ANALYTICALBIOCHEMISTRY151,182-187 (1985) High-Performance Liquid Chromatography of Water-Soluble Choline Metabolites MORDECHAI LISCOVITCH, ANDREW FREESE, JAN K. BLUSZTAJN, AND RICHARD J. WURTMAN Department o/ttpplied Biological Sciences. Massachusetts Institute o/Technology. Bldg. £25-604. Cambridge. Massachusells 02139 Received June 7, 1985 We have developed a new method for the separation of [H)choline metabolites by high-per- formance liquid chromatography. Using this method it is possible to separate, in one step. all of the known major water-soluble choline metabolites present in crude acid extracts of cells that have been incubated with [Hlcholine, with baseline or near-baseline resolution. We use a gradient HPLC system with a normal-phase silica column as the stationary phase. and a linear gradient of increasing polarity and ionic strength as the mobile ptw. The mobile phase is composed of two buffers: Buffer A, containing acetonitrile/water/ethyl alcohol/acetic acid/0.83 totsodium acetate (8oo/127/68/2/3), and buffer B (400/400/68/53/79), pH 3.6. A linear gradient from 0 to 100% buffer B. with a slope of 5%/min, is started 15 min after injection. At a flow rate of 2.7 ml/min and column temperature of 45°C, typical retention times for the following compounds are (in min): betaine, 10; acetylcholine, 18; choline, 22; glycerophosphocholine. 26: CDP-choline, 31; and phosphorylcholine. 40. This procedure has been applied in tracer studies of choline metabolism utilizing the neuronal NGl08-15 cell line and rat hippocampal slices as model systems. While the compounds labeled in the NO 108-15 cells were primarily ph05phorylcholine and glycero- phosphocholine. reflecting high rates of phospholipid turnover, in the hippocampal slices choline and acetylcholine were the major labeled species. Identification ofindividual peaks was confirmed by comparing the elution profiles of untreated cell extracts with extracts that had been treated with hydrolyzing enzymes of differing specificities. This HPLC method may be useful in studies of acetylcholine and phosphatidylcholine metabolism, and of the possible interrelationships of these compounds in cholinergic cells. 0 1915A_it IIas. lno:. KEY WORDS: HPLC: metabolites: choline; acetylcholine: phosphatidylcholine. The metabolicfateofcholine in cholinergic (11,12)either are time consuming or require neurons and other cells may be investigated multiple extraction or prepurification steps. by tracing the incorporation of choline into and thus are inconvenient.Moreover.none of its various cellular metabolites. Indeed. ra- these methods is capable of separating all of dioisotopicallylabeledcholine has been widely the known cellularmetabolitesof choline and used in studies of acetylcholine turnover and thus all invariably provide poor resolution of release and of the metabolism of membrane some of these compounds. choline-containing phospholipidS. This ap- HPLC methods for the separating and proach requires a method for the separation quantifyingof compounds of biologicalorigin of intact radiolabeledcholine from the.com- have become popular in the last decade be- pounds to whichit isconverted. Existingtech- cause they offer the inherent advantages of niques such as paper chromatography (1,2). liquid chromatography in a highly efficient high-voltagepaper electrophoresis(3-7), var- implementation. Numerous methods. involv- ious modes of ion-exchangechromatography ing ion-pairedreverse-phaseHPLC.have been (2.7-10), and thin-layer chromatography described for separating quaternary amines OOOJ.2697/SS $3.00 182 Copyn,110 115 by A_~ lne. Alln,lls of rq>rod~ In any ronn ~.- -- - -- --- ----
CHROMATOGRAPHIC SEPARATION OF CHOLINE METABOLITES 183 12 "0 -100 (13-16). including one for separating choline GPCII from acetylcholine (17). However. these assays C .J r I provide inadequate retention of phosphory- 51 lated choline metabolites and thus are unsuit- I I CII .,.:::..8: able for studies requiring a complete separa- 2 tion of choline metabolites. We have therefore developed a new normal- phase HPLC method for the one-step sepa- ration of all of the major water-soluble choline !t .= 4 u 0 51 IIAD PO ... metabolites present in crude cellular extracts. 0 a:: 2! MATERIALS AND METHODS HPLC system. Our gradient HPLC system (Rainin, Wobum, Mass.) is equipped with two FIG. J. HPLC separation of radioJabeJedstandard water- Rabbit HP pumps, a pressure monitor, and soluble choline metabolites. A mix of six radioisotopically an LC-22 temperature controller (the latter labeled standards (20 ,d) was cbromatographcd asdesaibed from Bioanalytical Systems, West Lafayette, under Materials and Methods. Briefly. two buffers were used: BufferA (acetoniuiJe/Water/cthyJ aIcohoI/acetic IJ:idI Ind.). The pumps are controlled by an Apple 0.83 Msodium acetate, 800/127/68/2/3) and buffer B (400/ lie microcomputer running a gradient control 4OO/68/S3/19). A linear gradient from 0 to 100% buffer program (Gradient Manager, Model 702 ver- B (slope = S~/min) was started IS min after injection. sion 1.2, (c) by Gilson International). A nor- Flow rate was 2.7 ml/min and column temperature was mal-phase Rainin Microsorb-silica column (5 4S.C. The standards utilized in this separation were: Bet, "m spherical particle size, 4.6 mm i.d. X 250 (H]betaine; ACh, (4C]acetyJcboline; Ch, (H]choline; GPCb, (HJgJycerophosphocbotine; CDP-Ch, [I~ mm length) has been used as the stationary dincdiphospbocboline; PCb, (HJphosphorylc:boline. The phase. The mobile phase was composed of two dashed line delineates the gradients profile. buffers: Buffer A, containing acetonitrile/wa- ter/ethyl alcohol/aceticacid/O.83 M Sodium acetate (800/127/68/2/3), pH 3.6, and buffer choline was prepared in our laboratory from B (400/400/68/53/79), pH 3.6. A linear gra- [methyPH]phosphatidylcholine (20-40 Ci/ dient from 0 to 100%buffer B, with a slope of mmol, New England Nuclear) by mild alkaline 5%/min, was started 15 min after injection. hydrolysis as described (18). [3H]Betaine was The gradient profile is shown in Fig. I. Flow prepared from eH]choline in the reaction ca- rate was 2.7 ml/min and column temperature talysed by choline oxidase. Sodium phosphate was maintained at 45°C. Fractions (2.7 mt) of buffer (0.1 ml, 0.2 M pH 7.8), containing I the eluate were collected in scintillation vials, "mol of (3H]choline (lOCi/mol) and 0.08 10 ml of Hydrotluor (National Diagnostics, units of choline oxidase (Sigma, St. Louis, Somerville N. J.) was added to each vial, and Mo.) was incubated for I hat J00c. This pro- radioactivity was then determined by liquid cedure resulted in the total conversion of cho- scintillation spectrometry in a Beckman LS- line to betaine, as HPLC analysis of the prod- 7500 spectrometer with an efficiency of 25- uct revealed only one radioactive peak and no 35% for 3H and 70% for 14c. counts at the position of a [3H]choline stan- Radioactive standards. [methyPH]Choline dard. (melhyPH]Phosphorylcholine was syn- chloride (80 Ci/mmol), [methyl-4C)cytidine thesized in the reaction catalyzed by choline diphosphocholine (42 Ci/mol), and [acetyl- kinase in the presence of ATP and magnesium. 4C)acetylcholine iodide (2.3 Ci/mol) were all A solution (0.1 ml) containing 7.5 mM gly- purchased from New England Nuclear, Boston cylglycine (pH 8.5), II mM magnesium chlo- Massachusetts. [methyl-3H]Glycerophospho- ride, 0.6 mM ATP, I nmol of[3H]choline (100---- - - -- -------
184 LISCOVITCH ET Al.Ci/mol), and 0.0 I unit of choline kinase Labeling of hippocampal slices. Male rats(Sigma) was incubated for 30 min at 30°C. (300-400 g) were purchased from Charles Unreacted choline was extracted to 1.5% tet- River, Wilmington, Massachusetts. The ratsraphenylboron in 3-heptanone. As shown by were killed by decapitation and the brains re- HPLC, the aqueous phase was free of choline moved into ice-cold Krebs-Ringer bicarbon-radioactivity. The radioactivity present in this ate buffer (KRB) containing (in mM): NaCI,phase eluted as a single peak which could be 120; KCI, 3.5; CaCh, 1.3; MgS04, 1.2;completely converted to choline by alkaline NaHC03, 25; and D-glucose, 10. The bufferphosphatase (data not shown). was gassed for at least 30 min prior to use with Cell cltlwre and labeling. NG 108-15, a a mixture of 95% O2/5% C02. Hippocampineuroblastoma X glioma hybrid cell line, was were excised and slices (300 Itm thick) werekindly provided by Dr. M. Nirenberg, NIH, prepared with a Mcilwain tissue chopper.Bethesda, Maryland. Cells (passages 17-25) Slices (15/tube) were rinsed twice with 4 mlwere subcultured in 6 X 35-mm multiwell of KRB and incubated in 0.5 ml of KRB con-plates at a concentration of 150,000 cells/well taining 5 ItCi of (3H]choline (125 nM) for 30and grown under an atmosphere of 90% air/ min at 31°C under an atmosphere of95% 021 10%C02 in Dulbeccos modified Eagles me- 5% C02. Following incubation the mediumdium (GIBCO, Grand Island N. Y.) contain- was discarded, slices were rinsed twice with 4ing 0.1 mM hypoxanthine, IItM aminopterine, ml of ice-cold KRB, and I ml of 5% perchloric 16 ItM thymidine (all from Sigma), and 5% acid was added to each tube. Acid extractionfetal calf serum (GIBCO). After 24 h growth then proceeded as described above.medium was replaced with a serum-free N2medium (19) with 10 ItMcholine and the cells RESULTS AND DISCUSSIONwere further incubated for 48 h. The cells werethen labeled by incubation with 60 ItM The suitability of the normal phase silicaPH]choline (2 IlCi/ml, 33 ItCi/llmol) in N2 matrix to serve as a stationary phase in HPLCmedium for 48 h. Following incubation the separation of choline metabolites was affirmedradioactive medium was removed and each in preliminary experiments in which a clearwell was rinsed with 2 ml of N2. separation of phosphorylated choline metab- Acid extraction and sample preparation. olites from nonphosphorylated ones was foundOne milliliter of 5% perchloric acid containing to be critically dependent on the polarity of 10 ItMeserine salicylate (Sigma) was added to the mobile phase. The separation of betaine,each well and the cells were scraped with a acetylcholine, and choline from the phos-rubber policeman, transferred to Eppendorf phorylated choline metabolites was thereforemicrotubes. and sonicated. Acid-insoluble effected using a low-polarity (high acetonitrilematerial was precipitated by centrifugation in concentration) buffer. under these conditionsan Eppendorf centrifuge and kept for phos- the phosphorylated metabolites were retained.pholipid extraction and protein determination. These metabolites were subsequently elutedPerchloric acid present in the supernatant was by a gradient ofincreasing polarity. In furtherprecipitated by the addition of an equimolar experiments. the effects of changing ionicconcentration of potassium acetate and cen- strength on the retention times of the varioustrifugation. A 0.8-ml aliquot of the superna- compounds were characterized. It was foundtant was dried under vacuum and the dried that the retention times of the standards. mostextracts wen: kept at -15°C. For analysis the notably those of acetylcholine and CDP-cho-extracts were routinely dissolved in 0.1 ml of line. were sensitive to the ionic strength of the 18 mM sodium acetate and filtered through a mobile phase. Both acetylcholine and CDP-0.2-ltm Acro LCI3 filter (Gelman. Ann Arbor, choline were greatly retarded under conditionsMich.). and 20-1t1aliquots were injected. of low ionic strength. and this property was
CHROMATOGRAPHIC SEPARATION OF CHOLINE METABOLITES 185utilized to resolve acetylcholine from choline CIIand CDP-choline from phosphorylcholine.Thus, the introduction of a gradient of in- c .creasing polarity and ionic strength allowed --the complete separation of the major cellular Q Mcholine-containing compounds. The chro- Ematography of choline metabolites on a nor- -mal phase silica column appears therefore tobe of the combined partition/adsorption type. .. .. .;; .. a::; 4! Figure I illustrates the separation of a mix-ture of six radioisotopically labeled water-sol-uble choline metabolites by HPLC under the Iconditions described under Materials andMethods. Retention times of these compoundsare summarized in Table I. All peaks are re- FIG. 2. Elfect of alkaline phosphatase on the radiochro-solved with baseline or near-baseline resolu- matographic profile of acid extracts from IH]choline.la.tion. An unidentified peak eluting at the sol- beled NGiOS-IS cells. NGiOS-IS cells were labeled withvent front probably results from degradation IH]choline and extracted as described under Materia1sof the standards and was found to increase in and Methods. For alkaline phosphatase digestion. two ex- tracts ~ dissolved each in 6S "I ofbulfer containing 0.2size upon prolonged storage. Similarly, the Mglycine. 2 mM MgClz, and 2 mM znS04, pH 10.7. Therelatively high background occurring between pH was readjusted to 10.7 with 1 N KOH and the volumethe betaine and acetylcholine peaks appears adjusted to 13O"J with HzO. The two extracts were thento constitute a degradation product(s) of the combined, filtered. and divided into two equal 100..,,1por- tions. 10"J of a 100 U/ml alkaline phosphatase solutionglycerophosphocholine standard. We have (Sigma) was added to one portion while 10 "I of assayobserved that the chromatographic behavior bulfer was added to the control tube. Tubes ~ incubatedof choline metabolites may vary slightly for 4 h at 37.C. and the reaction was terminated by heatingamong different columns. In these cases, full in a boiling water bath for 2 min. Mixtures were thenresolution could be restored by minor adjust- filtered and a 20-"J aliquot of each tube was analyzed by HPLC. (-) Untreated control extract: (...) alkaline phes-ments in the gradients profile. phatase-treated extract. The ability of this HPLC procedure to sep-arate metabolically labeled choline-containingcompounds derived from both a neuronal cell line and hippocampal slices incubated with PH)choline in vitro is demonstrated in Figs. 2 and 3. As shown in Fig. 2, labeling of NG 108- TABLE 1 IS cells with [3H)choline under the conditions RETENTION TIMES OF STANDARD RADIOL4BELEO used results in the labeling of two major peaks, CHOLINE METABOUTES exhibiting the retention times of phosphoryl- choline (40 min) and glycerophosphocholine Retention time (26 min). and a minor peak of [3H)choline. Standard (min) Alkaline phosphatase treatment of the extract IHJBetaine lO.1 %0.4 resulted in the complete disappearance of the 1~JAcetyk:holine 17.7 %O.S peak eluting at 40 min. and a quantitative IH)Choline 22.0 % 0.0 transfer of the counts to the position of cho- [JH)Glyccrophosphoc:hline 2S.9 % 0.4 line. These findings establish the presence of [I~P-choline 30.6 % 0.8 40.1 % 0.9 a phosphomonoester bond in this compound IHIPhospborylcboline and identify it as phosphorylcholine. The peak Not~. Values represent means %SD of raWcs obtained eluting at 26 min was not affected by thisin seven separate runs. treatment as would be expected of glycero-
186 LlSCOVITCH ET AL. 15 Ch GPCh j 12 II PCh ~ ~..... .§ ~..... Q 212 M .. E E Q. Q. :2 :2 ,... » ACh U <> o ij <> a: Ch ~ j I I I I I I o o 10 20 30 50 Retention (nin) Tune FIG. 3. Effect of acetylcholinesterase on the radiochro-matographic: profile of acid extracts from [3H]choline-la. FIG. 4. Stimulation of [JH)gIycerophosphocholine la- beling by the Caz+ ionophore A23187. Cells were subcul-beled hippocampal slices. Hippocampal slices were labeled tured and grown essentially as described under Materialswith [JH]choline and extracted as described under Mate- and Methods, with minor modifications. The cells wererials and Methods. For AChE digestion. an extract wasdissolved in 0.1 Msodium phosphate buffer. pH 7.4. The labeled by incubation with 5 ,rC"l/weII [JH]choline in ofpH was readjusted to 7.5 with I N KOH and the extract N2 medium without choline, methionine. tryptophan.was filtered and divided into two equal ponions of 90 ,d. serine and linoleic acid (N2(-» for I hour. After incu-Electric eel AChE (10,d. 1000 units/mL Sigma) was added bation the cells were rinsed with N2(-) and chased for 8 hours in normal growth medium (c:boline concentrationto one tube and 10,d of buffer was added to the controltube. Tubes were incubated for 120 min at 23°C. Reaction of 30 $&M),n the absence or presencIeof 5 $&M i A23187was terminated by the addition of 10 $&1 bovine serum of (Sigma). This was followed by extraction and HPLC anal.albumin (10 mg/ml. Sigma) and precipitation of protein ysis as described in Materials and Methods. (-) Untreated control culture; (---) A23187-treated culture.by diluting the mixture with 110,d of buffer A. Mixtureswereclearedofprecipitateby filtrationand a 20-$&1liquot aof each tube was analyzed by HPLC. (-) Untreated con-trol extract; (---) acetylcholinesterase-treated extract. Ch 100phosphocholine. However. no further attempt iwas made to establish the identity of this peak. u o .... In marked contrast to the distribution of ........ glabeled choline metabolites in NG 108-15 cells. ..incubation of hippocampal slices with rH]- !choline (Fig. 3) resulted primarily in the la- .-=- 41 ..beling of the free choline pool and of a radio- ~ 9active peak exhibiting the retention time of ... <> a:acetylcholine (18 min). Treatment of theseextracts with acetylcholinesterase, an enzymethat specifically hydrolyzes the acetylcholine I -.----------. , 20 30 40 50ester bond, caused the virtual disappearance RetentionTime (men)of the peak elutina at 18 min: the lost countswere quantitatively recovered at the position FIG. 5. lsocratic:separation of [I~lc:holine andof choline. identifying this peak as acetylcho- (lH)choIinestandards.A mixof[l~ine (ACh)line. and [lH]choIine(Ch) standards wasseparated utilizinga mixture of 95" buffer A and Sft bufIerB. Other details The HPLC method presented here has been of the chromatopapbic conditions,indudina buffercom-applied in a variety of studies concerned with position are describedin the Iqend 10 FI. I and underacetylcholine and phosphatidylcholine me- Materialsand Methods.
CHROMATOGRAPHIC SEPARATION OF CHOLINE METABOLITES 187 tabolism in which eH]choline was utilized as ACKNOWLEDGMENTS a precursor. In one of these studies, for ex- This work was supported by grants from the National ample, the effectsof increasing the intracellular Institute of Mental Health. the Department of Defense, concentration of Ca2+on phosphatidylcholine and the Centre for Brain Sciences and Metabolism Char- metabolism were investigated. To illustrate the itable Trust. M.L. is a recipient of the Myron A. Bantrell Post-doctoral Fellowship. utility of the present HPLC method, some of the results.obtained in this study are shown in Note added in proof We have recently found that the mo- Fig. 4. In this experiment, NG 108-15 cells bile phase can be made totally volatile by substituting am- monium acetate for sodium acetate (at the same concen- were pulse labeled with (3H]choline for I h tration) without affecting retention times. This modifica- and then chased for 8 h in the presence of tion facilitates the use of the present HPLC method in A23187, a Ca2+ionophore. HPLC analysis of preparative applications. the acid-soluble choline metabolites revealed a marked increase in the levels of two radio- REFERENCES active peaks exhibiting the retention times of I. PIagemann, P. G. W. (1969) J. Cell Bioi. 42, 766- 782. eH]glycerophosphocholine and (3H]choline, 2. Diamond, I.. and Kennedy, E. P. (1969) J. Bioi. Chem. suggesting that A23187 stimulated phospha- 244, 3258-3263. tidylcholine catabolism in these cells. This 3. Heilbronn, E.. and Carlsson. B. (1960) J. Chromalogr. finding is consistent with previous reports on 4, 257-259. the Ca2+dependence of phospholipase A2and 4. Hildebrand, J. G.. Barker. D. L.. Herbert, E.. and glycerophosphocholine phosphodiesterase Kravitz, E. A. (1971) J. Neurobiol. 2, 231-246. 5. Abdel-Latif. A. A.. and Smith. J. P. (1972) Biochem. (29,21), and shows that these enzymes can Pharmacol. 21,3005-3021. probably be activated in intact cells by ele- 6. Mas1and, R. H.. and Mills. J. W. (1979) J. Cell Bioi. vating intracellular :a2+ concentrations. 83, 159-178. Figure 5 demonstrates the possible utiliza- 7. Sleight. R.. and Kent. C. (1980) J. Bioi. Chem. 255, 10646-10650. tion of isocratic elution conditions, selected 8. Gardiner. J. E.. and Whittaker. V. P. (1954) Biochem. on the basis of standard gradient conditions, J. 58. 24-29. to separate a partial set of the choline metab- 9. Stein. R. L. (1981) J. Chromalogr. 214, 148-15 I. olites. In this case, choline and acetylcholine 10. Roridi. A.. Palmerini. C. A.. Fini. c.. Goracci. G.. standards were separated from each other Porcellati. G.. and Trovarelli. G. (1979) J. Liq. Chromalogr.2, 1375-1392. within 25 min with a mobile phase consisting II. Yavin. E. (1976) J. Bioi. Chern. 251, 1392-1397. of95% buffer A/S% buffer B. lsocratic elution 12. Vance. D. E.. Trip. E.. and Paddon. H. B. (1980) J. conditions that will allow the rapid separation Bioi. Chem. 255. 1064-1069. of any pair of choline metabolites may be sim- 13. Yakatan.G. J..and Tien.J.-Y.(I979)J. Chromalogr. ilarly and easily developed from the gradient 164, 399-403. conditions described in this presentation. Such 14. Buchanan. D. N.. Fucelt. F. R.. and Domino. E. F. (1980) J. Chromalogr. 181, 329-335. isocratic conditions could be useful in the sep- IS. De Ruyter. M. G. M.. Cronnelly. R.. and Castagnoli. aration of products from substrates in assays N. (1980) J. ChromaJogr. 183, 193-201. of various enzymes involved in choline me- 16. Kneake. M. (1980) J. ChromaJogr. 198. 529-533. tabolism. 17. Potter. P. E.. Meek. J. L.. and Neff. N. H. (1983) J. Neurochem. 41, 188-194. 18. Dittmer. J. C.. and Wells. M. A. (1969) in Methods CONCLUSIONS in Enzymology (J. M. Lowenstein. ed.). Vol. 14. pp. 482-S3O. Academic: Press. New York. 19. Bottenstein, J. E.. and Sato. G. H. (1979) Proc. Nal/. We have presented a new method for sep- Acad. Sci. USA 76. S I4-S 17. arating choline metabolites by HPLC. This 20. van Den Bosch. H. (1982) in New Comprehensive method may be useful in studies of acetylcho- Biochemistry (HaWthorne. J. N.. and Ansell. G. B.. cds.). Vol. 4, Phospholipids. pp. 313-357. Elsevier line and phosphatidylcholine metabolism, and Biomedic:al Press. Amsterdam. of the possible interrelationships of these 21. Spanner. S.. and Ansell. G. B. (1982) Biochem. J. 208. compounds in cholinergic cells. 845-850.- - -- -