Identification and Quantitation of the Fatty Acids Composing the CoA Ester Pool of Bovine Retina, Heart, and Liver*

Several proteins found in retinal photoreceptor cells (guanylate cyclase activating protein, protein kinase A, recoverin, and transducin) are N-terminally modified with the fatty acids 12:0, 14:0, 14:1n-9, and 14:2n-6, whereas similar proteins in other tissues contain only 14:0. It has been hypothesized that the acyl-CoA pool of the retina contains amounts of 12:0, 14:1n-9, and 14:2n-6 elevated over 14:0, in comparison to other tissues, and this accounts for the specificity of N-terminal fatty acylation. To test this hypothesis, we performed fatty acid analysis on total acyl-CoAs purified from bovine retina (light-adapted), heart, and liver. We also examined theN- and S-linked fatty acid composition of the total protein pools from these tissues. Acyl-CoAs were prepared from heart, liver, and retina and separated by high performance liquid chromatography (HPLC). Identities of peaks were based on HPLC of standard 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs. Total protein was subjected to base hydrolysis followed by acidic methanolysis to release S- and N-linked fatty acids, respectively, and fatty acid phenacyl esters were prepared for HPLC analysis. Retina had levels of 12:0 (2.7 ± 2.1%), 14:1n-9 (2.9 ± 2.2%), and 14:2n-6 (1.6 ± 0.7%) CoAs below that of 14:0 CoA (7.0 ± 1.8%). Likewise, heart levels of 14:2n-6 CoA (3.7 ± 0.1%) were near and 12:0 (2.6 ± 0.6%) and 14:1n-9 (0.7 ± 0.3%) CoAs were below that of 14:0 CoA (3.8 ± 1.0%). Liver had levels of 12:0 (16.1 ± 5.7%) and 14:2n-6 (8.1 ± 1.2%) CoAs above and 14:1n-9 CoA (1.2 ± 0.6%) below that of 14:0 CoA (5.9 ± 0.8%). Fatty acid analysis of total protein showed that all tissues contained S-linked 16:0, 18:0, and 18:1n-9. Retina proteins contained N-linked 14:0, 14:1n-9, and 14:2n-6, whereas heart and liver had only 14:0. Our findings do not support the hypothesis that the CoA ester pool of the retina is enriched with 12:0, 14:1n-9, and 14:2n-6 over 14:0, in comparison to other tissues. This suggests that alternative models must be considered for the regulation of N-terminal fatty acylation of proteins in photoreceptor cells.

It can be hypothesized that the retina contains elevated amounts of 12:0, 14:1n-9, and 14:2n-6 CoAs over 14:0 CoA compared with other tissues, accounting for the specificity of heterogenous N-terminal acylation (6). We describe here experiments designed to test this hypothesis by analyzing the fatty acid composition of the acyl-CoA pools from bovine heart, liver, and retina. As a complementary study, we determined the 12:0, 14:0, 14:1n-9, and 14:2n-6 content of proteins from bovine heart, liver, and retina. Our results show that all three tissues contain CoA derivatives of 12:0, 14:1n-9, and 14:2n-6, yet these fatty acids are N-acylated only to retinal proteins. Thus, the heterogeneous N-terminal fatty acylation in retina may reflect unusual synthesis and utilization of 12:0, 14:1n-9, and 14:2n-6 CoAs, rather than acyl-CoA pool composition. Tissue Preparation--Bovine eyes, heart, and liver were obtained from Mikkelson Beef Inc. (Oklahoma City, OK). Eyes were obtained within 2 h of slaughter and kept at 4°C. Heart and liver were obtained within 30 min of slaughter and finely sectioned and chilled rapidly in wet ice. For final processing, heart and liver were cut into 1-2-g cubes, washed in ice-cold phosphate-buffered saline (138 mM NaCl, 2.7 mM KCl, 5.4 mM Na 2 HPO 4 , and 1.8 mM KH 2 PO 4 , pH 7.2), and frozen in liquid nitrogen. Eyes were light-adapted on ice for 3-4 h in room light. Retinas were dissected from eyes, washed in ice-cold phosphate-buffered saline, and frozen in liquid nitrogen.
Fatty Acid Analysis of Total Protein (Base Hydrolysis and Acidic Methanolysis)-In the analysis, two retina protein pellets and one heart or liver protein pellet were used per sample. To remove free lipids, the protein pellets were washed two times with 10 ml of MeOH, two times with 10 ml of CHCl 3 /MeOH/12 N HCl (100:100:1 v/v/v), four times with 10 ml of CHCl 3 /MeOH (2:1 v/v), and two times with 10 ml of EtOH. The protein was dried in a Speed Vac concentrator (Savant, Hicksville, NY).
For release of thioester-linked fatty acids, the protein was saponified in 6 ml of 2% KOH/EtOH (w/v) at 37°C for 4 h. Following addition of 2 ml of H 2 O and 300 l of 12 N HCl, the hydrolysate was extracted three times with 4 ml of hexane. The hexane extracts were combined and used to prepare fatty acid phenacyl esters (FAPEs) by the method of Wood and Lee (28). FAPEs were separated on HPLC using a Supelco (Bellefonte, PA) Supelcosil TM LC-18 column (25 cm ϫ 4.6 mm inner diameter), with elution (2 ml/min) by a linear gradient of acetonitrile/H 2 O started at 80/20 (v/v), increased to 92:8 after 45 min, held at 92:8 for 10 min, and returned to 80:20 after 5 min. FAPE elution was monitored by absorbance at 242 nm. After adding 6 ml of EtOH to the aqueous phase to decrease polarity, the protein was pelleted by centrifugation, and the supernatant was discarded. The protein pellet was washed once with 10 ml of EtOH, twice with 10 ml CHCl 3 /MeOH (2:1 v/v), and three times with 10 ml MeOH. The protein was dried in a Speed Vac concentrator and subjected to acidic methanolysis as described below.
To release amide-linked fatty acids, the proteins were hydrolyzed in 6 ml of 2 N HCl/83% MeOH at 100°C for 6 h, under nitrogen. The hydrolysate was extracted three times with 4 ml of hexane. Fatty acid methyl ester extracts were dried under nitrogen and saponified in 2 ml of 2% KOH/EtOH at 100°C for 60 min. After adding 1 ml H 2 O and 100 l of 12 N HCl, the hydrolysate was extracted with three times with 2 ml of hexane. FAPEs were prepared and subjected to HPLC as described above.
Catalytic Hydrogenation of FAPEs-FAPE peaks from fatty acids released during base hydrolysis and acidic methanolysis were collected, 20 nmol of 17:0 FAPE was added as carrier, and the total FAPEs were extracted with hexane. The FAPEs were saponified with 2% KOH/ EtOH as described above. Free fatty acids were solubilized in 2 ml of EtOH/hexane (2:1) and bubbled with hydrogen (20 min) in the presence of ϳ10 mg of Pt 2 O (Matheson, Norwood, OH). Spent Pt 2 O was removed by centrifugation. Fatty acids were phenacylated and separated on HPLC as described above, with the modification of a 30-min elution  3. Acyl-CoA species making up the CoA ester pool of bovine retina, heart, and liver. A, bar graph summarizing the relative percentages for each identified acyl-CoA species found in HPLC profiles for retina (n ϭ 12). B, bar graph for the identifiable acyl-CoA species found in heart (n ϭ 2) and liver (n ϭ 2).
FIG. 4. HPLC of heart acyl-CoA extract before and after mixing with standards and after base hydrolysis. A, HPLC profiles for the extract made from ϳ1.5 g of heart. Identities for peaks were based on comparison to R F values for standards (not shown) and simultaneous HPLC with standards. B, HPLC profile for the same heart extract after mixing with standards. C, HPLC profiles for heart extract after subjecting to mild alkaline hydrolysis.
A HPLC profile for heart acyl-CoA extracts after base hydrolysis is shown in Fig. 4C. Complete disappearance was seen for all peaks except 17:0 CoA, supporting the idea that they are thioester-linked lipids. The resistance of the 17:0 CoA candidate suggests that it may not be an acyl-CoA.
The average percentage of each identifiable acyl-CoA species in the extracts (n ϭ 2) is shown in Fig. 3B. The most abundant is 16:0 CoA (37.8 Ϯ 1.  (Table I). Total acyl-CoA yield was estimated to be 4.6 Ϯ 0.8 nmol/g of wet weight.
Acyl-CoA Composition of Bovine Liver--A HPLC profile for acyl-CoAs from liver is shown in Fig. 5A. Identification of the peaks was carried out using R F values as described previously. FIG. 5. HPLC of liver acyl-CoA extract before and after mixing with standards and after base hydrolysis. A, HPLC profiles for acyl-CoA extract made from ϳ1.5 g of liver. Identities for peaks were based on comparison to R F values for standards (not shown) and simultaneous HPLC with standards. B, HPLC profile for the same liver extract after mixing with standards. C, HPLC profiles for liver extract after subjecting to mild alkaline hydrolysis.
Peaks near the appropriate R F positions for all the acyl-CoA standards were discernible. The HPLC profile for the extract after mixing with the standards is shown in Fig. 5B. Co-elution with the acyl-CoA standards was seen for the peaks found in the extract, lending support to their identity assignments. There were also several prominent peaks in the trace that did not match the R F value or co-elute with any standard, making identity assignment impossible.
A HPLC profile for liver acyl-CoA extracts after base hydrolysis is shown in Fig. 5C. Complete disappearance was seen for all the acyl-CoA candidate peaks, as well as those that were unidentifiable, supporting the idea that they are thioesterlinked lipids.
FIG. 7. Acidic methanolysis of total protein from bovine retina, heart, and liver. HPLC profiles of FAPEs for fatty acids released from retina, heart, and liver total protein by acidic methanolysis. Control represents contaminant fatty acids in reagents. Identities of all peaks were based on retention times for standard FAPEs (profile not shown). min, the identities of which are unknown. The short chain fatty acids 12:0, 14:1n-9, and 14:2n-6 were not found in any of the base hydrolysates.
Based on the catalytic efficiency (V max /K m ) data for the human NMT (20), there should be at least 2.6-and 3.1-fold higher amounts of 12:0 and 14:1n-9 CoAs, respectively, relative to 14:0 CoA for equal utilization. Activity (V max ) and dissociation constant (K d ) data for the yeast NMT toward 14:0 and 14:2n-6 CoAs (22) suggest that, minimally, 18.9-and 5.0-fold higher amounts, respectively, than 14:0 CoA would be required for equal incorporation. Although the amounts of 12:0, 14:1n-9, and 14:2n-6 CoAs in heart are consistent with the acylation of heart proteins with only 14:0, the levels of these CoAs in the retina are definitely not high enough to lead to the heterogeneous acylation pattern seen for photoreceptor proteins. Likewise, the relative amounts of 12:0 and 14:2n-6 CoAs in liver suggest that these fatty acids should be found on liver proteins in addition to 14:0 (104 and 7-28% of 14:0, respectively), but this is not the case (35). Consequently, it is very unlikely heterogeneous N-terminal fatty acylation is determined by differences in the acyl-CoA composition of the tissues.
Our results suggesting the presence of 12:0, 14:1n-9, and 14:2n-6 CoAs in heart and liver led us to question whether there might exist heterogeneously acylated proteins in these tissues that have yet to be identified. We performed fatty acid analysis on the total protein precipitates from heart, liver, and retina. Base hydrolysis released considerable 16:0 from all of the proteins, which is consistent with the thioester modification of proteins by 16:0 (15). We also observed the release of 14:0, 18:0, 18:1n-9, and 18:2n-6, which can also be S-linked to proteins (36,37). Although the representative palmitoylated proteins in heart and liver are uncertain, the major modified protein in retina is likely the photoreceptor protein rhodopsin (38). Acidic methanolysis of heart and liver total protein showed release of 14:0 but not of 12:0, 14:1n-9, and 14:2n-6. Retina total protein released 14:0, 14:1n-9, and 14:2n-6, consistent with the presence of these fatty acids in photoreceptor proteins. Thus, whereas all three tissues may contain 14:1n-9 and 14:2n-6 CoAs, only retinal proteins contained these two fatty acids in an amide linkage.