Glycerolipid Biosynthesis in Mouse Liver and Ehrlich

Proceedings of the National Academy of Sciences Vol. 68, No. 2, pp. 411415, February 1971 The Acyl Dihydroxyacetone Phosphate Pathway for Glycerolipid Biosynthesis in Mouse Liver and Ehrlich Ascites Tumor Cells BERNARD W. AGRANOFF AND ATMIYA K. HAJRA Department of Biological Chemistry and Mental Health Research Institute, Ann Arbor, Mich. 48104 Communicated by J. L. Oncley, December 7, 1970 University of Michigan, ABSTRACT The glycerol portion of lipids may be derived biosynthetically by reduction of dihydroxyacetone phosphate to glycerolphosphate, and then be acylated with fatty acids or, alternatively, dihydroxyacetone phosphate may first be acylated and then reduced to I-acyl sn glycerol- 3-phosphate. Since the former pathway utilizes NADH for reduction of the C-2 carbonyl, while the latter requires NADPH, we were able to compare the relative participation of the two pathways for phospholipid synthesis by measuring the incorporation of radioactivity from tritiumlabeled NADH and NADPH into C-2 of lipid glycerol. The acyl-dihydroxyacetone phosphate pathway plays a significant role in glycerolipid synthesis in mouse liver homogenates and a clearly dominant one in Ehrlich ascites tumor cell homogenates. This finding is related to a reported lack of glycerol-3-phosphate dehydrogenase in tumor cells and to their high glycerol ether lipid content. Long-chain acyl-dihydroxyacetone phosphate (DHAP) was first observed in experiments on rapid labeling of mitochondrial lipids with [32P]Pi and [7-"2P]ATP (1). After identification of the novel lipid and its chemical synthesis (2), we found enzymic activity in subcellular particulate preparations from guinea pig liver for its biosynthesis (3) from acyl-coa and DHAP and for the enzymic reduction of the acyl-dhap formed to-1-acyl an glycerol-3-phosphate (lysophosphatidate) by NADPH but not by NADH (4). Liver mitochondrial and microsomal preparations catalyzed further acylation of the lysophosphatidate to phosphatidate (5). A pathway was proposed whereby phosphatidate is synthesized from DHAP without the participation of glycerolphosphate dehydrogenase (GPDH, L-glycerol-3-phosphate:NAD oxidoreductase EC 1.1.1.8), or of i-glycerol-3-phosphate (6). Estimates of the extent of participation of the two pathways are inferential. The level of enzymes for the alternative acyl-dhap pathway is comparable to that for those of the established glycerolphosphate pathway (3, 5). An appealing aspect of the new pathway lies in its selectivity for saturated fatty acids for the acylation of DHAP, which might account for the observed preponderance of saturated fatty acids on C-1 of glycerol of natural phospholipids (5). Acylation of glycerol-3-phosphate in titro in the presence of saturated and unsaturated acyl-coas produces a random distribution between carbons 1 and 2 (7). Despite such indirect suggestions that the acyl-dhap pathway might be important, there has been no direct or more quantitative evidence for its role. Abbreviations: GPDH, glycerol-3-phosphate dehydrogenase (EC 1.1.1.8); DHAP, dihydroxyacetone phosphate. 411 To clarify this question, we have prepared specifically labeled pyridine-nucleotide cofactors. Since GPDH requires NADH, while the acyl-dhap reductase requires NADPH, the labeled substrates should be specific for the two pathways. In the present report, we have analyzed the radioactivity in the hydrogen at carbon-2 of glycerol obtained from lipids after they were incubated with the appropriate 3H-labeled coenzyme. MATERIALS AND METHODS [B-3H]NADPH* was prepared by the method of Pastore and Friedkin (9). Two mci of [1-'H]glucose (New England Nuclear, 792 Ci/mol in 2.0 ml of water) were added to 0.375 ml of 0.1 M glycyl-glycine buffer (ph 7.5), 0.15 ml of 0.1 M MgC12, 0.3 ml of 0.1 M ATP neutralized with KHCO3, 0.3 ml of NADP (Boehringer, 10 mg/ml), 0.3 ml of hexokinase (Boehringer, diluted with 4 volumes of water), and 0.06 ml of glucose- 6-phosphate dehydrogenase (EC 1.1.1.49) (Boehringer). After 2 hr of incubation at room temperature, the mixture was diluted with 19 volumes of ice-cold distilled water and applied directly to a DEAE-cellulose column (9). The radioactive fractions corresponding to ['H]NADPH were pooled and then divided into 2-ml aliquots and stored frozen at -76 C. Individual tubes were thawed as needed and 1 ml was generally used for quantitative conversion to [B-'H ]NADH as follows. To 1 ml of the labeled NADPH were added 0.06 ml of 0.32 M ethanolamine* HCl buffer (ph 10.1) and 0.01 ml of calf intestinal alkaline phosphatase (EC 3.1.3.1) (Sigma, 20,ug in 0.3 ml of the ethanolamine buffer containing 0.01 M magnesium acetate). After incubation for 10 min at 37 C, the mixture was heated at 100 C for 2 min, then chilled and used as such for further incubations. Purity of the labeled NADPH and NADH were periodically checked by high-voltage electrophoresis. Samples (10,Al) were applied to Whatman No. 1 paper and allowed to migrate in 0.05 * When tritium is introduced into NADPH via enzymic reduction of NADP by [1-3H]glucose-6-phosphate, the isotope is present in the B-position. We found the [B-'H]NADPH to enzymically reduce acyl-dhap with the production of [2-3HJ-1- acyl en glycerol-3-phosphate, establishing utilization of the B hydrogen in this reduction. GPDH is a "B" specific enzyme (8). As expected, the [3H]NADH produced from [3H]NADPH by alkaline phosphatase labeled glycerolphosphate from DHAP in the presence of GPDH.

412 Biochemistry: Agranoff and Hajra FIG. 1. Comparison of radioactivity in lipids after 10-min incubations of ascites cell and liver homogenates from the same mice with [B-3H]NADPH (1.85 X 106 cpm, 2.65 nmol) (open bars) or [B-3H]NADH of the same specific activity (striped bars). Unlabeled NADPH and NADH (360 nmol of each) were present in all tubes (see text). Additions as indicated were GIc, D-glucose 1.0,umol; FDP, fructose-1,6-diphosphate 1.0 smol; DHAP, dihydroxyacetone phosphate, 1.2,umol. The liver homogenate incubation contained 14.8 mg of protein and the ascites cell homogenates contained 6.63 mg of protein. M sodium tetraborate buffer (ph 8.7) for 25 min at 4500 V. Carrier nucleotides were detected by ultraviolet absorption and by phosphate spray. Samples were incinerated in a Packard tritium oxidizer and the radioactive water produced was counted by scintillation techniques. Acyl-DHAP was prepared as described (2). Ehrlich ascites cells were the generous gift of Dr. H. N. Christensen and were a hypotetraploid strain. They were harvested 1 week after injection into male Swiss strain mice and purified by a modification of the method of Horvat and Touster (10). Cells from two donor mice were combined, centrifuged at 80 X g for 5 min, and then washed 3 times by centrifugation at 80 X g for 5 min with 4 volumes of 0.9% NaCl at 2 C to remove red cells. The tumor cells were packed at 800 X g for 10 min and then suspended in 10 volumes of 0.25 M sucrose. Cells were disrupted by sonication at 2 C in a Branson sonifier at position 4 for two 30-sec periods. The supernatant mixture obtained after centrifugation at 800 X g for 10 min was used for incubation. Livers (3.8 g) from the 2 mice were homogenized in 4 volumes of 0.25 M sucrose; the supernatant homogenate obtained after centrifugation at 800 X g for 10 min was used for incubation. For both ascites cells and liver homogenates, incubations were in air at 37 C with shaking in 20 X 150 mm test tubes. Each incubation vessel contained 0.5 ml of homogenate, 42 mm potassium phosphate buffer (ph 7.0), 0.3 mm NADH, 0.3 mm NADPH, 1.2 mm ATP, 1 mm MgC12, 0.035 mm coenzyme A, 9 mm cysteine, and 70.5 mm NaF (11) in a total volume of 1.2 ml. After incubation for 10 min, 4.5 ml of methanol-chloroform 2: 1 was added to each tube, followed by 0.1 ml of 6 N HCl. Lipid extraction, counting, column chromatography, and alkaline methanolysis are described elsewhere (12). Partition after alkaline methanolysis distributed glycerol and glycerol phosphoesters into the upper (aqueous) phase, while uncleaved lipids and methyl esters of fatty acids were distributed into the lower (organic) phase. For assignment of positional distribution of radioactivity in glycerol, the upper phase obtained after alkaline methanolysis was subjected to periodate degradation (13). To aliquots of deionized upper phase (12) in 2.5 ml were added 50,; of an aqueous solution containing 5 mg of glycerol, 1.5 ml of 0.1 N sodium acetate buffer (ph 5), and 1.0 ml of 0.2 M sodium metaperiodate. The mixture was kept in the dark at room temperature for 1 hr, then 0.5 ml of 0.2 M sodium arsenite was added and the mixture agitated. 2 ml of Dimedone (20 mg/ml in 50% ethanol) were then added and the mixture was kept at 20C overnight. Absolute ethanol, 0.5 ml, was layered over the surface of each tube and the Dimedone precipitate was sedimented at 10,000 X g for 15 min in Corex tubes in the SS-34 rotor of a Sorvall RC2B centrifuge. The pellet was dissolved in 2 ml of toluene and was backwashed with 2 ml of 0.01 M glycerolphosphate, dried, then counted after the addition of scintillation fluors. The backwash was combined with the Dimedone supernatant fraction, neutralized with KOH, and a portion was dried and counted in 10 ml of toluene scintillant containing 1.0 ml of BBS-3 solubilizer (Beckman). Formation of labeled lipid RESULTS Mouse liver and ascites cell homogenates were incubated for 10 min with either [B-3H]NADPH or [B-3H]NADH, and the additions indicated in Fig. 1. The presence of glycolytic precursors of glycerol differentially stimulated tritium uptake from NADPH, yet had a smaller and opposite effect on incorporation of radioactivity from NADH. Labeled NADPH uptake was relatively high in ascites cell homogenates, although significant uptake from both precursors was seen. To further establish the nature of the labeled lipid formed and the location of the radioactivity within each lipid class, the incubation products shown in Fig. 1 were separated by silica gel chromatography (Table 1) into those eluted with chloroform (fraction I), with chloroform-methanol 7:3 (fraction II), or with chloroform-methanol 1:2 (fraction III). Recovery into the three fractions was 65-76% of the starting extract of lipids from Ehrlich ascites cell homogenates and 77-85% for lipids from mouse liver. Almost all of the recovered radioactivity was in fractions I and II. Fraction I contained free fatty acids in addition to neutral lipids (mainly cholesterol, diglyceride, triglyceride, and glycerol ethers). In general, radioactivity was predominantly in fraction II, and was attributable to the presence of labeled phosphatidate, but also, when present, to labeled lysophosphatidate and acyl-dhap. Fraction III contained phosphatidylcholine, phosphatidyl ethanolamine, etc., and was not extensively labeled during the 10-min incubation. Alkaline methanolysis Proc. Nat. Acad. Sci. USA After methanolysis of liver or of ascites cell lipids labeled by incubation with [B-3H]NADPH, small but significant amounts of radioactivity were found in the organic phase. Thin-layer chromatography indicated that the radioactivity was almost entirely in a region in which methyl esters of fatty acids migrated. The [B-'H]NADPH probably labels fatty acids effectively because both de novo synthesis of fatty acids and partial reversal of 3-oxidation may utilize NADPH. In liver homogenates, labeling from [B-'H]NADH was almost entirely in the upper phase. 95% or more of the recovered radioactivity was water-soluble after methanolysis, and no radioactivity was found to comigrate on thin-layer chromatographs with the methyl esters of fatty acids.

Vol. 68, 1971 Acyl Dihydroxyacetone Phosphate Pathway 413 TABLE 1. Incorporation of radioactivity into lipid fractions from [B-3H]NADPH or [B-3H]NADH Ehrlich ascites cells lipid, cpm Liver lipid, cpm Labeled precursor Addition F I F II F III F I F II F III NADPH 750 (18)* 2,900 (43) 450 5,288 (42) 8,413 (86) 3000 NADH - 1238 (47) 4,825 (78) 300 15,850 (95) 25,238 (99) 3538 NADPH Glucose 1463 (33) 5,738 (64) 863 5,363 (39) 9,313 (80) 3450 NADH Glucose 1388 (47) 4,563 (62) 388 16,175 (94) 27,488 (98) 3863 NADPH FDPt 3188 (75) 9,563 (92) 863 7,588(57) 11,325 (90) 3175 NADH FDP 1213 (53) 4,050 (68) 288 11,525 (94) 19,687 (98) 3063 NADPH DHAP 4400 (75) 11,400 (87) 1563 11,963 (67) 18,450 (89) 4613 NADH DHAP 1263 (57) 3,238 (65) 350 10,400 (94) 18,650 (98) 2525 Lipids from incubations described in Fig. 1 were separated on a silica gel column. The various fractions were then subjected to alkaline methanolysis. * Numbers in parentheses are % of recovered radioactivity in the aqueous phase after partition of the methanolysate. t Fructose-1,6-diphosphate. Degradation of glycerol The question then remained, what was the nature of the tritium-labeled material in the upper phase? Via either the glycerolphosphate or the acyl-dhap pathway, radioactivity from the labeled pyridine nucleotides should reside in the hydrogen linked to the C-2 of glycerol. In order to confirm this, the aqueous phases after methanolysis of fraction I from the liver and ascites cell incubations, in which either glucose or DHAP had been added, were degraded with periodate (Table 2). This phase should contain free glycerol, which periodate oxidizes to form formaldehyde from C-1 and C-3, while formate is formed from C-2. With [B-'H]NADPH, 95% of the recovered radioactivity in fraction I from Ehrlich ascites cell lipids was, as expected, at C-2, but considerable label from [B-$H]NADH appeared to be in position 1 and/or 3. With the upper-phase material obtained from alkaline methanolysis of fraction I from liver, there was even more marked labeling in C-1 + 3 (Table 2). Upper layers from fraction II methanolysates contain mainly labeled glycerolphosphate. Periodate converted only C-1 and its hydrogen to formaldehyde, which was precipitated with Dimedone, while the supernatant solution contained both C-2 and C-3. A separate portion of the upper layer from fraction II was treated with alkaline phosphatase to release glycerol, which was subsequently degraded with periodate. In this instance, C-1 and C-3 are separated from C-2. The combined periodate oxidations before and after alkaline phosphatase treatment of the fraction II methanolysates permitted us to evaluate the radioactivity in C-1, C-2, and C-3 separately. No radioactivity was present in C-3. While [B-3H]- NADH labeled C-2 predominantly, all of the remaining radioactivity appeared to be in C-1. The net incorporation of [B-8HJNADH and [B-1H]NADPH into C-2 of glycerol (Table 2) can be established from the total radioactivity in the upper layer, calculated from Table 1, and the fraction of the counts in C-2 shown in Table 2. Dehydrogenase activities GPDH was measured by study of the disappearance of absorbance at 340 nm in incubations with NADH, and DHAP, and was found to be 2.2 nmol/min per mg protein in the 100,000 X g supernatant fraction from Ehrlich ascites cells and 119 nmol/min per mg protein for liver homogenate supernatant. Presence of the acyl-dhap:nadph reductase in Ehrlich ascites cells was inferred from studies in which acyl- [82P]DHAP was incubated in the presence and absence of NADPH, and the radioactive lipids present after incubation were identified by thin-layer chromatography. The NADPH-dependent formation of phosphatidate was observed. TABLE 2. Distribution of recovered radioactivity in lipid glycerol Distribution of radioactivity in % Labeling IFraction II Labeled FractionI FractionII Total radioactivity (cpm) in [3HI NADPH Source of lipid precursor Addition C-2 C-1 + C-3 C-1 0-2 C-3 0-2 from F I + F II 13H]NADH ~~~~~~~~~~ratio Ehrlich ascites NADPH Glucose 94.8 5.2 7.8 92.2 0 3,928 1.59 NADH Glucose 79.2 20.8 31.0 69.0 0 2,468 NADPH DHAP 99.3 0.7 3.4 96.6 0 12,858 8.18 NADH DHAP 57.9 42.1 45.2 54.8 0 1,571 Liver NADPH Glucose 87.6 12.4 4.0 96.0 0 8,985 0.266 NADH Glucose 73.5 26.5 16.3 83.7 0 33,723 NADPH DHAP 99.0 1.0 1.5 98.5 0 24,110 1.24 NADH DHAP 59.1 40.9 25.0 75.0 0 19,486 Formaldehyde produced by periodate oxidation was precipitated with Dimedone. Fraction II was, in addition, oxidized after alkaline phosphatase treatment (see text).

414 Biochemistry: Agranoff and Hajra Proc. Nat. Acad. Sci. USA Possible interconversion of nucleotide coenzymes After a 10-min incubation, 20-1ul aliquots of homogenate were applied directly to paper previously impregnated with borate buffer and subjected to high-voltage electrophoresis. No significant amount of conversion of labeled NADH to NADPH, either by transhydrogenase or lyase-kinase activity, was noted. Studies on whole cells Because of a report that NADH might enter Ehrlich ascites cells (14), washed whole cells were incubated with the two labeled pyridine nucleotides. Results similar to that seen in homogenates were obtained for the lipid fractions, but there was much reduced incorporation. Labeled oxidized or reduced pyridine nucleotides were not detected inside the cells, indicating no penetration by the labeled nucleotides. Identification of labeled lipids Over two-thirds of the labeled lipid was present in fraction II; thin-layer chromatography showed that most of the radioactivity comigrates with phosphatidate. A few additional percent migrated in an area corresponding to lysophosphatidate + acyl-dhap, which could not be separated in the solvent system used (chloroform-methanol-acetic acid-water 100:40:12:4). Fraction I lipids were separated on silica gel G with diethyl ether-benzene-ethanol-acetic acid 40:50:2:0.2 and, in each instance, counts were found corresponding to migration rates for triglyceride and for the region corresponding to diglyceride + fatty acid. No evidence for labeled glycerol ether was found in these experiments, although added batyl alcohol is eluted in fraction I. DISCUSSION Alternate biosynthetic pathways in biological systems are difficult to evaluate quantitatively. Their relative contribution to carbon flow in vivo is always highly inferential. Substrate concentration, as well as turnover rates, can effect the ultimate participation of alternative pathways, and homogenates are of necessity unphysiological, since cellular components and enzymes are diluted. Because of the inability of the substrates used to enter whole cells, we were obliged to use broken cell preparations. It seemed that the ascites tumor cell was a good candidate for involvement of the acyl-dhap pathway because of its known deficiency of GPDH (15). This was borne out in the present study. By using substrates that would each selectively reflect activity of one of the two pathways, and by degrading the labeled lipids formed, we were able to demonstrate participation of the acyl-dhap pathway as distinguished from the glycerolphosphate pathway. Under the conditions used, no complicating transhydrogenase activity was noted. NADPH and NADH both contribute radioactivity to the aqueous and to the organic phase obtained after alkaline methanolysis of lipids. Much of the activity in the organic phase is present in fatty acids. NADPH is probably a superior precursor of fatty acids because both de novo synthesis of fatty acid and partial reversal of a-oxidation can lead to NADPH radioactivity entering fatty acid. The lipid phase of fraction I methanolysates might also contain labeled cholesterol and labeled glycerol ether. Since the alkyl-dhap is reduced by NADPH in brain (16), its dephosphorylation could lead to the presence of labeled glycerol ether in the lower layer, although none was observed in the present studies. Almost no radioactivity from [B-3H]NADPH entered C-1 or C-3. In contrast, [B-3H]NADH, particularly in liver, was extensively incorporated into C-1. [B-3H]NADH may have entered glycerol via glyceraldehyde-3-phosphate in a direction reverse of glycolysis. The combined triosephosphate isomerase (EC 5.3.1.1) and glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.9) equilibria favor formation of DHAP, and the [1-3H]DHAP could then be converted to [1-3H]glycerolphosphate via either the glycerolphosphate or the acyl-dhap pathways. Experiments from several laboratories (17, 18) in which [1,3-14C]- and [2-3H]glycerol were incubated with various bacterial preparations, have given no indication of relative loss of 3H in the resultant lipid. This might have been expected had glycerol been converted to DHAP prior to incorporation into lipid. In fact, an increase in the 3H/'4C ratio is seen in the resulting labeled lipid. These experiments require that glycerol be phosphorylated and oxidized via enzymes not involved in the acyl-dhap pathway, i.e., glycerolkinase (EC 2.7.1.30) and glycerolphosphate oxidase (EC 1.1.2.1), so they would not appear to bear directly on the question. Our present finding that NADH radioactivity can be incorporated into C-1 of glycerol suggests that the glycerol in such double-label experiments should be degraded to establish the location of the tritium in the lipid glycerol. The present experiments appear to give direct evidence for the acyl-dhap pathway in the mouse. Several possible complicating factors in interpreting our findings should nevertheless be considered. Despite the very small amount of GPDH present in Ehrlich ascites cells, they have been reported to contain substantial amounts of glycerolphosphate (19). It may be that while lipids are made de novo via acyl-dhap, there is considerable recycling of glycerolphosphate generated by lipid degradation. Alternatively, glycerolphosphate may be formed in these cells by phosphorylation of free glycerol furnished to the cell exogenously via the peritoneal fluid as free glycerol or as glycerolipid. Nevertheless, the present results indicate that de novo lipid glycerol derived via glucose, generally considered to be the usual source of intracellular lipid glycerol, is formed in part via the acyl-dhap pathway in mouse liver, and as the major route in the Ehrlich ascites cells. The dominanceof the acyl-dhap pathway in this tumor cell is of interest partly because it was predicted by the low concentration of glycerolphosphate dehydrogenase present. Perhaps of greater interest is the possible relationship of this dominance to the presence of glycerol ether lipids in many tumor cells, including the Ehrlich ascites cell (20, 21). It has recently been shown in brain that acyl-dhap is a precursor of alkyl- DHAP and that the latter is reduced to alkyl-glycerolphosphate via NADPH (16). Brain also contains significant amounts of glycerol ethers (22) and is relatively deficient in GPDH (15). The present findings thus affirm the relationship of the acyl-dhap pathway and the presence of glycerol ethers. A possible relationship between the diminished GPDH and the presence of the acyl-dhap pathway and lipid ethers in a tumor cell remains to be elucidated. The effects of altered membrane properties on cell growth have been the subject of considerable thought (23). The methods reported here for the preparation of [B-3H]NADH from [B-3H]NADPH make available a convenient tool for the further study of the glycerolphosphate and the acyl-dhap pathways for the forma- tion of phosphatidate and hence, all of the glycerolipids.

Vol. 68, 1971 We thank Mr. Edward Seguin for his skilled technical assistance. This investigation was supported by USPHS grant NB- 03101. 1. Hajra, A. K., and B. W. Agranoff, J. Biol. Chem., 242, 1074 (1967). 2. Hajra, A. K., and B. W. Agranoff, J. Biol. Chem., 243, 1617 (1968). 3. Hajra, A. K., J. Biol. Chem., 243, 3458 (1968). 4. Hajra, A. K., and B. W. Agranoff, J. Biol. Chem., 243, 3542 (1968). 5. Hajra, A. K., Biochem. Biophys. Res. Commun., 33, 929 (1968). 6. Agranoff, B. W., J. A. Benjamins, and A. K. Hajra, in Cyclitols and Phosphoinositides, ed. G. Hauser, Ann. N.Y. Acad. Sci., 165, 755 (1969). 7. De Jimenez, E. S., and W. W. Cleland, Biochim. Biophys. Acta, 176, 685 (1969). 8. Dixon, M., and E. C. Webb, Enzymes, 2nd Ed. (Academic Press, New York, 1964). 9. Pastore, E. J., and M. Friedkin, J. Biol. Chem., 236, 2314 (1961). 10. Horvat, A., and 0. Touster, Biochim. Biophys. Acta, 148, 725 (1967). Acyl Dihydroxyacetone Phosphate Pathway 415 11. Vaughan, M., J. Lipid Res., 2, 4 (1961). 12. Hajra, A. K., E. B. Seguin, and B. W. Agranoff, J. Biol. Chem., 243, 1609 (1968). 13. Hanahan, D. J., and J. N. Olley, J. Biol. Chem., 231, 813 (1958). 14. Letnansky, K., and G. M. Klc, Arch. Biochem. Biophys., 130, 218 (1969). 15. Boxer, G. E., and C. E. Shonk, Cancer Res., 20, 85 (1960). 16. Hajra, A. K., Biochem. Biophys. Res. Commun., 39, 1037 (1970). 17. Hill, E. E., and W. E. M. Lands, Biochim. Biophys. Acta, 202, 209 (1970). 18. Plackett, P., and A. W. Rodwell, Biochim. Biophys. Acta, 210, 230 (1970). 19. Hess, B., and B. Wright, eds., Control Mechanisms in Respiration and Fermentation (Ronald Press, New York, 1963), p. 333. 20. Snyder, F., E. A. Cress, and N. Stephens, Lipids, 1, 381 (1966). 21. Bollinger, J. N., Lipids, 2, 143 (1967). 22. Rapport, M. M., and W. T. Norton, Annu. Rev. Biochem. (Annual Reviews, Inc., Palo Alto, Calif., 1962), p. 103. 23. Wallach, D. F. H., New Eng. J. Med., 280, 761 (1969).