Oleoyl-CoA Is the Major de Novo Product of Stearoyl-CoA Desaturase 1 Gene Isoform and Substrate for the Biosynthesis of the Harderian Gland 1-Alkyl-2,3-diacylglycerol*

1-Alkyl-2,3-diacylglycerol (ADG) is a unique neutral lipid found in the eyeball-associated Harderian gland (HG) of the mouse and acts as a lubricant to facilitate eyelid movement. We found that the HG of the mice with a disruption in the gene for stearoyl-CoA desaturase 1 (SCD1) (SCD1−/−) is deficient in ADG. The amount of C20:1n-9, which is a major fatty acid of ADG, was reduced by greater than 90% despite normal elongase enzyme activity proposed to elongate it from C18:1n-9. HG from SCD1−/− mice exhibited high desaturase activity toward C16:0-CoA as substrate but had very low desaturase activity toward C18:0-CoA. Feeding diets containing high levels of oleate to the SCD1−/− mice did not increase the levels of C18:1n-9 or C20:1n-9 in the HG and failed to restore the ADG to the levels found in the HG of the wild-type mouse. De novo ADG synthesis as measured by the incorporation of [3H]glycerol and [14C]glucose was high in the SCD1+/+ mouse but was reduced by greater than 90% in the HG of SCD1−/− mouse. The deficiencies in the levels of ADG and C20:1n-9 were not compensated for by the expression of SCD2 and SCD3 isoforms in the HG of the SCD1−/− mouse. These observations demonstrate that SCD1-synthesized oleoyl-CoA is a major substrate required for the biosynthesis of normal levels of ADG and that the SCD isoforms present in the HG have different substrate specificity.

Stearoyl-CoA desaturase (SCD) 1 is a rate-limiting enzyme in the biosynthesis of monounsaturated fatty acids. It catalyzes the ⌬9-cis desaturation of acyl-CoA substrates, the preferred substrates being palmitoyl-CoA and stearoyl-CoA, which are converted to palmitoleoyl-CoA and oleoyl-CoA, respectively (1). The resulting monounsaturated fatty acids are substrates for incorporation into membrane phospholipids, triglycerides, and cholesterol esters (2). Several isoforms of SCD exist in the mouse genome. SCD1, SCD2, and SCD3, which are products of different genes, are the most well characterized (3)(4)(5). Most organs of different mouse strains express both SCD1 and -2 with the exception of liver, which expresses mainly the SCD1 isoform (3). SCD2 is constitutively expressed in the brain (3) and like SCD1 is expressed at high levels in livers of mice that overexpress the truncated nuclear form of sterol regulatory element-binding protein (SREBP)-1a (6). Despite the fact that the mouse SCD1, SCD2, and SCD3 genes are structurally similar, sharing ϳ87% nucleotide sequence identity in the coding regions, their 5Ј-flanking regions differ somewhat resulting in divergent tissue-specific gene expression. However, in some tissues such as the adipose and eyelid both SCD1 and SCD2 genes are expressed whereas in the skin all the three SCD gene isoforms are expressed (5). The reason for having two or more SCD isoforms in the same tissue is not known but could be related to the substrate specificity of the isomers and their regulation through tissue-specific expression.
The existence of multiple SCD genes in mice and rat tissues makes it difficult to determine the role of each gene in lipid metabolism. Most previous studies have assessed SCD gene function by measuring mRNA expression but have not differentiated which SCD isoform is responsible for the altered total SCD activity. The clue as to what the physiological role of the SCD1 gene and its endogenous products (the monounsaturated fatty acids) is has come from recent studies of the asebia mutant mouse strains (ab j and ab 2j ) that have a natural mutation in the SCD1 gene (5, 8 -9). Most recently we have generated a mouse model with a targeted disruption of the SCD1 gene (7) and have shown that it has phenotypes similar but not identical to those present in the natural models. The 5Ј-boundary of the natural deletion in the SCD1 gene of the asebia mutant mouse has not been mapped within a 10-kilobase promoter region suggesting that this deletion could extend into other genes. Consequently, it has not yet been possible to generate appropriate DNA primers to distinguish the heterozygous asebia mice from the wild-type mice by genotyping. The basis for the difference in some phenotype between the laboratory and the natural mutant mouse strains could also be because of strain background modifying gene effects.
The Harderian gland (HG) that was first described by Johann J. Harder in 1694 (10) occurs in most terrestrial vertebrates and is located within the orbit of the eye where in most species it is the largest structure. The chief products of the gland vary between different groups of vertebrates. In rodents the gland synthesizes lipids, porphyrins, and indoles (11). The lipids are excreted by an exocytotic mechanism (12). The major secretory lipid of the HG of the mouse has been identified as 1-alkyl-2,3-diacylglycerol (ADG) whereas in the rat the chief products are wax esters (11). The ADG acts as a lubricant of the eyeball and is therefore important in facilitating the movement of the eyelid (13). In addition it has been proposed that the HG can act as a site of immune response, a source of pheromones, a source of thermoregulatory lipids, a photoprotective organ, a part of a retinal-pineal axis, a site of osmoregulation, and a source of growth factors (11,12). The HG is also a target of hormonal agents including gonadal, thyroid, and pituitary hormones, in addition to the many neuropeptides that have been shown to regulate the activity of this gland (12).
In this study, we found that the HG of the mouse with a targeted disruption of the SCD1 gene isoform has decreased levels of ADG, the major lipid of this gland. The amount of C20:1n-9, which is the major fatty acid in the ADG, was reduced by greater than 90% despite the presence in both SCD1Ϫ/Ϫ and SCD1ϩ/ϩ mice of normal elongase activity, the enzyme proposed to elongate C18:1n-9 to C20:1n-9. HG microsomes isolated from SCD1Ϫ/Ϫ mice had very low desaturase activity toward C18:0-CoA as a substrate compared with C16: 0-CoA. Feeding diets containing high levels of oleate to the SCDϪ/Ϫ mice did not increase the levels of C18:1 or C20:1n-9 in the HG and failed to restore the levels of ADG to the levels found in the HG of the wild-type mouse. De novo ADG synthesis as measured by the incorporation of [ 3 H]glycerol and [ 14 C]glucose was reduced by greater than 80% in the HG of SCD1Ϫ/Ϫ mouse. Taken together, the observations demonstrate that SCD1 isoform has a preference of C18:0-CoA as a substrate of desaturation compared with C16:0-CoA and that SCD1-synthesized oleoyl-CoA is the major substrate required for the biosynthesis of very long chain monounsaturated fatty acids of HG 1-alkyl-2,3-diacylglycerol.
Materials-Radioactive [ 32 P]dCTP (3000 Ci/mmol) was obtained from Dupont. Thin layer chromatography plates (TLC Silica Gel G60) were from Merck. Rabbit anti-rat SCD, which was raised against purified rat liver SCD, was a gift from Dr. Juris Ozols, University of Connecticut Health Center, Farmington, CT. All other chemicals were purchased from Sigma.
Lipid Analysis-Total lipids were extracted from HG according to the method of Bligh and Dyer (14) and were separated by silica gel TLC using petroleum ether:diethyl ether:acetic acid (80:30:1) as the developing solvent. The lipids were visualized by cupric sulfate in 8% phosphoric acid. The lipids were scraped, methylated, and analyzed by gas-liquid chromatography on a capillary column coated with DB-225 (30-m length, 0.25 mm, internal diameter, 0.25 m; Agilent Technologies, Inc., Wilmington, DE). Column temperature was kept at 70°C for 1 min, increased to 180°C at a rate of 20°C/min and then to 220°C at a rate of 3°C/min. The temperature was kept at 220°C for 15 min. 20:1n-9 and 20:1n-7 fatty acids were identified by comparison of retention times with authentic standards (Sigma). A heptadecanoic acid and a 1,2-dihepatadecanoyl-L-␣-phosphatidylcholine (Sigma) were added as internal standards for the quantitation of 1-alkyl-2,3-diacylglycerol and phospholipids, respectively.
Immunoblotting-Total protein was prepared from pooled HG from three mice of each group. The HG (100 mg) was rinsed with phosphatebuffered saline and homogenized in 2 ml of 50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.25% sodium deoxycholate, 0.1% SDS, 1 mM Na 3 VO 4 , 10 mM Na 2 MoO 4 , 40 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, and 1 g/ml leupeptin. The homogenate was clarified by centrifugation at 10,000 ϫ g for 10 min at 4°C. The same amount of protein (25 g) from each fraction was subjected to 10% SDS-polyacrylamide gel electrophoresis and transferred to Immobilon-P transfer membranes at 4°C. After blocking with 10% non-fat milk in Tris-buffered saline buffer (pH 8.0) plus Tween at 4°C for 2 h, the membrane was washed and incubated with rabbit anti-rat SCD as primary antibody and goat anti-rabbit IgG-horseradish peroxidase conjugate as the secondary antibody. Visualization of the SCD protein was performed with an enhanced chemiluminescence Western blot detection kit (18).
In Vivo Assay for 1-Alkyl-2,3-diacylglycerol Synthesis-[ 3 H]Glycerol or [ 14 C]glucose was dissolved in 0.9% NaCl at a concentration of 5 Ci/0.2 ml and injected into mice 1 h before being sacrificed (19 -21). HG lipids were extracted as described by Bligh and Dyer (14) and separated by TLC using hexane:ether:acetic acid (90:30:1) as developing solvent. The 1-alkyl-2,3-diacylglycerol fraction was scraped off the plate, and the radioactivity was measured using a liquid scintillation counter. Fig. 1 shows a Northern blot of total RNA isolated from various tissues of wild-type mice and analyzed for the expression of SCD1, SREBP-1 and FAS mRNAs. Compared with several mouse lipogenic tissues including liver and white adipose tissue, the HG expresses much higher mRNA levels of SCD1, SREBP-1, and FAS indicating that the HG is highly lipogenic. 28 S mRNA expression used as loading control was similar in the tissues tested. Fig. 2 shows TLC analysis of lipids extracted from HG of SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. As reported previously (22,23), the main lipid present in the HG of the mouse is ADG. The ADG was markedly reduced in HG of the SCD1Ϫ/Ϫ mice compared with the SCD1ϩ/ϩ control mice. The free cholesterol levels were not changed. Triglyceride levels were very low, and cholesterol esters were undetectable. ADG and phospholipids were also measured quantitatively by GLC using heptadecanoic acid as an internal standard. Table I shows that the total ADG content in HG of SCD1Ϫ/Ϫ mice was decreased by 53% whereas the phospholipid content was decreased by 24%. Fig. 3 shows the contents (mg/g) of the major fatty acids measured in the total lipid, 1-alkyl-2,3-diacylglycerol, and phospholipid fractions of the HG of SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. The most abundant monounsaturated fatty acid in the total lipid of the SCD1ϩ/ϩ mice is eicosenate (C20:1n-9) (148 mg/g), which was decreased by greater than 90% in the SCD1Ϫ/Ϫ mice. The content of C20:1n-7 was decreased by 34% whereas that of C18:1n-9 was decreased by 60%. In the ADG fraction, C20:1n-9 and C20:1n-7 were decreased by 93 and 44%, respectively. The phospholipid fraction did not contain C20: 1n-9 or C20:1n-7, but the content of C18:1n-9 was decreased by 57%. The HG lipids contained very low levels of w-6 and w-3 polyunsaturated fatty acids.

RESULTS
To determine whether dietary oleate could substitute for the endogenously synthesized oleate and restore the HG ADG levels of the SCD1Ϫ/Ϫ mice to the levels observed in the SCD1ϩ/ϩ mice, we supplemented the semipurified diet with high levels of C18:1n-9 (75% of total fat) as high oleate oil and fed it to the SCD1Ϫ/Ϫ mice for 1 month, a long enough feeding regimen to ensure equilibration of lipid pools. The food intake was not very different between the SCD1ϩ/ϩ mice controls and the SCD1Ϫ/Ϫ mice. Total HG extracts were prepared, the lipid fractions were analyzed by TLC, and the fatty acid composition was analyzed by GLC. Feeding diets supplemented with 5 or 30% oleate-rich diet to the SCD1Ϫ/Ϫ mice did not result in an increase in the levels of ADG (Fig. 4A). The content in (mg/g tissue) of ADG was not increased (Fig. 4B). GC analysis showed that content s of C18:1n-9 and 20:1n-9 were not increased in the ADG fraction of the SCD1Ϫ/Ϫ mouse (Fig. 4C). The content

FIG. 2. TLC of lipid extracts from HG of SCD1؉/؉ and SCD1؊/؊ mice.
Total lipids were extracted from livers of SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice, and the extracts were pooled and analyzed by TLC. Equivalent amounts of lipid extract (from 0.1 mg of tissue homogenate) were loaded in each lane. 1-Octadecyl-2,3-dipalmitoylglycerol, cholesteryl oleate, and triolein were used as standards for ADG, cholesteryl ester, and triglycerides, respectively. Std, standards. of C20:1n-9 was not increased in the total lipid or ADG fraction when SCD1Ϫ/Ϫ mice were fed diets supplemented with high levels of C20:1n-9 as trieicosenoin (data not shown). These observations suggest strongly that the normal levels of C20: 1n-9 in 1-alkyl-2,3-diacylglycerol are largely dependent on endogenously synthesized C18:1n-9.
The low levels of ADG observed in the SCD1Ϫ/Ϫ mice could have resulted from reduced levels of the elongase activity, the enzyme that would catalyze the elongation of C18:1n-9 to C20: 1n-9. Fig. 5A shows that the elongase activity as measured by rate of conversion of [ 14 C]malonyl-CoA to labeled C20:1n-9 was almost equal in HG microsomes of both SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. Compared with elongase activities in liver and white adipose tissue, the HG elongase activity was highest when either palmitoleoyl-CoA or oleoyl-CoA were used as substrates (Fig. 5B). These observations demonstrate the requirement of palmitoleoyl-CoA and oleoyl-CoA as substrates for elongation into very long monounsaturated fatty acids that subsequently become incorporated into ADG and the HG.
To establish that the low levels of ADG in the SCD1Ϫ/Ϫ mice is because of lower de novo synthesis rates, we used [ 3 H]glycerol or [ 14 C]glucose as precursors of lipid synthesis to measure directly newly synthesized ADG in the HG of SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. Fig. 6 shows that the ADG synthetic rate both from glucose and glycerol was decreased by greater than 80% in the HG of the SCD1Ϫ/Ϫ mice. These results indicate that the normal levels of ADG in the HG are dependent on SCD1 gene expression. To determine whether the HG of the SCD1Ϫ/Ϫ mouse expresses other SCD isoforms in addition to SCD1, total RNA isolated from SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice was analyzed by Northern blot using as specific DNA probes SCD2 and SCD3 isoforms. Fig. 7A shows that the HG of SCD1Ϫ/Ϫ mice as expected did not express SCD1 but that SCD2 and SCD3 isoforms are expressed in both SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. Interestingly, SCD2 mRNA level in SCD1ϩ/ϩ was lower than that in the SCD1Ϫ/Ϫ mouse whereas that of SCD3 was higher in the SCD1ϩ/ϩ than in the SCD1Ϫ/Ϫ mouse. FAS, SREBP-1, and SREBP-2 mRNA levels were similar between SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. The 28 S mRNA expression used as a loading control was also similar in the SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice. Consistent with the Northern blot results of Fig. 7A, Western blot analysis showed high immunoreactive SCD protein in HG from the SCD1ϩ/ϩ with a decreased protein level in the SCD1Ϫ/Ϫ mice (Fig. 7B). To determine whether the SCD isoforms have different substrate specificity, we assayed microsomes from HG of SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice for desaturase activity toward the main substrates of SCD, C16:0-CoA, and C18:0-CoA. Fig. 7C shows that SCD enzyme activity in the HG as measured by the rate of conversion of [1-14 C]palmitoyl-CoA to [1-14 C]palmitoleoyl-CoA was higher in the wild-type mice and was reduced by 35% in the SCD1Ϫ/Ϫ mouse. However, the SCD activity as measured by the rate of conversion of [1-14 C]stearoyl-CoA to [1-14 C]oleate revealed that whereas the SCD1ϩ/ϩ microsomes had high SCD activity, the SCD activity was reduced by greater than 95% in microsomes of SCD1Ϫ/Ϫ mice. These results indicate that SCD1 can use C16:0-CoA (the 35%) as a substrate but uses C18:0-CoA (the 95%) as the preferred substrate for desaturation. On the other hand SCD2 or SCD3 isoform uses C16:0-CoA as a specific substrate of desaturation. These results suggest that SCD isoforms present in the HG have different substrate specificity. DISCUSSION We found that the eyeball-associated HG of the mouse has a higher expression of SCD1 mRNA than in other tissues analyzed previously, including liver and white adipose tissue (Fig.  1). The expression of other lipogenic genes (FAS and SREBP-1) was also higher than the levels observed previously in liver and white adipose tissue. These observations indicated to us that the HG is a highly lipogenic gland. The major neutral lipid found in the HG of the mouse is ADG with lower levels of phospholipid and free cholesterol. As indicated in Fig. 2 and Table I, the SCD1Ϫ/Ϫ mice have a deficiency in ADG and C20:1n-9; the main monounsaturated fatty acid in ADG was reduced by greater than 90% (Fig. 3). The phospholipid levels were also decreased in the SCD1Ϫ/Ϫ mice. Feeding diets supplemented with high levels of oleate did not result in an increase in the levels of C18:1 or C20:1n-9 and could not repair the deficiency in 1-alkyl-2,3-diacylglycerol (Fig. 4). Feeding diets supplemented with high levels of C20:1n-9 itself to the SCD1Ϫ/Ϫ mice did not correct the deficiency in ADG either. Further, we demonstrated that the deficiency of ADG in the SCD1Ϫ/Ϫ mice was because of a decreased rate of de novo synthesis (Fig. 5). However, the ADG synthesis occurred at a much lower level (20%) whereas the level of ADG was reduced by 50% suggesting a possible defect in the secretion of ADG from the HG of the SCD1Ϫ/Ϫ mice. The deficiencies in the levels of ADG and C20:1n-9 were not compensated for by the expression of SCD2 and SCD3 isoforms in the HG of the SCD1Ϫ/Ϫ mouse. These observations suggest that endogenously synthesized C18:1 as a result of SCD1 gene expression is required for the biosynthesis of ADG in the HG. The endogenously synthesized oleate is also required for synthesis of normal levels of phospholipids.
The mouse genome contains three well characterized structural genes (SCD1, SCD2, and SCD3) that are highly homologous at the nucleotide and amino acid level and encode the same functional protein (5). Although the difference in physiological function between SCD1, -2, and -3 has not been well addressed, we suggested previously that the SCD1 and SCD2 isoforms might exhibit different specificity for substrates (18), in addition to exhibiting tissue-specific expression (5,7). We found that the mouse HG expresses three SCD isoforms (SCD1, SCD2, and SCD3) and, in addition, noted that there were differences in the levels of their expression between the SCD1ϩ/ϩ and SCD1Ϫ/Ϫ mice (Fig. 7). Although SCD2 was expressed at lower levels in the SCD1ϩ/ϩ mice its expression was elevated in the SCD1Ϫ/Ϫ mice. Expression of SCD3 was higher in the SCD1ϩ/ϩ but decreased in the SCD1Ϫ/Ϫ mice. Similarly in the skin of asebia mouse lacking SCD1, SCD3 expression was decreased (6). In addition, the three SCD isoforms were expressed in different cell types in the skin (6). SCD1 is located in pre-sebocytes of the sebaceous gland of the skin whereas the mature sebocytes expressed SCD3. On the other hand SCD2 is expressed in hair follicles. It is not known whether the HG SCD isoforms are expressed in different cell types, as well, but it is possible that their pattern of expression is a reflection of HG cell types in different stages of differentiation. Despite the expression of SCD2 and SCD3 in the HG of SCD1Ϫ/Ϫ, the deficiencies in the levels of ADG and C20:1n-9 were not compensated for suggesting a distinct role of each SCD isoform in the synthesis of monounsaturated fatty acids of the HG.
The observation that microsomes isolated from HG of SCD1Ϫ/Ϫ mice had very low desaturase activity toward C18: 0-CoA compared with C16:0-CoA (Fig. 7) strongly suggests that C18:0-CoA is the main substrate of SCD1 isoform. The other isoforms (SCD2 or -3) preferentially utilize C16:0-CoA as the substrate of desaturation. Consistent with this notion the levels of C18:1n-7 and C20:1n-7 derived from elongation and desaturation of C16:0-CoA were decreased by 30% in the SCD1Ϫ/Ϫ mouse compared with a decrease of greater than 90% in the levels of C20:1n-9 that would be derived from the elongation of C18:1n-9. These studies, along with a recent report of a palmitoyl-CoA-specific ⌬9 desaturase from Caenorhabditis elegans (24), strongly suggest that the SCD isoforms have different substrate specificity and may explain why there are several SCD isoforms in the mouse genome. The differences in the catalytic selectivity of the SCD isoforms may also contribute to the establishing of the lipid composition of the cell. A finer control can be provided by regulated expression of several isoforms with differing selectivity than by expression of either one or two with the same substrate selectivity.
We propose, as depicted in Fig. 8, that in the HG palmitate is synthesized de novo by FAS from acetyl-CoA producing palmi-FIG. 6. The rate of 1-alkyl-2,3-diacylglycerol synthesis in HG. Mice (n ϭ 4 -5) fed nonpurified diet were intraperitoneally injected with 5 Ci of [ 3 H]glycerol and [ 14 C]glucose 1 h before being sacrificed. After extraction of the lipids, ADG was separated using TLC, and the radioactivity was determined by liquid scintillation counting. *, p Ͻ 0.001 versus SCD1ϩ/ϩ mice (Student's t test). tate (C16:0) as the major end product. Palmitate then serves as a substrate for the microsomal malonyl-CoA-dependent elongase to produce stearate, which then serves as the main substrate of SCD1. The C16:1n-7 synthesized by either SCD2 or SCD3 is converted first to C18:1n-7 by a general elongase and then both C18:1n-9 and C18:1n-7 are converted to C20:1n-9 and C20:1n-7, respectively, by the HG elongase. Desaturation of C20:0 to C20:1 by any of the SCD isoforms is unlikely, because C20:1n-11, the product of this reaction, was not detected. The elongase may be specific to the HG, because we cannot detect C20:1n-9 or C20:1n-7 fatty acids in other tissues. In addition, the elongase in HG exhibited higher substrate specificity for C18:1n-9-CoA and C16:1n-7-CoA, whereas that in other tissues such as liver and adipose tissues showed high substrate specificity for 16:0-CoA (Fig. 5B). The C20:1n-9 and C20:1n-7 then become the preferred substrates for esterification to the 2,3 positions of alkyl glycerol by unidentified 1-alkylglycerol: acyltransferase for the synthesis of ADG. The reasons for incorporating specific very long chain monounsaturated fatty acids in the 1-alkyl-2,3-diacyglycerol of the mouse HG are not clear, but it may be that they are required to maintain the correct physical properties of the fluids for normal eye function. The narrow eye fissure phenotype of mice with a targeted disruption of the stearoyl-CoA desaturase 1 gene (7), as well as in the asebia mouse with a natural deletion in the SCD1 gene (9), may be because of a deficiency in endogenous oleoyl-CoA.
We found very low levels of n-3 and n-6 polyunsaturated fatty acids in the HG indicating that these fatty acids are either not required or do not accumulate in the lipids of the Harderian gland. It seems that the lipids of this gland are composed mainly of saturated fatty acids and monounsaturated fatty acids of the n-9 series. Because of the numerous roles cited for the mouse HG (12) in different animal species, further studies of this gland could reveal previously undescribed roles that monounsaturated fatty acids play in diverse physiological processes. Humans do not have the HG, but because this structure is part of the retina axis of the mouse, it is likely that it has evolved into the retina in humans. Consistent with this stipulation high SCD expression has been reported in human retinal pigment epithelial cells (25), and its expression may play an important role in the pathophysiology of these cells.
In conclusion, the studies have revealed that SCD1 gene expression is required for the synthesis of another class of neutral lipid, the ADG. The Harderian-specific elongase that synthesizes the very long chain monounsaturated fatty acids of ADG requires C18:1 and C16:1, the endogenous products of SCD isoforms, as substrates. The failure of dietary fatty acids to repair the deficiency in phospholipids and ADG levels strongly suggests the HG relies on de novo monounsaturated fatty acids for lipid synthesis. Presently, the SCD1 knockout mouse may be a useful model to study the role of endogenous oleoyl-CoA in lipid metabolism, and the studies described here may have broad implications for the potential use of SCD1 as a target in the treatment of some eye diseases.