Retinol Esterification by Rat Liver Microsomes EVIDENCE FOR A FATTY ACYL COENZYME A: RETINOL ACYLTRANSFERASE*

To explore the nature of retinyl ester synthesis by liver microsomes, membranes prepared from rat or cat liver were incubated under various conditions with [’H] retinol dispersed in dimethyl sulfoxide. When [‘HJreti-nol, buffer, and microsomes were incubated together (basal conditions), some [‘HJretinol esterification was consistently observed. However, the rate of esterifica- tion could be increased 6- to 11-fold by addition of either palmitoyl-CoA (100 PM) or a fatty acyl CoA-gen- erating system. To determine whether the fatty acid used to esterify [3H]retinol under basal conditions might be derived from an endogenous pool of fatty acyl-CoA associated with the microsomal preparation, mi- crosomes were pretreated at pH 7.4 with 0.5 M hydroxylamine, a reagent that reacts with coenzyme A thioes-ters to form hydroxamates. ”his pretreatment reduced the basal reaction by 69%. However, hydroxylamine-treated microsomes still retained acyltransferase activ- ity, as shown by a 24- to 40-fold increase in retinyl ester synthesis after addition of palmitoyl-CoA. When microsomes were incubated with both C3H]retinol and [“C] palmitoyl-CoA of known specific radioactivities, the ratio of 14C to ‘H in newly synthesized retinyl palmitate was essentially equal to that of its putative substrates, indicating that [14C]palmitate did not undergo signifi- cant isotope dilution prior ison with a standard solution of all-trans-retinol. The acidic lipid fractions were concentrated, a mixture of carrier lipids including palmitic acid was added, and the samples were applied to a plastic-backed thin layer plate of silica gel and developed with a solvent system of hexane/diethyl ether/glacial acetic acid (70:30:1). Aft,er development, each lane was cut into zones which were transferred to counting vials containing ScintiLene for determination of radioactivity.

To explore the nature of retinyl ester synthesis by liver microsomes, membranes prepared from rat o r cat liver were incubated under various conditions with ['H] retinol dispersed in dimethyl sulfoxide. When ['HJretinol, buffer, and microsomes were incubated together (basal conditions), some ['HJretinol esterification was consistently observed. However, the rate of esterification could be increased 6-t o 11-fold by addition of either palmitoyl-CoA (100 PM) or a fatty acyl CoA-generating system. To determine whether the fatty acid used to esterify [3H]retinol under basal conditions might be derived from an endogenous pool of fatty acyl-CoA associated with the microsomal preparation, microsomes were pretreated at pH 7.4 with 0.5 M hydroxylamine, a reagent that reacts with coenzyme A thioesters to form hydroxamates. "his pretreatment reduced the basal reaction by 69%. However, hydroxylaminetreated microsomes still retained acyltransferase activity, as shown by a 24-t o 40-fold increase in retinyl ester synthesis after addition of palmitoyl-CoA. When microsomes were incubated with both C3H]retinol and ["C] palmitoyl-CoA of known specific radioactivities, the ratio of 14C t o 'H in newly synthesized retinyl palmitate was essentially equal to that of its putative substrates, indicating that [14C]palmitate did not undergo significant isotope dilution prior to acylation of [3H]retinoL These experiments provide direct evidence for retinol esterification catalyzed by a microsomal acyl-CoA:retinol acyltransferase and indirect evidence for a pool of fatty acyl-CoA in isolated liver microsomes that is available to react with [3H]retinol to form esterified retinol.
The major storage organ for vitamin A is the liver where this essential nutrient is found primarily as retinyl palmitate and other long chain fatty acid esters of retinol (1,2). It is known from physiological studies that most newly absorbed vitamin A is transported from the intestine to the liver in the form of esterified retinol as one component of the intestinal chylomicron (3-6). Uptake of chylomicron retinyl esters by rat liver was shown by Goodman et al. (7) to proceed by a sequence of steps that included ester bond hydrolysis and subsequent re-esterification of retinol in liver. The hepatic content of esterified retinol can vary widely with dietary intake, ranging from greater than 100 mg of retinol/liver in cases of frank vitamin A toxicity (8) to undetectable levels in animals maintained on a retinol-deficient diet (9)(10)(11).
Despite interest in the relationship of vitamin A status to hepatic content, only a few studies have addressed the enzymology of vitamin A esterification in liver or in other target organs. In 1964, Futterman and Andrews (2) reported in this journal on the esterification of retinol by tissue fractions of cat liver and bovine retina (12). Enzymatic activity was localized primarily to the microsomal fractions from both sources. It was reported that retinol esterification by microsomes proceeded in the absence of an exogenous source of ATP or coenzyme A (2,12) and in experiments with microsomes from bovine retina, it was noted that addition of preformed palmitoyl-CoA did not increase retinyl ester yield (12). No other source of the fatty acyl moiety was identified. These authors concluded that the esterification of retinol was not dependent on energy or activated fatty acid and, hence, that the reaction by which retinol was esterified differed in kind from the esterification of other cellular lipids such as cholesterol (2,12).
During a recent study of retinyl ester synthesis by the lactating mammary gland of the rat, I found that the rate of esterification of retinol by microsomes was markedly enhanced after inclusion of ATP and reduced coenzyme A or fatty acyl-CoA esters (34). These observations have prompted a reinvestigation of retinyl ester synthesis by liver microsomes.
I report here evidence that microsomes prepared from livers of the cat and rat catalyze the esterification of retinol by a fatty acyl-CoA:retinol acyltransferase reaction.

Preparation of Sub~trates-[l-~H(N)]Vitamin A1 (all-trans) from
New England Nuclear was mixed with all-trans-retinol (type X, Sigma Chemical Co.) to a specific radioactivity of 6 to 20 pCi/pmol and a concentration of 3 m~ in ethanol. This stock solution was kept under argon at -20 "C in the dark. Its purity was checked periodically by chromatography on aluminum oxide (below); it was repurified when more than 0.1% of 3H eluted in the retinyl ester fraction from alumina columns. [l-'4C]Palmitoyl coenzyme A (P-L Biochemicals, Inc.) was mixed with palmitoyl-CoA to approximately 8 pCi/pmol and 0.7 m~ in 0.01 M potassium phosphate buffer, pH 7.0, and stored in the same manner. Concentrations were determined by UV spectrophotometry after dilution using the molar absorption coefficients of 52,480 at 324 nm for retinol (13) and 15,400 at 259.5 nm for palmitoyl-CoA as supplied by P-L Biochemicals, Inc.
Other Materials-Coenzyme A (Li') and fatty acid esters of CoA were purchased from P-L Biochemicals, Inc.; ATP and sodium taurocholate were obtained from Calbiochem-Behring Corp., and bovine serum albumin (BSA' (essentially fatty acid-free)), dithiothreitol, butylated hydroxytoluene, and hydroxylamine hydrochloride were obtained from Sigma. Progesterone was a product of Stearaloids, Inc. Hexanes (certified) were purchased from Fisher Scientific Co. Diethyl ether containing butylated hydroxytoluene as preservative and high performance liquid chromatography grade acetonitrile were products of J. T. Baker. Liquid scintillation fluids (ScintiLene and Scintiverse) Preparation of Liver Microsomes-Livers of normal male or female Sprague-Dawley rats that had been fasted overnight were perfused in situ with cold 0.25 M sucrose, minced in 3 volumes of 0.25 M sucrose buffered with 0.01 M potassium phosphate, pH 7.4, and homogenized with a motor-driven Teflon pestle. Cat liver was processed without perfusion. Conditions for centrifugal preparation of microsomes were chosen with reference to Amar-Costesec et al. (14). After fdtering the liver homogenates through gauze, they were centrifuged at 4 "C for 10 min at 725 X g ; each supernatant was then centrifuged at 13,000 X g for 13 min in a Sorvall SS-34 rotor after which floating fat was removed by skimming. Supernatants were decanted and then centrifuged for 6.2 X lo6 g-min in the 60 Ti rotor of a Beckman ultracentrifuge. The resulting supernatant (cytosol) was removed and frozen. Each pellet was suspended in phosphatebuffered sucrose at approximately 10 mg/ml and recentrifuged as above. The resulting microsomal pellets were suspended in 0.15 M potassium phosphate buffer, pH 7.4 (referred to below as phosphate buffer), with 1 mM dithiothreitol at 15 to 25 mg of protein/& and divided into glass vials before freezing in Dry Ice/acetone and storage at -70 "C. Under these conditions, esterifying activity remained stable and nearly equivalent to that of fresh microsomes for at least 2 months.
In some experiments, microsomes were treated with hydroxylamine before incubation with ['Hlretinol. Hydroxylamine hydrochloride in water was adjusted to pH 7.4 with NH40H and diluted to 2.5 M. Frozen microsomes were thawed and diluted to 2.5 mg of protein/ml in phosphate buffer. One ml of this microsomal suspension and 0.25 ml of hydroxylamine solution, or phosphate buffer, were mixed and incubated at 37 "C for 20 min, then diluted with 7 ml of cold phosphate buffer, and centrifuged at 4 "C for 30 min at 40,000 rpm in a Beckman 40 rotor. Supernatants were discarded and the pellets were rinsed with 2 ml of phosphate buffer which were also discarded. Each pellet was then homogenized in 5 ml of cold phosphate buffer in a Dounce homogenizer and recentrifuged as above. The final washed pellets were homogenized in small volumes of phosphate buffer before incubation with substrates. In one experiment (Table II), other reagents replaced hydroxylamine in the same treatment protocol.
Protein was measured by the dye-binding method of Bradford (15), using bovine y-globulin as the reference standard.
Incubation with ['H]Retinol-All procedures with vitamin A were carried out under dim light. An aliquot of ['H]retinol from the stock solution above was taken to dryness under argon and redissolved in a small volume of dimethyl sulfoxide. Microsomes, which had been pipetted into glass screw-cap incubation tubes, were transferred to a shaking water bath in dim light at 37 "C to warm. After 50 s, any additional reagents were added so that the final volume was 0.25 ml.
Ten seconds later, ['Hlretinol was added in 5 pl of dimethyl sulfoxide, the mixture was vortexed briefly, and the tube was returned to the water bath for the specified incubation time. Each incubation condition was tested in duplicate or triplicate; a blank containing microsomes that had been boiled for 1 min was incubated under identical conditions.
Extraction and Chromatography-Incubations were stopped by addition of 1 ml of ethanol containing butylated hydroxytoluene (100 pg/ml). Neutral lipids were quantitatively partitioned into 4 ml of hexane containing butylated hydroxytoluene from 44% ethanol (16). After centrifugation, an aliquot of the upper phase (hexane) was counted to determine total isotope recovery, which usually equaled 94 to 98% for "H. A second aliquot was dried under nitrogen in preparation for separation of ['Hlretinyl esters from ["Hlretinol by chromatography on columns of aluminum oxide as described previoasly (17). ["HIRetinyl esters were eluted with hexane containing diethyl ether (3%, v/v); solvents were evaporated, and 'H was measured after addition of ScintiLene liquid scintillation fluid. Counting efficiencies were determined using calibrated ["HIor ["C]toluene standards as either internal or external standards. Data are expressed as picomoles of ["Hlretinol esterified after subtraction of values for appropriate blank incubations that contained boiled microsomes.
Analysis of J'H1Retinyl Esters by HPLC-The object of one experiment was to isolate and count labeled retinyl palmitate separately from other esters of retinol. To accomplish this, retinyl estercontaining eluates from aluminum oxide columns were concentrated under nitrogen, dissolved in acetonitrile, and subjected to HPLC on a Supelcosil LC-8 column as recently described (18). The mobile phase, pumped at 3.8 ml/min, consisted of acetonitrile/water (8812, v/v), followed by acetonitrile/water (98:2), as noted in the legend to Fig. 1. Fractions were collected every 30 s and portions were counted in Scintiverse liquid scintillation fluid. Portions of the fractions with the elution volume of retinyl palmitate were pooled, solvents were evaporated, and lipids were subjected to alkaline hydrolysis followed by sequential extraction of neutral lipids and acidic lipids after acidification (19). The neutral lipid fractions were rechromatographed by HPLC as described above using acetonitrile/water (88:12) in comparison with a standard solution of all-trans-retinol. The acidic lipid fractions were concentrated, a mixture of carrier lipids including palmitic acid was added, and the samples were applied to a plasticbacked thin layer plate of silica gel and developed with a solvent system of hexane/diethyl ether/glacial acetic acid (70:30:1). Aft,er development, each lane was cut into zones which were transferred to counting vials containing ScintiLene for determination of radioactivity.

RESULTS
Requirements for Retinyl Ester Synthesis-Liver microsomes were incubated with r3H]retinol and various additions to determine what exogenous substrates or supplements are needed for retinyl ester synthesis. In confirmation of previous reports (2, 12), some esterification was consistently observed when retinol and buffer alone were incubated with microsomes. As shown in Table I, this basal reaction equaled 108 pmol of retinyl ester/l5 min/0.25 mg of protein for rat liver microsomes and 65 pmol of retinyl ester/l5 min/0.25 mg of protein for microsomes prepared from cat liver. However, retinol esterification by both rat and cat liver microsomes was increased 6t o 11-fold upon addition of palmitoyl-CoA (100 p~) together with a reducing agent (dithiothreitol, glutathione, or cysteine) and BSA ( Table  1). Addition of a n acyl coenzyme A-generating system (20) containing ATP and coenzyme A also increased [3H]retinol esterification by more than 4-fold above the basal level when rat liver microsomes were incubated and more than %fold for microsomes from cat liver ( Table 1). Addition of exogenous fatty acid was not required, and addition of palmitic acid to r a t liver microsomes produced only a small increase in retinyl ester yield. These results suggested that ATP, coenzyme A, and an endogenous fatty acid provided substrates for synthesis of a fatty acyl-CoA thioester by microsomal fatty acid/CoA ligase (21) and that either this endogenous activated fatty acid or an exogenous source such as palmitoyl-CoA could be used as one substrate for retinyl ester synthesis by microsomes.
Other fatty acid esters of coenzyme A also significantly increased retinyl ester synthesis by rat liver microsomes. When added at an initial concentration of 100 p~ to 200 pg of microsomal protein (with BSA and dithiothreitol as described for palmitoyl-CoA in Table l Effect of Pretreatment of Microsomes on Retinol Esterification-An explanation was sought for the consistent observation that some ['Hlretinol was esterified by microsomes in the absence of either added fatty acyl-CoA or a generating system. Since fatty acyl-CoA has been isolated from rat liver microsomes (22) and appears to interact with microsomes in vitro (23,24), it seemed possible that the basal reaction might be due to endogenous fatty acyl-CoA either bound to or trapped within the microsomal vesicles. Because CoA thioesters are known to react with hydroxylamine a t neutral pH to form hydroxamic acids (21, 55), a series of washing experiments was conducted to determine whether pretreatment of microsomes with hydroxylamine might reduce their ability to esterify retinol in the absence of exogenous fatty acyl-CoA. Rat liver microsomes were pretreated by incubation for 20 min at 37 "C with 0.5 M hydroxylamine, pH 7.4, or with potassium phosphate buffer only. After incubation, microsomes were pelleted by ultracentrifugation, washed by homogenization in buffer, and recentrifuged. After resuspension in buffer, microsomes were incubated with r3H]retinol under either basal conditions or in the presence of palmitoyl-CoA, dithiothreitol, and BSA.
Experiment 1 of Table I1 shows that retinol esterification by untreated microsomes under basal conditions equaled 49. 3 pmol/min/mg of protein; after treatment with hydroxylamine, esterification was decreased to 15.9 pmol/min/mg of protein.
However, hydroxylamine-treated microsomes supplied with exogenous palmitoyl-CoA were still able to esterify [:'H]retinol and, indeed, showed a 48% increase in specific activity. Control experiments demonstrated that incubation of palmitoyl-Coh with hydroxylamine in the absence of microsomes, but 0therwise under the same conditions used to pretreat microsomes, was sufficient to chemically inactivate palmitoyl-CoA, whereas incubation without hydroxylamine had little effect on the ability of palmitoyl-CoA to stimulate the subsequent esterification of retinol by microsomes.
In a second experiment, hydroxylamine was compared with other agents (BSA, KC1, and NH4C1) that might bind or solubilize endogenous fatty acyl-CoA. After treatment withhydroxylamine, the basal reaction equaled only 19 pmol/min/ mg of protein; this rate was significantly lower than that observed after treatment with any of the other washing agents tested (50 to 67 pmol/min/mg of protein, Table 11).
These experiments provide indirect evidence that ["Hlretinyl ester synthesis under basal conditions is supported by an endogenous supply of activated fatty acid. By treating microsomal membranes with neutral hydroxylamine, this endogenous source is apparently depleted, thereby making subsequent esterification of ['HH]retinol nearly totally dependent on the addition of fatty acyl-CoA esters.
Product Formation after Incubation with r3H]Retinol a n d ['4C]Pa2mitoyE-CoA-The observation that retinol esterification is increased by either exogenous fatty acyl-CoA esters or a generating system is consistent with catalysis by a fatty acyl-CoA:retinol acyltransferase in which palmitoyl-CoA is the immediate donor of fatty acid to retinol. An alternative possibility is that added palmitoyl-CoA might fist acylate another molecule that subsequently acts as fatty acid donor to retinol. If this is the case, it is likely that the specific radioactivity of palmitic acid from [ l-14C]palmitoyl CoA would be reduced by dilution in the microsomal pool of this endogenous fatty acid donor before its esterification with retinol.
T o determine whether or not the palmitic acid moiety of ['4C]palmitoyl-CoA undergoes isotope dilution before acylation of retinol, an experiment was conducted in which both ["Hlretinol and [14C]palmitoyl-CoA of known specific radioactivities were added to treated rat liver microsomes. After incubation, lipids were extracted and the fraction containing  Rat liver micro-Cat liver microsomes somes Additions to incubation mixture Pmol pH] % of Pmol pH) % of retinol es-con-retinol es-   Fig. 1B) were subjected to alkaline hydrolysis followed by sequential extraction of neutral lipids and acidic lipids; these fractions were then analyzed by HPLC and TLC, respectively. As shown in Table III

crosomes-Studies
were conducted to determine the effects of pH, substrate concentrations, microsomal protein, and incubation time on the esterification of retinol. Using buffers from pH 6.2 to 10, maximum esterification was observed between pH 7 and 8, and potassium phosphate buffer at pl-3 7.4 was used for all further studies. When microsomes were treated with hydroxylamine and incubated in the presence of palmitoyl-CoA, the amount of esterified r3HJretinol increased with r3H]retinol concentration up to approximately 100 nmol/ ml ( Fig. 2A). The basal reaction (without added pahnitoyl-CoA) showed little dependence on retinol concentration.
After hydroxylamine treatment, esterification under basal conditions was greatly reduced at all concentrations of retinol tested ( Fig. 2A).
Retinol esterification also increased as fatty acyl-CoA concentration was increased. Significant differences were measured even when as little as 1 nmol/ml of pabnitoyl-CoA or oleoyl-CoA was added to untreated microsomes. Stimulation by palmitoyl-CoA did not require the presence of BSA; however, the extent of stimulation could be modified by the amount of BSA in the incubation mixture (Fig. 2 B ) .  Fig. 3 shows the dependence of retinyl ester synthesis on the amount of microsomal protein incubated and on the time of incubation. Esterification was linear with microsomal protein up to approximately 135 pg when hydroxylamine-treated microsomes were incubated with palmitoyl-CoA. When untreated rat liver microsomes were incubated under similar conditions or with the fatty acyl-CoA-generating system of Table I, esterification was linear to approximately 200 pg and 450 pg of protein, respectively. Retinol esterification increased linearly with time of incubation for a t least 8 min for both hydroxylamine-treated microsomes (Fig. 3B) as well as untreated microsomes that were incubated either with palmitoyl-CoA or with the fatty acyl-CoA-generating system. Inhibitors of Retinol Esterification-Esterification of [3H] retinol was measured in the presence of several agents which are known to inhibit other microsomal acyltransferases. The sulfhydryl-blocking reagent p-chloromercuribenzoate abolished retinol esterification when present at a concentration of 5 mM, indicating that reduced sulfhydryl groups are necessary either on the enzyme or its thioester substrate. Bile salt detergents, sodium deoxycholate and taurocholate (0.5%), also inhibited esterification when present during incubation with [3H]retinol. However, pretreatment with 0.5% sodium taurocholate (under conditions described for hydroxylamine) resulted in a 2.3-fold increase in specific enzyme activity (578 pmol/min/mg) as compared to microsomes pretreated with buffer only (249 pmol/min/mg). This increase appeared to be due to solubilization and loss of other microsomal protein rather than to activation of the membrane-associated retinolesterifying activity.
Enzyme activity of untreated microsomes was also measured in the presence of progesterone, one of several polar steroids that have been shown to inhibit microsomal acyl-CoA:cholesteroI acyltransferase (26)(27)(28). At concentrations Retinol Esterification by Rat Liver Microsomes above 40 mnol/ml, progesterone also inhibited the esterification of retinol in a dose-dependent manner (Fig. 4). In contrast, cholesterol at 80 nmol/ml had no significant effect (96.3% of control). DISCUSSION The primary goal of this work was to determine whether the esterification of retinol by liver microsomes is catalyzed by a fatty acyl-CoA:retinol acyltransferase, analogous to the esterification of cholesterol, glycerides, and phospholipids and in agreement with a recent study of retinyl ester synthesis by mammary gland microsomes (34), or whether the esterification of retinol proceeds by a different type of reaction, perhaps a transacylation as suggested by the earlier reports of Futterman and Andrews (2,12 were also designed to explore further the basal esterification reaction, which might be due either to the presence of an endogenous pool of fatty acyl-CoA substrate for an acyltransferase or to a second type of reaction that is independent of fatty acyl-CoA. Long chain fatty acyl-CoA has been isolated from rat liver microsomes by Garland et al.  22) was not reported. Since these authors have also reported that the content of total fatty acyl-CoA in rat liver doubled after starvation (29), it is plausible that the content of fatty acyl-CoA bound to microsomes was sufficiently greater in our fasted rats to account for the observed basal reaction. Further direct measurements are planned. In the present study, the reagent hydroxylamine that has been used to "trap" newly synthesized fatty acyl-CoA esters (21) was chosen as a potential means of depleting microsomes of endogenous fatty acyl-CoA and, indeed, pretreatment with hydroxylamine was successful in nearly eliminating the basal reaction ( Fig. 2A), yet these treated microsomes still retained acyltransferase activity as demonstrated by their response to exogenous palmitoyl-CoA. This experiment provides strong, albeit indirect, evidence for endogenous fatty acyl-CoA as a component of the microsomal preparation.
Apparently, this pool of fatty acyl-CoA may react either with hydroxylamine or with [3H]retinol to yield retinyl esters under basal incubation conditions. It is not clear why earlier reports failed to show stimulation of retinol esterification by ATP and coenzyme A (2, 12) or pahnitoyl-CoA (12). Unfortunately, Futterman and Andrews (2, 12) did not report the concentrations of the supplements that they tested. Clearly, in experiments reported herein, the magnitude of the effect of palmitoyl-CoA depended on the concentration of retinol ( Fig. 2A) and, hence, the ratio of retinol to microsomes. Earlier studies had employed larger amounts of cat liver microsomes and a lower ratio of retinol to membrane (2). It seems plausible that sufficient fatty acyl-CoA was contained in these microsomes to mask an effect of exogenous supplements.
More recently, Berman et al. (30) have suggested that endogenous fatty acyl-CoA in bovine pigment epithelium microsomes provided fatty acid for the esterification of [3H]retinol. Another point of controversy is whether or not the inclusion of cytosol enhances retinol esterification by microsomes. Futterman and Andrews (12) reported that addition of cytosol prolonged retinyl ester synthesis by retinal microsomes and increased retinol esterification by liver microsomes up to 2fold (2), while Krinsky (31) had reported that supernatant from bovine retina contained a heat-stable and dialyzable substance that enhanced retinyl ester synthesis by microsomes. In contrast, Berman et al. (30) did not find an effect with retinal cytosol and, in this laboratory, addition of rat liver cytosol (0.2 mg of protein) to 0.2 mg of microsomes produced little to no effect on esterification under basal conditions and a slight negative effect in the presence of added palmitoyl-CoA.* Because tissue sources and incubation conditions have differed in each laboratory, interest should remain open in the possibility that certain cytosolic factors, perhaps cofactors or soluble proteins (30), can modulate the esterification of retinol.

Results of this investigation
indicate that the enzymatic esterification of retinol is analogous to that of other cellular lipids, such as cholesterol. Cholesterol esteritication in various cells is catalyzed by acyl-CoAcholesterol acyltransferase (EC 2.3.1.26), known as ACAT (32,33). By analogy then, the activity responsible for retinol esterification is acyl-CoA:reti-no1 acyltransferase or ARAT. Both activities are membranebound and sensitive to detergents (32,33) and both are inhibited in the presence of the polar steroid progesterone (26)(27)(28). Whether a unique enzyme catalyzes retinyl ester synthesis is not yet known. A complete description of the complex processes leading to retinyl ester synthesis and storage will require more detailed information about this membranebound enzyme, as well as an understanding of the mode of delivery of its lipid substrates and the disposition of esterified retinol after synthesis.