Changes in the Concentration and Fatty Acid Composition ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 264, No. 5, Issue of February 15, pp. 2574-2580,1989

0 1989 by The American Society for Biochemistry end Molecular Biology, Inc.

Printed in U.S.A.

Changes in the Concentration and Fatty Acid Composition of Phosphoinositides Induced by Hormonesin Hepatocytes* (Received for publication, April 20, 1988)

Guy Augert, Peter F. Blackmore*, and JohnH. Extons From the Howard HuPhes Medical Institute and the Devartment of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nmhville, Tennessee 37232 ~~

The hormonal regulation of phosphoinositide levels in isolated hepatocytes wasstudied using chemical means. Extracted inositol phospholipids were adsorbed to neomycin-coated glass beads and then eluted and quantitated by charring afterseparation by thin layer chromatography on silica gel. The amounts (in nanogramslmg wet weight) of phosphatidylinositol 4,5-bisphosphate (PIPz),phosphatidylinositol 4-phosphate (PIP), and phosphatidylinositol (PI) were 20 f 1, 16 f 1, and 1790 f 140, respectively). Incubation of the cells with 100 nM vasopressin decreased the value for PIP2 to 10 & 0.2 at 15 s, 12 f 1.5 at 1 min, and 14 f 2.1 at 5 and 30 min. In contrast, the hormone increased 1,2-diacylglycerol plus phosphatidate by over 200 nglmg wet weight at 5 min under similar conditions (Bocckino, S. B., Blackmore, P. F., Wilson, P. B., and Exton, J. H. (1987) J. Biol. Chern. 262, 15309-15315). PIPz was also significantly decreased at 15 s by angiotensin I1 (100 nM), ATP (100 p ~ ) and , epinephrine (1p ~ ) .In contrast, PIP was not significantly changed, and PI was significantlydecreased (by -15%) at later times (15 and 30 min). The changes in phosphoinositide mass were well correlated with changes in labeled phosphoinositides in hepatocytes previously incubated with [3H]inositolfor 90 min. The amounts of inositol phospholipids in liver plasma membranes (in micrograms/mg protein) were2.1 f 0.2 for PIP2, 0.24 f 0.03 for PIP, and 23 f 4 for PI. Comparison of these values with those for whole cells suggests that PIP2is enrichedin the plasma membrane, whereas PIP is present elsewhere in the cell. The fatty acid composition of whole cell PIP2 showed significant differences fromthat of PI. The percentages of palmitic, stearic, linoleic, and arachidonic acids were, respectively, 14, 41, 10, and 25 for PIP2 and 10, 34,7 , and 37 for PI. Vasopressin treatment for 15 s did not alter the fatty acid composition of PIP2. The corresponding fatty acid percentages for liverplasma membranes were 13, 41, 11,and 21 forPIP2 and 8 , 3 4 , 0 , and 40 for PI. The fatty acid composition of PIP in whole cells and plasma membranes resembled that of PIP2. These data provide the first detailed chemical quantitation of inositol phospholipid changes in cells and indicate: (a)that vasopressin acts on a single pool of PIP2 in hepatocytes, which is presentpredominantly in theplasma membrane; ( 6 )that the fatty acids of PIP and PIP2are probably remodeled following synthesis

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Associate Investigator, Howard Hughes Medical Institute. § Investigator, Howard Hughes Medical Institute. To whom reprint requests should be addressed.

from PI; and ( c ) that PIP2 represents a quantitatively modest source of the 1,2-diacylglycerol and phosphatidate accumulating in hormone-stimulated hepatocytes, suggesting that othermechanisms, e.g. phosphatidylcholine hydrolysis, are more important.

Hormones, neurotransmitters, and otherligands which use the Ca2+ion as an intracellular messenger stimulate inositol lipid metabolism in their target cells. It is widely accepted that activation of the receptors for these agonists causes the breakdown of membrane PIP: to produce IP3, which increases intracellular Ca2+,and DAG, whichstimulates protein kinase C , thereby producing biological effects (1-3). In thevast majority of published studies, phosphoinositide metabolism has been investigated by analysis of the changes in radioactivity incorporated into these phospholipids from 32Piand/or [3H]inositol or by measuring alterations in the radioactivity of their products, e.g. inositol phosphates. However, there are some problems of interpretation associated with such use of isotopes to analyze metabolic changes. For example, in few cases was it established that theexperiments were carried out under conditions of isotopic equilibrium, i.e. when all phosphoinositide pools were labeled to equal and constant specific radioactivity. This is a particular problem for the interpretation of studies of polyphosphoinositide breakdown utilizing changes in 32Plabeling, since the label can be incorporated into these lipids by different routes, and most of the studies using this labeling have not distinguished between the individual phosphate groups in each lipid molecule. In other reports, [3H]inositol labeling has been favored over 32Plabeling since changes in specific activity and reincorporation of label are less of a problem. However, isotopic equilibrium is usually only slowly reached with this isotope, especially in isolated cells. Despite these potential or real problems in data interpretation, the assumption has been made in numerous studies that changes in the amount of radioactivity reflect changes in the total amount of these phospholipids and inositol phosphates. However, there is evidence in some cells of two different pools of phosphoinositides, one which is small and sensitive to hormone and a larger one which is insensitive to hormone (4-8). The major reason why chemical measurements of polyphosphoinositides have not been undertaken is that they represent less than 0.1% of the total phospholipids in most cells (9). In this manuscript, we report the changes in phosphoinositide mass in isolated hepatocytes induced by a number of agonists. We have used a modification of a method of The abbreviations used are: PIP2, phosphatidylinositol 4,5-bisphosphate; DAG,1,Z-diacylglycerol; PI, phosphatidylinositol; PIP, phosphatidylinositol 4-phosphate; IP3, inositol 1,4,5-triphosphate.

2574

Hormonal Changes in Phosphoinositide Concentrations Schacht (10) to separate polyphosphoinositides on neomycin columns and have quantitated them by charring after thin layer chromatography (11). Phosphoinositide levels in liver plasma membranes were also measured, and fatty acid analyses of these lipids and also those of whole cells were performed. EXPERIMENTALPROCEDURES

Materials-Angiotensin 11, (-)-epinephrine, butylated hydroxytoluene, o-nitrophenylhydrazine, fatty acids, l-ethyl-3-(3-dimethylaminopropyl)carbodiimide, [8-argininelvasopressin,ATP, neomycin sulfate, and phosphoinositides (bovine brain) were from Sigma. Collagenase type I was from Cooper Biomedical. Silica Gel 60 plates were from Merck. Methanol, acetonitrile, and chloroform were from Burdick and Jackson. Other reagents and organic solvents were from Fisher. ~-myo(2-~H]Inositol was purchased from American Radiolabeled Chemicals. Other radiochemicals werefromDu Pont-New England Nuclear. Porous reactive glass beads (Glycophase G/CPG200,200-400 mesh) were obtained from Pierce Chemical Co. Polyphosphoinositide Extraction from Hepatocytes-Hepatocytes were prepared from fed male rats (250-300 g) by perfusing their livers with Ca2+-freeKrebs-Henseleit bicarbonate buffer containing 0.3 mg/ ml collagenase (Type I, Worthington) as described previously (12). The digested livers were cut into small pieces, shaken vigorously in perfusion buffer (gassed with 95% 02,5% C02), and filtered through nylon mesh (12). The cells were washed 3 times in collagenase-free buffer containing 2.5 mM Ca2+and 15 mg/ml gelatin (12). In typical experiments, the hepatocyte suspensions (80-100mg/ml wet weight) were shaken and gassed with 95% 0 2 , 5% COZfor 15 min at 37 “C before addition of hormones. Samples (2.5 ml) were removed and the phosphoinositides were extracted with 7.5mlof ice-cold CHC13/ MeOH (1:2). Butylated hydroxytoluene (0.05%,w/v) was included in samples for fatty acid analysis. After addition of 2.4 N HC1 (2.5 ml) and chloroform (2.5 ml) and centrifugation (2000 X g for 10 min), the infranatant was removed and the supernatantwas reextracted twice with chloroform (5 ml). Extracts were pooled and washed twice with 7.5 ml of MeOH/1 N HCl (1:l) (13). The chloroform phase was dried under N, and the samples were resuspended in CHCldMeOH (1:l) before loading in neomycin columns. Aliquots were kept for PI determination (see below). Neomycin Columns-Neomycin sulfate was reductively coupled to reactive porous glass beads as described by Schacht (10). The neomycin-coated glass beads (1 ml)were packed in small columns, converted to therequired salt form, and equilibrated with the starting solvent by eluting in order with 3 column volumes each of 0.5 M ammonium formate in CHC13/MeOH/H20 (5:10:2), CHC13/MeOH/ HZ0 (5:10:2), and CHC13/MeOH (1:l). The lipid extracts in CHC13/ MeOH (1:l)(1 ml) were passed through the columns followed by 3 column volumes each of CHCI3/MeOH (1:l)and CHC13/MeOH (E!). The phospholipids, except PIP andPIP2, were eluted with 1 2 column volumes of 0.2 M ammonium formate in CHCI3/MeOH/H20 (5:10:2). PIP and PIP2were eluted together with 10 column volumes of 1.2 M ammonium formate in CHC13/MeOH/H20 (5:10:2) (14). The eluates containing PIP and PIP, were washed by adding 5 ml of 2.4 N HC1/ 10 mlof eluate. After centrifugation (2000 x g for 5 min) the resulting lower phase was removed and the supernatantwas reextracted twice with chloroform. Extracts were pooled and washed twice with the same volume of MeOH/1 N HCI (1:l) and dried under NP.The dried aliquots of the CHC13/MeOH extracts of the cells kept for PI (see above) and of the PIP and PIP2 eluted from the neomycin column were resuspended in CHCI3 and separated by thin layer chromatography on Silica Gel 60 plates (impregnated with 1%potassium oxalate) using a developing solution of CHCI3/acetone/MeOH/acetic acid/H20 (12:45:393624) (15). After development, the plates were dried, dipped in 10% CuS04in 8% H3P04(11)and charredby heating at 180 “C for 15 min. The phosphoinositides were localized by their comigration with [3H)phosphoinositidesfrom Du Pont-New England Nuclear and were quantitated using a LKB2202 Ultrascan densitometer using PI, PIP, andPIP2 as standards. These were prepared from mixed phosphoinositides by chromatography on neomycin columns as described above. PIP and PIP2were eluted separately with 0.4 M ammonium formate (PIP) and 1.2 M ammonium formate (PIP,) in CHCla/MeOH/H20 (5102). Their concentrations were estimated by Po4 determination (16). The recovery of [3H]PIP2from the cell extracts after elution from the neomycin column was 99 k 3 ( n = ll), 13H]PIPwas 97 t 8 ( n =

2575

lo), and [3H]PI from the cell extracts was 98 k 4 ( n = 11). Since these phosphoinositides were quantitated directly on the thin layer plates by comparison with standards using laser densitometry, no further recovery factor is involved.However, in the experiments employing my~-[~Hlinositol or in the analyses of fatty acid composition (see below), the phospholipids were extracted from the plates and the recoveries were 75 k 3 ( n = 8),76 k 5 ( n = 6), and 100 f 5 (n = 13) for PIP,, PIP, and PI, respectively. Fatty Acid Analysis-Polyphosphoinositides were purified by column chromatography on neomycin columns and thin layer chromatography as described above. Plates were dried under NP and lipids were located by exposure to iodine vapor. Located spots were scraped off the plates and thelipids extracted three timeswith 2 ml of CHC13/ MeOH/0.2 N HCI (1:2:0.8). 3 ml each of chloroform and water were added to form a two-phase system. The chloroform extract was dried under Nz, saponified and the fatty acids converted to 2-nitrophenylhydrazides by the method of Miwa et al. (17). The derivatized fatty acids were separated by high pressure liquid chromatography (17). Since it was possible that the iodine used to locate the phospholipids on the thin layer plates might have oxidized the polyunsaturated fatty acids, some analyses were performed using Coomassie Blue to identify the spots. However, the fatty acid analyses of PIP and PIP2 identified in this way were not significantly different from those in which iodine vapor was used (data not shown). Fatty acids were also analyzed by gas liquid chromatography (18) in some experiments, but the results were no different from those obtained using high pressure liquid chromatography. To eliminate a possible error arising from the selective extraction of certain fatty acid species of PIP and PIP2 from the silica gel of the thin layer plates, these phospholipids were also analyzed after elution from the neomycin columns, i.e. prior to application to the plates. However, no differences in fatty acid composition were noted before and after the thin layer chromatography (data not shown). Liver Plosrna Membranes-Rat liver plasma membranes were prepared by the method of Prpic etal. (191, and thelipids were extracted and analyzed as described above. The characteristics of these membranes (enzyme markers, hormone receptors) have been reported elsewhere (19-23). RESULTS

Phosphoinositide Content of Isolated Hepatocytes and Rat Liuer Plasma Membranes-Since polyphosphoinositides are present in cells in minor amounts (2, 3, 9), it is necessary to purify them after extraction. The antibiotic neomycin exhibits a specific affinity for PIP and PIPz compared with other phospholipids (lo), and this is the basis for their chromatographic separation from the other lipids. As described under “Experimental Procedures,” the recoveries for PI, PIP, and PIP, from rat liver extracts were 97-99% as assessed using standard [3H]phosphoinositides. Their quantitation by photodensitometric analysis of thin layer plates after charring (11)was linear from 1 to 10 Kg and was not different between PIP and PIPz. Theconcentrations of polyphosphoinositides extracted from isolated hepatocytes and liver plasma membranes are given in Table I. It was calculated from this table that PIP and PIP2 account for 0.9 and 1.1%, respectively, of the total inositol containing phospholipids in whole cells with a PIPz/ TABLEI Phosphoinositide content of isolated hepatocytesand ratliver plasma membranes Phosphoinositides from isolated hepatocytes and rat liver plasma membranes were extracted and analyzed as described under “Experimental Procedures.” Values are the mean k S.E. of 12-20 determinations. plasma liver Rat Isolated hepatocytes

PI PIP PIP2

membranes

ng/mg wet weight

p g / m g protein

1790 & 140 16 k 1.2 20 2 1

22.6 k 4 0.24 k 0.03 2.1 -e 0.2

2576

Phosphoinositide Hormonal Changes in Concentrations

PIP ratio of 1.25. The picture was different in rat liver plasma In order to demonstrate the possible presence of multiple membranes, where PIP and PIP, represented, respectively, pools of phosphoinositides in hepatocytes, cells were incu1.1 and 9.2% of the total inositol containing phospholipids bated with [3H]inositol for 90 min as described previously (24). Fig. 2 shows the time course of the effect of vasopressin with a ratio PIP2/PIP of 8.7. Effect ofVasopressin on Phosphoinositide Levels in Isolated ( M ) on [3H]inositol-labeledphosphoinositides in isolated Hepatocytes-Fig. 1 illustrates the time course of the effect of hepatocytes. The hormone induced a rapid andtransient vasopressin on the levels of phosphoinositides in isolated decrease of [3H]PIPz (lower panel),and the time course and hepatocytes. Vasopressin at a maximally effective concentra- the amplitude of the response werevery similar to those tion obtained when PIPz mass was measured. Labeled PIP (middle M, seeFig. 3) induced a rapid breakdown of PIP, (lower panel).This response was significant at 5 s (data not panel) decreased by 10% within 15 s, but returned to control shown) and was largest at 10-15 s (40-45% decrease). It values within 5 min. Vasopressin did not significantly affect declined gradually over 15 min, but was still present at 30 the labeling in PI over a period of 10 min, but decreased it min ( p < 0.005 for all time points except 15 min where p < 0.05). In control incubations without hormone, there were no significant changes in thephosphoinositide (data notshown). COMROL In contrast, no alterations in PIP were observed with vaso? 220000 / T pressin (middle panel).The hormone did not change the PI content during the first 10 min of incubation, but it induced small but significant ( p < 0.05) decreases (-15%) in the level of this lipid at 15 and 30 min. The initialrate of PIP, breakdown represents the netdisappearance of approximately 2.6% PIP2/s or 0.5 ng/s. mg wet weight.

I ’

I

1

A 160000

-

5000

-

4600 5 4400 -

-

E 01

1

1

22 . .

c ’

c

0

3

18-

a

I VWRESSIN

;s/

16CONTROL

2 12 C

E

3600 -

E 3800

V6OPRESSIN

F14--

-

‘ \

icn 2 0 -

.g

F 4200

g 4000 -

1

,CONTROL

2 4800

10

-

-6

-

5 2300

-

.- 2500

n

t

8-

1

2 2100 0 01

7 CONTROL

g” 1900 -

\ v

a

1700

-

1500

0

I

’

0

I 5

10

15

20

25

30

TIME min

FIG. 1. Time course for vasopressin-stimulated changes in phosphoinositides in isolated hepatocytes. Hepatocytes (90-110 mg/ml wet weight) were incubated with vasopressin M ) for the indicated times. Incubations were stopped in ice-cold CHC13/MeOH (1:2). Phosphoinositides were extracted and analyzed as described under “Experimental Procedures.” Results are the mean f S.E. of six different experiments performed in triplicate. The earliest time point shown after vasopressin addition is 15 s.

5

10

15

20

25

30

TIME rnin

FIG. 2. Time course for vasopressin-stimulated changes in [3H]inositol-labeledphosphoinositides in isolated hepatocytes. Rat hepatocytes (50 mg wet weight/ml) were prelabeled with 0.1 mM my0[2-~H]inositol(25 pCi/ml cell suspension) for 90 min, washed once, and resuspended in medium without my0-[2-~H]inositol.At the indicated times after the addition of vasopressin (lo” M), aliquots (0.5 ml) were removedinto tubes containing 1.5 ml of ice-cold CHCb/ MeOH (1:2). Phosphoinositides were extracted and separated as described under “Experimental Procedures.” The results shown are the mean 2 S.E. for two different experiments performed in triplicate.

in Phosphoinositide Concentrations

Hormonal Changes

after this time (upper panel). The effect was about 15% after 30 min. The results of Figs. 1 and 2 strongly suggest the existence of a single hormone-sensitive pool of PIP,. Dose Response for Vasopressin-Fig. 3 shows the concentration dependency of the effect of vasopressin on PIPzlevels in isolated hepatocytes measured after 15s. The half-maximal concentration of vasopressin was -lo-’ M and the response was maximalwith M. This dose responsecurve is similar to that obtained measuring [3H]IP3 formation (24). At this time point, PIP and PI levels were not affected at any concentration tested (data not shown). Effects of Various Agonists on Phosphoinositide Breakdown in Isolated Hepatocytes-The maximal effects of various agonists on PIPzlevels measured after 15 s are shown in Fig. 4. Among the various agonists tested, vasopressin was the most efficacious (45% decrease). The order of efficacity of these agonists was vasopressin > ATP > angiotensin I1 > epinephrine. When cells were challenged with maximally effective M ) plus angiotensin ( concentrations of vasopressin M), the decrease in PIPz was 72%.Achallengewith the

L

-cd’-lO

-9

-8

-7

A -6

log concentration FIG. 3. Dose response of vasopressin on PIPz levels in isolated hepatocytes. For experimental details see the legend to Fig. 1. The effect of various concentrations of vasopressin on PIP2 breakdown were analyzed 15 s after the addition of the hormone. The results presented are the mean f S.E. of 2 different experiments performed in triplicate.

2577

combination of these two hormonesplus ATP M ) and epinephrine (10 WM) (shown as “ALL” in the figure) did not further increase this effect. Under these conditions, the level of PIP was slightly decreased.Neither glucagon nor epidermal growth factor had any significant effect on PIP2 levels (21.7 & 2.3 and 19.3 5 2.6, respectively, versus 20.7 f 0.7 ng/mg wet weight ( n = 6)), and glucagon did not alter the effect of a maximal dose of vasopressin (12.0 f 1.8 versus 12.2 f 1.4 ng/mg wet weight, n = 5 ) ) .These resultssuggest that glucagon and epidermal growth factor do not cause a significantbreakdown of PIPZ. However, it should be kept in mind thateffects smaller than 5% would be hard to detect with the methods used. The levels of PIP and PI were also not affectedby glucagon or epidermal growth factor (data not shown). fatty Fatty Acid Composition of Phosphoinositides-The acid compositions of phosphoinositides from isolated hepatocytes (A) and ratliver plasma membranes (B) are presented in Table 11. The major fatty acids were arachidonic, linoleic, oleic, stearic, and palmitic acids. No other fatty acid represented more than 4% of the total (data not shown). The fatty acid composition of PI differed from that of PIP and PIPz, particularly with respect to arachidonic and palmitic acids. Arachidonic acid was present in PIP and PIPz to a much lesser extent (-20-25%) than in PI (37-40%). On the other hand, palmitic acid was present in PIP and PIP,(14-17%) to a greater extent than in PI(8-10%). The fatty acid compositions of the phosphoinositides identified on the thin layer plate using Coomassie Blue were not significantly different from those found when iodine vapor was used and when gas liquid chromatography was used instead of high pressure liquid chromatography to analyze the fatty acids (data not shown). In isolated hepatocytes or plasma membranes, no significant differences between the fatty acid compositions of PIP and PIPzwere observed. Interestingly, comparison of the fatty acid composition of PI in isolated cells (A) with that in the plasmamembranes (B) showed a marked differencewith respect to linoleic acid. This fatty acid was not detected in the membranePI, whereas it represented7% of the total fatty acids in PI from isolated cells. The great similarity of the fatty acid compositions of the whole cell and membrane PIP, supports theexistence of a single homogeneous pool of PIP2. TABLEI1 Fatty acid composition (mole %) of phosphoinositides from isolated hepatocytes (A) and rat liver plasma membrane (B) Phosphoinositides were extracted and separated as described in Table I. Fatty acid analysis was performed as described under “EXperimental Procedures.” Values presented are the mean & S.E. of determinations made in three(A) or seven (B) different experiments. PI

PIP

PIP,

% of total fatty acids

A. Whole cells C204

c,,:, C,,:, Cl80

C16” CON

VASO

ANGIO

ATP

€PI

ALL

FIG. 4. Effect of various agonists on PIPz levels in isolated see legend to Fig. 1. The hepatocytes. Forexperimentaldetails agonists were: vasopressin (VASO, M ) , angiotensin I1 (ANGZO, M ) , ATP M ) , and epinephrine (EPI, M). Control incubations were with saline designated CON and incubations with all the agonists together are designated ALL. PIP2 levels in hepatocytes were analyzed 15 s after the additionof the agonists. The results oresented remesent the mean f S.E. of two different exneriments performed in triplicate.

37.3 t 3.6“ 7.0 f 1.8 6.0 f 1.5 33.7 & 5.9 9.8 f 0.7“

22.0 k 1.5 9.4 t 1.0 4.3 t 0.6 38.5 t 5.0 13.9 f 0.2

25.0 f 3.5 10.3 f 0.8 3.6 f 0.7 40.5 f 4.0 14.0 f 2.3

40.0 f 4”

20.5 f 3.3 7.8 t 1.0 6.1 t 4.0 35.4 -C 2.6 16.8 f 2.6

23.2 f 2.8 9.4 f 1.8 4.6 f 0.2 39.6 & 3.0 13.5 f 1.8

B. Rat liver plasma membranes G O 4

CI,:, C,,, C18 0

C160

ND 4.1 f 1.0 34.0 f 2.0 8.2 t 0.4”

Significantly different ( p < 0.05) from correspondingvalue P I P or PIP,. I, ND, not detected.

in

2578

Hormonal Changes


The fatty acid composition of the phosphoinositides in isolated hepatocytes was also measured after 15 s of stimulaM). Fig. 5 shows a representative tion with vasopressin profile of the fatty acid analyses of PIP2. The effect of vasopressin on PIP2 content was reflected by the large decrease in the total fatty acid content of PIP2.All the fattyacids were decreased to about the same extent. The conclusion that the fatty acid composition of PIP2 was not significantly affected by vasopressin is borne out by the combined results from several incubations presented in Table I11 and when Coomassie Blue was used instead of iodine to identify the lipids. These results further suggest the existence of a single pool of PIP2. As expected, the fatty acid composition of PI or PIP was not changed after 15 s of stimulation with vasopressin (data notshown).

absence of sufficiently sensitive and accurate methods for their measurement. We have established conditions under which phosphoinositide mass can be measured relatively easily in both cells and plasma membranes. The method is rapid and more sensitive than those involving phosphate determination. Using this technique we have found concentrations of PIP2 and PIP of 20 f 1 and 16 f 1 ng/mg wet weight in isolated rat hepatocytes. These results arein agreement with the approximation of total polyphosphoinositide levels in rat liver made by Michell et al. (9). Based on the measurement of acid extractable inositol lipids, these workers reported a value of 20-40 ng/mg wet weight. However, our results disagree with the chemical measurements of phosphoinositides based on phosphate determinations which were published by Litosch et al. (25) for hepatocytes. They reported that PIP represented 10% of the total inositol lipids, cf. our value of DISCUSSION 0.9%.Furthermore, theirPIP values were almost %fold higher The purpose of this study was to obtain quantitative infor- than those of PIP2, cf. our value of 0.8. These authors demation about the changes in phosphoinositide metabolism scribed two unidentified labeled lipids that migrated closely occurring in response to various Ca2+mobilizing agonists in to PIP, and it is possible that these may have contributed to hepatocytes. Despite recent intense interest in the role of the high phosphate content measured in PIP. Direct comparpolyphosphoinositide breakdown in receptor-controlled gen- ison of phosphoinositide masses between the two studies is eration of second messengers, almost no direct measurements not possible due to the lack of experimental details in the of these lipids have been made in any tissue. This is because report of Litosch et al. (25). Comparison of the ratios between the phosphoinositide of their very low concentrations in most cells (1-3,9) and the levels found in our study with those of the radioactivities found in the liver phosphoinositides in rats injected with [3H] inositol 18-22 h previously (25-28) show similarities in both rat liver plasma membrane and isolated hepatocytes, except that in [3H]inositol-labeledhepatocytes [3H]PIPwas proportionally greater than in our study (26). These data suggest that, during this relatively long labeling period, isotopic nearequilibrium is achieved. Our phosphoinositide measurements on whole cells and liver plasma membranes are consistent with the idea that PIP, is enriched in plasma membranes, whereas PIP is present elsewhere in the cell. A higher PIP2 level of rat liver plasma membranes has also been suggested on the basis of labeling studies (27-39). As shown in Fig. 1, the large initial decrease in PIP, mass induced in hepatocytes by vasopressin was not accompanied by any significant change in PIP or PI, at least during the first few minutes. This establishes that the primary effect in these cells is stimulation of PIP2hydrolysis. The absence of modification of PIP levels can be explained if FIG. 5. Changes in fatty acid content of hepatocyte PIPz it is accepted that the PIP in the plasma membranes is only following vasopressin stimulation. Hepatocytes (90-110 mg/ml) a small fraction of the total cell PIP and turns over very wereincubatedfor 15 s inthe absence (-1 or presence (---) of rapidly. Studies of the production of IPS and other inositol vasopressin M ) . The reaction was stopped in ice-cold CHC13/ MeOH (1:2). The phosphoinositides were extracted, separated by thin phosphates induced by vasopressin (24, 30-32) indicate that layer chromatography, and their fatty acid content was analyzed as PIP, continues to break down beyond 15 s. Thus its failure described under“Experimental Procedures.”The profile is represent- to continue falling (Fig. 1) indicates that itis being resyntheative of four different experiments. sized from PI via PIP. This is generally assumed to be the case because of the high activities of the two kinases involved. TABLE111 It is likely that the plasma membrane PIP content declines Effect of vasopressin on fatty acid composition (mole %) of PIP, in in parallel with PIP2, but the decrease cannot be detected isolated hepatocytes against the large background of PIP in othercell components. Hepatocytes were incubated for 15 s with or without vasopressin the very large content of PI compared to PIP2 ex(lo-’ M) and the phosphoinositides extracted and separated for fatty Finally, plains why no changes in this lipid were observed during the acid analysis as described under “Experimental Procedures.”Results are the mean ? S.E. of determinations made in three different first minutes of stimulation. The decrease in PI mass at 30 min (-300 ng/mg wet weight) is in accord with the view that experiments. the agonist-induced loss in PI is due to its phosphorylation Control Vasopressin to form PIP and PIP2 (30-32). % of total fatty acids Interestingly, the changes in phosphoinositide mass were 23.0 k 4.0 C204 25.0 f 3.5 well correlated with changes in labeled phosphoinositides in CIS2 10.3 k 0.8 9.6 f 0.9 Cl&l 3.6 f 0.7 5 f 0.8 hepatocytes previously incubated with [3H]inositolfor 90 min 38.4 f 1.5 CIS0 40.5 f 4.0 (Figs. 1and 2) (25, 32,35). The small effect observed on [3H] Cl60 14.0 +. 2.3 15.6 f 2.3 PIP (Fig. 2) at 15 s in this study can again be explained if the

Hormonal Changes


plasma membrane PIP representsa small fraction of the total. Additionally, the phosphoinositides may not have reached isotopic equilibrium during the 90-min labeling period. Lack of this equilibrium will also explain the smaller effect of vasopressin observed on PIP, breakdown measured in most of the studies using [3H]inositol-labeled hepatocytes (25, 26, 30). Our results are similar in some aspects to those obtained with horse platelets where the decreases in 32P-lipids matched the changes in lipid mass (36). However, they are in opposition to the results of Litosch et al. (25) which showed that the levels of PIP, in isolated hepatocytes were significantly increased after 30 s of treatment with vasopressin. As discussed above, there were problems in the determination of PIP content in this study. Since cellular PI is believed to be in rapid metabolic equilibrium with PIP and PIP,, it is expected that these lipids would share thesame fatty acid profile. Although this appears to be the case in brain (37) and erythrocytes (38), it is not entirely true for hepatocytes and liver plasma membranes (Tables I1 and 111). The high level of arachidonic acid in PI in isolated rat hepatocytes and rat liver plasma membranes (Tables I1 and 111) and reported elsewhere (39,40) in rat liver contrasts with the results of Colbeau et al. (41), showing that in rat liver plasma membranes arachidonic acid represented only 8% of the total PI fatty acids. Since antioxidants were not included in this study, the difference can bedue to autooxidation. Another explanation relates to the different methods utilized for plasma membrane isolation. The method of Prpic et al. (19) used in the present study is more rapid than that used by Colbeau et al. (41) and yields membranes well enriched in 5”nucleotidase and hormone receptors (is23). Our results on PI fatty acid composition show for the first time that the composition is different between plasma membrane PI and hepatocyte PI, i.e. PI in other cellular membranes. Although it must be recognizedthat it is possible that changes in the fattyacid composition of the PI of the plasma membranes could have occurred during their isolation, this result suggests that the fatty acids of PI could be remodeled in theplasma membrane. Alternatively, an independent pathway of PI synthesis could exist in rat liver plasma membrane as reported by Imai et al. (42) inGHs cells. Another possibility is that theproteins which transport PI from the endoplasmic reticulum (43) where it is synthesized (44) may be specific for certain species of PI. The absence of linoleic acid in membrane PI and its presence in PIP, and PIP and thelower representation of arachidonic acid in PIP and PIP, compared with PI suggest that PIP, synthesis does not occur exclusively from the phosphorylation of membrane PI but also from PI present in other cellular membranes. This is compatible with the localization of PI kinase at other intracellular sites (45). However, the possibility that the fatty acids of PIP and PIP, are remodeled following synthesis of PI must also be considered. We report here for the first time that PIP, is not almost exclusively constituted of arachidonic and stearic acids as has always been assumed based on fatty acid analysis of membrane PI. This probably explains in part why the breakdown of PIP, induced by vasopressin in isolated hepatocytes results in the formation of other species ofDAG (18), although evidence is accumulating that the breakdown of phosphatidylcholine also occurs (46,47). One can speculate that perhaps only a few species of DAG generated from the breakdown of PIP:!and phosphatidylcholine are important inthe activation of the protein kinase C or are specific activatorsfor the

2579

different forms of protein kinase C described recently (see Ref. 48 for review). A major point of this study is the demonstration of the existence of a single pool of PIP, in hepatocytes that is hormone-sensitive. This conclusion is based on three observations. First, the [3H]inositol-labeling experiments gave identical results to those utilizing mass analysis. Second, the fatty acid composition of PIP2is identical in rat liver plasma membranes and whole hepatocytes. Third, the vasopressin treatment does not modify the fattyacid composition of PIP2, suggesting that the different species constituting the pool of PIP, are equally subject to hydrolysis. This pool of PIP2 appears to be located mainly in plasma membrane (27, 29, 45). However, it must be emphasized that these resultsshould not be extrapolated to other cell types until the relevant analyses have been performed. The existence of multiple pools of phosphoinositides is suggested by the work of Fain et al. (4) in blowfly salivary glands, and thecoexistence of two nonmixing inositol lipid pools within a single cell has been reported by Monaco et al. (5) in vasopressin-stimulated WRK2 cells, by Rebecchi et al. (8) in thyrotropin-releasing hormone-stimulated GH3 cells (8), by Rana et al. (6) in pancreatic islets and by Muller et al. (7) in erythrocytes. The determination of the changes in PIPz concentration in isolated hepatocytes permits calculations of the minimum levels of IP3 and DAG generated in response to vasopressin. The breakdown of PIP, after 15 s represents a decrease of 9 1). Although this is probably an ng/mg wetweight(Fig. underestimate because of resynthesis of PIP, from PI, it corresponds to theformation of 9 pmol of IP3/mg wet weight. Assuming a cell water content of 0.5 pl/mg wet weight, this . studies wouldyield an increase in IP3 of 18 p ~ However, utilizing prelabeling with [3H]inositol indicate lower values (24, 30), and the concentration of IP3 giving maximal intracellular Ca2+release in in vitro studies is in the range of 0.52 p~ (2, 49). The discrepancy is probably due to the very kinase rapid metabolism of IP3 by bothphosphataseand pathways. The breakdown of PIP, apparently beyond that needed to generate IP3 for Ca2+ mobilization could be related to the supply of DAG. Vasopressin has been shown to elicit a 2-3fold increase in DAG content during 6 min (18).This represents an accumulation of 260-400 ng of DAG/mg wet weight above control levels.According to our calculations and as concluded by Bocckino et al. (18),this increase is far greater than the breakdown of PIPz (maximum of 9 ng/mgwet weight). Although there is undoubtedly resynthesis of PIP,, this cannot explain the discrepancy since the PI level wasnot detectably decreased at 5 min (Fig. 1).Furthermore, it is likely that some further metabolism ofDAG occurred. For these reasons, the present findings support the view that DAG must be coming from another source, e.g. phosphatidylcholine breakdown. Acknowledgments-We express our gratitude to Dr. S. J. Taylor for help in thedevelopment of the assay of the polyphosphoinositides and the supply of phosphoinositide standards, to Jeff Smith and Annette Ross for expert technical assistance, and to Carolyn Sielbeck for secretarial assistance. REFERENCES 1. Exton, J. H. (1988) FASEB J. 2, 2670-2676 2. Berridge, M. J. (1987) Annu. Rev. Biochm. 56, 159-193 3. Downes, C. P., and Michell, R. H. (1985) in Molecular Mechanism of MembraneTransignulling (Cohen, P., and Houslay, M. D., eds) pp. 3-56, Elsevier Science Publishers, New York 4. Fain, J. N., and Berridge, M. J. (1979) Biochem. J. 180,655-661

2580

Hormonal Changes


5. Monaco, M. E., and Woods, D. (1983) J. Biol. Chem. 258,1512515129 6. Rana, R. S., Kowluru, A., and McDonald, M. J. (1986) Arch. Biochem. Biophys. 245,411-416 7. Muller, E., Hegewald, H., Jaroszewicz, K., Cumme, G.A., Hoppe, H., and Frunder, H.(1986) Biochem. J. 235, 775-783 8. Rebecchi, M. J., and Gershengorn, M. C. (1983) Biochem. J . 216, 299-308 9. Michell, R. H., Hawthorne, J. N., Coleman, R., and Karnovsky, M. L. (1970) Biochim. Biophys. Acta 210,86-91 10. Schacht, J. (1978) J. Lipid Res. 1 9 , 1063-1067 32. 11. Baron, C.B., Cunningham, M., Strauss, J . F., 111, and Coburn, R. F. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,6899-6903 12. Blackmore, p. F., and Exton, J. H. (1985) Methods Enzymol. 109,550-558 13. Schacht, J . (1981) Methods Enzymol. 72,626-631 14. Palmer, F. B. St. C. (1981) J. Lipid Res. 22, 1296-1300 15. Jolles, J., Zwiers, H., Dekker, A., WirtZ, K. W. A,, and Gispen, W. H. (1981) Biochem. J. 1 9 4 , 283-291 16. Baginski, E, s.,F ~ p.~p.,, and Zak, B. (1966) clin, chim,~~t~ 15,155-158 17. ~ i H., yamamoto, ~ ~ M., , and Nishida, T, (1986) clin, chim, Acta 155,95-102 18. Bocckino, S. B., Blackmore, P. F., and Exton, J. H. (1985) J. Bid. Chem. 260,14201-14207 19. PrpiC, V., Green, K. C., Blackmore, P. F., and Exton,J. H. (1984) J. Biol. Chem. 259, 1382-1385 20. Lynch, c. J.9 Charest, R.1 Blackmore, p. F., and Exton, J. H. (1985) J . Biol. Chem. 260, 1593-1600 21. c. J., Charest, R., s. J. H., and Blackmore, P. F. (1985) J. Bid. Chem. 2 6 0 , 2844-2851 7971 22. Lynch, c . J.1 black more^ p. F Y Charest, R.7 and Extony J. H. (1985) Mol. Pharmacol. 2 8 , 93-99 23. Lynch, c. J.2 Prpi;, v.9 Blackmore, p. F.9 and Exton, J. H. (1986) Mol. Pharmacol. 3443 29, 196-203 24. Charest, R., PrpiC, v., Exton, J. H., and Blackmore, p. F. (1985) Biochem. J. 2 2 7 , 79-90 25. Litosch, I., Lin, S. H., and Fain, J. N. (1983) J. Bid. Chem. 258, 13727-13732 26. Rhodes, D., PrpiC, V., Exton, J. H., and Blackmore, P. F. (1983) J.Chem. Biol. 258,2770-2773 B.l

Extony

27. Taylor, D., Uhing, R. J., Blackmore, P. F., PrpiC, V., and Exton, J. H. (1985) J . Biol. Chem. 2 6 0 , 2011-2014 28. Uhing, R. J., Prpil, V., Jiang, H., and Exton, J. H. (1986) J. Biol. Chem. 261,2140-2146 29. Seyfred, M. A., and Wells, W. W. (1984) J . Biol. Chem. 259, 7659-7665 30. Thomas, A. P., Alexander, J., and Williamson, J. R. (1984) J . Bid. Chem. 259, 5574-5584 31. Hansen, c. A., Mah, s., and Williamson, J. R. (1986) J. Bid. Chem. 261,8100-8103 Palmer, S., Hawkins, P. T., Michell, R. H., and Kirk, C. J. (1986) Biochem. J .2 3 8 , 491-499 33. Berridge, M. J. (1983) Biochem. J. 212,849-858 34. C r e k J. A.9 Down'% c. p., Hawkins, p. T.9 Brewster, G., Michell, R. H., and Kirk, C. J. (1983) Biochem. J. 2 1 2 , 733-747 35. Thomas, A. P., Marks, J. S., Coll, K. E., and Williamson, J. R. (1983) J . Biol. Chem. 258, 5716-5725 36. M. M., and Lapetina, E. G. (1982) J. Bioi. C k m . 2 5 7 9 12705-12708 37. Holub, B. J., Kuksis, A., and Thompson, w . (1970) J . Lipid Res. 11,558-564 38. Allan, D., and Cockcroft, s. (1983) Biochem. J. 2 1 3 9 555-557 39. Holub, B. J., and Kuksis, A. (1971) J. Lipid Res. 12,669-705 40. Patton, G. M., Fasulo, J. M., and Robins, S. J. (1982) J . Lipid Res. 23, 190-196 41. Colbeau, A., Nachbaur, J., and Vignais, P. M. (1971) Biochim. Biophys. Acta 249,462-492 42. Imai, A., and Gershengorn, M. C. (1987) Nature 3 2 5 , 726-728 43, Lumb, R. H., Kloosterman, A. D., Wirtz, K.W. A., and Van Deenen, L. L. M. (1976) Eur. J. Biochem. 69, 15-22 44. Jelsema, C. L., and Morre, D. J. (1978)J. Biol. Chem. 253,796045. Cockcroft, S., Taylor, J. A., and Judah, J. D. (1985) Biochim. Biophys. Acta 845, 163-170 46. Irving, H. R., and Exton, J. H. (1987) J . Biol. Chem. 2 6 2 , 344047. Bocckino, S. B., Blackmore, P. F., Wilson, P. B., and Exton, J. H. (1987) J. Biol. Chem. 262,15309-15315 48. Ono, Y., and Kikkawa, U. (1987) Trends Biochem. Sci. 12, 421423 49. Joseph, S. K., Thomas, A. P., Williams, R. J., Irvine, R. F., and Williamson, J. R. (1984) J. Bid. Chem. 259, 3077-3081