Phosphocholine synthesis in spinach: Characterization of phosphoethanolamine N-methyltransferase

Phosphocholine is a precursor for phosphatidylcholine or it may be hydrolysed to choline. Choline can be oxidized to form the compatible osmolyte glycine betaine which is accumulated by many plants under conditions of osmotic stress. In Spinacia oleracea phosphocholine is synthesized by 3 sequential N‐methylations of phosphoethanolamine with the first step catalysed by the enzyme phosphoethanolamine N‐methyltransferase (EC 2.1.1.103). This enzyme has been partially purified 5400‐fold from spinach leaves using a combination of ammonium sulphate fractionation, followed by chromatographic separations on DEAE‐Sepharose, phenyl‐Sepharose, Ω‐aminohexyl‐agarose, Mono Q and adenosine‐agarose. Sodium dodecyl sulphate‐polyacrylamide gel electrophoresis (SDS‐PAGE) separation and silver‐staining of the final preparation revealed several polypeptides present, only one of which with an estimated molecular mass of 54 kDa could be photoaffinity cross‐linked to the substrate [³H] S‐adenosyl‐l ‐methionine. HPLC gel permeation chromatography was used to obtain an estimate for the native molecular mass of 77 kDa. Enzyme activity was optimal at pH 7.8 in HEPES‐KOH buffer, it was inhibited by S‐adenosyl‐l ‐homocysteine, phosphocholine, phosphate, Mn²⁺ and Co²⁺ but not by ethanolamine, methylethanolamine, dimethylethanolamine, choline, glycine betaine or Mg²⁺. Using phosphoethanolamine as substrate, the final preparation had a specific activity of 189 nmol mg^?1 protein min^?1. The reaction products were identified and their relative abundance estimated following separation by TLC as phosphomethylethanolamine (87%), phosphodimethylethanolamine (10%) and phosphocholine (2%). Thus, a highly purified preparation of phosphoethanolamine N‐methyltransferase was shown to catalyse 3 successive N‐methylations of phosphoethanolamine. Photoaffinity cross‐linking of proteins extracted from leaves of spinach followed by SDS‐PAGE and autoradiography shows that a 54‐kDa radiolabelled polypeptide was more prominent in extracts from salinized plants and barely visible in extracts from plants exposed to prolonged dark periods, a pattern which corresponds to the salt and light‐responsive changes in phosphoethanolamine N‐methylating activity. Thus, the production of phosphocholine for glycine betaine accumulation in spinach can be mediated by a single phosphobase N‐methyltransferase which is more abundant in salt‐stressed plants.     

Effect of choline deficiency on the enzymes that synthesize phosphatidylcholine and phosphatidylethanolamine in rat liver

Activities have been determined in subcellular fractions of livers from choline-deficient and normals rats for the enzymes that convert choline and ethanolamine to phosphatidylcholine and phosphatidylethanolamine respectively, that methylate phosphatidylethanolamine to yield phosphatidylcholine, and that oxidize choline to betaine. The activities of ethanolamine kinase, phosphoethanolamine cytidylyltransferase, and CDP-ethanolamine: 1,2-diacylglycerol phosphoethanolaminetransferase are not changed in the livers from choline-deficient rats for at least 18 days. Similarly, the activities of choline kinase and CDP-choline: 1,2-diacylglycerol phosphocholine transferase were unaffected by choline depletion. A decrease of 30-41% was observed, however, in the mitochondrial oxidation of choline to betaine. Also, the activity of the phosphocholine cytidylyltransferase was reduced in the choline-deficient livers to 60% olf the control values. The only observed increase in enzyme activity was a 62% elevation of the phosphatidylethanolamine-S-adenosylmethionine methyltransferase activity after 2 days of choline deficiency. This increased activity was maintained for at least 18 days of choline deprivation. The results suggest a lack of adaptive change in the levels of these phospholipid biosynthetic enzymes as a result of choline deficiency.     

Synthesis of methylated ethanolamine moieties: regulation by choline in lemna

The results of experiments in which intact plants of Lemna paucicostata were labeled with either l-[(3)H(3)C]methionine, l-[(14)CH(3)]methionine, or [1,2-(14)C]ethanolamine support the conclusion that growth in concentrations of choline of 3.0 micromolar or above brings about marked decreases in the rate of biosynthesis of methylated forms of ethanolamine (normally present chiefly as phosphatidylcholine, with lesser amounts of choline and phosphocholine). The in vivo locus of the block is at the committing step in the biosynthetic sequence at which phosphoethanolamine is methylated by S-adenosylmethionine to form phosphomethylethanolamine. The block is highly specific: flow of methyl groups originating in methionine continues into S-adenosylmethionine, S-methylmethionine, the methyl moieties of pectin methyl ester, and other methylated metabolites. When choline uptake is less than the total that would be synthesized by control plants, phosphoethanolamine methylation is down-regulated to balance the uptake; total plant content of choline and its derivatives remains essentially constant. At maximum down-regulation, phosphoethanolamine methylation continues at 5 to 10% of normal. A specific decrease in the total available activity of AdoMet: phosphoethanolamine N-methyltransferase, as well as feedback inhibition of this enzyme by phosphocholine, and prevention of accumulation of phosphoethanolamine by down-regulation of ethanolamine synthesis may each contribute to effective control of phosphoethanolamine methylation. This down-regulation may necessitate major changes in S-adenosylmethionine metabolism. Such changes are discussed.     

Enzymes of Choline Synthesis in Spinach (Response of Phospho-Base N-Methyltransferase Activities to Light and Salinity)

In spinach (Spinacia oleracea L.), choline is synthesized by the sequential N-methylation of phosphoethanolamine -> phosphomono- -> phosphodi- -> phosphotrimethylethanolamine (i.e. phosphocholine) followed by hydrolysis to release choline. Differential centrifugation of spinach leaf extracts shows that enzymes catalyzing the three N-methylations are cytosolic. These enzymes were assayed in leaf extracts prepared from plants growing under various light/dark periods. Under a diurnal, 8-h light/16-h dark photoperiod, the activity of the enzyme catalyzing the N-methylation of phosphoethanolamine is highest at the end of the light period and lowest following the dark period. Prolonged dark periods (exceeding 16 h) lead to a further reduction in the activity of this enzyme, although activity is restored when plants are reexposed to light. In contrast, the activity of the enzyme(s) catalyzing the N-methylations of phosphomono- and phosphodimethylethanolamine does not undergo comparable changes in response to light/dark treatments. Salt shock of plants with 200 mM NaCl results in a 2-fold increase in all three N-methylation activities relative to nonsalinized controls but only in plants exposed to light. Thus, light is required for the salt-responsive up-regulation of choline synthesis in spinach.     

Phosphoethanolamine bases as intermediates in phosphatidylcholine synthesis by lemna

The pathway for synthesis of phosphatidylcholine, the dominant methyl-containing end product formed by Lemna paucicostata, has been investigated. Methyl groups originating in methionine are rapidly utilized by intact plants to methylate phosphoethanolamine successively to the mono-, di-, and tri-methyl (i.e. phosphocholine) phosphoethanolamine derivatives. With continued labeling, radioactivity initially builds up in these compounds, then passes on, accumulating chiefly in phosphatidylcholine (34% of the total radioactivity taken up by plants labeled to isotopic equilibrium with l-[(14)CH(3)]methionine), and in lesser amounts in soluble choline (6%). Radioactivity was detected in mono- and dimethyl derivatives of free ethanolamine or phosphatidylethanolamine only in trace amounts. Pulse-chase experiments with [(14)CH(3)]choline and [(3)H] ethanolamine confirmed that phosphoethanolamine is rapidly methylated and that phosphocholine is converted to phosphatidylcholine. Initial rates indicate that methylation of phosphoethanolamine predominates over methylation of either phosphatidylethanolamine or free ethanolamine at least 99:1. Although more studies are needed, it is suggested this pathway may well turn out to account for most phosphatidylcholine synthesis in higher plants. Phosphomethylethanolamine and phosphodimethylethanolamine are present in low quantities during steady-state growth (18% and 6%, respectively, of the amount of phosphocholine). Radioactivity was not detected in CDP-choline, probably due to the low steady-state concentration of this nucleotide.     

Lipid metabolism in T47D human breast cancer cells: 31P and 13C-NMR studies of choline and ethanolamine uptake.

31P and 13C-NMR were used to determine the kinetics of choline and ethanolamine incorporation in T47D clone 11 human breast cancer cells grown as small (150 microns) spheroids. Spheroids were perfused inside the spectrometer with 1,2-13C-labeled choline or 1,2-13C-labeled ethanolamine (0.028 mM) and the buildup of labeled phosphoryl-choline (PC) or phosphorylethanolamine (PE) was monitored. Alternatively the PC and GPC pools were prelabeled with 13C and the reduction of label was monitored. 31P spectra were recorded from which the overall energetic status as well as total pool sizes could be determined. The ATP content was 8 +/- 1 fmol/cell, and the total PC and PE pool sizes were 16 and 14 fmol/cell, respectively. PC either increased by 50% over 24 h or remained constant, while PE remained constant in medium without added ethanolamine but increased 2-fold within 30 h in medium containing ethanolamine, indicating a dependence on precursor concentration in the medium. The 31P and 13C data yielded similar kinetic results: the rate of the enzymes phosphocholine kinase and phosphoethanolamine kinase were both on the order of 1.0 fmol/cell per h, and the rate constants for CTP:phosphocholine cytidyltransferase and CTP:phosphoethanolamine kinase were 0.06 h-1 for both enzymes. The kinetics of choline incorporation did not alter in the presence of 0.028 mM ethanolamine indicating that they have non-competing pathways.     

Effect of salt stress on the metabolism of ethanolamine and choline in leaves of the betaine-producing mangrove species Avicennia marina

Glycinebetaine synthesis from [methyl-14C]choline and [1,2-14C]ethanolamine in leaf disks of Avicennia marina, was increased by salt stress (250 and 500 mM NaCl). After 18 h incubation with [methyl-14C]choline, phosphocholine and CO(2) were found to be heavily labelled. Phosphocholine contained 39% of the total radioactivity taken up by non-salinised (control) leaf disks and 15% of the total for salinised leaf disks stressed with 500 mM NaCl. Eighteen and 49% of the radioactivity absorbed by control and salinised disks, respectively, were released as CO(2). Metabolic studies of [1,2-14C]ethanolamine revealed that the radioactivity taken up by the leaf disks was recovered as the following compounds after 18 h: phosphorylated compounds (mainly phosphoethanolamine, phosphodimethylethanolamine and phosphocholine) (40-50%); choline (1-2%); glycinebetaine (3-5%); lipids (20-28%); CO(2) (6-10%). Unlike glycinebetaine, incorporation into phosphorylated compounds and lipids were reduced by salt stress. Incorporation of [methyl-14C]S-adenosyl-L-methionine (SAM) into choline, phosphocholine and glycinebetaine in leaf disks was stimulated by salt stress. In vitro activities of adenosine kinase and adenosine nucleosidase, which are implicated in stimulating the SAM regeneration cycle, increased after the leaf disks were incubated with 250 and 500 mM NaCl for 18 h. Changes in metabolism involving choline and glycinebetaine due to salt stress are discussed.     

Synthesis of methylated ethanolamine moieties: regulation by choline in soybean and carrot

The results of experiments in which intact plants of Lemna paucicostata were labeled with either l-[³H₃C]methionine, l-[¹⁴CH₃]methionine, or [1,2-¹⁴C]ethanolamine support the conclusion that growth in concentrations of choline of 3.0 micromolar or above brings about marked decreases in the rate of biosynthesis of methylated forms of ethanolamine (normally present chiefly as phosphatidylcholine, with lesser amounts of choline and phosphocholine). The in vivo locus of the block is at the committing step in the biosynthetic sequence at which phosphoethanolamine is methylated by S-adenosylmethionine to form phosphomethylethanolamine. The block is highly specific: flow of methyl groups originating in methionine continues into S-adenosylmethionine, S-methylmethionine, the methyl moieties of pectin methyl ester, and other methylated metabolites. When choline uptake is less than the total that would be synthesized by control plants, phosphoethanolamine methylation is down-regulated to balance the uptake; total plant content of choline and its derivatives remains essentially constant. At maximum down-regulation, phosphoethanolamine methylation continues at 5 to 10% of normal. A specific decrease in the total available activity of AdoMet: phosphoethanolamine N-methyltransferase, as well as feedback inhibition of this enzyme by phosphocholine, and prevention of accumulation of phosphoethanolamine by down-regulation of ethanolamine synthesis may each contribute to effective control of phosphoethanolamine methylation. This down-regulation may necessitate major changes in S-adenosylmethionine metabolism. Such changes are discussed.     

Phorbol esters modulate the turnover of both ether- and ester-linked phospholipids in cultured mammalian cells

The effects of 12-O-tetradecanoylphorbol 13-acetate (TPA) on the metabolism of ester- and ether derivatives of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were studied in HeLa and HEL-37 cells. TPA stimulated the incorporation of [3H]choline into diacyl-, alkylacyl- and alkenylacy/PC in HeLa cells, but inhibited the incorporation of [3H]ethanolamine into the corresponding derivatives of PE. TPA also stimulated the incorporation of [3H]ethanolamine into lysoPE and the release of labelled ethanolamine and phosphoethanolamine from HeLa cells prelabelled with [3H]ethanolamine. All responses to TPA were abolished in HeLa cells preincubated with the phorbol ester and which were deficient in protein kinase C. In HEL-37 cells TPA stimulated label incorporation into both ester- and ether-forms of PE. The marked effects of TPA on ether-lipid metabolism raises the possibility that hydrolysis products of this class of lipid are important in transmembrane signalling pathways.     

Several lines of evidence demonstrating that Plasmodium falciparum, a parasitic organism, has distinct enzymes for the phosphorylation of choline and ethanolamine

In Plasmodium falciparum-infected erythrocyte homogenates, the specific activity of ethanolamine kinase (7.6 +/- 1.4 nmol phosphoethanolamine/10(7) infected cells per h) was higher than choline kinase specific activity (1.9 +/- 0.2 nmol phosphocholine/10(7) infected cells per h). The Km of choline kinase for choline was 79 +/- 20 microM, and ethanolamine was a weak competitive inhibitor of the reaction (Ki = 92 mM). Ethanolamine kinase had a Km for ethanolamine of 188 +/- 19 microM, and choline was a competitive inhibitor of ethanolamine kinase with a very high Ki of 268 mM. Hemicholinium 3 inhibited choline kinase activity, but had no effect on ethanolamine kinase activity. In contrast, D-2-amino-1-butanol selectively inhibited ethanolamine kinase activity. Furthermore, when the two enzymes were subjected to heat inactivation, 85% of the choline kinase activity was destroyed after 5 min at 50 degrees C, whereas ethanolamine kinase activity was not altered. Our results indicate that the phosphorylation of choline and ethanolamine was catalyzed by two distinct enzymes. The presence of a de novo phosphatidylethanolamine Kennedy pathway in P. falciparum contributes to the bewildering variety of phospholipid biosynthetic pathways in this parasitic organism.     

Rate-limiting steps in the cytidine pathway for the synthesis of phosphatidylcholine and phosphatidylethanolamine.

An analysis of the available data on the cytidine pathway for the synthesis of phosphatidylcholine and phosphatidylethanolamine, by the logic derived from the theoretical principles of metabolic regulation, shows that the first two reactions catalysed by choline (ethanolamine) kinase and phosphocholine (phosphoethanolamine) cytidylyltransferase are rate-limiting, whereas the phosphocholine (phosphoethanolamine) transferase step is near equilibrium in rat liver.     

A comparison of the reversibility of phosphoethanolamine transferase and phosphocholine transferase in rat brain microsomes

The reversibility of phosphoethanolamine transferase (EC 2.7.8.1) in rat brain is demonstrated in this paper. Microsomal ethanolamine glycerophospholipids were prelabeled with an intracerebral injection of [3H]ethanolamine 4 h before killing young rats. Labeled CDPethanolamine was produced by incubation of the microsomes with CMP, although to a lesser extent than for the previously observed release of CDPcholine. Ethanolamine and choline glycerophospholipids were labeled with [2-3H]glycerol by incubation with primary cultures of rat brain. Microsomes from rat brains, with diisopropyl phosphofluoridate for inhibition of lipases, were incubated with the labeled glycerophospholipids separately, and labeled diacylglycerols were produced. The kinetic parameters of phosphoethanolamine transferase and phosphocholine transferase (EC 2.7.8.2) were compared by incubating rat brain microsomes with [3H]CMP. Inclusion of AMP in the reaction mixture was necessary in order to inhibit the hydrolysis of CMP by an enzyme with the properties of 5'-nucleotidase (EC 3.1.3.5). For phosphoethanolamine transferase and phosphocholine transferase respectively, the Km values for CMP were 40 and 125 microM and the V values were 2.3 and 21.6 nmol/h per mg protein. The reversibility of both enzymes permits the interconversion of the diacylglycerol moieties of choline and ethanolamine glycerophospholipids. During brain ischemia, a principal pathway for degradation of ethanolamine glycerophospholipids may be by reversal of phosphoethanolamine transferase followed by hydrolysis of diacylglycerols by the lipase.     

Choline kinase and ethanolamine kinase activity in the cytosol of nerve endings from rat forebrain.

Both choline kinase and ethanolamine kinase are present in the cytosol of nerve endings prepared from rat brain are the products of their action, phosphocholine (84 nmol/g fresh wt. of brain) and phosphoethanolamine (190 nmol/g fresh wt. of brain). In contrast with the enzymes from the cytosol of whole brain, both are as equally active at pH 7.5 as 9.0. Determination of kinase activity in membrane-containing tissue samples at pH9 gives low values because of the activity of alkaline phosphatase. Choline kinase, but not ethanolamine kinase, requires Mg2+ in excess of that required for the formation of the MgATP complex and is inhibited by an excess of free ATP. The Km for choline is 2.6mM and for ethanolamine is 2.2mM. The differing requirements for ATP and Mg2+ and the inhibition of choline kinase, but not ethanolamine kinase, by hemicholinium-3 suggest either the presence of two separate enzymes or two different active sites on the same enzyme.     

Synthesis of Ethanolamine and Its Regulation in Lemna paucicostata

The metabolism of ethanolamine and its derivatives in Lemna paucicostata has been investigated, with emphasis on the path-way for synthesis of phosphoethanolamine, a precursor of phosphatidylcholine in higher plants. In experiments involving labeling of intact plants with radioactive serine, ambiguities of interpretation due to entry of radioactivity into methyl groups of methylated ethanolamine derivatives were mitigated by pregrowth of plants with methionine. Difficulties due to labeling of diacylglyceryl moieties of phospholipids were avoided by acid hydrolysis of crucial samples and determination of radioactivity in isolated serine or ethanolamine moieties. The results obtained from such experiments are most readily reconciled with the biosynthetic sequence: serine → ethanolamine → phosphoethanolamine → phosphatidylethanolamine. A possible alternative is: serine → phosphatidylserine → phosphatidylethanolamine → ethanolamine → phosphoethanolamine. Cell-free extracts of L. paucicostata were shown to produce CO₂ from the carbon originating as C-1 of serine at a rate sufficient to satisfy the demand for ethanolamine moieties. A number of experiments produced no support for a hypothetical role for phosphoserine in phosphoethanolamine formation. Uptake of exogenous ethanolamine commensurately down-regulates the synthesis of ethanolamine moieties (considered as a whole, and regardless of their state of derivatization at the time of their formation). In agreement with previous observations, uptake of exogenous choline down-regulates the methylation of phosphoethanolamine, without being accompanied by secondary accumulation of a marked excess of ethanolamine derivatives.     

cDNA cloning of phosphoethanolamine N-methyltransferase from spinach by complementation in Schizosaccharomyces pombe and characterization of the recombinant enzyme

The N-methylation of phosphoethanolamine is the committing step in choline biogenesis in plants and is catalyzed by S-adenosyl-L-methionine:phosphoethanolamine N-methyltransferase (PEAMT, EC ). A spinach PEAMT cDNA was isolated by functional complementation of a Schizosaccharomyces pombe cho2(-) mutant and was shown to encode a protein with PEAMT activity and without ethanolamine- or phosphatidylethanolamine N-methyltransferase activity. The PEAMT cDNA specifies a 494-residue polypeptide comprising two similar, tandem methyltransferase domains, implying that PEAMT arose by gene duplication and fusion. Data base searches suggested that PEAMTs with the same tandem structure are widespread among flowering plants. Size exclusion chromatography of the recombinant enzyme indicates that it exists as a monomer. PEAMT catalyzes not only the first N-methylation of phosphoethanolamine but also the two subsequent N-methylations, yielding phosphocholine. Monomethyl- and dimethylphosphoethanolamine are detected as reaction intermediates. A truncated PEAMT lacking the C-terminal methyltransferase domain catalyzes only the first methylation. Phosphocholine inhibits both the wild type and the truncated enzyme, although the latter is less sensitive. Salinization of spinach plants increases PEAMT mRNA abundance and enzyme activity in leaves by about 10-fold, consistent with the high demand in stressed plants for choline to support glycine betaine synthesis.     

Regulation of mitogenesis by water-soluble phospholipid intermediates.

Many recent observations implicate choline and ethanolamine kinases as well as phosphatidylcholine-specific phospholipase C in the regulation of mitogenesis and carcinogenesis. For example, human cancers generally contain high concentrations of phosphoethanolamine and phosphocholine, and in different cell lines various growth factors, cytokines, oncogenes and chemical carcinogens were all shown to stimulate the formation of phosphocholine and phosphoethanolamine. In addition, other reports have appeared showing that both extracellular and intracellular phosphocholine as well as ethanolamine and its derivatives can regulate cell growth. This area of research has clearly arrived at a stage when it becomes important to examine critically the feasibility of water-soluble phospholipid intermediates serving as potential regulators of cell growth in vivo. Accordingly, the goal of this review is to summarise available information relating to the formation and mitogenic actions of intracellular and extracellular phosphocholine as well as ethanolamine and its derivatives.     

Radiotracer evidence implicating phosphoryl and phosphatidyl bases as intermediates in betaine synthesis by water-stressed barley leaves

In barley, glycine betaine is a metabolic end product accumulated by wilted leaves; betaine accumulation involves acceleration of de novo synthesis from serine, via ethanolamine, N-methylethanolamines, choline, and betaine aldehyde (Hanson, Scott 1980 Plant Physiol 66: 342-348). Because in animals and microorganisms the N-methylation of ethanolamine involves phosphatide intermediates, and because in barley, wilting markedly increases the rate of methylation of ethanolamine to choline, the labeling of phosphatides was followed after supplying [¹⁴C]ethanolamine to attached leaf blades of turgid and wilted barley plants. The kinetics of labeling of phosphatidylcholine and betaine showed that phosphatidylcholine became labeled 2.5-fold faster in wilted than in turgid leaves, and that after short incubations, phosphatidylcholine was always more heavily labeled than betaine. In pulse-chase experiments with wilted leaves, label from [¹⁴C]ethanolamine continued to accumulate in betaine as it was being lost from phosphatidylcholine. When [¹⁴C]monomethylethanolamine was supplied to wilted leaves, phosphatidylcholine was initially more heavily labeled than betaine. These results are qualitatively consistent with a precursor-to-product relationship between phosphatidylcholine and betaine.     

C Tracer Evidence for Synthesis of Choline and Betaine via Phosphoryl Base Intermediates in Salinized Sugarbeet Leaves

Like other chenopods, sugarbeets (Beta vulgaris L. cv Great Western D-2) accumulate glycine betaine when salinized; this may be an adaptive response to stress. The pathway of betaine synthesis in leaves of salinized (150-200 millimolar NaCl) sugarbeet plants was investigated by supplying [¹⁴C]formate, phosphoryl[¹⁴C]monomethylethanolamine ([¹⁴C][unk] MME) or phosphoryl[¹⁴C]choline ([¹⁴C][unk] choline) to leaf discs and following ¹⁴C incorporation into prospective intermediates. The ¹⁴C kinetic data were used to develop a computer model of the betaine pathway.

When [¹⁴C]formate was fed, [unk] MME, phosphoryldimethylethanolamine ([unk] DME) and [unk] choline were the most prominent methylated products at short labeling times, after which ¹⁴C appeared in free choline and in betaine. Phosphatidylcholine labeled more slowly than [unk] choline, choline, and betaine, and behaved as a minor end product. Very little ¹⁴C entered the free methylethanolamines. When [¹⁴C][unk] MME was supplied, a small amount was hydrolyzed to the free base but the major fate was conversion to [unk] DME, [unk] choline, free choline, and betaine; label also accumulated slowly in phosphatidylcholine. Label from supplied [¹⁴C][unk] choline entered choline and betaine rapidly, while phosphatidylcholine labeled only slowly and to a small extent.

These results are consistent with the pathway [unk] MME →[unk] DME → [unk] choline → choline → → betaine, with a minor side branch leading from [unk] choline into phosphatidylcholine. This contrasts markedly (a) with the pathway of stress-induced choline and betaine synthesis in barley, in which phosphatidylcholine apparently acts as an intermediate (Hitz, Rhodes, Hanson 1981, Plant Physiol 68: 814-822); (b) with choline biogenesis in mammalian liver and microorganisms. Computer modeling of the experimental data pointed strongly to regulation at the [unk] choline → choline step, and also indicated that the rate of [unk] choline synthesis is subject to feedback inhibition by [unk] choline.