Analogues of choline and related compounds

• Main Author: Edwards, R. G.
• Published: University of Oxford 1973
• Subjects: 547
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.454443

The biological role of choline (I), and the metabolic pathways involved in the synthesis and degradation of lecithin (II) and other phosphatides, have been discussed. Reference has been made to current ideas on the properties of phospholipids, both in cell membranes and in isolation, and to the physical methods required for their evaluation. The published literature on the biological properties of choline analogues has been reviewed. Although the earliest investigations, mainly by Hunt and Renshaw, were concerned solely with pharmacological properties, the observation by Best et al. of such consequences of choline deficiency as fatty liver and renal haemorrhage stimulated the search for substances which could effectively replace choline in the diet, and. these lipotropic compounds included the phosphorus (III) and arsenic (IV) analogues. Welch claimed to have isolated the arsenic analogue of lecithin from the tissues of rats fed a choline-deficient arsenocholine (lV)-supplemented diet, but his evidence was not conclusive by modern standards. Moreover, knowledge of the lipotropic role of phosphocholine (III) was confined solely to the fact that its administration allayed the symptoms of choline deficiency. It was therefore decided to undertake a thorough chemical and biological investigation of this compound. The tri-n-butylphosphonium analogue (V) was used in all exploratory synthetic work, however, for reasons of convenience and economy. The chemical investigation involved the synthesis of the phosphorus analogues of those derivatives of choline involved in the cytidine pathway to lecithin, namely phosphorylphosphocholine (VI), cytidine-5andprime;-diphosphate phosphocholine (CDP-phosphocholine) (VII) and phosphatidyl-phosphocholine ("phospholecithin") (VIII). Of these, the synthesis of phosphorylphosphocholine proved the most difficult, as the standard methods of choline phosphorylation were unsuccessful when applied to the phosphorus analogue. Use was eventually made of a novel phosphorylating agent, a complex (IX) formed by methylphosphorodichloridate (X) in pyridine, discovered, by chance, by Smart and Catlin. Thus both phosphocholine and its tri-n-butyl homologue were phosphorylated in fair yield (31-40%) in a single-stage process. The synthesis of CDP-choline, in 24% yield, was a straightforward adaption of the method of Kennedy, and involved the condensation of phosphorylphosphocholine (VI) with, cytidine-5andprime;-monophosphate in the presence of dicyclohexylcarbodiimide (XI). Well-established methods of lecithin synthesis were considered unsuitable for the phosphonium compound, but a recent method of Aneja and Chadha, involving the condensation of choline and phosphatidic acid (XII) in the presence of triisopropylbenzenesulphonyl chloride (XIII) proved to be satisfactory, and in this way both 1,2-distearoyl-DL-phosphatidylphosphocholine (VIII, R ?(CH₂)₁₆CH₃) and its tri-n-butyl homologue were prepared in reasonable yields (33-35%). Some of the properties of these synthetic "phospholecithins" were compared with those of distearoylphosphatidylcholine; thus the phosphorus analogues were found to exhibit lower transition temperatures, and remarkably lower melting points (ca. 135°), than the nitrogen compound (m.p. 233°). Both the trimethylphosphonium and trimethylammonium compounds appeared to form similar spherical or ellipsoidal single-bilayer liposomes in aqueous dispersion, but those of the tri-n-butylphosphonium compound were unusually long and thin, like collapsed vesicles; this is possibly due to the less hydrophilic nature of the tri-n-butylphosphonium group. Both trimethyl compounds were attacked to a similar extent by phospholipase C, while the enzyme showed a far lower affinity for the tri-n-butylphosphonium lecithin. Since the phosphorus-31 nucleus has a spin of andfrac12;, and couples with the proton, both proton magnetic resonance (PMR) and ³¹P nuclear magnetic resonance (³¹P-NMR) studies of the phosphonium analogues were found to be of interest. Thus the methyl proton signal in the PMR spectrum of phosphocholine (III) is a doublet with a coupling constant of ca. 14.5 Hz, and is characteristic of all the trimethylphosphonium derivatives. This property later proved to be a useful method of structure confirmation when phosphatidylphosphocholine (VTII), synthesized in vivo, was isolated from rat tissues. As phosphorylphosphocholine (VI) and its derivatives were probably the first compounds to be prepared containing both a phosphonium and a phosphate group, they provided a unique opportunity for the simultaneous study of two very different phosphorus groups by ³¹P-NMR. The use of ³H-labelled choline and ¹⁴C-labelled phosphocholine provided a means of investigating the relative affinity of the enzymes of the cytidine pathway for phosphonium compounds. Thus it was found that phosphocholine was at least as acceptable as choline to the enzyme choline phosphokinase: ATP + choline → phosphorylcholine + ADP and was even the preferred substrate in the initial stages of competitive incubations near enzyme-saturation level. Phosphorylcholine-cytidyl-transferase: CTP + phosphorylcholine and ? CDP-choline + pyrophosphate was far more specific, and initial rates of conversion of the phosphorylated bases to the CDP derivatives indicated a threefold preference for the natural substrate. In the case of phosphorylcholine-diglyceride-transferase: CDP-choline + diglyceride and ? lecithin + CMP both phosphorus and nitrogen substrates were equally acceptable in the forward reaction, but phosphatidylphosphocholine, once formed, appeared far less susceptible to degradation by the reverse reaction. The in vitro investigation was extended to cell culture, using the P815Y line of neoplastic mouse mast cell. The cells showed similar growth characteristics in choline-supplemented and in phosphocholine-supplemented media, and extraction of lecithin from cells grown in the latter led to the conclusion that both bases were equally acceptable; radioactivity readings showed similar choline/phosphocholine ratios in the medium and in the extracted lecithin. Moreover, measurement of the relative areas of phosphate and phosphonium peaks in the ³¹P-NMR spectrum of the extract confirmed this finding. Finally, an in vivo study was undertaken using choline-deficient weanling rats. One group of animals was provided with a choline-free ¹⁴C-phosphocholine-supplemented diet ("P-diet"), while a control group received food containing a full complement of choline (N-diet); the two analogues were provided in equimolecular quantities to their respective groups. Both diets contained ³H-labelled choline of negligible molarity and high specific activity. The animals were killed, and their livers, kidneys, lungs and brains removed, after a period of 7 weeks. Neutral phospholipid was extracted, from the organs, and preparative thin-layer chromatography (TLC) of these extracts led to the isolation of biosynthesized phosphatidylphosphocholine (VIII), free from phosphatidylcholine and identified by NMR and ³¹P-NMR; the latter gave a spectrum bearing two signals of equal intensity, from phosphonium and phosphate. Measurement of relative peak areas in the ³¹P-NMR spectra of some of the extracts gave a measure of the level of phosphocholine incorporation, varying from 33% in lung lecithin to 6% in kidney sphingomyelin. Although the effect of lecithin already present in the tissues at the start of the experiment was unknown, it was suggested that much of the phosphatidylcholine in the P-diet rats originated from sequential methylation of phosphatidylethanolamine, a route well-established in the liver but claimed to be of little significance in other tissues. It was also suggested that sphingomyelin was being synthesized mainly by an indirect route, from CDP-choline formed reversibly from this methylation-derived phosphatidylcholine, and not by a direct pathway involving CDP-phosphocholine. Additional signals in the ³¹P-NMR spectra of extracts from P-rats (but not N-rats) were attributed to phosphobetaine (XIV) produced by oxidation of phosphocholine in the liver, and trimethylphosphine oxide (XV) in the kidney, originating from phosphocholine oxidised by the bacteria of the intestinal tract. Similar liver oxidation of arsenocholine had been observed by Mann et al., while the oxidation of choline to trimethylamine oxide in the intestine, and its excretion in the urine, was first reported by Davies. Comparison of ³H/¹⁴C ratios in the P-diet itself and in the phospholipid extracts gave a possible indication of the differing preferences of some enzyme systems for choline and phosphocholine, although it was necessary to make several basic assumptions. Comparison of the relative fatty acid composition in the lecithin and "phospholecithin" of the four organs investigated showed little difference between the two in liver, lung and brain, but a profound difference in the kidney. Thus renal phosphatidylphosphocholine contained greatly elevated levels of stearic and polyunsaturated acids, and it was suggested that this might be due to a need to maintain a critical optimum membrane fluidity, since it had already been noted that a change from ammonium to phosphonium produced a lecithin with a lover transition temperature. Thus a change from palmitic to stearic acid increased the transition temperature, while a change from oleic and linoleic acids to polyunsaturates increased the membrane stability.