INTRODUCTION

Coenzyme A (CoA) is the major fatty acid carrier molecule in plant and animal cells. Fatty acids are bound to CoA through a thioester linkage, the formation of which is carried out in mammalian cells by long chain acyl-CoA synthetases (1, 2). Fatty acylated CoAs (acyl-CoAs) play many diverse functional roles in cells (for reviews, see Refs 1 and 3-5); in this investigation, acyl-CoAs served as substrates for enzymes that acylate proteins. Several proteins found in retinal photoreceptor cells (guanylate cyclase-activating protein, protein kinase A, recoverin, and transducin) are N-terminally acylated with 12:0, 14:0, 14:1n-9, and 14:2n-6, whereas similar proteins in other tissues contain only 14:0 (6). In the case of guanylate cyclase-activating protein, recoverin, and transducin, the percentages of 12:0, 14:1n-9, and 14:2n-6 present are near or greater than that of 14:0 (6). Consequently, it appears modification with 12:0, 14:1n-9, and 14:2n-6 is a retina-specific phenomenon (6) with potential functional significance as described for recoverin (7).

N-terminal acylation of proteins is carried out by N-terminal myristoyltransferase (NMT)¹ (for reviews, see Refs. 8-19). Human and yeast NMTs are known to have a high substrate affinity for 14:0 CoA (2, 20, 21). Although 14:2n-6 CoA has not been tested, 12:0 and 14:1n-9 CoAs are utilized by the human NMT at a 3.1- and 1.9-fold lower catalytic efficiency, respectively (V_max/K_m), than 14:0-CoA (20). A recent examination of acyltransferase activity for yeast NMT toward 12:0, 14:1n-9, and 14:2n-6 CoAs, compared with 14:0 CoA (22), showed equal activity for 14:1n-9 CoA and significantly lower activity for 12:0 and 14:2n-6. Overall, the studies for NMTs show that 12:0, 14:1n-9, and 14:2n-6 CoAs are not superior substrates compared with 14:0 CoA. Consequently, to compete for the NMT, levels of available 12:0 and 14:2n-6 CoAs must be higher and 14:1n-9 at least equal to that of 14:0 CoA.

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.

EXPERIMENTAL PROCEDURES

12:0, 14:0, 16:0, 17:0, 18:0, 18:1n-9, 18:2n-6, 18:3n-3, and 20:4n-6 CoA standards were from Sigma-Aldrich. [³H]14:0, 14:1n-9, 14:2n-6, [³H]16:0, and 22:6n-3 CoAs were synthesized according to the method of Hajra and Bishop (23, 24). [³H]14:0, [³H]16:0, and 22:6n-3 used in the synthesis were from NEN Life Science Products, American Radiolabeled Chemicals, Inc. (St. Louis, MO), and Sigma-Aldrich, respectively. 14:1n-9 and 14:2n-6 were generously provided by Dr. H. Sprecher (Department of Medical Biochemistry, Ohio State University, Columbus, OH). All synthesized acyl-CoAs were judged to be ~100% pure by HPLC (described below), thin layer chromatography (silica gel 60 plates (EM Science, Philadelphia, PA) using a butanol/HOAc/H₂O (50:30:20, v/v/v) solvent system), and liquid chromatography/mass spectroscopy (courtesy of Dr. K. Jackson, University of Oklahoma Health Sciences Center). KH₂PO₄, acetonitrile, CHCl₃, hexane, isopropanol, and MeOH were HPLC grade, and 12 N HCl was Optima^TM grade (Fisher). Ethanol was HPLC grade, and KOH was semiconductor grade (Sigma-Aldrich). All other chemicals were of reagent grade (Fisher or Sigma-Aldrich).

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 Na₂HPO₄, and 1.8 mM KH₂PO₄, 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.

The extraction procedure was a modification of the methods of Woldegiorgis et al. (25) and Corkey (26). Frozen tissue was finely powdered by grinding in a mortar chilled by dry ice; the amount per extraction was ~1.5 g of heart, ~1.5 g of liver, and two retinas. Powdered frozen tissue was suspended in 5 ml of isopropanol, 50 mM KH₂PO₄, pH 7.2 (1:1 v/v) and warmed to 4 °C. For some samples, extraction efficiency was determined by adding 25,000-100,000 dpm (0.05-0.2 nmol) [³H]14:0 CoA or [³H]16:0 CoA. In some retina samples, 29,000 dpm (0.06 nmol) [³H]14:0 CoA and 5 nmol each of 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs were added. The samples were acidified with 100 µl of glacial acetic acid, warmed to 23 °C, and extracted twice with 4 ml of hexane/isopropanol (2:1 v/v) and then twice with 3 ml of hexane. The hexane phases containing nonpolar lipids were discarded. To the aqueous phase was added 200 µl of saturated (NH₄)₂SO₄ and 10 ml MeOH/CHCl₃ (2:1 v/v) to form a monophase. After 20 min, the precipitate was removed by centrifugation and washed once with 4 ml of MeOH/CHCl₃/H₂O (5:2.5:1 v/v/v). Protein pellets were saved for analysis of covalently bound fatty acids. To the pooled supernatant, 4 ml of water was added to form a stable bilayer. The CHCl₃ phase containing glycerolipids was removed and combined with subsequent CHCl₃ extracts. The aqueous phase was extracted twice more with 3 ml of CHCl₃, the pooled CHCl₃ phases were back-extracted twice with 1.5 ml of H₂O, and the water phases added to the previous aqueous phase. The CHCl₃ phases were then discarded. After adding ~10 mg of butylated hydroxy toluene, the aqueous pool was taken to dryness by nitrogen gas evaporation and lyophilization (VirTis lyophilizer, VirTis, Gardiner, NY). The residue, containing acyl-CoAs, was solubilized in 0.5-1.0 ml of 25 mM KH₂PO₄ (pH 5.3) and passed through a 0.45 µm filter (nylon-66 Microfilterfuge^TM, Rainin Instruments, Inc., Woburn, MA); this material was used for HPLC analysis.

HPLC was carried out following the modified methods of Woldegiorgis et al. (25) and Corkey (26). Standards or acyl-CoA extracts from tissues (entire extract injected) were separated using a Waters (Milford, MA) Nova-Pak^TM C-18 column (60 A, 4 µm, and 3.9 mm × 15 cm inner diameter). Elution (2 ml/min) was done using a linear gradient of 25 mM KH₂PO₄ (pH 5.3)/acetonitrile started at 70:30 (v/v), decreased to 54:46 after 20 min, decreased again to 38:62 after 5 min, held at 38:62 for 20 min, and returned to 70:30 after 5 min. Acyl-CoA elution was monitored by 254 nm absorbance. Recovery of [³H]14:0 or [³H]16:0 CoAs was quantitated with an on-line scintillation counter (Flo-One Radioactive Flow Detector A-200, Radiomatic, Tampa, FL) using Ultima-Flo M mixture (Packard Instrument Co., Inc., Meriden, CT) at a 2.5:1 (v/v) ratio with the column eluant.

The procedure used was a modification of the methods of Corkey (27). Acyl-CoA extracts from heart, liver, and retina were adjusted to pH 12.5 by addition of 10 M KOH and heated for 10 min at 56 °C, with agitation. After cooling on ice, the pH was readjusted to 5.3 by addition of 70% (w/v) HClO₄. After centrifugation to remove the KClO₄ precipitate, the extract was filtered and subjected to HPLC as described above. This procedure was carried out on a mixture of 12:0, 14:0, 14:1n-9, 14:2n-6, 16:0, 17:0, 18:0, 18:1n-9, 18:2n-6, 18:3n-3, 20:4n-6, and 22:6n-3 CoAs; the posthydrolysis HPLC profile (data not shown) showed complete disappearance of these standards.

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 CHCl₃/MeOH/12 N HCl (100:100:1 v/v/v), four times with 10 ml of CHCl₃/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 H₂O and 300 µl 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 × 4.6 mm inner diameter), with elution (2 ml/min) by a linear gradient of acetonitrile/H₂O 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 CHCl₃/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 H₂O and 100 µl 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.

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 ~10 mg of Pt₂O (Matheson, Norwood, OH). Spent Pt₂O was removed by centrifugation. Fatty acids were phenacylated and separated on HPLC as described above, with the modification of a 30-min elution with 100% acetonitrile immediately following 98:2 acetonitrile/H₂O, to elute saturated fatty acids larger than 18 carbon atoms.

RESULTS

Fig. 1A is a HPLC profile for standard acyl-CoAs, chromatographed in parallel with the retina acyl-CoA extracts. Under our HPLC conditions, 12:0, 14:0, 14:1n-9, 14:2n-6, 16:0, 18:0, 18:1n-9, and 18:3n-3 acyl-CoAs eluted as individual peaks with moderate to complete base line separation. While retaining complete separation from the other standards, 18:2n-6, 20:4n-6, and 22:6n-3 acyl-CoAs co-eluted as an unresolved peak. Detection response at 254 nm was estimated (12:0, 14:0, and 16:0 CoAs, 2 nmol each) to be 250 ± 51 mV·s/nmol. For each standard, R_F values were calculated relative to 16:0 CoA because of its inclusion as an internal reference standard ([³H] 16:0 CoA) in the extractions.

[View Larger Version of this Image (31K GIF file)]

To determine the efficiency of our method, we added [³H]14:0 (retina) or [³H]16:0 (heart, liver, and retina) CoAs prior to extraction. Percentage of recovery of the [³H]16:0 CoA was quantitated from the obtained HPLC radioactivity profiles. A radioactivity tracing from the retina extract is shown in Fig. 1C. Recovery of [³H]16:0 CoA was 12.5 ± 1.8, 34.4 ± 1.7%, and 35.9 ± 11.0% for heart (n = 2), liver (n = 2), and retina (n = 7), respectively.

We also examined the efficiency of the extraction of 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs from the retina by adding 5 nmol each of 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs prior to extraction. Recovery was calculated by comparing the HPLC absorbance at 254 nm (data not shown) to that of the original mixture. The recoveries for 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs were 53.8 ± 4.0, 46.9 ± 1.4, 41.7 ± 7.1, and 39.5 ± 2.9%, respectively.

HPLC profiles of the acyl-CoAs from retina are shown in Figs. 1B and 2A. Identification of the peaks was made by comparing R_F values to those calculated for standard acyl-CoAs. Peaks near the R_F for 12:0, 14:0, 14:1n-9, 14:2n-6, 16:0, 18:0, 18:1n-9, 18:2n-6, 18:3n-3, 20:4n-6, and 22:6n-3 CoAs are discernible. 14:0 and 16:0 CoAs were also identified by their radioactivity profiles (Fig. 1C); both [³H]CoAs were recovered intact and aligned exactly with the corresponding absorbance peaks. To further confirm identity assignments, a sample of retina extract was subjected to HPLC after mixing with 12:0, 14:0, 14:1n-9, 14:2n-6, 16:0, 17:0, 18:0, 18:1n-9, 18:2n-6, 18:3n-3, 20:4n-6, and 22:6n-3 CoA standards (Fig. 2B). Co-elution with the standards lends support to their identity assignments.

[View Larger Version of this Image (37K GIF file)]

A HPLC profile for retina extracts after base hydrolysis is shown in Fig. 2C. Although disappearance was uncertain for the 12:0, 14:1n-9, and 14:2n-6 CoA candidates, complete disappearance was seen for all of the other peaks except 17:0 CoA, suggesting that they are thioester-linked lipids. The resistance of the 17:0 CoA candidate suggests that it may not be an acyl-CoA.

Fig. 3A shows the average percentage of each identifiable acyl-CoA found in the retina extracts (n = 12). The most abundant acyl-CoA is 16:0 (27.3 ± 2.6%), followed by 18:2n-6/20:4n-6/22:6n-3 (23.3 ± 5.5%), 18:0 (18.7 ± 6.2%), 18:1n-9 (14.0 ± 2.0%), 14:0 (7.0 ± 1.8%), 14:1n-9 (2.9 ± 2.2%), 12:0 (2.7 ± 2.1%), 18:3n-3 (2.6 ± 0.9%), and 14:2n-6 (1.6 ± 0.7%). The most noteworthy finding is that the relative amounts (Table I) of 12:0, 14:1n-9, and 14:2n-6 CoAs are less than that of 14:0 CoA.

[View Larger Version of this Image (25K GIF file)]

Table I. Relative ratios for 12:0, 14:1n-9, and 14:2n-6 CoAs to 14:0 CoA in acyl-CoA extracts from bovine retina, heart, and liver Percentages for 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs are found in Fig. 3. Tissue 12:0 CoA 14:1n-9 CoA 14:2n-6 CoA 14:0 CoA Retina 0.4 0.4 0.2 1.0 Heart 0.7 0.2 1.0 1.0 Liver 2.7 0.2 1.4 1.0

The total acyl-CoA content of the retina was estimated from HPLC profiles (n = 7). Total identifiable peak areas from each profile were adjusted to a theoretical 100% yield, based on [³H]16:0 recovery, and divided by the sample wet weight; the resulting values were averaged and divided by the detection response at 254 nm. The yield was estimated to be 1.9 ± 0.7 nmol/g of wet weight or 1.6 ± 0.5 nmol/retina.

A HPLC profile of acyl-CoAs from heart is shown in Fig. 4A. Identification of the peaks was carried out using R_F values as described previously. Peaks near the R_F positions for 12:0, 14:0, 14:1n-9, 14:2n-6, 16:0, 18:0, 18:1n-9, 18:2n-6, 18:3n-3, 20:4n-6, and 22:6n-3 CoA were discernible. Shown in Fig. 4B is the HPLC profile for the extract after mixing with the standards. Co-elution with the acyl-CoA standards lends support to their identity assignments. Co-elution with the 14:1n-9 CoA standard was not seen, making its existence in heart uncertain.

[View Larger Version of this Image (33K GIF file)]

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 = 2) is shown in Fig. 3B. The most abundant is 16:0 CoA (37.8 ± 1.5%), followed by 18:2n-6/20:4n-6/22:6n-3 (30.0 ± 0.6%), 18:0 (13.4 ± 1.0%), 18:1n-9 (5.3 ± 1.3%), 14:0 (3.8 ± 1.0%), 14:2n-6 (3.7 ± 0.1%), 18:3n-3 (3.0 ± 0.2%), 12:0 (2.6 ± 0.6%), and 14:1n-9 (0.7 ± 0.3%) CoAs. The levels of 12:0 and 14:2n-6 CoAs are both near that of 14:0 CoA, and the level of 14:1n-9 CoA is lower than that of 14:0 CoA (Table I). Total acyl-CoA yield was estimated to be 4.6 ± 0.8 nmol/g of wet weight.

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. 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.

[View Larger Version of this Image (32K GIF file)]

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 thioester-linked lipids.

The average percentage of each identifiable acyl-CoA species in the extracts (n = 2) is shown in Fig. 3B. The most abundant was 18:2n-6/20:4n-6/22:6n-6 CoA (35.3 ± 4.0%), followed by 16:0 (20.1 ± 0.9%), 12:0 (16.1 ± 5.7%), 18:3n-3 (9.3 ± 1.6%), 14:2n-6 (8.1 ± 1.2%), 14:0 (5.9 ± 0.8%), 18:1n-9 (2.4 ± 0.1%), 18:0 (1.8 ± 1.3%), and 14:1n-9 (1.2 ± 0.6%) CoAs. The levels of 12:0 and 14:2n-6 CoAs were greater than that of 14:0 CoA, and the level of 14:1n-9 CoA was lower than that of 14:0 CoA (Table I). Total acyl-CoA yield was estimated to be 5.3 ± 1.4 nmol/g of wet weight.

Shown in Fig. 6 are the HPLC profiles for FAPEs released by base hydrolysis of retina, heart, and liver total protein. Controls consisted of blank tubes on which all procedures were performed. All controls contained 16:0 and 18:0; their averaged areas were subtracted from those of the tissues prior to percentage calculations. Retina proteins released 16:0 (66.1 ± 1.1%), 18:1n-9 (11.4 ± 0.6%), and 18:0 (16.7 ± 0.9%), as well as trace amounts (<2%) of 14:0, 18:2n-6, and 22:6n-3. Heart proteins released 16:0 (40.5 ± 0.7%), 18:0 (31.0 ± 0.4%), 18:1n-9 (15.4 ± 0.4%), and 18:2n-6 (9.2 ± 0.3%), as well as a trace amount (<3%) of 14:0. Liver proteins released 16:0 (29.9 ± 2.4%), 18:0 (46.1 ± 3.2%), and 18:1n-9 (8.8 ± 0.8%), as well as trace amounts (<2%) of 14:0 and 18:2n-6. The liver protein also showed four peaks (11.3 ± 0.6% total) at 40-50 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.

[View Larger Version of this Image (29K GIF file)]

The HPLC profiles of the FAPEs released by acidic methanolysis of retina, heart, and liver total protein are shown in Fig. 7. All controls contained 14:0, 16:0, and 18:0; their averaged peak areas were subtracted from those of the tissues prior to percentage calculations. Retina proteins released 14:2n-6 (4.3 ± 1.3%), 14:1n-9 (4.0 ± 1.0%), 14:0 (20.4 ± 2.9%), 16:0 (31.1 ± 1.0%), and 18:0 (40.3 ± 6.1%). Heart protein, although devoid of detectable 14:1n-9 and 14:2n-6, released 14:0 (22.0 ± 9.3%), 16:0 (31.5 ± 2.4%), and 18:0 (46.5 ± 11.2%). The small peak in the area of 14:1n-9 did not yield 14:0 after catalytic hydrogenation (data not shown). Liver protein was devoid of detectable 14:1n-9 and 14:2n-6 released 14:0 (27.4 ± 5.5%), 16:0 (25.9 ± 2.5%), and 18:0 (45.3 ± 6.2%).

[View Larger Version of this Image (31K GIF file)]

HPLC profiles of the hydrogenated 14:1n-9 and 14:2n-6 peaks from methanolysis of retina total protein are shown in Fig. 8. Control samples, consisting of blank tubes on which all hydrogenation procedures were performed, were found to contain 14:0, 16:0, and 18:0. Identities of both 14:2n-6 and 14:1n-9 peaks were confirmed by production of 14:0 greater than that seen in the controls (2.5 ± 0.2 and 3.8 ± 0.4 times the control level, respectively); however, the 16:0, 17:0 (carrier), and 18:0 levels were not above those of the controls. The 14:0 released from retina proteins by acidic methanolysis was also hydrogenated (data not shown) and produced only 14:0 and 20:0 at levels 20.9 ± 2.0 and 2.7 ± 0.1 times, respectively, those found in the control. The presence of 20:0 suggests that the original 14:0 peak contained some underlying 20:4n-6 (1.4% of 14:0), which has been found to modify proteins on lysine residues (29).

[View Larger Version of this Image (24K GIF file)]

DISCUSSION

Although bovine tissues have not been examined, acyl-CoA compositional analyses have been carried out for many other mammals (25, 30-33). Only rat and pig heart (30) and rat liver and skeletal muscle (25, 30) were reported to have 12:0, 14:0, and 14:1 CoAs. No reports of 14:2 CoA have been made. Levels of 12:0 CoA in pig heart and 12:0 and 14:1 CoA in rat liver and skeletal muscle were below that of 14:0 CoA, whereas 14:1 CoA in pig heart and 12:0 and 14:1 CoA in rat heart were higher than that of 14:0 CoA. Pig and rat heart results are noteworthy because protein kinase A is exclusively modified with 14:0 in heart (6, 34), whereas in photoreceptors the same protein is N-terminally acylated with 12:0, 14:0, 14:1n-9, and 14:2n-6 (6).

Our examination of bovine heart, liver, and retina for 12:0, 14:0, 14:1n-9, and 14:2n-6 CoA content was essential in explaining how heterogeneous N-terminal fatty acylation arises. In heart and retina, 12:0, 14:0, 14:1n-9, and 14:2n-6 CoAs were sometimes difficult to distinguish from background; therefore, the reported percentages should be taken as an upper limit for their presence. Although our data are relative percentages, there was no obvious indication in heart and retina that 12:0, 14:1n-9, and 14:2n-6 CoAs exceed 14:0 CoA by a substantial amount. On the other hand, liver acyl-CoA levels were all significantly above background levels, showing amounts of 12:0 and 14:2n-6 CoAs slightly higher than that of 14:0 CoA, whereas 14:1n-9 CoA was somewhat lower. As in heart and retina, there was no dramatic enrichment of 12:0, 14:1n-9, and 14:2n-6 CoAs over 14:0 CoA in liver.

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.

In light of our results, alternative models must be considered for regulation of N-terminal fatty acylation in photoreceptors. Previously, we showed that the retina synthesizes 14:1n-9 and 14:2n-6 by retroconversion of 18:1n-9 and 18:2n-6 or 20:4n-6, respectively (39). We hypothesized that retina peroxisomes might be more active in producing 14:1n-9 and 14:2n-6. It is also possible that the retina acyl-CoA synthetase is more efficient at generating the required 18:1n-9, 18:2n-6, and 20:4n-6 CoAs. Interestingly, the retina has very high 20:4n-6 CoA synthetase activity (40, 41). Alternatively, the retina may contain an unique NMT isozyme that prefers 12:0, 14:1n-9, and 14:2n-6 CoAs. Consistent with this, bovine brain has been shown to contain NMT isoforms (42-44), and bovine brain and spleen (17) NMTs differ by 10 kDa in mass. Another option is a regulatory protein that makes the NMT more selective for 12:0, 14:1n-9, and 14:2n-6 CoAs. Proteins controlling the enzymatic rate of NMT have been identified (45-47). Finally, although NMTs are cytosolic proteins (48-51), the retina NMT could be compartmentalized where it is more accessible to 14:1n-9 and 14:2n-6 CoAs than 14:0 CoA. Our previous studies showed that 14:2n-6 could not be provided to the retina NMT through a direct cytoplasmic route (39), whereas 14:0 was utilized. It is likely that one or more of the above mentioned mechanisms for controlling NMT activity is at work in photoreceptors.