D-amino acid levels in human physiological fluids

CHIRALITY 5375-378 (1993)

D-Amino Acid Levels in Human Physiological Fluids DANIEL W. ARMSTRONG, MARY GASPER, SUN HAING LEE, JANUSZ ZUKOWSKI, AND NURAN ERCAL Department of Chemkby, University of Missouri-Rolla, Rolla, Missouri

ABSTRACT Plasma, urine, cerebrospinal fluid (CSF), and amniotic fluid were examined to determine whether free D-amino acids were present and if so at what levels. It was found that D-amino acids exist in all physiological fluids tested, but that their level varied,considerably. The lowest levels of D-amino acids were usually found in amniotic fluid or CSF (almost always go%). Correspondingly high levels of D-pipecolic acid were not found in plasma. Some of the trends found in this work seemed to be analogous to those found in a recent rodent study. 0 1993 Wiley-Liss, Inc.

KEY WORDS: urine, plasma, cerebrospinal fluid, amniotic fluid, D-pipecolic acid

The distribution of free amino acids in human body fluids has been of considerable interest for years.’” There have been many studies on the plasma level of amino renal excretion and absorption of amino and the level of amino acids in blood, cells,g amniotic fluid, lo and other fluids. Advances in the quantitative analysis of amino acids have made it possible to ascertain what are normal distributions and variations in amino acid levels in healthy human subjects.3p5 The level and disposition of certain amino acids can be of clinical interest. It is generally acknowledged that normal amino acid levels must be known if deviations due to disease or fluctuations resulting from normal physiological changes are to be recognized. It is well known that the naturally occurring amino acids in humans are of the L-configuration. Indeed, L-amino acids predominate in all known living systems. However, there are a few documented cases of the presence of mirror image D-amino acids in certain cyclic peptides and as structural elements in the cell walls of a few microorganisms. ”*” Recently, several studies have appeared that report f i t e levels of free D-amino acids in mammals’s’7 as well as marine invertebrates. Common laboratory rodents were shown to excrete surprisingly high levels of D-amino acids in some cases. 2o However, the relative levels of D-amino acids in their plasma were generally much lower (frequently by an order of magnitude). Early data show that D-amino acids are present in human urine as well. l7 In this work, we examine the level of some D-amino acids in various physiological fluids. AU samples are from healthy adults. No claims are made as to the meaning or importance of free D-amino acids in regard to pathology, clinical usefulness, age, etc. However, as in the case of total amino acid levels, knowledge of normal D-amino acid levels may be useful if we are to recognize perturbations caused by either natural or pathological conditions. 0 1993 Wiley-Liss, Inc.

MATERIALS AND METHODS Materials Amino acids were purchased from Sigma Chemical Company (St. Louis, MO). Acetonitrile, water, acetic acid, and triethylamine were of Omnisolv grade and supplied by EM Science (Gibbstown, NJ). The derivatiziig agent, 9-fluorenyl methylchloroformate, was purchased from Sigma. Perchloric acid, mercaptoethanol, boric acid, o-phthalaldehyde, and DLamino acids were obtained from Aldrich (Milwaukee, WI). Methanol, ethanol, and potassium hydroxide were supplied by Fisher (St. Louis, MO). Plasma samples were donated by Dr. Grannemann of the Phelps County Regional Medical Center (Rolla, MO). They were prepared by centrifuging citrated blood (1 ml of 3.8% sodium citrate: 9 ml blood) for five minutes. Sterile cerebrospinal fluid (SFC) samples were also donated by Dr. Grannemann. Amniotic fluid samples (first trimester) were provided by Dr. Sau w. Cheung (Director of the Cytogenetics Laboratory, Washington University School of Medicine, St. Louis, Missouri). Urine samples were obtained from University staff volunteers. Methods D-proline and pipecolic acid analysis Derivatization was performed according to reference 21. After derivatization with 9-fluorenylmethyl chloroformate (FMOC-Cl), 5-50 p.L of the sample was injected into a C,, reversed phase column (100 x 4.6 mm) supplied by Advanced Separation Technol-

Received for publication November 6, 1992; accepted December 17, 1992. Address reprint requests to Dr. Daniel W. Armstrong. Department of Chemistry, University of Missouri-Rolla, Rob, MO 65401. Sun Haing Lee’s afiiliition is with the Kyungpook National University, Taegu 702-701, Korea.

Pump A 200 fieCN 800 water 2 HOAC

Pump B 800 MeCN 200 water 2 WAC

Pump C 300 MeCN 380 MeOH 320 water

Pump D 800 MeCN 200 MeOH 2 HOAC

Fig. 1. Block diagram showing the achiral-achiral-chiral,double column switching set-up for measuring pipecolic acid in urine and plasma. The sample work-up and injection volumes were the same as for proline. All mobile phase compositions are given in this figure. MeCN, acetonitrile; MeOH, methanol; HOAC, glacial acetic acid TEA, triethylamine; SPE, solid phase extraction; Inj, injector; SV, switching valve: UV, ultraviolet detector (wavelength of detection given below).

ogy (Whippany, NJ). The mobile phase consisted of acetoni-

cence” chromatographic system also included a Rainin (San Carlos, CAI model A-30-S pump (Scientific Systems, Inc., min. A W wavelength of 266 nm was used to monitor the State College, PA), pulse dampener model LP-21, and two effluent. The column switching value was turned for 2-15 Rheodyne injectors (models 7125 and 7010) for post-column seconds after the signal reached the maximum of the standard derivatization. retention time. A small portion of the eluting peak of FMOCThe chromatographic columns used in these studies inproline was introduced to the RN-p-CD chiral column cluded Crownpak CR(+) and Crownpak CR(-), supplied by (100 x 4.6 mm; Advanced Separation Technology, Whip- Michael Henry at J.T. Baker (Phillipsburg, NJ). Crownpak pany, NJ), which was eluted with acetonitrile + glacial acetic columns used for post-derivatization were used at 5°C. All the acid (1,OOO + 6, v/v) at 1ml/min. The effluent from the chiral samples and mobile phase solutions were filtered using 0.2 column was monitored using a fluorimetric detector (Aex 265 pm filters supplied by Alltech Associates, Inc. (Deerfield, IL). nm, A,, 315 nm). It is important to note that in proper aqueA CIScolumn was used to isolate the individual amino acids. ous buffer, no detectable racemization of the p r o h e occurs The mobile phase consisted of 95:5, water:methanol (v:v). during the derivatization process. Furthermore, once the The flow rate was 1 or 1.5 mVmin. The column was at room FMOC derivative is formed, no change in enantiomeric ratios temperature and the UV detector setting was at 195 nm. A were found even after several weeks. column switching or coupled column method was used previThe pipecolic acid was by far the most challenging separa- ously for the analysis of tyrosine, phenylalanine, or tryptophan tion because of its low level. Derivatization with FMOC-Cl in rat or mouse urine.2o After separation on the CI8 coIumn, was done as in the case of proline. However, a double column individual amino acids were switched to the chiral crown ether switching procedure was needed to separate FMOC-pipecolic column in order to determine their enantiomeric ratio. acid from all impurities. This set-up is shown in Figure 1. The Post-column OPA derivatization of the amino acids was first achiral column was Ca, the second achiral column was required for fluorescence detection. The excitation waveCIS, and the third column was chiral Cyclobond I. All mobile length was 340 nm, and the emission wavelength was 450 nm. phase conditions are shown in Figure 1. The Rainin pump was the post-column pump for this system. D-phenylalanine, D-tyrosine, and D-tryptophan The OPA solution was prepared as follows. Seven hundred analyses The LC system that utilized post-column fluores- milligrams of o-phthalaldehyde was dissolved in 15 ml of ethacence detection was made up of the following Shimadzu de- nol. Three hundred milliliters of mercaptoethanol was added vices: two LC-6A pumps, a SCL-6B system controller, a to this solution. The solution was then added to 1liter of 3% CR601 chromatopac recorder, SPD-6AV UVNIS spectro- boric acid solution, which had been adjusted to pH = 10.0 photometric detector, and a RF-535 fluorescence high-perfor- with KOH. Teflon tubing (0.5 mm id) was used with the mance liquid chromatography (HPLC) monitor. The “fluores- post-column apparatus, and the post-column reactor was 3 m trile

+ water + acetic acid (380 + 620 + 2, v/v) at 0.5 mV

377

D-AMINO ACID LEVELS

TABLE 1. Summary of free amino acid data for urine samples

Amino acid

Phenylalanine Pipecolic acid Proline Tryptophan Tyrosine

No. of Range for Average % samples total amino % D-amino D-amino analyzed acids (pn/L.)" acid range acid 14 7 12 10 12

35-130 0.3-3.7 20-90 46-280 45-210

0.3-9.5 56.0-95.0 1.4-12.9 0.06-0.4 0.1-6.6

2.4 86.0 3.9 0.2 0.8

TABLE 2. Summary of free amino acids data for plasma

Amino acid

Phenylalanine Pipecolic acid Proline Tyrosine

No. of Range for samples total amino analyzed acids (@)" 126 8 6

15-92 0.3-1.8 90-180

7b

-

% D-amino

acid range 0.044.O b 0.62.2 0.5-1 0.05-4.ob

Average % D-amino acid

0.B6 1.3 0.7 0.6'

"These numbers include L-plus D-amino acids. 'These were serum samples rather than plasma samples.

"These numbers include L-plus D-amino acids.

TABLE 3. Summary of free amino acids data for human cerebrospinal fluid

long. In order to venfy further that no racemization occurred during the post-column method, the FMOC-gly-Cl precolumn derivatization method was also applied. Because FMOCgly-Cl is not commercially available, it was synthesized in our laboratory by essentially the same method as that described for FMOC-proline above. In order for any results to be considered valid, the same answer had to be obtained (within experimental error) by each of the two methods (i.e., precolumn and post-column). In addition, the opposite chiral crown ether column was used to reverse the retention order of the amino acids of interest. In order to venfy this further, the sample was divided and run on chiral columns of the opposite configuration. This reverses the retention order of the phenylalanine enantiomers but does not affect the retention of any achiral impurities. In all measurements done on biological samples (vide supra), care must be taken to prevent chemical racemization andlor microbiological contamination. The procedures outlined in this section have been shown to produce little or no chemical racemization down to the parts per ten thousand level, or lower levels in some cases. 1722 More often microbiological contamination of the sample can be a problem. Contamination can cause an increase or decrease in relative, D-amino acid levels (depending on the type of sample and length or type of microbial contamination). In some cases levels of both D- and L-amino acids decrease dramatically. Only sterile samples frozen immediately after collection could be stored for later analysis. In this study all underivatized samples were analyzed immediately after collection.

RESULTS AND DISCUSSION Recently it has been reported that low but detectable levels of several D-amino acids can be found in human urine.17 A more extensive compilation of analogous data is given in Table 1 for five amino acids. In addition to the relative amount of D-amino acids, the range of total amino acid excretion is given as well. In all cases the level of total amino acid excretion fell within previously reported ranges for healthy adults. The single most interesting aspect of this data concerned pipecolic acid. The absolute amount of pipecolic acid excreted is much less than the other amino acids (i.e., one to two orders of magnitude). However, the majority of all pipecolic acid excreted seemed to be of the D-configuration. Indeed most of the specimen samples tested exceeded 90%D-pipecolic acid,

Amino acid

Phenylalanine Proline

No. of Range for Average % samples total amino % D-amino D-amino analyzed acids (pm/L)" acid range acid 6 6

7-25 1-7

0.06-0.7 0.144

0.3 0.2

"These numbers include L-plus D-amino acids.

TABLE 4. Summary of amino acid data for human amniotic fluid"

Amino acid Phenylalanine Proline

No. of Range for Average % samples total amino % D-amino D-amino analyzed acids (mrdJbacids range acid 5 5

80-170 30-150 ~

0. ow.9 0.064).3 ~~~

0.4 0.15

~

"All samples were from healthy women in their first trimester. bThese numbers include L- plus D-amino acids.

and only one sample contained less than 84% of the D-enantiomer. This is in stark contrast to the other amino acids, in which much lower relative levels of the D-enantiomers were found, i.e., tenth percent to low percent levels (Table 1). Unlike the other compounds in Table 1, pipecolic acid is not one of the 20 amino acids commonly found in proteins. Pipecolic acid is a homologue of the imino acid proline and is thought to be derived from the metabolism of ly~ine.'~Although its biological role is not completely understood, it has been reported as a possible neurotransmitter in the central nervous system. 24 Tables 2-4 summarize the free amino acid data for plasma, cerebrospinal fluid (CSF), and amniotic fluid, respectively. Unfortunately, there was not a large enough volume of these samples to test for as many amino acids as was done for urine. However, it is clear that the relative amounts of D-amino acids found were significantly lower than those found in urine. Generally 4% of the amino acids found in CSF and amniotic fluid were of the D-configuration (Tables 3, 4). The plasma and serum results (Table 2) indicate that the relative level of D-amino acids can range at least over two orders of magnitude in the blood. However the L-enantiomer predominated in all

378

ARMSTRONG E T AL.

cases. Interestingly, the relative level of D-pipecolic acid in plasma also was low and comparable to that of the other D-amino acids investigated. Apparently, the high relative levels of D-pipecolic acid found in urine are not the result of unusually high levels in the blood.

CONCLUSIONS D-amino acids were present in all human physiological fluids tested. The levels were very low in CSF and amniotic fluid. The amount of D-amino acids excreted in the urine were signhcantly higher than other fluids but vaned considerably with the amino acid tested. Only in the case of pipecolic acid was the D-enantiomer excreted predominantly. Relative levels of D-amino acids in plasma were usually lower than those found in urine and often slightly higher than those found in CSF or amniotic fluid. The relative level of D-pipecolic acid found in the plasma was comparable to those found for other amino acids. Hence pipecolic acid showed the greatest change in enantiomenc ratio when comparing plasma to urine. There was a considerable range of values for the total amino acid level and the relative D-amino acid level in all fluids tested. In an analogous study on laboratory rodents, similar trends were found in regard to D-amino acid levels. ” In general, the relative D-amino acid levels in the plasma were about ten times lower than those found in urine. The reason for this was thought to be that the kidney reabsorbs useful L-amino acids preferentially to D-amino acids2’ If so, it is possible that an analogous process occurs in humans. Unfortunately, pipecolic acid was not considered in the rodent study. The relative level of D-amino acids excreted by rodents was somewhat higher than those found in this work for humans. ” However, none of the rodent samples approached the high (D/L) enantiomeric ratios found for pipecolic acid in this work. A sigdkant portion of the D-amino acids found in the rodent study were thought to be dietetic in origin. 2” ACKNOWLEDGMENTS Support of this work by the National Institute of General Medical Science (BMT lROl GM36292-04) is gratefully acknowledged. Also, we thank Dr. Neal Grannemann of the Phelps County Regional Medical Center for providing CSF and plasma samples as well as Dr. Sau W. Cheung, Director of the Cytogenetics Laboratory, Washington University School of Medicine, St. Louis, Missouri, for providing the amniotic fluid samples. LITERATURE CITED 1. Van Slyke, D.D., Meyer, G.M. The absorption of amino acids from the blood by the tissues. 1. Biol. Chem. 1 6 147-212, 1913. 2. Holden, J.T. (ed.) Amino Acid Pools. Distribution, Formation and Function of Free Ammo Acids. New York Elsevier, 1962.

3. Scriver, C.R., Rosenberg, L.E. In: Amino Acids, Metabolism and Its Disorders. WB Saunders, New York 1973. 4. Feigin, R.D., Klainer, A.S., Beisel, W.R. Circadian periodicity of blood amino acids in adult men. Nature 215512-514, 1967. 5. Scriver, C.R., Clow, C.L., Lamm, P. Plasma amino acids: screening, quantitation and interpretation. Am. J. Clin. Nutr. 24:876-890, 1971. 6. Raisner, S.H., Aranda, J.V., Colle, F., Papageorgiou, A., Schiff, D., Scriver. C.R., Stem, L. An effect of intravenous glucagon on plasma amino acids in the newborn. Pediatr. Res. 7 184-191. 1973. 7. Baerlocher, K., Scriver, C.R., Mohyuddin, F. Ontogeny of amino acid transport in mammalian kidney. Proc. Natl. Acad. Sci. USA 6510091016, 1970. 8. Christensen. H.N., Jones, J.C. Amino acid transport models: renal resorption and resistance to metabolic attack. J. Biol Chem 237 12034206,1962. 9. Levy, H.L., Barkin. E. Comparison of amino acid concentrations between plasma and erythrocytes. Studies in normal human subjects and those with metabolic disorders. J. Lab. Clin. Med. 78517-523, 1971. 10. Butterfield, L.J.. OBrien, D. The effect of maternal toxaemia and diabetes on transplancental gradients of free amino acids. Arch. Dis. Child. 38326327, 1963. 11. Schleifer, K.H., Kandler, 0. Peptidoglycan types of bacterial cell walls and their taxonomic implications. Bacteriol. Rev. 36407477, 1972. 12. Wheat, R. W. In: Microbiology, 17th ed. Zmsser, ed. New York Appleton-Century-Crofts, 1988: a 7 6 . 13. Fisher, G.J., Garcia, N.M., Payan, LL., Cadilla-Perezrois, R., Sheremata. W.A., Man, E.H. D-Aspaxtic acid in purified myelin and myelin basic protein. Biochem. Biophy. Res. Commun. 1 3 5 6 8 S 7 , 1986. 14. Nagata, Y., Akino, T., Ohno, K., Kataoka, Y., Ueda, T., Sakura, T., Shiroshita, K., Tasuda, T. Free D-amino acid in human plasma in relation to senescence and renal diseases. Clin. Sci. 7 3 105-108, 1987. 15. Konno, R., Nagata, Y., Niwa, A., Yasumura, Y. Spontaneous excretion of D-alanine in urine in mutant mice lacking D-amino acid oxidase. Biochem. J. 261:285287, 1989. 16. Nagata, Y., Konno, R., Yasumura, Y., Akino. T. Involvement of D-amino acid oxidase in e l i n a t i o n of free D-amino acids in mice. Biochem. J. 257291-292, 1989. 17. Armstrong, D.W., Duncan, J.D., and Lee, S.H. Evaluation of D-amino acid levels in human urine and in commercial L-amino acid samples. Amino Acids 1:97-106, 1991. 18. Preston, R.L. Occurrence of D-amino acids in higher organisms: a survey of the distribution of D-amino acids in marine invertebrates. Comp. Biochem. Physiol. 87B:55-62, 1987. 19. Preston, R.L. Sodium amino acid cotransport systems in marine invertebrates. In: Comparative aspects of sodium cotransport systems, Comp. Physiol. Series. Ed. R.K.H. Kinne, Karge Press, Basel 7 1-129, 1990. 20. Armstrong, D.W., Gasper, M.P., Lee, S.H., Ercal, N., Zukowski, J. Factors controlling the level and accurate determination of D-amino acids in the urine and plasma of several laboratory rodents. Amino Acids 4402413, 1993. 21. Einarsson, S., Josefsson, B., Lagerkvist, S. Determination of amino acids with 9-fluorenylmethyl chloroformate and reversed-phase high-performance liquid chromatography. J. Chromatogr. 282609418. 1982. 22. Zukowski, J., Palowska, M., Armstrong, D. W. Efficient enantioselective separation and determination of trace impurities in secondary amino acids (i.e., irnimo acids). J. Chromatogr. 62333-41, 1992. 23. Kelley, R.I. Quantification of pipecolic acid in plasma and urine by isotopediluytion gas chromatography mass spectrometry. In: Techniques in Diagnostic Human Genetics. New York Wiley-Liss, 1991: 205-218. 24. Chang, Y.-F. Lysine metabolism in the rat brain: the pipecolic acid forming pathway. J. Neurochem. 30347-354, 1977.