Exp. Eye Res. (1990) 51, 167-176

Phosphoinositide HONG-GWAN aDivision

CHOE”“,

of Neuroscience

(Received

Metabolism ABDALLAH

in Frog Rod Outer

J. GHALAYINI

AND

ROBERT

Segments E.ANDERSON”

and Cullen Eye institute, Baylor College of Medicine, Houston, TX 77030, U.S.A.

9 August

1989 and accepted in revised form 13 December

One Baylor Plaza,

1990)

Previous studies have shown that vertebrate rod outer segments (ROS) have a light activated phospholipase C which hydrolyzes phosphatidylmositol-4,5-bisphosphate (PIP,). Three different experimental approaches have been used to test the hypothesis that the phosphatidylinositol (PI) biosynthetic cycle is present in ROSand that PIP, can be regenerated from DGindependentof rod inner segments.In the first study, enzyme activities of the PI cycle were assayedsimultaneously in the presence of CTP, myo-inositol and [y-3ZP]ATP using endogenous lipids as substrates. Under these conditions, broken (leaky) ROS prepared by continuous sucrose gradient centrifugation showed PI, PIP and DG kinase activities similar to those found in intact ROS and non-ROS membranes, whereas PI synthetase activity was much lower in the leaky ROSthan in the other two fractions. The relative distribution of PI synthetase specific activity in the three membrane preparations was similar to that of the microsomal enzyme marker cytochrome c reductase. ROSprepared by discontinuous sucrose gradient centrifugation showed only 2-3 Y0of whole homogenate PI synthetase or phosphatidyl : cytidyl transferase activities. and the distribution of activities was the same as for microsomal and mitochondrial marker enzymes. In the second study, whole retinas were incubated with myo-[2-3H]inositol or [2-aH]glycerol in vitro, and the time course of incorporation of radioactivity into PI and other phospholipids was determined for ROSand three other retinal fractions. Over a lo-hr period, the rate of incorporation of myo-[2-3H]inositol or [2-3H]glycerol into PI in ROSwas lowest among the various retinal fractions. In the third study, chemical analysis of the molecular speciescomposition of PI, DG and phosphatidic acid (PA) from ROSshows that PA is substantially different from PI and DG,the latter two being quite similar. Theseresults are consistent with a precursor-product relationship between PI and DG, but not with the conversion of DG to PA or of PA to PI. Taken together, these three studies indicate that ROS do not have PI synthetase or phosphatidyl : cytidyl transferase activities. but do have DG, PI and PIP kinase activities. Thus, the PI in ROSlost through rapid turnover must be replaced with molecules derived from de novo synthesis in the inner segment of the photoreceptor cell. Key words: rod outer segments; phosphatidylinositol : phosphatidic acid ; lipid metabolism.

1. Introduction

Rod outer segments (ROS)are specializedorganelles of rod photoreceptor cells that capture light and initiate visual excitation. The visual pigment rhodopsin is an integral component of disc membranes. Following photobleaching, rhodopsin undergoes a series of conformational changes that leads to the activation of cGMP phosphodiesterase (Fung, 1987) which is important in visual transduction. Another activity increased by light is phospholipase C, which hydrolyzes phosphatidylinositol-4,5bisphosphate (PIP,) (Ghalayini and Anderson, 1984; Hayashi and Amakawa, 1985; Millar and Hawthorne, 1985; Millar et al., 1988). The hydrolysis products, 1,2-diacylglycerol (DG) and 1,4,5inositol trisphosphate, have been shown to have second messenger functions in other cells (Abdel-Latif, 1986), although their role in vertebrate photoreceptors is not known (Anderson and Brown, 1989). PIP, is the product of sequential phosphorylations of phosphatidylinositol (PI) by PI and phosphatidyli* For correspondence at: Departmentof Ophthalmology,Baylor Collegeof Medicine,OneBaylor Plaza,Houston,TX 77030, U.S.A. 00144835/90/080167+

10 %03.00/O

nositol-4-phosphate (PIP) kinases. PI is a minor component in ROS membranes found at the level of l-2 % of total phospholipids, and the polyphosphoinositides are even less plentiful (Fliesler and Anderson, 1983). Since ROS contain such small amounts of PI, the rapid turnover of PIP, could lead to the depletion of PI unless there is a replenishment pathway. A rapid turnover of PI in ROSwas first observed by Hall, Basinger and Bok (1973). They demonstrated that the specific radioactivity of PI in frog ROS increased for the first 5 days, then dropped rapidly, while that of other phospholipids increased slowly until 34 weeks after the injection of 33P0,. When [2-3H]glycerol was used as a precursor, similar results were obtained, suggesting a more rapid turnover rate of PI than the other phospholipids in ROS (Anderson, Maude and Kelleher, 1980b; Wetzel and O’Brien, 1986). Schmidt (1983a,b,c) showed that the synthesis of PI and phosphatidic acid (PA) were enhanced during the incubation of rat retinas with [2-3H] glycerol or 32P0, in light. ROS isolated from the retinas by microdissection indicated that lightenhanced labeling of PI occurred in ROS.However, it is not clear whether this stimulation of PI labeling is 0 1990 Academic Press Limited

H.-G. CHOE

168

due to the increase of de novo synthesis of PI in ROS or is a result of rapid translocation of newly synthesized PI from the inner segments. The existence of DG, PI and PIP kinase activities has been reported in vertebrate ROS(Seyfred et al., 1984 ; Giusto and Ilincheta de Boschero, 1986). However, activities of the two key enzymes of PI synthesis, phosphatidyl : cytidyl transferase and PI synthetase, have not been demonstrated in ROS. Thus, it is not known if PI can be synthesized do novo in ROS or is translocated from the inner segments as are the other phospholipids. In this study, three different experimental approaches were used to test the hypothesis that all enzymes of the PI biosynthetic cycle are present in ROS. Phosphatidyl : cytidyl transferase and PI synthetase activities were assessedin frog retinal ROSprepared by two different procedures. In addition, the incorporation rates of myo-[2-3H]inositol and [2-3H]glycerol into PI in various retinal fractions including ROS were measured. Finally, the molecular speciescomposition of DG, PI and PA from ROS was compared. The results of these three types of experiments indicate that PI synthesis does not occur in ROS. 2. Materials and Methods Preparation of Retinal Fractions Continuous sucrose gradient centr$ugation. Frog ROS

were prepared using the method of Zimmerman and Godchuax (1982) with a slight modification. Darkadapted frogs were decapitated under dii red light and the retinas were dissectedin Ringer’s-bicarbonatepyruvate buffer (RBP) containing 120 mM NaCl, 2 mM KCl, 2 mM MgCl,, 2 mM CaCl,, 25 mM NaHCO, and 1 mM pyruvate (pH 7.4). After removal of the pigment epithelium, about 14 retinas were placed in a test tube containing 2-3 ml of 70% (w/v) sucrose solution in 5 mM Tris-HCl (pH 7.5) containing 1 mM dithiothreitol (D’JT) and 0.5 mM EGTA. ROS were detached by gentle shaking of the sucrose solution for l-2 min. The suspension of ROS was transferred to the bottom of a continuous sucrose gradient (2 S-50 %, w/v) and centrifuged at 25 000 rpm for 1.5 hr in a Sorvall AH-627 (Swinging Bucket) rotor. This procedure gave three distinct bands which are labeled as follows (Fig. l), from the top to the bottom: band I (broken ROS),band II (intact ROS),and band III (nonROS membranes). The bands were harvested individually, mixed with 20 ml of 25% sucrose, and centrifuged at 15 000 rpm for 30 min in a Sorvall SS 34 rotor. The pellets were resuspended in 5 mM TrisHCl (pH 7.5) containing 1 mM DTT and 0.5 mM EGTA. sucrose gradient centri,fugation Discontinuous (Papermaster, 1982). Frog retinas dissected in RBP

were homogenized briefly in 1.17 g ml-l sucrose solution (3 retinas ml-l) with a glass homogenizer and

ET AL.

Contllous Sucross Gradimt 25% Band1 @rdm ROS) Band1

(htact ROS) Band1 (Nell-Ros

FIG. 1.

msmbratlss)

Diagram showing retinal fractions used in the

study. separated into four retinal fractions as described (Choe and Anderson, 1990). Enzyme Assays

Endogenous lipid phosphorylation assay. The assay mixture contained 20 mu Tris-HCl (pH 75), 10 mM magnesium acetate, 100 PM [y-32P]ATP(10 ,Li), and membrane preparation containing 70-90 ,ug protein in a f?nal volume of 50 yl. In some experiments, 32Plabeling in CDP-DGand PI was enhanced by adding 100 ,UM CTP and 20 ,UM myo-inositol to the assay mixture. Incubation was carried out at room temperature for 10 min and terminated by the addition of HCl 0.5 ml chloroform-methanol-concentrated (100 : 100 : 6, v/v/v). The phospholipids were extracted and separated by TLC as described by Ghalayini and Anderson (1984) with slight modification. PI synthetase . The enzyme activity was measured by the procedures of Benjamins and Agranoff (1969) and Wallace and Fain (1985), with modification. The reaction mixture contained 20 mM Tris-HCl (pH 8-O), 1 mM EGTA, 3 mM MgCl,, 1 mM myo-[2-3H]inositol (1 ,Li), 100 ,UM didecanoyl-CDP-DG, and lo-20 ,ug protein in a total volume of 120 ~1. Incubation was carried out for 30 min at room temperature and terminated by the addition of 1 ml chloroformmethanol-concentrated HCl(lO0 : 100 : 6, by volume). The phospholipids were extracted as described above and the lipid residues were dried in a vial under a stream of nitrogen. Ten millilitres of ACS was added to the vial and the radioactivity was determined. The difference in values between sampleswith and without exogenous CDP-DG was used to calculate PI synthetase activity. PhosphatidyI:cytidyl transferuse. The incubation medium contained 20 mM Tris-HCl (pH 8*0), 5 mM

PI METABOLISM

IN FROG

169

ROS

MgCl,, 1 mM [~z-~~P]CTP (4 &i), 0.5 mM dioleoyl-PA and 5-10 pug protein. After 10 min incubation at room temperature, the reaction was terminated by adding 1 ml chloroform-methanol-concentrated HCl (100 : 100 : 6, by volume). The amount of 32Plabel in the lipid residues was determined as described above. This method is a modification of the procedures described by Bishop and Strickland (19 75) and Wallace and Fain (1985). NADPH-cytochromec reductase.The enzyme activity was measured by the method of Omura and Takesue (1970), with a modification. The reaction mixture contained 50 mM potassium phosphate buffer (pH 7.5), 0.1 mM cytochrome c, 02 mu NADPH,, and 1 mM KCN in a final volume of 1 ml. The reaction was started by the addition of 20 ,ug protein, and the reduction of cytochrome c was followed by measuring absorbance at 550 nm in a Varian DMS 100s spectrophotometer at room temperature. Cytochrome c oxidase. The activity was measured spectrophotometrically by following the oxidation of reduced cytochrome c at 550 nm. Reduced cytochrome c was generated by the addition of a freshly prepared solution of sodium hydrosulfite to a cytochrome c solution (Cooperstein and Lazarow, 1951). The reaction mixture contained 20 mM potassium phosphate (pH 7.5), 40 ,UM reduced cytochrome c, 0.5 mu FDTA, and 10 pg protein as the enzyme source in a final volume of 1 ml. incubation Proceduresand Lipid Separation myo-[2-3H]lnositol incorporation into retinal fractions. Isolated retinas were incubated in RBP (30 retinas per 15 ml) containing myo-[2-3H]inositol (70 ,&i ml-l). The incubation medium was under a continuous stream of 9 5 % O,-5 % CO, and kept in the dark. At each time point, five retinas were removed, washed with ice-chilled medium without radioactive myoinositol, and homogenized immediately in 1.17 g ml-’ sucrose solution for the discontinuous sucrosegradient centrifugation described above. Following centrifugation, fractions were collected in polypropylene tubes, the sucrose solution was diluted at least two ties with 20 mM Tris-HCl, pH 8.0, and membranes were harvested by centrifugation at 5000 rpm for 1 hr in a Sorval SS-34 rotor in the presence of 5% trichloroacetic acid. After removing the supernatant, the pellet was washed three times with ice cold distilled water to remove the trichloroacetic acid. Lipids were extracted by the method of Ghalayini and Anderson (1984). In vitro pulse chase labeling of retinas with [2-3Hj glycerol. Retinas were incubated in RBP (30 retinas

per 5 ml) containing [2-3H]glycerol (200 ,uCiml-l) for 30 min. Following the pulse labeling, retinas were

washed two times with RBP containing 1 mM unlabeled glycerol and incubated further to chase the label. At each time point, five retinas were removed from the incubation medium and homogenized in 1.17 g ml-’ sucrose solution for the discontinuous sucrose gradient centrifugation. Procedures used for the lipid extraction of the fractions were the same as those described above. Lipid separation. Two-dimensional, three-step TLC was used to separate the lipid extract into classes as described (Choe and Anderson, 1990). Lipid spots were located by exposing the plate to iodine vapor and were scraped for phosphorus assay (Rouser, Fleischer and Yamamoto, 1970). An aliquot of the assay solution was counted for radioactivity. Molecular SpeciesAnalysis

ROSlipids were separated into molecular speciesby reverse-phaseHPLC as described (Choe, Wiegand and Anderson, 1989). 3. Results Phosphorylation of Endogenous Lipids

The effects of CTP and myo-inositol on 32P incorporation into endogenous lipids are shown in Fig. 2. 32P was incorporated into four endogenous phospholipids, PA, PIP, PIP, and CDP-DG,following the incubation of band II (intact ROS)with [Y-~~P]ATP in the absenceof CTP and myo-inositol. No significant amount of label was found in PI, although further metabolism of 32P-labeled PA and CDP-DG could potentially give rise to labeled PI. In the presence of CTP, there was about 2*5-fold increase in 32P-labeling of CDP-DG,while the addition of myo-inositol alone did not affect the degree of 32P incorporation into any other lipid. However, the amounts of 32Plabel in CDPDG and PI were significantly increased when both CTP and myo-inositol were present in the incubation mixture. Figure 3 shows an autoradiogram representing the radioactivities of 32Pin the phospholipids from three diierent retinal fractions prepared by continuous sucrose gradient centrifugation and incubated in the presence of [yJ2P]ATP, CTP and myo-inositol. Both bands I (broken ROS) and II (intact ROS) showed a similar extent of 32P label in PA, PIP and PIP, compared to band III (non-ROS membranes). However, the radioactivity of PI from band I was very low compared to band II and band III. PI Synthetase, Phosphatidyl: Cytidyl Transferase and the Marker Enzyme Activities in Retinal Fractions

The specific activities of PI synthetase, cytochrome c reductase (microsomal marker), and cytochrome c

H.-G. CHOE

170

ET AL.

PA

COP-DG

PIP PIP*

Origin l

Basal (-

+CTP (...........)

1

+Inositol ( --m-v---

- 1

+CTP and I nositol ( -----

Intensity

)

FIG. 2. TLC autoradiogram of phospholipids showing the effects of CTP, myo-inositol, and CTP plus myo-inositol on endogenous lipid phosphorylation with [y-32P]ATP. Band II (Intact ROS)was incubated with 100 ,LAM [yJ*P]ATP (10 ,&i) for 10 min. Lipid extracts of the incubation mixtures were separated into classes by TLC in a solvent system of chloroform-methanol-H,0/-40% methyl amine (60 : 36 : 5 : 5, by volume) and plates were exposedto X-ray film overnight. The concentrations of CTP and myo-inositol were 100 ,uM and 20 ELM,respectively. The intensities of the bands of the autoradiograms were measured by MegaVision 1024XM image analysis system and converted to the traces shown in the figure.

PA

PI

COP-DG

PIP

PIP2

,.

Origin

l

Band I (-

1

Bond II (-----)

Bond III (-----)

Intensity

FIG. 3. TLC autoradiogram of phospholipids extracted from band I (broken ROS), band II (intact ROS), and band III (nonROS), after incubation with [yJzP]ATP, CTP and myo-inositol as described under Materials and Methods. Lipid extracts were separated into classesby TLC as described in the legend to Fig. 2.

oxidase (mitochondrial marker) in bands I, II and III (Fig. 1) are shown in Table I. PI synthetase activity was determined with exogenous didecanoyl-CDP-DG as a substrate by following the incorporation of myo[2-3H]inositol into PI. Band III showed the highest PI

synthetase specific activity, while the activity found in band I (broken ROS) was minimal. The extent of PI synthetase specific activity paralleled the specific activities of microsomal (cytochrome c reductase) and mitochondrial (cytochrome c oxidase) marker

PI METABOLISM

IN FROG

171

ROS

TABLE I

PI synthetase and marker enzyme activities in fractions prepared by continuous sucrose gradient centrijugation Cytochrome c Reductaset (nmol min-’ mg-‘)

PI Synthetase* (pm01min-l mg-‘)

Fraction Band I (Broken ROS) Band II (Intact ROS) Band III (Non-ROS)

0.3 (4.1%) 1.8 (24.3 %) 74 (100%)

11.1+0.9 (3.7%) 82.95 14.0 (27.4%) 302.4*27,9 (100%)

Cytochrome c Oxidaset (nmol min-l mg-I) 0.7 (0.7%) 28.0 (28.5 %) 98.1 (100%)

The data in parentheses are % of specific activities in band III. * The values are the means + S.D.from one experiment performed in triplicate. Two additional experiments gave similar results. t The values represent the average of two independent experiments performed in duplicate.

TABLE II

Distribution of PI synthetase, phosphatidyl :cytidyl transjerase and marker enzyme activities in retinal homogenates. Whole retinal homogenates were fractionated by discontinuous sucrosegradient centrijugation, and enzyme assays were carried out as described under Materials and Methods. Typical total activities of the enzymes in 32.1 mg of whole homogenate are as follows: PI synthetase, 4282.1 pmol min-‘; phosphatidyl :cytidyl transjerase, 507.2 pmol min-’ : cytochrome c reductase,808.9 nmol min-’ ; cytochromec oxidase, 1094.6 nmol min-‘. % of total in whole homogenate Phosphatidyl :

Fraction Whole Homogenate

PI Synthetase

Protein*

100.0

100.0

1.11/1.13 1.13/1.15

lO.S_+l*O 8.1 k4.1

Interband Pellet

34.4 +4.7 38.8 f 8.4

Values are expressed as the means f

S.D.

2.3 f 1.5 20.0+ 13.9 12.3 + 2.4 39.7513.3

cytidyl TransferaseS 100.0

1.4 14.4

11.4 51.5

Cytochrome c Reductase*

Cytochrome c Oxidaset

100.0

100.0

1.5 f0.7 7.1 f2.3

53.4 f 6.9 38.1 f 11.9

2.3 + 0.6 10.4 + 5.2

17.3 * 3.0 62.5k25.0

of % values from four*, threet. or two+ separate fractionation experiments assayed in duplicate.

enzymes. Table II shows the distribution of PI synthetase and phosphatidyl : cytidyl transferase activities determined with exogenous substrates. In this study, retinas were homogenized and separated into four membrane fractions by discontinuous sucrose gradient centrifugation (see fig. 1 in Choe and Anderson, 1990). The marker enzyme activities observed in the 1.11/1.13 g ml-’ interface fraction (ROS) were minimal, whereas significant amounts of those marker enzyme activities were found in other retinal fractions. Compared to the whole retinal homogenate, only about 2-3X of PI synthetase and l-2 % of phosphatidyl : cytidyl transferase activities were recovered in 1*11/1*13 g ml-l interface fraction (ROS), which represents lo-11 % of the whole homogenate protein. The 1.13/1*15 g ml-l interface fraction was enriched with PI synthetase and the highest microsomal marker enzyme activity was observed in the interband fraction. The mitochondrial marker enzyme was very active in the pellet fraction.

In Vitro incorporation of myo-[2-3H]lnositol into Retinal Fractions

Retinas were incubated in vitro with myo-[2-3H] inositol for various times and then were fractionated by discontinuous sucrose gradient centrifugation. Figure 4 shows the time-course of incorporation of the label into PI. As indicated, there was a time-dependent increase in the specific radioactivity of PI over the incubation period in all membrane fractions examined. However, only a trace amount of radioactivity was present in ROS PI after 2 hr of incubation, and it increased only slightly over the next 6 hr. Even after 10 hr of incubation, the specific radioactivity of ROS PI was still much less than that of PI in other membrane fractions. The highest rate of incorporation of the labeled inositol into PI was observed in the pellet fraction followed by the 1.13/1*15 g ml-l fraction. In Vitro Pulse Chase Labeling of Retinas with [2-3H]GlyceroZ

Figure 5 shows the specific radioactivities of PI, phosphatidylethanolamine (PE) and phosphatidyl-

H.-G. CHOE

b-

60-

3 3

50. 50.

2.5

+

specific radioactivities of ROSphospholipids increased steadily, while those of other fractions, in general, reached maximal specific radioactivities after 9-S hr of chase incubation. Moreover, the spectic radioactivity of PI from all fractions was about ten times higher than that of PE or PC, indicating a more active

1~11/1~13 l~ll/l~l3

---vs.-- l./3/l.l5 A A Interband --[3--

ET AL.

Pellet

synthesis of PI in every fraction compared to that of

the other phospholipids.

-!

12

Molecular Speciesof PA, PI and DG in ROS

Time (hr)

PA, PI and DG of ROSwere isolated and converted to their corresponding 1,2-DGAC. The molecular species composition of 1,2-DGAC was determined by HPLC (Fig. 6). The HPLC profiles of PI and DG show that they have similar molecular speciesdistribution, which is quite diierent from that of PA. The two major molecular speciesfound in PI and DG were 16 : O-22 : 6 and 18 : O-22 : 6. Only very small amounts of 22: 6-22: 6 molecular species were found in these

4. myo-[2JH]Inositol incorporation into retinal fractions prepared by discontinuous sucrose gradient centrifugation. Thirty retinas were incubated with 15 ml RBP containing 1 mCi mgo-[2-3H]inositol. At each time point, five retinas were removed from the incubation mixture, homogenated, and fractionated by discontinuous sucrose gradient centrifugation. Lipids were extracted and separatedinto classesby TLC, as described in Materials and Methods. FIG.

choline (PC) from the four retinal fractions during a 9.5-hr chase period following 30 min pulse labeling with [2-3H]glycerol. After a 30-min labeling period, ROSshowed very low specific radioactivities of PI, PE and PC compared to other fractions. However, the

7oor------

!

600

PI

lipids. However,

PA from ROS showed significant

amounts of 22 : 6-22 : 6 and other dipolyunsaturated species, which are mostly 22:5-22:6, 22:4-22:6 and 20: 4-22 : 6. These latter specieseluted between 22~6-2216 and 18:1-22:6.

PE 50.

0

0

0 4.

rCC____--------0 /’

f--l0

2

4

6

8

IO

0

-1 2

4

,. 6

I 8

IO

0

2

4

6

8

IO

Time (hr) + ----A

A

1.11/1~13 ----

1.13/1.15

--•-

,“+erb(j”d -

Pellet

FIG. 5. In vitro pulse chase labeling of retinas with [2-aH]glycerol. Following 30 min of incubation, the label was chased with RBP buffer containing 1 mM unlabeled glycerol. The retinal fractions were prepared by discontinuous sucrose gradient centrifugation.

PI METABOLISM

IN FROG

ROS

173 CTP

PA \/

my0 - lnositol

69 Y

DG,

ATP PIP

ATP

FIG. 6. Separation of 1,2-DGAC molecular species from

ROSDC, PI and PA by reverse-phaseHPLC. Lipids extracted from ROSwere resolved into classesby TLC. Each lipid class was converted to 1.2-DGAC by acetolysis. The 1,2-DGACs were resolved into molecular speciesby HPLC, as described in Materials and Methods. The molecular species were detected by measuring absorbance at 210 nm. The fatty acids comprising the molecular species of each peak were identified by GLC. Only the peaks composed of a single molecular specieswere designated in the figure. Fatty acids are denoted as number of carbons : number of double bonds.

FIG. 7. PI cycle. Enzymes involved are as follows: 1, PI kinase; 2, PIP kinase; 3. phospholipase C; 4, DG kinase: 5, phosphatidyl : cytidyl transferase; 6, PI synthetase.

et al., 1980b), compared to 18-23 days for the other phospholipids (Anderson et al., 1980a,c). A physiological role for the rapid PI turnover and replacement was not apparent until the demonstration of lightstimulated PIP, hydrolysis in vertebrate ROS (Ghalayini

4. Discussion When newly synthesized radioactive proteins are translocated from the inner segments and are incorporated into ROS, they form a distinctive band at the baseof the ROSwhen observedby autoradiography. With time, this band migrates apically as ROStips are shed and new membranes are added at the base (Young, 1967). However, newly synthesized lipids are distributed in ROScompletely differently from integral proteins. Phospholipids seem to move freely between discs (Bibb and Young, 1974) and turn over faster than would be predicted if they were renewed only through disc shedding (Anderson, Kelleher and Maude, 1980a; Anderson et al., 1980b,c). More recent

studies

have

shown

that

ROS contain

phospholipases A, (Jelsema,1987) and C (Tarver and Anderson, 1988) and actively incorporate fatty acids into phospholipids

presumably through

deacylation-

reacylation reactions (Zimmerman and Keys, 1988). Thus, an active metabolism of lipids exists in ROS. The unique metabolism of PI in ROS was 6rst reported by Hall and coworkers in 1973, when they demonstrated that the specific radioactivity of PI in frog ROS increased rapidly for the ilrst 5 days, then dropped quickly, while that of the other phospholipids increased slowly for 3-4 weeks after the injection of 33P0,. When [2-3H]glycerol was used as a precursor,

and

Anderson,

1984 ; Hayashi

and

Amakawa, 1985; Millar and Hawthorne, 1985; Das et al., 1986: Millar et al., 1988). Since PIP, is derived from PI, some mechanism must be available to replenish these phosphoinositides. In the present study, we have tested the hypothesis that enzymes exist in ROSfor the regeneration of PI through the PI cycle (Fig. 7). Some of the enzyme activities involved in the PI cycle are found in both cytosol and membrane fractions, except for phosphatidyl : cytidyl transferase (Petzold and Agranoff, 1967) and PI synthetase (Benjamins and Agranoff, 1969), which are strictly membrane bound enzymes. Therefore, in order to test whether ROS contain all of the enzymes required to recycle DG to PIP,, it was necessary to use an intact ROS preparation that had both cytosolic and membrane proteins. During the phosphorylation studies using band II (intact ROS), it was found that CDP-DG and PI, as well as PA, PIP and PIP,, could be labeled by [Y-~~P]ATP if CTP and myo-inositol were added to the incubation medium. Therefore, it was possible to simultaneously assay all of the enzyme activities involved in the synthesis of PIP, from DG (Fig. 7) using endogenous lipids as substrates. Under these conditions, 32Plabel found in PI of band I (broken ROS)

was much lower than that of band II (intact ROS)(Fig. 3). A reason for the difference in PI synthetase activity

1980b) and rats (Wetzel and O’Brien, 1986). In frogs, ROSPI derived from labeled glycerol reached a specific

between broken and intact ROS preparations under this condition was not clear at ilrst, because the intact ROS preparation was assumed to have more soluble proteins than broken ROS and PI synthetase is a

radioactivity 10 times that of the other phospholipids and turned over with a half-life of 3.5 days (Anderson

membrane hound enzyme (in brain) that requires the presence of Triton X-100, GDP-DG, and either PC or

similar results were obtained in frogs (Anderson et al.,

174

asolectin to be eluted from the membrane (Ghalayini and Eichberg, 1985). Therefore, the purity of band II (intact ROS)was questioned and activities of enzymes known to be present in other retinal cellular organelles were assessed.Band II (intact ROS)was found to have relatively high microsomal and mitochondrial enzyme activities, making the results obtained using these membranes difficult to interpret. However, the extent of 32P-labelingin PA, PIP and PIP, was similar in the three fractions, indicating that ROSmembranes have DG, PI and PIP kinase activities, in agreement with previous reports (Seyfred et al., 1984; Giusto and Ilincheta de Boschero, 1986). The distribution of PI synthetase, phosphatidyl : cytidyl transferaseand the marker enzyme (cytochrome c reductase and oxidase) activities clearly showed that ROS prepared by discontinuous sucrose gradient centrifugation have only residual levels of any of the four enzyme activities, accounting for l-3% of the whole homogenate levels. Since the marker enzyme activities were still detectable, although minimal, we believe that the low levels of PI synthetase and phosphatidyl : cytidyl transferase activities in ROS are due to some small amount of contamination of our preparation, and that neither of these enzymes are integral components of ROS discs. Alternatively, PI synthetase activity reportedly is present in pituitary plasma membranes (Imai and Gerschengorn, 1987) so we cannot rule out the possibility that the ROS PI synthetase activity is associated with ROS plasma membranes, which would make up about 2-3 % of the ROS membranes harvested at the 1.1 l/1*1 3 g ml-* sucrose interface. However, the existence of PI synthetase in the plasma membrane is still controversial (Michell et al., 1988). Also, phosphatidyl : cytidyl transferase specific activity has been demonstrated in microsomal (van Golde et al., 1974) and mitochondria (Vorbeck and Martin, 1970) fractions of rat liver but not in plasma membranes. If PI in ROS comes from recycling of DG, then both PI synthetase and phosphatidyl : cytidyl transferase would be present in these membranes. We feel that the low levels of both enzymes in ROS are the result of minor contamination with other non-ROSmembranes, and that neither contributes to the production of PI in ROS through the PI cycle (Fig. 7). Both precursors of PI, myo-[2-3H]inositol and [2-3H] glycerol, are incorporated into PI in in vitro incubation (Figs 4 and 5). However, the specific radioactivity of PI in ROS was much lower than that of PI isolated from other membranes following in vitro incubation of retinas. In fact, the specificradioactivities of not only PI but also of PC and PE in ROSwere very low following the incubation of retinas with [2-3H] glycerol. Thus, it is likely that the time-dependent increase of radioactivity of PI (and also PE and PC) in ROS shown in Fig. 5 is due to translocation of PI synthesized de novo in the inner segments. Schmidt ( 1983 b, c) also reported that autoradiograms of

H.-G. CHOE

ET AL.

retinas show active incorporation of [3H]cytidine or [3H]inositol into CDP-DG or PI, respectively, of the photoreceptor inner segments.In all four fractions, the specific radioactivity of PI was much higher than that of PE or PC when retinas were labeled with [2-3H] glycerol. Since PI is synthesized from PA, while PE and PC are synthesized from DG, this may be due to the difference in specific radioactivity between the precursors PA and DG. However, considering the fact that PI turns over more rapidly than do other phospholipids in ROS (Anderson et al., 1980b). we believe that PI must be synthesized more rapidly in the inner segments to keep up with its greater demand in the ROS. A glycerophospholipid class consists of various molecular speciesthat have the same head group, but different fatty acids on their C-l and C-2 positions. Assuming that lipids with closemetabolic relationships should have similar molecular species distribution (R&tow et al., 1988 a,b), one would expect the PI, DG and PA in ROS to have similar molecular species composition if the three lipids were closely linked metabolically. PI and DG of ROSdo have a very similar molecular species distribution (Fig. S), indicating a close metabolic relationship, which would be an expected consequenceof the light-activated hydrolysis of PIP,. In both lipids, 16 : O-22 : 6 and 18 : O-22 : 6 are two major molecular species, and species containing arachidonic acid (20:4) are minor. This latter observation is rather unusual since 18 : O-20 : 4 is the most dominant species of PI in many tissues (Holub and Kuksis, 1978). However, Aveldano and Bazan (1983) demonstrated that the percent of 22:6 in PI was higher in bovine ROSthan that observed in other retinal membranes or reported for any other mammalian tissue. Furthermore, it is well known that amphibian retinal DG, especially DG of ROS,contains higher levels of 22: 6 than 20:4 (Fliesler and Anderson, 1983). Although a relatively high content of 22:6 (about 65%) in PA of bovine ROS was reported by Bazan et al. (1982) the richness of 22 : 6-22 : 6 molecular speciesis first demonstrated by the data in Fig. 5. The striking differences in molecular species distribution between PA and PI suggest that PA is not converted to PI in the ROS. This suggestion is indeed strengthened by our studies in various retinal membrane fractions. which demonstrate no significant enzyme activities involved in the synthesis of PI in ROS.Although the presence of DG kinase in ROS has been demonstrated (Seyfred et al., 1984; Giusto and Ilincheta de Boschero, 1986) there seems to be no close metabolic relationship between DG and PA of ROS, according to the molecular species distribution analysis. Indeed, the specific radioactivity of DG is lower than that of PA in all molecular species examined (for example, specific radioactivity of 22 ~6-22 :6 of DG and PA was 339 and 814 dpm nmol-’ , respectively, following a 6-hr incu-

PI METABOLISM

IN FROG

ROS

175

bation), indicating that most of ROSPA must originate from a metabolic pathway other than one involving DG kinase. Based on fatty acid composition analysis, Bazan et al. (1982) also suggested that PA is not metabolically related, in any major fashion, to DG in ROS.Thus, the physiological role of DG kinase in ROS is not clear at this moment. All of the experiments carried out in the present study - distribution

of the enzyme activities,

incor-

poration of labeled myo-inositol into PI in various retinal membranes, and molecular species analysis - support our contention that no significant PI synthesis

takes place in frog ROS through

the

pathways of DG+ PA + CDP-DG+ PI. Thus, PI in ROS must arise from synthesis in the inner segments.

Acknowledgments We would like to thank Janice Wilson for her assistance in preparing this manuscript and Creston A. Gay for his contribution in doing image analyses of the autoradiograms. This work was supported by grants from the Retinitis Pigmentosa Foundation Fighting Blindness, Research to Prevent Blindness Inc., the National Eye Institute, and the Retina Research Foundation.

References Abdel-Latif, A. A. (1986). Calcium-mobilizing receptors, polyphosphoinositides, and the generation of second messengers.Pharmacol. Rev. 38, 227-72. Anderson, R. E. and Brown, J, E. (1989). Phosphoinositides in the retina. In Progress in Retinal Research (Eds Osborne, N. and Chader, G.) Pp. 221-28. Pergamon Press: Oxford. Anderson, R. E., Kelleher, P. A. and Maude, M. B. (1980 a). Metabolism of phosphatidylethanolarnine in the frog retina. Biochim. Biophys. Acta 620, 227-35. Anderson, R. E., Maude, M. B. and Kelleher, P. A. (1980 b). Metabolism of phosphatidylinositol in the frog retina. Biochim. Biophys. Acta 620, 23646. Anderson, R. E., Maude, M. B., Kelleher, P. A., Maida, T. M. and Basinger. S. F. (1980~). Metabolism of phosphatidylcholine in the frog retina. Biochim. Biophys. Acta 620, 212-26. Aveldano, M. I. and Bazan, N. G. (1983). Molecular species of phosphatidylcholine, -ethanolamine, -serine, and -inositol in microsomal and photoreceptor membranesof bovine retina. J. Lipid Res. 24. 620-7. Bazan, N. G., de Escalante, S. M.. Careaga, M. M.. Bazan, H. E. and Giusto, N. M. (1982). High content of 22:6 (docosahexaenoate) and active [2-3H]glycero1 metabolism of phosphatidic acid from photoreceptor membranes. Biochim. Biophys. Acta 712, 702-6. Benjamins, J. A. and Agranoff, B. W. (1969). Distribution and properties of CDP-diglyceride: inositol transferase from brain. I. Neurochem. 16, 513-27. Bibb, C. and Young, R. W. (1974). Renewal of glycerol in the visual cells and pigment epithelium of the frog retina. J. Cell BioZ. 62, 378-89. Bishop, H. H. and Strickland, K. P. (1975). Studies on the formation by rat brain preparation of CDP-diglyceride from CTP and phosphatidic acids of varying fatty acid composition. Can. J. Biochem. 54, 249-60. Choe, H.-G. and Anderson, R. E. (1990). Unique molecular species composition of glycerolipids of frog rod outer segments. Exp. Eye Res. 51. 159-65.

Choe. H.-G., Wiegand, R. D. and Anderson, R. E. (1989). Quantitative analysis of retinal glycerolipid molecular speciesacetylated by acetolysis. 1. Lipid Res. 30,454-7. Cooperstein, S. J. and Lazarow, A. (1951). A microspectrophotometric method for the determination of cytochrome oxidase. 1. BioZ. Chem. 189, 665-70. Das, N. D., Yoshioka, T., Samuelson, D. and Schichi, H. localization of (1986). Immunocytochemical phosphatidylinositol-4,5-bisphosphate in dark- and light- adapted rat retinas. CeZZ Struct. Funct. 11, 53-63. Fliesler. S. J. and Anderson, R. E. (1983). Chemistry and metabolism of lipids in the vertebrate retina. Prog. Lipid Res. 22, 79-13 1. Fung. B. K.-K. (198 7). Transducin : Structure, function, and role in phototransduction. In Progress in Retinal Research. (Eds Osborne, N. and Chader. G.). Vol. 6. Pp. 15 l-7 7. Pergamon Press: Oxford. Ghalayini, A. J. and Anderson, R. E. (1984). Phosphatidylinositol4,5-bisphosphate : light-mediated breakdown in the vertebrate retina. Biochem. Biophgs. Res. Commun. 124, 503-6. Ghalayini. A. and Eichberg, J. (1985). Purification of phosphatidylinositol synthetase from rat brain by CDPdiacylglycerol affinity chromatography and properties of the purified enzyme. 1. Neurochem.44, 175-82. Giusto. N. M. and Ilincheta de Boschero, M. G. (1986). Synthesis of polyphosphoinositides in vertebrate photoreceptor membranes. Biochim. Biophys. Acta 877. 440-6. Hall. M. O., Basinger. S. F. and Bok, D. (1973). Studies on the assembly of rod outer segment disc membranes. In Biochemistry and Physiology of Visual Pigments (Ed. Langer, H.). Pp. 319-26. Springer-Verlag: New York. Hayashi, F. and Amakawa, T. (1985). Light-mediated breakdown of phosphatidylinositol-4.5-bisphosphate in isolated rod outer segments of frog photoreceptor. Biochem. Biophys. Res. Commun. 128, 954-9. Holub, B. J. and Kuksis, A. (1978). Metabolism of molecular species of diacylglycerophospholipids. Adv. Lipid Res. 16, 1-125. Imai. A. and Gerschengorn. M. C. (1987). Independent phosphatidylinositol synthesis in pituitary plasma membrane and endoplasmic recticulum. Nature 325, 726-8. Jelsema,C. L. (198 7). Light activation of phospholipase A, in rod outer segmentsof bovine retina and its modulation by GTP-binding proteins. 1. BioZ. Chem. 262, 163-8. Michell, R. H., Kirk, C. J., Maccallum, S. H. and Hunt, P. A. (1988). Inositol lipids: receptor-stimulated hydrolysis and cellular lipid pools. Philos. Trans. R. Sot. Lord. (Biol.) 320. 23946. Millar. F. A. and Hawthorne. J. N. (198 5). Polyphosphoinositide metabolism in response to light stimulation of retinal rod outer segments. Biochem. Sot. Trans. 13. 1X5-6. Millar, F. A., Fisher, S. C.. Muir, C. A.. Edwards, E. and Hawthorne, J. N. (1988). Polyphosphoinositide hydrolysis in response to light stimulation of rat and chick retina and retinal rod outer segments. Biochim. Biophgs. Acta 970, 205-l 1. Omura. T. and Takesue, S. (1970). A new method for simultaneous purification of cytochrome b, and NADPH-cytochrome c reductase from rat liver microsomes. 1. Biochem. 67, 249-57. Papermaster, D. S. (1982). Preparation of retinal rod outer segments. Methods EnzgmoZ. 81, 48-52. Petzold, G. L. and Agranoff, B. W. (1967). The biosynthesis of cytidine diphosphate diglyceride by embryonic chick brain. 1. BioZ. Chem. 242. 1187-91. Rouser, G., Fleischer. S. and Yamamoto, A. (1970). Two dimensional thin layer chromatographic separation of

176

polar lipids and determination .of phospholipids by phosphorous analysis of spots, Lipids 5, 494-6. R&stow, B., Nakagawa, Y., Rabe, H., Reichmann, G., Kunze, D. and Waku, K. (1988a). Comparison of the HPLCseparated species patterns of phosphatidic acid, CDPdiacylglycerol and diacylglycerol synthesized de novo in rat liver microsomes (a new method). Biochim. Biophys. Acta 961, 364-9. Riistow, B.. Nakagawa, Y., Rabe, H., Waku, K. and Kunze, D. (1988 b). Species pattern of phosphatidylinositol from lung surfactant and a comparison of the species pattern of phosphatidylinositol and phosphatidylglycerol synthesized de novo in lung microsomal fractions. Biochem.J. 254, 67-71. Schmidt, S. Y. (1983a). Cytidine metabolism in photoreceptor cells of the rat. J. Cell Biol. 97, 824-31. Schmidt, S. Y. (1983 b). Light- and cytidme-dependent phosphatidylinositol synthesis in photoreceptor cells of the rat. J. Cell Biol. 97, 832-7. Schmidt, S. Y. (1983 c). Phosphatidylinositol synthesis and phosphorylation are enhanced by light in rat retinas. J. Biol. Chem.258. 6863-8. Seyfred,M. A., Kohnken, R. E., Collins, C. A. and McConnell, D. G. (1984). Rapid diglyceride phosphorylation in isolated bovine rod outer segments. Biochem. Biophys. Res.Commun. 123, 121-7.

H.-G. CHOE

ET AL.

Tarver, A. P. and Anderson, R. E. (1988). Phospholipase C activity and substrate specificity in frog photoreceptors. Exp. Eye Res.46, 29-35. van Golde, L. M., Raben, J., Batenburg, J. J., Fleischer, B., Zambrano, F. and Fleischer, S. (1974). Biosynthesis of lipids in Golgi complex and other subcellular fractions from rat liver. Biochim. Biophys. Acta 360, 179-92. Vorbeck, M. L. and Martln, A. P. (19 70). Glycerophosphatide biogenesis. I. Subcellular localization of cytidime triphosphate : phosphatidlc acid cytidyl transferase. Biochem.Biophys. Res. Commun.40. 901-8. Wallace, M. A. and Fain, S. N. (1985). Analysis of hormoneinduced change of phosphoinositide metabolism in rat liver. Methods EnzRmol.109, 469-79. Wetzel, M. G. and O’Brien, P. J. (1986). Turnover of palmitate, arachidonate and glycerol in phospholipids of rat rod outer segments. Exp. Eye Res. 43, 941-54. Young, R W. (1967). The renewal of photoreceptor cell outer segments.J. Cell Biol. 33, 61-72. Zimmerman, W. F. and Godcheaux, W. III (1982). Preparation and characterization of sealed bovine rod cell outer segments. Methods Enzymol. 81, 52-7. Zimmerman, W. F. and Keys, S. (1988). Acylation and deacylation of phospholipids in isolated bovine rod outer segments. Exp. Eye Res.47, 247-60.