously described (18). Polyphosphoinositides were separated on silica gel 60 plates using the solvent system described above for separating radiolabeled PA, PIP, and PIP (4). Authentic phospholipid standards were run on both plates. The phospholipids were detected in an iodine chamber and scraped from the plates. Phosphorous content was determined (16). Blanks of similar area were also scraped from the plates and their absorbance was subtracted. RESULTS
In this investigation, PIP, PIPB, and PA were prelabeled with 32Pi by incubating freshly prepared rabbit ghosts for 20 min
QUIST
at 25°C with 5 mM MgC12 and 1 mM [y 32P]ATP. After incubation, the ratio of radiolabeled PA:PIP:PIP2 was approximately 0.‘75:1.0:0.35. It was found that this ratio could be greatly altered by the washing procedure used to remove labeled ATP after the phosphorylation step. In particular, the levels of labeled PA and PIP were reduced by as much as 25 and 66%, respectively, by the washing procedure. Labeled PIP2 usually was not affected. The mechanism for the loss of labeled PA and PIP2 is unknown, but this loss accounts for variations in the initial levels of prelabeled PA and PIP in different washed ghosts preparations used for further Ca2+ studies.
Ca2+-Stimulated Phospholipid Breakdown Incubation of prelabeled ghosts with Caz+ resulted in a rapid decrease in labeled PIP and PIP2 (Fig. 1). Representative time
1
I
5
10
Time
15
(min)
FIG. 1. Time courses of Ca’+-stimulated polyphosphoinositide breakdown: PIP in the presence of 10 PM Ca2+ (0) and 180 PM Ca” (m); PIP2 in the presence of 10 PM Ca” (0) and 180 pM Ca*+ (0). Other conditions: 1 rnM EGTA, 25 mM imidazole-HCl, pH 7.0, and 5 mm MgClz.
RABBIT
ERYTHROCYTE
MEMBRANE
courses for PIP and PIP2 breakdown at 10 and 180 PM Ca2+ demonstrated that maximal breakdown occurred within 10 min at either Ca2+ concentration. The rate and maximal amount of breakdown was greater at 180 PM Ca2+. However, halfmaximal breakdown in the presence of either 10 or 180 pM Ca2+ occurred after approximately 2.5 min. The results in Fig. 1 were obtained in the presence of 5 mM MgClz, but identical results were found in the absence of MgCl (not shown). At 10 and 180 pM Ca2+, MgCl, greatly reduced the rate and amount of breakdown of labeled PA (Fig. 2). In the presence of 5 mM MgC12, breakdown of PA by 10 PM Ca2+ was negligible after 15 min incubation at 25”C, but up to 40% of PA was lost after 5 min incubation in the presence of 180 PM Ca. In the absence of MgC12, loss of labeled PA was rapid and more complete at either 10 or 180 pM Ca2’.
3-
PHOSPHOLIPID
PHOSPHOESTERASES
143
MgC& Inhibition of Ca2+-Dependent Phospho1ip.d Breakdown Further studies showed that MgC12 decreased the rate of Ca2+ stimulated PIP and PIP2 breakdown at Ca2+ concentrations less than 10 PM (Fig. 3). In the absence of MgClz, half-maximal and maximal breakdown of PIP and PIP2 occurred at 1 and 2 pM Ca2+, respectively, after a 5-min incubation. In the presence of 5 mM MgC12, half-maximal and maximal breakdown occurred at 5 and 10 pM Ca2+, reafter 5 min incubation. spectively, Therefore, the apparent Ca2+ affinity of the enzymes involved in Ca2+-stimulated breakdown of PIP and PIP2 were reduced fivefold by MgC12. Additional studies showed the Ca2+ concentration curves for PIP and PIP2 breakdown were identical after 15 min incubation in the presence or absence of MgC12 (not shown). These curves were identical to the curve obtained after 5 min incubation in the absence of MgC12 (Fig. 3). Thus, MgClz reduces the rate of PIP and PIP2 breakdown but not the maximal amount of breakdown induced by Ca2+. MgC12 Inhibition of Ca2+-Dependent Release of Water-Soluble Products
5
IO Time
(min)
FIG. 2. Time course of phosphatidic acid breakdown: PA in the presence of 10 j&M Ca2+ and 5 mM MgClz (O), 180 pM ca2+ and 5 mM MgClz (m), 10 pM Ca*+ and 0 MgCl, (0), and 180 ELMCa’+ and 0 MgClz (0). Other conditions: 1 mM EGTA and 25 mM imidazoleHCl, pH ‘7.0.
Ca2+-dependent breakdown of 32P-labeled PA, PIP, and PIP2 was assayed indirectly by measuring the radiolabeled phosphates released from these phospholipids during their breakdown (5, 6). In the presence of 5 mM MgC12, half-maximal release of total radiolabeled water-soluble products occurred at 4.8 and 1.5 pM Ca2+ after 5- and 15-min incubation times, respectively (Fig. 4). In the absence of MgC12, half-maximal release of water soluble products occurred at 0.8 to 1.2 pM Ca2+ after 5- and 15-min incubation times, respectively (Fig. 4). Thus, the effect of MgC12 on the Ca2+-dependent release of water-soluble products corresponds to the effect on Ca2+-stimulated breakdown (Fig. 3). In both instances, in the presence of MgC12, the concentration of Ca2+ producing half-maximal breakdown decreased with increasing incubation time. This difference in the Ca2+ concentration dependence of release was not observed in the
144
EUGENE
I
0
I 2
4
6
a
1
10
[Ca*+l(ouM) FIG. 3. Effect of MgCIP at low concentrations of mM MgCla (0); PIP with PIP2 with 0 (A) and 5 was 5 min at 25°C.
on phospholipid breakdown Ca’+: PA with 0 (W) and 5 0 (0) and 5 mM MgC& (0); mM MgClz (A). Incubation
QUIST
isotonic conditions over a wide Ca2+ concentration range. An incubation time of 10 min was used to obtain maximal breakdown of PA, PIP, and PIP2 at all Ca2+ concentrations. Under both hypotonic (Fig. 6) and isotonic conditions (Fig. 7), Ca2+dependent breakdown curves of PIP and PIP2 were biphasic in the presence of 5 mM MgClz. A large initial breakdown phase, which accounted for 60% of the total decrease in PIP and PIP, was found between 1 and 10 PM Ca2+. In this initial phase, Ca2+stimulated breakdown was half-maximal at 2-3 PM Ca2+ under both ionic strength conditions. A second breakdown phase occurred between 10 and 100 PM Ca2+, which accounted for approximately 15% of the total loss of PIP and PIP2. The concentration of Ca2+which produced halfmaximal breakdown of PIP and PIP2 in the second phase was variable in different ghost preparations, but usually was be-
absence of MgC12, thus indicating that Mp decreased the rate of Ca’+-stimulated release of water-soluble products but not the maximal amount of release. With 0 Ca2+and 5 mM MgClz, the initial amounts of radioactivity in Fig. 4 represent 32Pi and [y-32P]ATP not removed by washing after the phosphorylation procedures. With 0 Ca2+ and 0 Mg2f, labeled Pi increased with time due to Mg2f- and Ca2+independent phosphomonesterase activities, which primarily breakdown PIP and PA to release Pi (not shown). In the absence of MgC12, low concentrations of Ca2+ were found to decrease labeled PA, and 5 mM MgC12 completely inhibited Ca2+-invoked PA breakdown (Fig. 3). At 10 pM Ca’+, 0.3 mM Mg half-maximally inhibited PA breakdown (Fig. 5). Eflect of Ionic Strength on Ca2’Dependent Phospholipid Breakdown To determine if the ionic strength would affect the affinity of the Ca’+-stimulated PIP and PIP2 phosphoesterase activities in rabbit ghost membranes, the Ca2+ dependence of polyphosphoinositide breakdown was examined under hypotonic and
ll 0
2
4
[Ca”l
6
a
1 10
bdvl)
FIG. 4. Effect of MgCla on the release of watersoluble products at low concentrations of Caa? After 5 min incubation with 0 (m) and 5 mM MgClz (0); after 15 min incubation with 0 (Cl) and 5 mM MgC& (0). Other conditions: 1 mM EGTA, 1 mM ATP, and 25 mM imidazole-HCl, pH 7.0.
RABBIT
L 0
1
ERYTHROCYTE
2 [MgCl$
MEMBRANE
I
I
I
3
4
5
mM
PHOSPHOLIPID
145
PHOSPHOESTERASES
down were examined. In the presence of 5 mM MgC12, Ca2+ stimulated the release of approximately equal amounts of labeled Pi and IPz (Fig. 8). An increase in labeled IPS was not observed under these conditions, even though labeled PIP2 was shown to decrease (Fig. 7). The Ca2+ concentration curves for labeled Pi and IP2 release were biphasic. Most of the label was released between 1 and 10 PM Ca2+, and further release occurred between 10 and 100 FM Ca2+.The Ca2+ concentration producing half-maximal release of labeled water-soluble products corresponded to the Ca2+ concentration required for half maximal breakdown of PIP, PIPz, and PA (Figs. 3, 6, and 7). It should be noted that, in the presence of 5 mM MgC12and 10 pM Ca2+,the breakdown of PA was negligible and, therefore, the labeled Pi and IP2 released must be associated with only PIP and PIP2 breakdown. In these studies, 1
FIG. 5. MgClz dependence of PA breakdown by Ca’+. Incubation was at 25°C for 5 min with 0 (0) and 10 FM Ca2+ (0).
3 tween 25 and 50 PM Ca2+. In the absence of MgC12, similar Ca2+ breakdown curves for PIP and PIP2 were found (not shown). It was found that the total percentage breakdown of both PIP and PIP2 by 180 PM ca2+ in different membrane preparations was approximately 7580% irregardless if the initial amounts of PIP were much higher or approximately the same as PIP2. In the presence of 5 mM MgC12, PA breakdown was significant only at Ca2+ concentrations >lO PM under hypotonic or isotonic conditions (Figs. 6 and 7). PA breakdown was highly significant at lower Ca2+ concentrations in the absence of MgC12(Figs. 2 and 3). Ca’+-Stimulated Release of Labeled Pi and Inositol Phosphates
To identify the enzyme pathways by which Ca2+ stimulates the breakdown of PA, PIP, and PIP2 in rabbit erythrocyte ghosts, the labeled water-soluble products released from these lipids during break-
2 F ii:: s cz 1
jk
7
I 5
I 6
I 4
I 3
p Ca*+ FIG. 6. Ca2+ dependence of phospholipid breakdown under low ionic strength conditions. PIP (O), PIP, (m), and PA (A) after 10 min incubation. Other conditions: 5 mM MgC&, 1 mM EGTA, and 25 mM imidazole-HCl, pH 7.0.
146
EUGENE
QUIST
(Fig. 8), the breakdown of all three labeled phospholipids contributed to the release of water-soluble products. Efect of Ca*’ on Phosphoinositide
OoLl
7I
6I
5
41
p Ca2+
FIG. 7. Ca” dependence of phospholipid breakdown under isotonic conditions. PIP (O), PIP2 (m), and PA (A) after 10 min incubation. Other conditions: 5 mM pH 7.0, MgClz, 1 mM EGTA, 25 mM imidazole-HCl, 80 mM NaCI. and 30 mM KCl.
mM cold ATP was included in the incubation medium to dilute traces of radiolabeled ATP which were not removed during washing (see Materials and Methods). Thus, breakdown of this residual ATP by Ca2+ + M$+-ATPase activity was prevented and hydrolysis of ATP by ATPase activity would not contribute to the Ca’+stimulated increase in labeled Pi. In control experiments, up to 2 mM cold ATP had no effect on Ca’+-stimulated phospholipid breakdown (not shown). The high concentration of labeled Pi observed in the presence of 1 mM EGTA and 0 CaClz was a result of incomplete removal by washing after the prelabeling period (Fig. 8). In the absence of MgCl,, labeled Pi, IP2, and IPQ were released in approximately equal amounts (Fig. 8). Half-maximal release of all three products occurred between 1 and 3 &LM Ca’+, and release was maximal at 10 I.LM Ca’+. Because PA, PIP, and PIP2 are broken down at Ca2+ concentrations
Content
Ghosts were incubated for 20 min at 25°C with 1 mM ATP and 5 mM MgC12 to increase membrane PIP and PIPa (4). After washing to remove ATP, the ghosts were incubated for 20 min in the presence of 5 mM MgClz f 10 PM Ca’+. In the absence of added Ca”+, there was no significant change in the contents of membrane PI, PIP, and PIP2 after incubation (Fig. 9). In the presence of 10 pM Ca2+, PI was increased 23% and PIP and PIP2 were decreased by 35 to 50%, respectively. The effect of Ca2+ on phosphoinositide content was variable in different ghosts preparations, but in all preparations there was a significant increase in PI content after incubation with Ca2+. DISCUSSION
The results of this investigation have revealed a number of characteristics of Ca2+-stimulated phospholipid phosphoes-
0
0
7
6
5
4
p Ca*’
FIG. 8. Effect of Ca’+ and MgClz on the release of water-soluble products from label phospholipids. Release of “Pi with 0 (A) and 5 mM MgClz (A); release of labeled IPZ with 0 (0) and 5 mM MgClz (0); release of labeled IP, with 0 (Cl) and 5 mM MgClz (m). Incubation was 10 min at 25OC in the presence of 1 mM EDTA and 25 mM imidazole-HCl, pH 7.0.
RABBIT
ERYTHROCYTE
MEMBRANE
FIG. 9. Effect of Ca2+ on phosphoinositide content. Incubation was 20 min at 25°C with 0 (open bars) and 10 pM Ca2+ (hatched bars). Other conditions: 5 mM MgC&, 1 mM EGTA, and 25 mM imidazole-HCl, pH 7.0.
terase activities in rabbit erythrocyte membranes which have not previously been described. In general, the enzyme pathways by which Ca2+ stimulates PA, PIP, and PIP2 breakdown appear to be more complex than those described previously in human erythrocyte membranes (5, 6, 20). There are also considerable differences in the responses of membrane phosphoesterases to MgClz and Ca2+ in rabbit erythrocyte membranes compared to those reported in human erythrocyte membranes (6,20). These differences, and the Ca’+-stimulated enzyme pathways involved in PA, PIP, and PIP2 breakdown, will be discussed below. Efect of MgC12 on Ca”-Dependent and PIP Breakdown
PIP
In the presence of Ca’+, 32P-labeled PIP and PIP2 in rabbit ghosts rapidly decrease by a M$+-independent mechanism (Figs. l-3). The kinetics of breakdown are unusual because both the rate of breakdown as well as the maximal total breakdown increase with increasing Ca2+ concentration (Figs. 1, 6, and 7). Thus, Ca2+ may act both as a cofactor to stimulate the activity of polyphosphoinositide phosphoesterase activity and to make more PIP and PIP2 available for attack by this activity. The mechanism for making more
PHOSPHOLIPID
PHOSPHOESTERASES
147
PIP and PIP2 available to phosphoesterase activity is unknown, but could be related to the ability of cations to decrease binding of highly acidic PIP and PIP2 to membrane proteins (4, 21). In agreement with the results of Downes and Michell (6), only a maximum of 65 to 80% of labeled PIP and PIP2 was broken down even at high Ca2+ concentrations (Figs. 6 and 7). The stability of the remainder suggests that part of the PIP and PIP2 synthesized was unavailable to Ca2+ phosphoesterase activities. In this study, Me was found to reduce the rate of Ca2+-stimulated PIP and PIP2 breakdown but only at Ca2+concentrations t10 I.LM (Figs. 3 and 4). MgC12 was found not to decrease the maximal Ca2+-stimulated PIP and PIP2 breakdown at any Ca2+ concentration studied (Figs. 6 and 7). Downs and Michell (20) previously reported that 1 mM MgC12 shifted the apparent K, of Ca2+for PIP and PIP2 breakdown by an order of magnitude under similar hypotonic conditions in human erythrocyte ghosts. These workers (20), however, examined breakdown after only 3 min incubation at 37°C and, therefore, the effect of Ca2+ on maximal breakdown after longer incubations was not determined. Ca2’ Dependence of PIP and PIP, Breakdown Under Hypotonic and Isotonic Conditions
Downes and Michell (20) reported that, under approximately physiological ionic conditions, polyphosphoinositide phosphodiesterase activity was stimulated only at Ca2+ concentrations greater than 100 PM in human erythrocyte membranes. Because intracellular Ca2+ concentrations would never be expected to increase to 100 pM in healthy erythrocytes, these workers concluded that Ca2+ polyphosphoinositide phosphodiesterase would not be activated in normal erythrocytes in the circulation. In contrast, Ca2+ was found here to stimulate PIP and PIP2 breakdown at l-3 pM Ca under hypotonic and isotonic conditions in rabbit erythrocyte membranes (Figs. 6 and 7). Half-
148
EUGENE QUIST
maximal breakdown occurred at approximately 2 PM Ca2+ under both ionic conditions. Because PIP and PIPa can be broken down at these low Ca2+ concentrations, these data suggest that Ca2+ polyphosphoinositide phosphoesterase activities in rabbit erythrocytes could possibly be activated in healthy erythrocytes in the circulation. The large differences in the Ca2+ affinities in these rabbit ghosts and those reported in human ghosts (6, 20) could be a result of species or experimental procedural differences. It is less likely that species differences are responsible because 5 pM Ca2+ was reported to inhibit membrane phosphorylation of human erythrocyte membranes in the presence of Mgaf and [y-32P]ATP under isotonic conditions (12,4). This decrease in labeling was due primarily to “inhibition” of polyphosphoinositide labeling as a result of Ca2+ stimulation of polyphosphoinositide phosphoesterase activity (Quist, published data). Furthermore, others have shown that ~10 PM Ca2+ in the presence of the ionophore A23187 decreased 32P-labeled PIP and PIP2 in intact human erythrocytes under physiological ionic strength conditions (10,ll). Because Ca2+ can stimulate PIP and PIP2 breakdown in human erythrocytes membranes and intact cells in some (lo-12), but not in other (20), studies, the decreased Ca affinity of Ca2+ polyphosphoinositide phosphodiesterase activity may be secondary to membrane preparation or other procedural differences.
Pi; however, enzymes (ii), (iv) and an IP
phosphomonoesterase (6, 22, 23) could also account for 32Pi release under most conditions. It is unlikely that PA phosphomonoesterase activity is active under physiological conditions, because this activity is completely inhibited by 1 mM MgC12 in the presence of ~10 pM Ca2+ (Figs. 3 and 4). In agreement with previous studies (5, 6), evidence for Ca2+-stimulated PIP and PIP2 phosphodiesterase activity was found here. In the absence of MgC12, both 32P-labeled IP2 and IP3 were released from PIP and PIP2 half-maximally by approximately 2 pM Ca2+ (Fig. 8). This corresponds to the Ca2+ concentration required for half-maximal breakdown of PIP and PIP2 (Fig. 7). The Ca2+ concentration required for half-maximal release of IP2 and IP3 were identical, which supports the contention of Downes and Michell (6) that one phosphodiesterase may hydrolyze PIP and PIP2. In the presence of MgC12 and Ca2+, breakdown of PIP and PIP2 occurred (Fig. 2), but only IP2 accumulated (Fig. 8). The absence of IP3 accumulation could be explained by a membrane-bound, Me-dependent IP3 phosphomonoesterase which degrades IP3 to IP2 and Pi (22, 23), or by a PlPz phosphomonoesterase activity. Strong evidence for the existence of IP3 phosphomonoesterase activity in human erythrocyte membranes was reported (22, 23). Because Ca2+-stimulated release of 32Plabeled Pi could be accounted for by PA, PIP, PIP2, and IP2 phosphomonoesterase activities, it was not possible to quantitate which activities were responsible for Ca2+Ca” Phosphommesterase and stimulated phospholipid breakdown by Phosphodiesterase Activities measurement of released water-soluble products only. It was also found that the Potential enzymes invoking PA, PIP, Pi, IP2, and IP3 and PIP2 degradation would include Ca2+- ratios of radiolabeled varied in different ghosts preparations, stimulated (i) PA phosphomonoesterase, which would further complicate quantifi(ii) PIP phosphomonoesterase, (iii) PIP cation of relative amounts of phosphophosphodiesterase, (iv) PIP2 phosphomonoesterase and phosphodiesterase acmonoesterase, and (v) PIP2 phosphodiestivities. Therefore, measurement of reterase activities. The presence of PA products can only phosphomonesterase in rabbit ghosts was leased water-soluble confirm the presence of polyphosphoinopreviously reported (5), and was confirmed here as evidenced by Ca2+-stimulated loss sitide phosphodiesterase activities in these of 32P-labeled PA (Figs. 2, 6, and 7). Ca2+ membranes. Evidence for the presence of polyphosphoinositide monoesterase activalso stimulated the release of 32P-labeled
RABBIT
ERYTHROCYTE
MEMBRANE
ity was found by measuring the effect of Ca2+ on phosphoinositide content. Ca2+ was found here to increase PI content (Fig. 9), and in erythrocyte membranes this result could only be explained by polyphosphoinositide phosphomonoesterase activity. This is because erythrocyte membranes cannot synthesize PI de nova (19) and, therefore, the only possible PI precursors could be PIP or PIP2. In human ghosts, Ca2+was reported not to have any effect on PI concentration, which suggests that this activity may not be present (5) or that differences in phosphorylation procedures or membrane preparation could affect this activity. Although both Ca2+ polyphosphoinositide phosphomonoesterase and phosphodiesterase activities are present in rabbit erythrocyte ghosts, no difference in their affinity for Ca2+ was found. Increases in labeled Pi and IP2 were half-maximally activated by approximately 2 PM Ca2+under physiological ionic conditions in the presence of 5 mM MgC12 (Fig. 8). This finding could imply that one enzyme is responsible for both polyphosphoinositide phosphomonoesterase and phosphodiesterase activities with an identical Ca2+dependency, or that there are two or more Ca2+ phosphoesterase activities. It is therefore obvious that further extensive studies will be required to elucidate the complex properties of these Ca-stimulated phospholipid phosphoesterase activities in erythrocyte membranes and intact cells.
M. R. (1964)
Biochim.
Biophys. Acta 84, 563-575. R. P., AND KIRSCHNER, L. B. (1970) 202, 283-294.
Biophys. Acta
J. T., AND HAWTHORNE,
J. N. (1972)
J.
BioL Chem 247, 7218-7223. 4. QUIST,
E. E., AND BARKER,
B&hem. 5. ALLAN,
Arch.
R. B. (1983)
Biophys. 222,170-178.
D., AND MICHELL,
R. H. (1978)
B&him.
Biophys. Acta 508, 277-286. 6. DOWNES,
R.
B&hem.
C.,
AND
MICHELL,
R.
H.
(1981)
J. 198, 133-140.
‘7. PETERSON,
S. C., AND KIRSHNER, L. B. (1970) Biophys. Acta 202.295-304. GARRETT, N. E., GARRETT, R. J. B., TALWALKER, R. T., AND LESTER, R. L. (1976) .I Cell. PhysioL 87.63-70. LANG, V., PRYHITKA, G., AND BUCKLEY, J. T. (1977) Canad J. Biochem 55,1007-1012. ALLAN, D., AND THOMAS, P. (1981) B&him J. 198,441-445. PANAPPA, B. C., GREENQUIST, A. D., AND SHOHET, C. B. (1980) B&him. Biophys. Acta 598, 494501. QUIST, E. E. (1980) B&hem. Biophys. Res. Commun. 92, 631-637. QUIST, E. E., AND REECE, K. L. (1980) B&hem. Biophys. Res. Commun 95, 1023-1030. QUIST, E. E., AND ROUFOGALIS, B. D. (1975) Arch. B&hem. Biophys. 168, 240-251. KATZ, A. M., REPKE, D. L., UPSHAW, J. E., AND POLASCIK, M. H. (1970) Biochim. Biophys. Acta 205,473-490. BARTLEIT, G. R. (1959) J. BioL Chem. 234, 449458. ELLIS, R. B., GALLIARD, T., AND HAWTHORNE, J. N. (1963) Biochem. J. 88,125-131. QUIST, E. E. (1982) B&hem. Pharmacol 31, 3130-3133. HOKIN, L. E., AND HOKIN, M. R. (1963) Biochim.
B&him.
8.
9. 10. 11.
12. 13. 14. 15.
16. 17. 18. 19.
Biophys. Ada 67, 472-484. 20. DOWNES,
C.
P.,
AND
MICHELL,
Biochem. J. 202,53-58. 21. QUIST, E. E. (1982) Arch. B&hem. C.
B&hem.
R.
H.
(1982)
Biophys.
219,
P.,
AND
MICHELL,
R.
H.
(1982)
J. 203,169-177.
23. DOWNES, C. P., HAWKINS, P. T., AND MICHELL, R. H. (1982) Biochem Sot. Trans. 10. 250-251. 24. QUIST, E. E. (1980) 123-133.
REFERENCES
2. SCHNEIDER, B&him.
3. BUCKLEY,
22. DOWNES,
I thank Ronnie Barker and Patricia Powell for their excellent technical assistance. This research was supported by NIH Grant HL28458.
L. E., AND HOKIN,
149
PHOSPHOESTERASES
58-64.
ACKNOWLEDMENTS
1. HOKIN,
PHOSPHOLIPID
25. DAWSON,
B&hem. 26. GRADO,
Arch. B&hem.
R. M. C., AND DITTMER, J 81,540-545.
C., AND BALLOU,
Chem. 236,54-60.
Biophys.
203,
J. C. (1961)
C. E. (1961)
J. BioL
Ca2+-Stimulated Phospholipid Phosphoesterase in Rabbit Erythrocyte Membranes EUGENE Department
of Pharnwxology,
Texas CoUege
Activities
QUIST
of Osteopathic Medicine, Fort Worth, Texas 76107
Received May 11, 1984, and in revised form August 21, 1984
The properties of the enzymes involved in Ca 2+-stimulated breakdown of phosphatidylinositol 4’-phosphate (PIP), phosphatidylinositol 4’,5’-bisphosphate (PIP2), and phosphatidic acid (PA) in rabbit erythrocyte ghosts were studied. At 25”C, 1 to 180 /LM Ca2+ rapidly stimulated the breakdown of PIP and PIPe, and maximal breakdown occurred within 10 minutes at all Ca2+ concentrations. The rate and the total amount of breakdown of PA, PIP, and PIP2 increased with Ca2+ concentration. MgCl, inhibited the rate of Ca2+-stimulated breakdown of PIP and PIP2 at Ca*’ concentrations less than 10 PM, but did not have any appreciable effects at higher Ca2+ concentrations. MgCl, also protected against Ca 2+-stimulated breakdown of PA. In the presence and absence of 5 mM MgC12, Ca2+ stimulated half-maximal breakdown of PIP and PIP2 at 2-3 ~.LMunder hypotonic and isotonic conditions. In the presence of 5 mM MgC12, Ca’+stimulated breakdown of PIP and PIP2 was associated with the release of Pi and inositol bisphosphate. In the absence of MgC12, Ca2+ stimulated the release of 32Plabeled Pi, inositol bisphosphate, and inositol trisphosphate from labeled PIP, PIPz, and PA. Ca2+ increased phosphatidylinositol content and decreased PIP and PIP2 content in these membranes. The results of this investigation suggest that Ca2+ stimulates the breakdown of polyphosphoinositides by stimulating polyphosphoinositide phosphomonoesterase and phosphodiesterase activities in rabbit erythrocyte ghosts. These activities were activated by less than 3 PM Ca2+ in the presence of MgCl, under hypotonic or isotonic conditions. These Ca2+-stimulated polyphosphoinositide phosphoesterase activities could therefore be active under physiological conditions in normal rabbit erythrocytes. o 19% Academic PRSS, IW.
In mammalian erythrocytes, phosphatidylinositol 4’-phosphate (PIP),l phosphatidylinositol 4’,5’-bisphosphate (PIP,), and phosphatidic acid (PA) are synthesized from membrane phosphatidylinositol and diacylglycerol, respectively, by Mgz+dependent kinase activities (l-4). Phosphoesterase activities are also present which hydrolyze these phospholipids, and together these enzymes account for the ’ Abbreviations used: EGTA, ethylene glycol bis(baminoethyl ether)-N,W-tetraacetic acid, PIP, phosphatidylinositol 4’-phosphate; PIP2, phosphatidylinositol 4’,5’-bisphosphate; IP2. inositol bisphosphate; IP,, inositol trisphosphate; PA, phosphatidic acid. 0003-9861/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.
rapid turnover of PA, PIP, and PIP2 in erythrocyte membranes (5-7). Few studies have been undertaken to study the phospholipid phosphoesterase activities; however, Ca2+ has been shown to stimulate the breakdown of PIP and PIP2 by a Mgz+independent phosphodiesterase activity (5, 6). This activity hydrolyzes PIP to diacylglycerol and inositol bisphosphate (IPz), and PIP2 to diacylglycerol and inositol trisphosphate (IPB) (5, 6). Although a number of workers have suggested that PIP and PIP, may also be broken down by a Ca2+-stimulated phosphomonoesterase activity (7-g), direct evidence for the existence of this enzyme activity in ery140
RABBIT
ERYTHROCYTE
MEMBRANE
thocyte membranes is lacking. Presently there are conflicts in the literature regarding Ca2+ affinity of polyphosphosphoinositide phosphosphodiesterase activity (5, 6, 10, 20). In rabbit (5, 13) and human (5) erythrocyte membranes, Ca2+ has been shown to stimulate polyphosphoinositide breakdown at l-10 PM Ca” under hypotonic conditions. Ca2+ concentrations as low as 10 PM in the presence of A23187 have also been reported to stimulate the breakdown of 32P-labeled PIP and PIP2 in intact erythrocytes (10, 11). In contrast, Downes and Michell (20) have reported that, under isotonic conditions in the presence of physiological MgC&, greater than 100 PM Ca2+ is required to stimulate polyphosphoinositide breakdown. Therefore, according to this report (20), activation of Ca2+ polyphosphoinositide phosphodiesterase would never occur in healthy human erythrocytes because intracellular Cast concentrations would never be expected to increase above 10 PM. However, if this enzyme were activated at ~10 PM Ca’+, it could possibly regulate erythrocyte function under various conditions in the circulation. In this study, the properties of Ca2+-dependent PA, PIP, and PIP phosphoesterase activities in freshly prepared rabbit erythrocyte ghosts were examined under a number of conditions. These studies provide evidence that rabbit erythrocyte membranes possess Ca2+ polyphosphoinositide phosphomonoesterase activity in addition to polyphosphoinositide phosphodiesterase activity. This study also shows that these activities are half-maximally stimulated by 2-3 I.LM Ca2+ under isotonic conditions in the presence of MgC12. MATERIALS
AND
METHODS
Materials. [y-JLp]ATP (25 Ci/mmol) and qi (carrier free) was purchased from ICN (Irvine, Calif.). Reference phospholipids, L-a-phosphatidylinositol (98% from soybean), L-a-phosphatidylinositol 4’,5’-diphosphate (98% from bovine brain), L-a-phosphatidylinositol 4’-monophosphate (98% from bovine brain), and Dowex-1 (1 X 8-400) were purchased from Sigma Chemical Company. Silica gel 60 thin-layer plates (0.25 mm) were purchased from E. Merck. CHC& and CH,OH were HPLC grade. Methylamine (40%) was purchased from MCB. All other chemicals were reagent grade.
PHOSPHOLIPID
PHOSPHOESTERASES
141
Preparation of erythrocyte membranes. Fresh heparinized blood was drawn from New Zealand white rabbits. Erythrocytes were washed free of plasma, and white cells in isotonic saline (4) and ghosts were prepared by washing packed erythrocytes three to four times with 10 vol of 20 mM Tris-HCI, pH 7.6, at 5°C (4). Ghosts were used within 1 h of preparation. Phosphmylatim studies To phosphorylate PA, PIP, and PIP2 with 32P, 3 ml ghosts was incubated for 20 min at 25”C, in a final volume of 7.5 ml, in 25 mM imidazole-HCI, pH 7.0,5 mM MgClz, 1 mM [r-J2PjATP, and 1 mM EGTA. After incubation, the ghosts were washed twice with 25 ml 10 mM Tris-HCl, pH 7.6, at 5°C and 20,OOOg.The washed ghosts were resuspended to 4.5 ml in the washing medium, and 0.25ml aliquots were incubated in a final volume of 0.5 ml with a final composition of 25 mM imidazole, pH 7.0, 1 mM EGTA, and usually 1 mM ATP. MgClz and CaCI, were varied in this medium, and the free Ca ion concentration was calculated using a CaEGTA stability constant of lO”-~ (14, 15). In some experiments 80 mM NaCl and 30 mM KC1 were also included to increase the ionic strength to 150 mM. The tubes were incubated at 25”C, and the reaction was stopped with 1.0 ml 10% trichloroacetic acid. The tubes were centrifuged at 2500 rpm at 5”C, and 1.0 ml of the supernatant was collected for further determination of radiolabeled water-soluble products (see below). The pellets were further washed with 3.0 ml HzO, and 2.0 ml CHCla:CHaOH:HCI (20:40:1) was added to each tube, and the tubes were left 20 min at 5°C. After this time, 0.75 ml CHCIB and Hz0 were consecutively added to the tubes, and the tubes were centrifuged at 25009 for 10 min. The upper phase and protein interface were removed, and a lml aliquot of the CHCls phase was dried under Nz and resuspended to 40 111with CHCI,:CH,OH:HCl (6:3:0.1). Twenty-microliter aliquots were spotted on silica gel 60 plates, and lo-p1 aliquots were analyzed for total lipid phosphorous (16). Thin-layer plates were developed in CHC1a:CH30H:Hz0:30% NH, (25:35:7.2:2.5), which completely separates radiolabeled PA, PIP, and PIP2 as previously described (4). Labeled phospholipids were detected by autoradiography, scraped, and counted in 8.0 ml Tritisol (4). Measurement of radiolabeled, water-soluble pm&As. The =P-labeled Pi, IPz, an IPa released from phospholipids in the presence of various concentrations of CaClz and MgClz were determined by anionexchange chromatography (6, 17). The supernatant obtained after stopping the reaction with trichloracetic acid was neutralized with NaOH, and 12 ml Hz0 was then added to each tube. Ten-milliliter aliquots were applied to l-ml Dowex-1 (formate form) columns, and radiolabeled Pi, IPz. and IPB were eluted with 10 ml 0.1 M formic acid:0.2M ammonium formate, 0.1 M formic acid:0.4 M ammonium formate, and 0.1 M formic acid:1 M ammonium for-
142
EUGENE
mate, respectively. One-milliliter aliquots of each fraction were counted in 10 ml Tritisol. Labeled PIP and PIP* were prepared and separated by phosphorylating ghosts and by thin-layer chromatography methods described above. IP, and IPI standards of unknown positional isomerism were prepared from 3aP-labeled PIP and PIP2 by hydrolysis for 7 min in 5 N HCl at 100°C (25). After removal of the organic products and unhydrolyzed polyphosphoinositides by extraction with chloroform (25), the aqueous phases, containing primarily IPz or IP, and other minor water-soluble products, were lyophilized. The identities of labeled IPz and IP3 obtained by anionexchange chromotagraphy or by hydrolysis of PIP and PIP2 were also confirmed by determining their relative mobilities on Whatman No. 1 chromatography paper developed in n-propranol:concentrated ammonia:water (5:4:1) as previously described by others (5, 26). “*Pi (carrier free) was also used as a standard. Lktermination
of phphoinositide
content
in ghosts.
Ghosts were incubated in the identical medium used above to label the phospholipids with 32P in a final volume of 5 ml ghosts (l-2 ml) were added to each tube. The reaction was stopped with 7.5 ml 5% trichloroacetic acid, and the tubes were centrifuged at 25009 for 10 min. The pellet was washed with 7.5 ml Hz0 at 30009 for 15 min. Lipids were extracted from the pellets with 7.5 ml CHC13:CHIOH:HCl (20:40:1) containing 0.01% butylated hydroxytoluene for 20 min at 5°C. CHC& (2.5 ml) and 1.5 ml HaO) were consecutively added, and the tubes were centrifuged for 10 min at 2500g. The HzO:CH,OH phase was discarded, and the CHCla phase was washed three times with 3 ml 0.1 N HCl. Three milliliters of the chloroform phase was dried under nitrogen, and the lipids were resuspended with 100 pl CHCl,: CH30H:HCl (63:O.l). For total phospholipid determination, lo-p1 aliquots were analyzed for phosphorus (16). To separate PI, 80-~1 aliquots were spotted over 1.5 cm on silica gel 60 thin-layer plates. These plates were developed in CHC13:CH,0H:CHaNH2 (65:35:10), which routinely separates phosphatidylinositol (Rf 0.25) from phosphatylserine and phosphatidic acid (H0.35) as previously described (18). Polyphosphoinositides were separated on silica gel 60 plates using the solvent system described above for separating radiolabeled PA, PIP, and PIP (4). Authentic phospholipid standards were run on both plates. The phospholipids were detected in an iodine chamber and scraped from the plates. Phosphorous content was determined (16). Blanks of similar area were also scraped from the plates and their absorbance was subtracted. RESULTS
In this investigation, PIP, PIPB, and PA were prelabeled with 32Pi by incubating freshly prepared rabbit ghosts for 20 min
QUIST
at 25°C with 5 mM MgC12 and 1 mM [y 32P]ATP. After incubation, the ratio of radiolabeled PA:PIP:PIP2 was approximately 0.‘75:1.0:0.35. It was found that this ratio could be greatly altered by the washing procedure used to remove labeled ATP after the phosphorylation step. In particular, the levels of labeled PA and PIP were reduced by as much as 25 and 66%, respectively, by the washing procedure. Labeled PIP2 usually was not affected. The mechanism for the loss of labeled PA and PIP2 is unknown, but this loss accounts for variations in the initial levels of prelabeled PA and PIP in different washed ghosts preparations used for further Ca2+ studies.
Ca2+-Stimulated Phospholipid Breakdown Incubation of prelabeled ghosts with Caz+ resulted in a rapid decrease in labeled PIP and PIP2 (Fig. 1). Representative time
1
I
5
10
Time
15
(min)
FIG. 1. Time courses of Ca’+-stimulated polyphosphoinositide breakdown: PIP in the presence of 10 PM Ca2+ (0) and 180 PM Ca” (m); PIP2 in the presence of 10 PM Ca” (0) and 180 pM Ca*+ (0). Other conditions: 1 rnM EGTA, 25 mM imidazole-HCl, pH 7.0, and 5 mm MgClz.
RABBIT
ERYTHROCYTE
MEMBRANE
courses for PIP and PIP2 breakdown at 10 and 180 PM Ca2+ demonstrated that maximal breakdown occurred within 10 min at either Ca2+ concentration. The rate and maximal amount of breakdown was greater at 180 PM Ca2+. However, halfmaximal breakdown in the presence of either 10 or 180 pM Ca2+ occurred after approximately 2.5 min. The results in Fig. 1 were obtained in the presence of 5 mM MgClz, but identical results were found in the absence of MgCl (not shown). At 10 and 180 pM Ca2+, MgCl, greatly reduced the rate and amount of breakdown of labeled PA (Fig. 2). In the presence of 5 mM MgC12, breakdown of PA by 10 PM Ca2+ was negligible after 15 min incubation at 25”C, but up to 40% of PA was lost after 5 min incubation in the presence of 180 PM Ca. In the absence of MgC12, loss of labeled PA was rapid and more complete at either 10 or 180 pM Ca2’.
3-
PHOSPHOLIPID
PHOSPHOESTERASES
143
MgC& Inhibition of Ca2+-Dependent Phospho1ip.d Breakdown Further studies showed that MgC12 decreased the rate of Ca2+ stimulated PIP and PIP2 breakdown at Ca2+ concentrations less than 10 PM (Fig. 3). In the absence of MgClz, half-maximal and maximal breakdown of PIP and PIP2 occurred at 1 and 2 pM Ca2+, respectively, after a 5-min incubation. In the presence of 5 mM MgC12, half-maximal and maximal breakdown occurred at 5 and 10 pM Ca2+, reafter 5 min incubation. spectively, Therefore, the apparent Ca2+ affinity of the enzymes involved in Ca2+-stimulated breakdown of PIP and PIP2 were reduced fivefold by MgC12. Additional studies showed the Ca2+ concentration curves for PIP and PIP2 breakdown were identical after 15 min incubation in the presence or absence of MgC12 (not shown). These curves were identical to the curve obtained after 5 min incubation in the absence of MgC12 (Fig. 3). Thus, MgClz reduces the rate of PIP and PIP2 breakdown but not the maximal amount of breakdown induced by Ca2+. MgC12 Inhibition of Ca2+-Dependent Release of Water-Soluble Products
5
IO Time
(min)
FIG. 2. Time course of phosphatidic acid breakdown: PA in the presence of 10 j&M Ca2+ and 5 mM MgClz (O), 180 pM ca2+ and 5 mM MgClz (m), 10 pM Ca*+ and 0 MgCl, (0), and 180 ELMCa’+ and 0 MgClz (0). Other conditions: 1 mM EGTA and 25 mM imidazoleHCl, pH ‘7.0.
Ca2+-dependent breakdown of 32P-labeled PA, PIP, and PIP2 was assayed indirectly by measuring the radiolabeled phosphates released from these phospholipids during their breakdown (5, 6). In the presence of 5 mM MgC12, half-maximal release of total radiolabeled water-soluble products occurred at 4.8 and 1.5 pM Ca2+ after 5- and 15-min incubation times, respectively (Fig. 4). In the absence of MgC12, half-maximal release of water soluble products occurred at 0.8 to 1.2 pM Ca2+ after 5- and 15-min incubation times, respectively (Fig. 4). Thus, the effect of MgC12 on the Ca2+-dependent release of water-soluble products corresponds to the effect on Ca2+-stimulated breakdown (Fig. 3). In both instances, in the presence of MgC12, the concentration of Ca2+ producing half-maximal breakdown decreased with increasing incubation time. This difference in the Ca2+ concentration dependence of release was not observed in the
144
EUGENE
I
0
I 2
4
6
a
1
10
[Ca*+l(ouM) FIG. 3. Effect of MgCIP at low concentrations of mM MgCla (0); PIP with PIP2 with 0 (A) and 5 was 5 min at 25°C.
on phospholipid breakdown Ca’+: PA with 0 (W) and 5 0 (0) and 5 mM MgC& (0); mM MgClz (A). Incubation
QUIST
isotonic conditions over a wide Ca2+ concentration range. An incubation time of 10 min was used to obtain maximal breakdown of PA, PIP, and PIP2 at all Ca2+ concentrations. Under both hypotonic (Fig. 6) and isotonic conditions (Fig. 7), Ca2+dependent breakdown curves of PIP and PIP2 were biphasic in the presence of 5 mM MgClz. A large initial breakdown phase, which accounted for 60% of the total decrease in PIP and PIP, was found between 1 and 10 PM Ca2+. In this initial phase, Ca2+stimulated breakdown was half-maximal at 2-3 PM Ca2+ under both ionic strength conditions. A second breakdown phase occurred between 10 and 100 PM Ca2+, which accounted for approximately 15% of the total loss of PIP and PIP2. The concentration of Ca2+which produced halfmaximal breakdown of PIP and PIP2 in the second phase was variable in different ghost preparations, but usually was be-
absence of MgC12, thus indicating that Mp decreased the rate of Ca’+-stimulated release of water-soluble products but not the maximal amount of release. With 0 Ca2+and 5 mM MgClz, the initial amounts of radioactivity in Fig. 4 represent 32Pi and [y-32P]ATP not removed by washing after the phosphorylation procedures. With 0 Ca2+ and 0 Mg2f, labeled Pi increased with time due to Mg2f- and Ca2+independent phosphomonesterase activities, which primarily breakdown PIP and PA to release Pi (not shown). In the absence of MgC12, low concentrations of Ca2+ were found to decrease labeled PA, and 5 mM MgC12 completely inhibited Ca2+-invoked PA breakdown (Fig. 3). At 10 pM Ca’+, 0.3 mM Mg half-maximally inhibited PA breakdown (Fig. 5). Eflect of Ionic Strength on Ca2’Dependent Phospholipid Breakdown To determine if the ionic strength would affect the affinity of the Ca’+-stimulated PIP and PIP2 phosphoesterase activities in rabbit ghost membranes, the Ca2+ dependence of polyphosphoinositide breakdown was examined under hypotonic and
ll 0
2
4
[Ca”l
6
a
1 10
bdvl)
FIG. 4. Effect of MgCla on the release of watersoluble products at low concentrations of Caa? After 5 min incubation with 0 (m) and 5 mM MgClz (0); after 15 min incubation with 0 (Cl) and 5 mM MgC& (0). Other conditions: 1 mM EGTA, 1 mM ATP, and 25 mM imidazole-HCl, pH 7.0.
RABBIT
L 0
1
ERYTHROCYTE
2 [MgCl$
MEMBRANE
I
I
I
3
4
5
mM
PHOSPHOLIPID
145
PHOSPHOESTERASES
down were examined. In the presence of 5 mM MgC12, Ca2+ stimulated the release of approximately equal amounts of labeled Pi and IPz (Fig. 8). An increase in labeled IPS was not observed under these conditions, even though labeled PIP2 was shown to decrease (Fig. 7). The Ca2+ concentration curves for labeled Pi and IP2 release were biphasic. Most of the label was released between 1 and 10 PM Ca2+, and further release occurred between 10 and 100 FM Ca2+.The Ca2+ concentration producing half-maximal release of labeled water-soluble products corresponded to the Ca2+ concentration required for half maximal breakdown of PIP, PIPz, and PA (Figs. 3, 6, and 7). It should be noted that, in the presence of 5 mM MgC12and 10 pM Ca2+,the breakdown of PA was negligible and, therefore, the labeled Pi and IP2 released must be associated with only PIP and PIP2 breakdown. In these studies, 1
FIG. 5. MgClz dependence of PA breakdown by Ca’+. Incubation was at 25°C for 5 min with 0 (0) and 10 FM Ca2+ (0).
3 tween 25 and 50 PM Ca2+. In the absence of MgC12, similar Ca2+ breakdown curves for PIP and PIP2 were found (not shown). It was found that the total percentage breakdown of both PIP and PIP2 by 180 PM ca2+ in different membrane preparations was approximately 7580% irregardless if the initial amounts of PIP were much higher or approximately the same as PIP2. In the presence of 5 mM MgC12, PA breakdown was significant only at Ca2+ concentrations >lO PM under hypotonic or isotonic conditions (Figs. 6 and 7). PA breakdown was highly significant at lower Ca2+ concentrations in the absence of MgC12(Figs. 2 and 3). Ca’+-Stimulated Release of Labeled Pi and Inositol Phosphates
To identify the enzyme pathways by which Ca2+ stimulates the breakdown of PA, PIP, and PIP2 in rabbit erythrocyte ghosts, the labeled water-soluble products released from these lipids during break-
2 F ii:: s cz 1
jk
7
I 5
I 6
I 4
I 3
p Ca*+ FIG. 6. Ca2+ dependence of phospholipid breakdown under low ionic strength conditions. PIP (O), PIP, (m), and PA (A) after 10 min incubation. Other conditions: 5 mM MgC&, 1 mM EGTA, and 25 mM imidazole-HCl, pH 7.0.
146
EUGENE
QUIST
(Fig. 8), the breakdown of all three labeled phospholipids contributed to the release of water-soluble products. Efect of Ca*’ on Phosphoinositide
OoLl
7I
6I
5
41
p Ca2+
FIG. 7. Ca” dependence of phospholipid breakdown under isotonic conditions. PIP (O), PIP2 (m), and PA (A) after 10 min incubation. Other conditions: 5 mM pH 7.0, MgClz, 1 mM EGTA, 25 mM imidazole-HCl, 80 mM NaCI. and 30 mM KCl.
mM cold ATP was included in the incubation medium to dilute traces of radiolabeled ATP which were not removed during washing (see Materials and Methods). Thus, breakdown of this residual ATP by Ca2+ + M$+-ATPase activity was prevented and hydrolysis of ATP by ATPase activity would not contribute to the Ca’+stimulated increase in labeled Pi. In control experiments, up to 2 mM cold ATP had no effect on Ca’+-stimulated phospholipid breakdown (not shown). The high concentration of labeled Pi observed in the presence of 1 mM EGTA and 0 CaClz was a result of incomplete removal by washing after the prelabeling period (Fig. 8). In the absence of MgCl,, labeled Pi, IP2, and IPQ were released in approximately equal amounts (Fig. 8). Half-maximal release of all three products occurred between 1 and 3 &LM Ca’+, and release was maximal at 10 I.LM Ca’+. Because PA, PIP, and PIP2 are broken down at Ca2+ concentrations
Content
Ghosts were incubated for 20 min at 25°C with 1 mM ATP and 5 mM MgC12 to increase membrane PIP and PIPa (4). After washing to remove ATP, the ghosts were incubated for 20 min in the presence of 5 mM MgClz f 10 PM Ca’+. In the absence of added Ca”+, there was no significant change in the contents of membrane PI, PIP, and PIP2 after incubation (Fig. 9). In the presence of 10 pM Ca2+, PI was increased 23% and PIP and PIP2 were decreased by 35 to 50%, respectively. The effect of Ca2+ on phosphoinositide content was variable in different ghosts preparations, but in all preparations there was a significant increase in PI content after incubation with Ca2+. DISCUSSION
The results of this investigation have revealed a number of characteristics of Ca2+-stimulated phospholipid phosphoes-
0
0
7
6
5
4
p Ca*’
FIG. 8. Effect of Ca’+ and MgClz on the release of water-soluble products from label phospholipids. Release of “Pi with 0 (A) and 5 mM MgClz (A); release of labeled IPZ with 0 (0) and 5 mM MgClz (0); release of labeled IP, with 0 (Cl) and 5 mM MgClz (m). Incubation was 10 min at 25OC in the presence of 1 mM EDTA and 25 mM imidazole-HCl, pH 7.0.
RABBIT
ERYTHROCYTE
MEMBRANE
FIG. 9. Effect of Ca2+ on phosphoinositide content. Incubation was 20 min at 25°C with 0 (open bars) and 10 pM Ca2+ (hatched bars). Other conditions: 5 mM MgC&, 1 mM EGTA, and 25 mM imidazole-HCl, pH 7.0.
terase activities in rabbit erythrocyte membranes which have not previously been described. In general, the enzyme pathways by which Ca2+ stimulates PA, PIP, and PIP2 breakdown appear to be more complex than those described previously in human erythrocyte membranes (5, 6, 20). There are also considerable differences in the responses of membrane phosphoesterases to MgClz and Ca2+ in rabbit erythrocyte membranes compared to those reported in human erythrocyte membranes (6,20). These differences, and the Ca’+-stimulated enzyme pathways involved in PA, PIP, and PIP2 breakdown, will be discussed below. Efect of MgC12 on Ca”-Dependent and PIP Breakdown
PIP
In the presence of Ca’+, 32P-labeled PIP and PIP2 in rabbit ghosts rapidly decrease by a M$+-independent mechanism (Figs. l-3). The kinetics of breakdown are unusual because both the rate of breakdown as well as the maximal total breakdown increase with increasing Ca2+ concentration (Figs. 1, 6, and 7). Thus, Ca2+ may act both as a cofactor to stimulate the activity of polyphosphoinositide phosphoesterase activity and to make more PIP and PIP2 available for attack by this activity. The mechanism for making more
PHOSPHOLIPID
PHOSPHOESTERASES
147
PIP and PIP2 available to phosphoesterase activity is unknown, but could be related to the ability of cations to decrease binding of highly acidic PIP and PIP2 to membrane proteins (4, 21). In agreement with the results of Downes and Michell (6), only a maximum of 65 to 80% of labeled PIP and PIP2 was broken down even at high Ca2+ concentrations (Figs. 6 and 7). The stability of the remainder suggests that part of the PIP and PIP2 synthesized was unavailable to Ca2+ phosphoesterase activities. In this study, Me was found to reduce the rate of Ca2+-stimulated PIP and PIP2 breakdown but only at Ca2+concentrations t10 I.LM (Figs. 3 and 4). MgC12 was found not to decrease the maximal Ca2+-stimulated PIP and PIP2 breakdown at any Ca2+ concentration studied (Figs. 6 and 7). Downs and Michell (20) previously reported that 1 mM MgC12 shifted the apparent K, of Ca2+for PIP and PIP2 breakdown by an order of magnitude under similar hypotonic conditions in human erythrocyte ghosts. These workers (20), however, examined breakdown after only 3 min incubation at 37°C and, therefore, the effect of Ca2+ on maximal breakdown after longer incubations was not determined. Ca2’ Dependence of PIP and PIP, Breakdown Under Hypotonic and Isotonic Conditions
Downes and Michell (20) reported that, under approximately physiological ionic conditions, polyphosphoinositide phosphodiesterase activity was stimulated only at Ca2+ concentrations greater than 100 PM in human erythrocyte membranes. Because intracellular Ca2+ concentrations would never be expected to increase to 100 pM in healthy erythrocytes, these workers concluded that Ca2+ polyphosphoinositide phosphodiesterase would not be activated in normal erythrocytes in the circulation. In contrast, Ca2+ was found here to stimulate PIP and PIP2 breakdown at l-3 pM Ca under hypotonic and isotonic conditions in rabbit erythrocyte membranes (Figs. 6 and 7). Half-
148
EUGENE QUIST
maximal breakdown occurred at approximately 2 PM Ca2+ under both ionic conditions. Because PIP and PIPa can be broken down at these low Ca2+ concentrations, these data suggest that Ca2+ polyphosphoinositide phosphoesterase activities in rabbit erythrocytes could possibly be activated in healthy erythrocytes in the circulation. The large differences in the Ca2+ affinities in these rabbit ghosts and those reported in human ghosts (6, 20) could be a result of species or experimental procedural differences. It is less likely that species differences are responsible because 5 pM Ca2+ was reported to inhibit membrane phosphorylation of human erythrocyte membranes in the presence of Mgaf and [y-32P]ATP under isotonic conditions (12,4). This decrease in labeling was due primarily to “inhibition” of polyphosphoinositide labeling as a result of Ca2+ stimulation of polyphosphoinositide phosphoesterase activity (Quist, published data). Furthermore, others have shown that ~10 PM Ca2+ in the presence of the ionophore A23187 decreased 32P-labeled PIP and PIP2 in intact human erythrocytes under physiological ionic strength conditions (10,ll). Because Ca2+ can stimulate PIP and PIP2 breakdown in human erythrocytes membranes and intact cells in some (lo-12), but not in other (20), studies, the decreased Ca affinity of Ca2+ polyphosphoinositide phosphodiesterase activity may be secondary to membrane preparation or other procedural differences.
Pi; however, enzymes (ii), (iv) and an IP
phosphomonoesterase (6, 22, 23) could also account for 32Pi release under most conditions. It is unlikely that PA phosphomonoesterase activity is active under physiological conditions, because this activity is completely inhibited by 1 mM MgC12 in the presence of ~10 pM Ca2+ (Figs. 3 and 4). In agreement with previous studies (5, 6), evidence for Ca2+-stimulated PIP and PIP2 phosphodiesterase activity was found here. In the absence of MgC12, both 32P-labeled IP2 and IP3 were released from PIP and PIP2 half-maximally by approximately 2 pM Ca2+ (Fig. 8). This corresponds to the Ca2+ concentration required for half-maximal breakdown of PIP and PIP2 (Fig. 7). The Ca2+ concentration required for half-maximal release of IP2 and IP3 were identical, which supports the contention of Downes and Michell (6) that one phosphodiesterase may hydrolyze PIP and PIP2. In the presence of MgC12 and Ca2+, breakdown of PIP and PIP2 occurred (Fig. 2), but only IP2 accumulated (Fig. 8). The absence of IP3 accumulation could be explained by a membrane-bound, Me-dependent IP3 phosphomonoesterase which degrades IP3 to IP2 and Pi (22, 23), or by a PlPz phosphomonoesterase activity. Strong evidence for the existence of IP3 phosphomonoesterase activity in human erythrocyte membranes was reported (22, 23). Because Ca2+-stimulated release of 32Plabeled Pi could be accounted for by PA, PIP, PIP2, and IP2 phosphomonoesterase activities, it was not possible to quantitate which activities were responsible for Ca2+Ca” Phosphommesterase and stimulated phospholipid breakdown by Phosphodiesterase Activities measurement of released water-soluble products only. It was also found that the Potential enzymes invoking PA, PIP, Pi, IP2, and IP3 and PIP2 degradation would include Ca2+- ratios of radiolabeled varied in different ghosts preparations, stimulated (i) PA phosphomonoesterase, which would further complicate quantifi(ii) PIP phosphomonoesterase, (iii) PIP cation of relative amounts of phosphophosphodiesterase, (iv) PIP2 phosphomonoesterase and phosphodiesterase acmonoesterase, and (v) PIP2 phosphodiestivities. Therefore, measurement of reterase activities. The presence of PA products can only phosphomonesterase in rabbit ghosts was leased water-soluble confirm the presence of polyphosphoinopreviously reported (5), and was confirmed here as evidenced by Ca2+-stimulated loss sitide phosphodiesterase activities in these of 32P-labeled PA (Figs. 2, 6, and 7). Ca2+ membranes. Evidence for the presence of polyphosphoinositide monoesterase activalso stimulated the release of 32P-labeled
RABBIT
ERYTHROCYTE
MEMBRANE
ity was found by measuring the effect of Ca2+ on phosphoinositide content. Ca2+ was found here to increase PI content (Fig. 9), and in erythrocyte membranes this result could only be explained by polyphosphoinositide phosphomonoesterase activity. This is because erythrocyte membranes cannot synthesize PI de nova (19) and, therefore, the only possible PI precursors could be PIP or PIP2. In human ghosts, Ca2+was reported not to have any effect on PI concentration, which suggests that this activity may not be present (5) or that differences in phosphorylation procedures or membrane preparation could affect this activity. Although both Ca2+ polyphosphoinositide phosphomonoesterase and phosphodiesterase activities are present in rabbit erythrocyte ghosts, no difference in their affinity for Ca2+ was found. Increases in labeled Pi and IP2 were half-maximally activated by approximately 2 PM Ca2+under physiological ionic conditions in the presence of 5 mM MgC12 (Fig. 8). This finding could imply that one enzyme is responsible for both polyphosphoinositide phosphomonoesterase and phosphodiesterase activities with an identical Ca2+dependency, or that there are two or more Ca2+ phosphoesterase activities. It is therefore obvious that further extensive studies will be required to elucidate the complex properties of these Ca-stimulated phospholipid phosphoesterase activities in erythrocyte membranes and intact cells.
M. R. (1964)
Biochim.
Biophys. Acta 84, 563-575. R. P., AND KIRSCHNER, L. B. (1970) 202, 283-294.
Biophys. Acta
J. T., AND HAWTHORNE,
J. N. (1972)
J.
BioL Chem 247, 7218-7223. 4. QUIST,
E. E., AND BARKER,
B&hem. 5. ALLAN,
Arch.
R. B. (1983)
Biophys. 222,170-178.
D., AND MICHELL,
R. H. (1978)
B&him.
Biophys. Acta 508, 277-286. 6. DOWNES,
R.
B&hem.
C.,
AND
MICHELL,
R.
H.
(1981)
J. 198, 133-140.
‘7. PETERSON,
S. C., AND KIRSHNER, L. B. (1970) Biophys. Acta 202.295-304. GARRETT, N. E., GARRETT, R. J. B., TALWALKER, R. T., AND LESTER, R. L. (1976) .I Cell. PhysioL 87.63-70. LANG, V., PRYHITKA, G., AND BUCKLEY, J. T. (1977) Canad J. Biochem 55,1007-1012. ALLAN, D., AND THOMAS, P. (1981) B&him J. 198,441-445. PANAPPA, B. C., GREENQUIST, A. D., AND SHOHET, C. B. (1980) B&him. Biophys. Acta 598, 494501. QUIST, E. E. (1980) B&hem. Biophys. Res. Commun. 92, 631-637. QUIST, E. E., AND REECE, K. L. (1980) B&hem. Biophys. Res. Commun 95, 1023-1030. QUIST, E. E., AND ROUFOGALIS, B. D. (1975) Arch. B&hem. Biophys. 168, 240-251. KATZ, A. M., REPKE, D. L., UPSHAW, J. E., AND POLASCIK, M. H. (1970) Biochim. Biophys. Acta 205,473-490. BARTLEIT, G. R. (1959) J. BioL Chem. 234, 449458. ELLIS, R. B., GALLIARD, T., AND HAWTHORNE, J. N. (1963) Biochem. J. 88,125-131. QUIST, E. E. (1982) B&hem. Pharmacol 31, 3130-3133. HOKIN, L. E., AND HOKIN, M. R. (1963) Biochim.
B&him.
8.
9. 10. 11.
12. 13. 14. 15.
16. 17. 18. 19.
Biophys. Ada 67, 472-484. 20. DOWNES,
C.
P.,
AND
MICHELL,
Biochem. J. 202,53-58. 21. QUIST, E. E. (1982) Arch. B&hem. C.
B&hem.
R.
H.
(1982)
Biophys.
219,
P.,
AND
MICHELL,
R.
H.
(1982)
J. 203,169-177.
23. DOWNES, C. P., HAWKINS, P. T., AND MICHELL, R. H. (1982) Biochem Sot. Trans. 10. 250-251. 24. QUIST, E. E. (1980) 123-133.
REFERENCES
2. SCHNEIDER, B&him.
3. BUCKLEY,
22. DOWNES,
I thank Ronnie Barker and Patricia Powell for their excellent technical assistance. This research was supported by NIH Grant HL28458.
L. E., AND HOKIN,
149
PHOSPHOESTERASES
58-64.
ACKNOWLEDMENTS
1. HOKIN,
PHOSPHOLIPID
25. DAWSON,
B&hem. 26. GRADO,
Arch. B&hem.
R. M. C., AND DITTMER, J 81,540-545.
C., AND BALLOU,
Chem. 236,54-60.
Biophys.
203,
J. C. (1961)
C. E. (1961)
J. BioL