Fluorometric determination of adenine nucleotides and adenosine by ion-paired, reverse-phase, high-performance liquid chromatography

ANALYTICAL

BIOCHEMISTRY

145, 9-l 3 (1985)

Fluorometric Determination Paired, Reverse-Phase,

of Adenine Nucleotides and Adenosine by lonHigh-Performance Liquid Chromatography’

ALFONSO RAMOS-SALAZAR AND ANDREW D. BAINES Department of Clinical Biochemistry, Banting Institute, University of Toronto, 100 College Street, Toronto, Ontario MSG lL5. Canada Received June 28, 1984 A sensitive and specific assay for measurement of adenine nucleotides and adenosine by paired-ion high-performance liquid chromatography is described. The I,N6-ethenoderivatives of ATP (c-ATP), ADP (+ADP), AMP (c-AMP). and adenosine (c-Ado), formed by reaction with chloroacetaldehyde at 37°C. were separated under isocratic conditions in 20 min. These compounds are strongly fluorescent at an emission wavelength of 280 nm, rendering a lowest detection limit of 2-5 pmol per injection. The detector responded linearly over the measured ranges (5-100 pmol for C-Ado and 5-4000 pmol for nucleotides). Specificity was confirmed enzymatically. a$-Methyleneadenosine 5’-diphosphate could be used as an internal standard for measurement of the nucleotides. Significant amounts of NADH appeared as a separate peak in hypoxic tissue. Recoveries from snap-frozen kidney were 88. 92, 76, and 63% for AMP, ADP, ATP, and adenosine, with SD for recovery of 1.O, 10.5, 8.3, and 5.6%, respectively. This method was successfully used to measure adenine nucleotides and adenosine in oxygenated and hypoxic perfused rat kidneys. G 1985 Academic PES. inc. KEY WORDS: adenine nucleotides: fluorescence; isocratic; HPLC; adenosine; kidney.

Several methods for analysis of nucleosides and nucleotides by HPLC with uv absorption detection have been reported in the literature. They require gradient elution resulting in baseline drift, long analysis time, low sensitivity, and poor reproducibility (l-3). The introduction of charged molecules with a hydrophobic moiety into the mobile phase (paired ion), has allowed the separation of adenine nucleotides under isocratic conditions (4). These conditions, combined with uv detection, give a lowest detection limit of 100 pm01 per injection (0.1 nmol/pl) for ATP, ADP, and AMP in cardiac tissue extracts (5,6). The reaction of adenosine with chloroacetaldehyde yields a strongly fluorescent 1,N6ethenoadenosine derivative (7). The reaction has been used to detect 100 fmol-100 pmol of adenosine by HPLC (8- 11). We have ’ Supported Canada.

by the Medical

Research Council

applied this reaction to the determination of ATP, ADP, and adenosine by HPLC under isocratic conditions. The technique has allowed us to study the effect of hypoxia on the metabolism of adenine nucleotides and the regulation of adenosine production by the isolated perfused rat kidney. While this paper was being prepared for publication a method using gradient-elution HPLC to separate fluorescent derivatives of adenine nucleotides and adenosine was reported (12); the conditions for derivatization described in that report produced low recoveries for ATP and ADP. Our method has the advantages of using an isocratic system and derivatization conditions that produce much higher recoveries. MATERIALS

AND

METHODS

Appurutzts. HPLC pump 6000 A with U6K injector (Waters, Assoc., Inc., Milford, Mass. 0 1757), with a C-18 Perkin-Elmer column

of

9

0003-2697/85 $3.00 Copyright 0 1985 by Academic Press. Inc. All rights of reproduction in any form reserved.

10

RAMOS-SALAZAR

40 mm long X 4.6 mm i.d., and 3-pm particle size (Perkin-Elmer Corp., Nowalk, Conn. 06856). Fluorescence was measured with an Aminco fluoromicrophotometer (American Instrument Co., Inc. Md.). The excitation wavelength was obtained with a Coming 754 bandpass filter that transmits from below 254 nm, with maximum transmission at 320 nm. The emission wavelength was obtained with a Schott KV sharp-cutoff filter at 399 nm. Reagents. Chemicals were obtained from the following sources: ATP, ADP, AMP, Ado, +ATP, t-ADP, C-AMP, C-Ado, NADH, CAMP, and tetrabutylammonium hydrogen sulfate [TBAHS2 (Sigma, St. Louis, MO.)]; chloroacetaldehyde (Fluka Chemical Corp., N.Y. 11787); potassium dihydrogen phosphate and methanol (Fisher Scientific, Springfield, N.J. 07410); D-gkOSC (BDH Chemicals, Toronto, Canada); Millipore filters, Type PH, 0.3~pm pore size (Millipore Corp., Bedford, Mass. 0 1730); hexokinase (EC 2.7.1.1.), myokinase (EC 2.7.4.3), SAMP deaminase (EC 3.5.4.6) and adenosine deaminase [EC 3.5.4.4, (Boehringer-Mannheim, Canada)]. Preparation of tissue extracts. Rat kidneys were perfused as previously described (13). At the end of perfusion, the kidneys were freeze-clamped between brass blocks chilled with liquid nitrogen and pulverized in a mortar precooled with liquid nitrogen. Tissue protein was precipitated by adding 0.6 M perchloric acid while the tissue was still frozen, followed by centrifugation at 6000 r-pm for 15 min. The supernatant was neutralized to pH 6.8 with 2 M KOH and immediately prepared for derivatization and measurement of adenine nucleotides and adenosine. Preparation of standards. A mixture of AMP, ADP, and ATP standards was diluted in 7 mM KH2P04 (pH 4.2) at 1, 2, and 4 mM concentrations, respectively, and stored frozen at -70°C until used for the prepara’ Abbreviation used: TBAHS, hydrogen sulfate.

tetrabutylammonium

AND BAINES

tion of the standard curves. As described by Riss et al. (14), this procedure prevented breakdown of the standards during storage. Adenosine was dissolved in water to produce a final concentration of 1 mM. Serial dilutions of the nucleotides and adenosine were prepared from the stock solution and perchloric acid was added in the proportion 1:3 for adenosine and 1:4 for nucleotides (approximately in the same proportion that perchloric acid was added to tissue). This solution was then neutralized with 2 M KOH before derivatization. Aliquots of 0.5 ml of extracts or standard solutions were incubated for 24 h at 37°C in the presence of 0.05 ml of 55% chloroacetaldehyde and 0.1 ml of 1 M acetate buffer, pH 4.5. Samples from 10 to 25 ~1 were injected into the HPLC system and run with a mobile phase of 0.1 M KH2P04, pH 6.5, and various concentrations of TBAHS. Recovery studies were carried out by dividing the pulverized frozen tissue in halves. One portion was extracted with 3 ml of perchloric acid alone, while known amounts of ATP, ADP, AMP, or adenosine were added with perchloric acid to the other. The amount recovered was calculated from the difference between the two aliquots, i.e., (extract with added nucleotides - extract without added nucleotides) (quantity of nucleotide added) = fractional recovery. RESULTS AND DISCUSSION Chromatographic separation. Figure la illustrates the chromatographic separation of a standard mixture of adenosine and adenine nucleotides. The optimum conditions for analysis were mobile phase, 100 mM KH2P04 buffer, pH 6.5, 0.2 mM TBAHS; flow rate, 1.5 ml/min; pressure, 1000 psi; excitation wavelength, 280 nm; emission wavelength, 400 nm; sensitivity setting, 0.03, equivalent to 0.001 /LA. To establish the optimum conditions for analysis, the effect of changes on the composition of the mobile phase was tested. In

FLUOROMETRIC/CHROMATOGRAPHIC

DETERMINATION

OF Ade NLJCLEOTIDES

( (b)

11

E-AMP

60-

70-

(a) 60

608 5 ,g

50-

g iz

40-

30-

zo-

I

I

0

,

5

10 Time

15

I

20

(min)

I

I

I

I

I

0

5

10

15

20

Time

(min)

FIG. 1. (a) Separation of a standard mixture of c-ATP, c-ADP, t-AMP, and e-adenosine. The concentration per 10 ~1 injected was 1, 0.5, 0.25, and 0.125 nmol, respectively. The sensitivity setting was 0.03, equivalent to 0.001 PA. Recorder speed 1 in./5 min. (b) Separation obtained from a partially hypoxic tissue perfused with sir/5% CO*. Note the NADH peak, which is not visible in extracts of tissue perfused with 95% 02/5% COz. 10 (~1was injected. A higher sensitivity setting and injection of 25 ~1 were required for accurate adenosine measurements.

the absence of TBAHS, the more polar compounds ATP, ADP, and AMP were not resolved and appeared in the solvent front, and the less polar adenosine was retained for 10 min. With constant ionic strength and pH of the mobile phase and gradually increasing concentrations of counter-ion from 0.05 to 2 tItM, the retention times of the nucleotides increased as expected from their respective polarities. With 0.2 mM TBAHS (Fig. 1) ATP had the longest retention time, followed by ADP, Ado, and AMP. Gradual increases in pH (5.0 to 7.0) or ionic strength (10 to 100 mM), while the concentration of TBAHS was kept constant, decreased the retention time of the nucleotides with consequent loss in resolution. The retention time of adenosine was only slightly affected by changes in pH or ionic strength; its retention time was decreased by an increase in TBAHS concentration. These observations agree with previous studies performed with uv detection of the nonderivatized nucleotides and aden-

osine ($6). Aging of the column may cause peak tailing and loss of resolution, which can be overcome by increasing the concentration of TBAHS. A variety of temperatures and incubation times have been used to derivatize adenosine (8- 11) and nucleotides ( 12). In our hands, incubation at 80°C for 40 min produced 42% hydrolysis of ATP to ADP and AMP, and 24% hydrolysis of ADP to AMP. To avoid this hydrolysis the conditions originally reported by Barrio et al. (7) were used; that is, incubation for 24 h at 37°C. These conditions reduced the hydrolysis of both ATP and ADP to only 10%. The extent of hydrolysis from ATP to ADP was not affected by changes in the pH (3.5 to 5.5) or ionic strength of the acetate buffer used for the derivatization reaction. Incubation at 37°C for 24 h produced 19, 5, and 9% less E-AMP, t-ADP, and t-ATP, respectively, than did incubation at 80°C for 40 min. Therefore, although the yield of the

12

RAMOS-SALAZAR

chloroacetaldehyde reaction may be lower at 37°C than at 80°C increased hydrolysis of ATP and ADP at 80°C results in similar apparent yields under both incubation conditions. Although the reaction with chloroacetaldehyde is specific for adenine related compounds (7), it was necessary to confirm this by incubation of the tissue extracts with specific enzymes before derivatization. Following this treatment, the peak for ATP was removed and the ADP peak increased after incubation of the extract with 20 U of hexokinase plus glucose at pH 8.5 and 25°C for 30 min. The peak for ADP disappeared after incubation with 20 U of myokinase at pH 7.6 at 37°C for 15 min. Incubation of the extract with 0.03 U adenylic acid deaminase at pH 6.5 and 37°C for 20 min eliminated the AMP peak. Adenosine was removed by addition of 40 mU adenosine deaminase and incubation for 1 h at 45°C and pH 7.5. NADH and CAMP were also derivatized, but did not interfere with the assay. The retention time of CAMP was 35 min and interference with subsequent injections was not a problem because of its low concentration in tissue. NADH was resolved and appeared between adenosine and ADP (Fig. lb). This peak, prominent in kidneys perfused under hypoxic conditions, was insignificant in normoxic kidneys. The response was linear for adenosine in the range of 5 to 100 pmol per injection and for the nucleotides from 5 to 4000 pmol per injection. The intraassay variation of the assay for the nucleotides and adenosine was calculated as the mean of the coefficient of variation obtained from three different kidneys perfused with 95% 02/5% COZ. Each kidney extract was assayed three or four times. The CVs were 3.3, 5.3, 3.7, and 3.0% for AMP, ADP, ATP, and adenosine, respectively. Recoveries between 90 and 95% for the nucleotides have been reported for uv-HPLC methods (5). However, these recovery values do not include the loss of nucleotides during

AND BAINES

the tissue-extraction procedure. To calculate recovery of tissue nucleotides we added the standard compounds to the perchloric acid before precipitation of the tissue protein; therefore, our recoveries reflect the loss of nucleotides throughout the whole procedure (Table 1). The limit of detection for ATP, ADP, and AMP was 5, and for adenosine, 2 pmol per injection (25 ~1). The detection limit reported for uv detection is 100 pmol per injection (1 ~1); therefore, our assay is about 20 times more sensitive than the uv-detection system for adenine nucleotides. Due to its low concentration (6 nmol/g wet wt) and interferences by other nucleotides present in kidney tissue, adenosine can not be detected by uv systems. The specificity conferred by the reaction with chloroacetaldehyde, and the strong fluorescence of t-Ado at 280 nm, enabled us to measure it in amounts of 1.5 nmol/g tissue. The sensitivity for adenosine could be increased twofold if no counterion were added to the mobile phase, but the nucleotides could not be resolved under these conditions. A recent publication of a similar system for the measurement of adenine nucleotides and adenosine by HPLC-fluorescence detection (12) reports a limit of detection of 1 pmol for the nucleotides. However, with the method described, despite the use of a gradient system, ATP and ADP peaks in the chromatograms of tissue extract were poorly resolved. Furthermore, the high temperatures used for derivatization caused extensive breakdown of ATP to ADP and of ADP to TABLE RECOVERY

1

OF ADENINE AND

NUCLEOTIDES

ADENOSINE”

AMP

ADP

ATP

Ado

88 f 1

93 + 10

16 f 8

63 f 5

a Percentage recovery; mean f SD. Triplicate measurements from each of four experiments in which different amounts of nucleotides were added to pulverized frozen kidney before extraction with perchloric acid.

FLUOROMETRRXHROMATOGRAPHIC

AMP. In calculating the results, the authors compensated for loss due to breakdown but did not consider the conversion of ATP and ADP to AMP. Two compounds not present in the kidney that are also derivatized by this reaction could be used as internal standards for the measurement of the nucleotides. These are 2-0-methyladenosine, which could not be completely resolved from the ATP peak under our chromatographic conditions, and a$methyleneadenosine 5’-diphosphate, which was completely resolved and appeared between the AMP and adenosine peaks. ACKNOWLEDGMENTS This work was done in partial fulfillment of the requirements for a Master of Science degree in Clinical Biochemistry at the University of Toronto. The authors thank Professor C. J. Porter and Rosa Drangova for their advice and assistance.

REFERENCES 1. Zakaria. M., and Brown, 226, 267-290.

P. R. ( 198 1) J Chromatogr.

DETERMINATION

OF

Ade

NUCLEOTIDES

13

2. Schweinsberg, P. D., and Loo, Ti Li (1980) J. Chromatogr. 181, 103-107. 3. D’Souza, L., Glueck, H. I. (1977) Thromb. Haemostasis 38, 990- 100 I. 4. Hoffman, N. E., and Liao, J. C. (1977) Anal. Chem. 49, 2231-2234. 5. Juengling, E., and Kammermeier, H. (1980) Anal. Biochem. 102, 358-36 1. 6. Ingebresten, 0. C., and Bakken. A. M. (1982) J. Chromatogr. 242, 119- 126. 7. Barrio, J. R., Secrist J. A., III, and Leonard, N. J. (1972) Biochem. Bioohvs. . , Rex Commun. 46. ;79-804. 8. Kuttesch, J. F., Schmalstieg, F. C., and Nelson, J. A. (1978) J. Liq. Chromatogr. l(l), 97-109. 9. Yoshioka, M., and Tamura, Z. (1976) J. Chromatogr. 123, 220-224. 10. Jacobson, M. K., Hemingway, L. M., Farrel, T. A., and Jones, C. E. (1983) Amer. J. Physiol. 245(5), H887-H890. I I. Wojcik, W. J., and Neff, N. H. (1982) J. Neurochem. 39, 280-282. 12. Levitt, B., Head, R. J., and Westfall, D. P. (1984) Anal. Biochem. 137, 93-100. 13. Baines, A. D., and Drangova. R. (1984) Cunad. J. Physiol. Pharmacol. 62, 272-276. 14. Riss, T. L., Zorich, N. L., Williams, M. D., and Richardson, A. J. Liq. Chromatogr. 3(l), 133158.