- Email: [email protected]
Simultaneous extraction and reverse-phase high-performance liquid chromatographic determination of adenine and pyridine nucleotides in human red blood cells
ANALYTICAL
BIOCHEMISTRY
118-124 (1985)
146,
Simultaneous Extraction and Reverse-Phase High-Performance Liquid Chromatographic Determination of Adenine and Pyridine Nucleotides in Human Red Blood Cells’ VILBERTO
STOCCHI, LUIGI PIERANGELA Istituto
CUCCHIARINI, MAURO MAGNANI, LAURA PALMA,* AND GIANCARLO CRESCENTINI*
di Chimica Biologica and *Istituto di Scienze Universitri degli Studi di Urbino, 61029 Urbino,
CHIARANTINI,
Chimiche, Italy
Received July 25, 1984 A simple and rapid method for the determination of ATP, ADP, AMP, NADP+, NAD+, NADPH, and NADH in human erythrocytes is described. A single-step extraction procedure employing alkaline medium and CF 50A Amicon ultrafiltration membranes allows a simultaneous and total recovery of the compounds of interest. Analysis is performed by reverse-phase highperformance liquid chromatography on a 5-pm Supelcosil LC-18 column and uv detection. Extraction and analysis require about 30 min. Levels of adenine and pyridine nucleotides in normal adults are also presented. 0 1985 Academic PESS, IX. KEY WORDS: reverse-phase high-pressure liquid chromatography; alkaline extraction; nucleotide separation; erythrocytes.
The adenine ribonucleotides (ATP, ADP, AMP) and pyridine nucleotides play a vital role in cellular metabolism and their involvement in differentiation and dedifferentiation processes has been reported (1). These compounds participate in biochemical reactions which are responsible for the energetic and redox state of the cell and are important cofactors in erythrocyte glycolysis. The body and accuracy of the information which can be obtained are obviously related to the precision, accuracy, and sensitivity of the analytical method employed. Speed of analysis is an important factor since routine determination is generally required and inconveniences due to possible decomposition can be minimized. However, quantitative analysis of erythrocyte nucleotides is a difficult and tedious task because of problems encountered during extraction and quantitation. Problems are caused in extraction by the
presence of hemoglobin (2), and oxidized and reduced pyridine nucleotides show different stabilities in acids and alkali, respectively (3). For this reason, a double-step procedure is generally used (4), although a single-step extraction has been proposed (5). Both methods are lengthy and require considerable analytical work. Furthermore, recoveries lower than 70% are obtained for the single coenzymes. As far as quantitative analysis is concerned, high-performance liquid chromatography is a more powerful tool than are spectrophotometric, fluorometric, or enzymatic cycling methods. Radioimmunoassay is more sensitive but more complicated (6). Ion-exchange HPLC is widely used for the separation and determination of nucleotides (7-9). However, the compounds of interest either are only partially separated (9) or require exceedingly long analysis times (N 1 hr). Moreover, it is necessary to use highconcentration buffers and/or a fast gradient elution. Reverse-phase liquid chromatography has been shown to be a simpler and more
’ This work was partially supported by the Cassa di Risparmio di Pesaro and Minister0 della Pubblica Istruzione. 0003-2697/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproductmn in any form reserved.
118
ALKALINE
EXTRACTION
AND
CHROMATOGRAPHY
OF
NUCLEOTIDES
119
Extraction. Perchloric acid (PCA)2 extracts of erythrocytes were obtained following the method reported by Beutler (15). slightly modified. Two milliliters of 7.5% (v/v) HClO4 was added to I ml of freshly drawn whole blood. After centrifugation at 3000 r-pm X 10 min, the supernatant was separated and neutralized with K2C03. The solution was again centrifuged at 3000 rpm X 10 min and the superna~nt was removed and brought to a final volume of 4 ml (pH 6.5). Alkaline extraction was performed as follows: 1 ml of freshly drawn blood was added to 1 ml of ice-cold 0.5 M KOH solution and immediately deproteinized under vigorous shaking on Vortex. After 3 min standing in ice, 2 ml MATERIALS AND METHODS of ice-cold HZ0 was added and the resulting solution was placed on a CF 50A Amicon C~~~~~~~s. Nucieotide standards of the membrane and centifuged at 2500 rpm X 10 highest grade available were purchased from min. One milliliter of clear ultrafiltered soSigma (St. Louis, MO.). The coenzymes lution was obtained and the pH was adjusted NAD’, NADH, NADP’, and NADPH of to 6.5 by adding 0.1 ml of a 1 M KH2P04 the highest purity were obtained from Boeh- solution. This volume was largely sufficient ringer (Mannheim, West Germany). for the subsequent HPLC determination. Analytical-grade potassium dihydrogen Nevertheless, a greater volume can be recovfrom Merck phosphate was obtained ered and used without any change in the (Darmstadt, W. Germany), HPLC-grade concentration of the compounds. A micromethanol was from Hoecst (Frankfu~, W. method employing 75 X 1-mm-i.d. capillary Germany) and Fluka (Buchs, Switzerland), tubes was used for hematocrit estimation. while double-distilled water was prepared in Chrmnatographic apparatus and condithe laboratory. CF 50A and PM 30 (4 = 22 tions. Two liquid chromatographic systems mm) ultrafiltration Amicon membranes were were used throughout this work. One was a purchased from Amicon (Lexington, Mass.) Varian 5000 liquid chromatograph (Varian, and Millipore filters (0.22 pm) were from Palo Alto, Calif.) equipped with a singleMillipore (Bedford. Mass.). Stock solutions piston reciprocating pump and a variable(1 mM) of ATP, ADP, AMP, NAD’, and wavelength uv detector (Model UV-SO) emNADP’, as well as of the other nuc~eotides ploying a lo-p1 flow cell. Samples were inand nucleosides, were prepared by dissolving jected by means of a Valco injector valve a weighed mass of the dried material in 0.1 and retention times and peak areas were M KH2P04 buffer solution (pH 6.5) and were obtained using a Shimadzu Chromatopack stored at -20°C. The only exception was R-IB electronic integrator. The other one xanthine, which was prepared as a 0.1 mM was a Beckman system (Beckman, Berkeley, standard solution. Standards of various con- Calif.) which consisted of two Model 1 12 centrations were prepared by proper dilution pumps, a 420 solvent programmer. a 2 10 of the stock solutions. Standard solutions of sample injection valve, and a 160 fixedNADH and NADPH were prepared before wavelength (254 nm) uv detector equipped use in KH2P04 buffer solution (pH 6.5) kept * Abbreviation used: PCA, perchloric acid. at 0°C.
versatile technique and has in most cases replaced ion-exchange HPLC in the analysis of compounds such as nucleotides, nucleosides, and related bases ( 10-l 3). We recently reported the separation of selected nucleotides present in the PCA extracts of human erythrocytes, utilizing a Supelcosil LC-18 reversephase column (14). In this paper we describe a simple and rapid (30-min) analytica method, based on alkaline extraction and RPLC quantitation, which allows a simultaneous and accurate determination of ATP, ADP, AMP, NAD+, NADP+, NADH, and NADPH in human red blood cells.
120
STOCCHI
with a 18,5-&l flow cell. Integration of peak areas was obtained by means of a 3390A HP (Hewlett-Packard, Avendale, Pa.) electronic integrator. In both cases a 5-pm Supelcosil LC-18 [25 cm X 4.6 mm i.d. (Supelco, Bellefonte, Pa.)] was the analytical column and a stainless-steel guard column (2 cm X 4.6 mm i.d.) packed with pellicular reversedphase material was used. The mobile phase consisted of two eluants: 0.1 M KH2P04 solution, pH 6, (Buffer A), and a 0.1 M KH2P04 solution, pH 6, containing 10% (v/v) of CH30H (Buffer B). All buffer solutions, after preparation and pH adjustment, as well as standards and sample solution, were filtered through a 0.22~pm Millipore filter. The chromatographic conditions used to obtain the chromatograms reported in Figs. 1 and 2 were the following: 9 min at 100% of Buffer A, 6 min at up to 25% of Buffer B, 2.5 min at up to 90% of Buffer B, 2.0 min at up to 100% of Buffer B, and hold for 6 min. The gradient was then returned to the first buffer and the initial conditions were restored in 5 min. The flow rate was 1.3 ml/
I 0
, 5
ET AL.
min and detection was performed at 254 nm. The chromatogram of Fig. 3 was obtained isocratically by using Buffer B as mobile phase. The flow rate was again 1.3 ml/min while detection was pe~o~ed at 340 nm. The temperature was 18-22°C in both cases. Peaks identities were confirmed by coelution with standards, enzymatic peak-shift, and detection at different wavelengths. Quantitative measurements were carried out by injection of standard solutions of known concentration. The following millimolar extinction values were used for standards calibration: 13.8 for ATP, ADP, and AMP; 16.6 for NAD+ and NADP+; and 13.0 for NADH and NADPH at 254 nm. The value of 6.2 was used for NADH and NADPH at 340 nm. RESULTS AND DISCUSSION
HPLC analysis. Figure 1 shows the separation of a standard mixture of 20 nucleotides and nucleosides where the compounds of interest are completely separated in 22 min. This is a noteworthy improvement with re-
10
15
20
28
min
FIG. 1. Reverse-phase HPLC on a S-pm (25-cm X 4.6mm-i.d.) Supelcosil LC-18 column of OS-2 nmol of a standard mixture of nucleotides and nucleosides using the chromato~phi~ conditions reported under Materials and Methods. The compounds were monitored at 254 nm.
ALKALINE
0
I 5
, 10
EXTRACTION
15
I 20
AND
25
CHROMATOGRAPHY
min
FIG. 2. Reverse-phase determination of adenine nucleotides and oxidized and reduced pyridine coenzymes in human erythrocytes of normal subjects. Fifty microliters of the alkaline extract obtained as described under Materials and Methods was injected. Chromatographic conditions were as in Fig. 1.
spect to reported separations, obtained in either the ion-exchange or the reverse-phase mode, which require 45-60 min. The choice of mobile phase parameters (buffer concentrations and pH, percentage of organic modifier), limits of detection, and precision of the method, together with the experimental procedure which should be followed to obtain reproducible and accurate measurements and to avoid routine problems (rapid column deterioration, lines and loops clogging, etc.), are described in detail elsewhere (14). We have shown that if one needs to determine, ATP, ADP, AMP, NAD, and NADP+ only, as in the case of PCA extracts, analysis can be completed in 13 min by simply changing the gradient program (14). In Fig. 2, the chromatogram of an erythrocyte alkaline extract obtained as described under “Materials and Methods” is reported. It should be pointed out that ATP, ADP, AMP, NAD+, NADP+, NADH, and NADPH are completely separated from interferences and the gradient used does not cause an excessive baseline drift. These factors, coupled to the baseline separation of the seven compounds, allows an easy and accurate quantitation of AMP and NADH, which are present at con-
OF
121
NUCLEOTIDES
centrations much lower than those of the other nucleotides. The chromatogram also shows that other important metabolites, i.e., xanthine and hypoxanthine, can be determined. NADPH and NADH can be separately determined at 340 nm (Fig. 3) in about 6 min by using the same column and isocratic conditions. Because of the lower extinction coefficients, larger volumes of sample have to be injected. The column shows a quite long lifetime since 3000 sample injections have been made up to now without our observing any irreversible deterioration effects. The guard column needs to be changed every 200-300 injections when analysing 20111aliquots. Extraction qf adenine and p.vridine nucleotides from human red blood cells. Preliminary experiments were carried out to find a method that would allow the simultaneous extraction of ATP, ADP, AMP, NAD+, NADH, and NADPH and yield extracts which could be readily analyzed by HPLC. Different extraction mediums were tried: (i) Hydroalcoholic KOH solution containing various amounts, ranging from 10 to 80%, of CH30H or C2HSOH; (ii) KOH containing O.l-0.2% Tri-
0
2.5
5
min.
FIG. 3. Reverse-phase HPLC separation of NADPH and NADH in extracts of human red blood cells. A 500~1 aliquot of a potassium hydroxide extract of erythrocytes, obtained as reported under Materials and Methods, was injected. Mobile phase: 0.1 M KH2P04, pH 6.0, containing 10% (v/v) methanol. The compounds were monitored at 340 nm.
122
STOCCHI
ET AL.
TABLE 1 RECOVERY
TABLE 3
OF ADENINE AND P~RIDINE NUCLEOTIDES FROM HUMAN ERYTHROCYTES”
Compound
ADENINE AND P~RIDINE HUMAN ERYTHR~CYTES
Recovery (W)
ATP ADP AMP NADP+ NADPH NAD+ NADH
96 97 94 94 92 101 92
+ f k t k +
NUCLEOTIDE LEVELS IN OF 10 NORMAL ADULTS
Mean ct SD (nmoI/ml RBC)
1.0 3.2 2.9 2.1 3.4 3.6
ATP ADP AMP NADP+ NADPH NAD+ NADH
+ 3.1
1488 1: 90 83.0 rt 9.0 10.6 + 2.2 26.0 xk 3.7 16.0 f. 3.5 48.0 rfr 6.4 1.40 It; 0.57
’ Mean and standard deviation of five experiments.
ton X- 100; (iii) KOH added with 0. l- 10 mM solutions of s~bilizing agents such as cysteine and dithiothreitol; (iv) KOH solutions of different concentrations (0. l- 1.O M). Different ultrafiltration membranes, i.e., PM 30 and CF 50A Amicon, were tested in each case. To evaluate the percentage of recovery under physiological conditions, a sample of fresh whole blood was divided into four aliquots and, to three of these, standard solutions of nucleotides at various concentrations were added. The four aliquots were immediately extracted and analyzed by HPLC. The method employing 0.5 M KOH and CF 50A Amicon membranes yielded the best results, as shown by Table 1, where recovery data are reported for the single compounds. Average recoveries better than 95% were ob-
tained for ATP, ADP, AMP, NAD+, and NADP+, while a slightly lower recovery (~92%) was obtained for NADH and NADPH. Reproducibility was tested by preparing four separate extracts from the same blood sample, and a coefficient of variation of 1% was found for all compounds. Stability was also checked by analyzing the same extract, maintained in ice at pH 6.5, over a period of 12 h. No significant variation was observed with the exception that NADPH and NADH showed an oxidation rate of about 3%/h. However, this did not affect the accuracy of their determination, since the overall analysis time was about 30 min. The advantages of the present method are seen in Table 2, where concentration levels,
TABLE 2 COMPARISON
BETWEEN THE CONCENTRATION OF ADENINE AND PVRIDINE NUCLEOTIDES IN HUMAN CELLS OBTAINED FROM PERCHLORIC ACID AND ALKALINE EXTRACTS’
PCA extract A+B+C+Db Total PCA’ extract Present method (alkaline extract) ’ pmol/ml
red blood
RED BLEND
ATP
ADP
AMP
NADP+
NADPH
NAD+
NADH
1.013
0.067 0.074 0.141
0.0076 0.0032
0.0283
0.0101
0.0398
-
0.0347
0.317 1.331
-
1.497
0.091
0.0101
0.0252
0.0174
0.0415
cell; N.D.
0.0115
0.0091 0.0440
0.002 1
= not detectable.
’ A, B: first and second wash, respectively, of the protein precipitate; C, D: first and second wash, respectively, of the precipitate after neutmli~tion. ’ PCA and alkaline extractions were carried out utilizing the same blood sample.
XTRACTION
AND
CHROMATOGRAPHY
I
I I I
I
I I I
I
I I I
I
I I I
I
I I I
OF
NUCLEOTIDES
123
124
STOCCHI ET AL.
obtained by using PCA extraction and our alkaline extraction, are compared. First, the PCA extraction obviously does not allow determination of the reduced coenzymes NADPH and NADH, while their conversion to the oxidized forms affects the actual determination of NAD+ and NADP+ (2,5). Second, appreciable amounts of all compounds remain in the protein and perchloric precipitates, so that operative conditions must be carefully followed each time to obtain reproducible results. Third, the PCA method is quite time consuming (~45 min) compared to our method, which requires only LO-12 min. Finally, it should be pointed out that the agreement on the ATP, ADP, and AMP levels is a further proof that good recoveries can be obtained with the alkaline extraction. In Table 3 the concentrations of adenine and pyridine nucleotides in human erythrocytes of 10 normal adults are reported. These results are then compared with those obtained from other authors (Table 4) by using different methods of extraction and analysis. It is known that oxidized coenzymes are unstable in alkali (2,5,6). However, this was not observed in our experiment, probably due to the fact that during extraction the compounds remained in the alkaline medium for a short time. Also, the highest values of the ATP/ ADP and ATPfAMP ratios, 18 and 135, respectively, show that no degradation occurs and are a good indication of the reliability of the method. As far as adenine nucleotides levels are concerned, good agreement is found with the data of Dean et al. (2 1), while large discrepancies are observed, mainly in ADP concentration, with respect to other authors. In the case of pyridine nucleotides, it should be noted that only the data reported by Reinauer and Bruns ( 19) and by Sander r?tal. (5), with the exception of NADP+, show good correlation with our results. In conclusion, the method described in this paper based on alkaline extraction and reverse-phase liquid chromatography allows a simuItaneous and accurate determination
of adenine and pyridine nucleotides in human red blood cells. The short analysis time makes it a very useful tool for the study of cellular metabolic pathways. Finally, the results obtained indicate that the method can be extended to determine other metabolites, i.e., xanthine and hypoxanthine, while the applicability to different biological systems needs to be tested. REFERENCES 1. Clark, J. B.. Greenbaum. A. L.. and McLean. P. (1966) B&hem. J. 98;546-5;6. Burch, H. B., Bradley, M. E., and Lowry, 0. H. (1967) .I Biol. Chem. 246,4546-4554. Lowry, 0. H., Passoneau. J. V., and Rock, M. K. (1961) .I. Biol. Chem. 236,2756-2759. Klingenberg, M. (1974) in Methods of Enzymatic Analysis (Bergmeyer, H. V., ed.), Vol. 4, pp. 2046-2077, Academic Press, New York/London. 5, Sander, B. J., Oelshlegel, F. J., Jr., and Brewer, G. J. (1976) Anal. Biochem. 71, 29-36. 6. Bredehorst, R., Lengeyel, H., and Hilz, H. (1979) Eur. J. Biochem. 99,40 1-4 11. 7. Hartwick, R. A., and Brown, P. R. (1975) J. Chromatogr. 112, 651-662. 8. Brown, E. G., Newton, P. R., and Shaw, N. M. (1982) Anal. B~ochem. 123, 378-388. 9. Perret, D. (1982) Chromatographia 16, 2 1 l-2 13. IO. Anderson, F. S., and Murphy, R. C. (1976) J. Chramatogr. 121, 25 I-262. Il. Hoffman, N. E., and Liao, T. C. (1977) Anal. Chem. 49,223 I-2234. 12. Hartwick, R. A., Assenza, S. P., and Brown, P. R. (1979) J. Chromatagr. 186, 647-659. 13. Brown, P. R., and Krstulovich, A. M. (1979) Anal. Biochem. 99, l-12. 14. Crescentini, G., and Stocchi. V. (1984) J. Ch~amatagr. 290, 393-399. 15. Beutler, E. (1975) Red Cell Metabolism, pp. 8-18, Grune & Stratton, New York. 16. Grass, R. T., Schroeder, E. A,, and Gabrio, 8. W. (1966) J. Clin. Invest. 45, 249-255. 17. Omachi, A.. Scott, C. B., and Hegarty, H. (1969) Bighorn. Siuphys. Acta 184, 139- 147. ts. Marshall, W. E., and Omachi, A. (1974) Biochim. Biophys. Acta 354, l- 10. 19. Reinauer, H., and Brims, F. H. (1964) Z. Physio!. Chem. 337,93- 100, 20. Scholar, E. M., Brown, P. R., Parks, R. E., Jr., and Calabresi, P. (1973) Blood 41, 927-936. 21. Dean, B. M., Perrett, D., and Sensi, M. (1978) Biochem. Biophys. Res. Commun. 80, 147-154. 22. Ericson, A., Niklasson, F., and de Verdier, C, H. (1983) Cl&. Chim. Acta 127, 47-59.
BIOCHEMISTRY
118-124 (1985)
146,
Simultaneous Extraction and Reverse-Phase High-Performance Liquid Chromatographic Determination of Adenine and Pyridine Nucleotides in Human Red Blood Cells’ VILBERTO
STOCCHI, LUIGI PIERANGELA Istituto
CUCCHIARINI, MAURO MAGNANI, LAURA PALMA,* AND GIANCARLO CRESCENTINI*
di Chimica Biologica and *Istituto di Scienze Universitri degli Studi di Urbino, 61029 Urbino,
CHIARANTINI,
Chimiche, Italy
Received July 25, 1984 A simple and rapid method for the determination of ATP, ADP, AMP, NADP+, NAD+, NADPH, and NADH in human erythrocytes is described. A single-step extraction procedure employing alkaline medium and CF 50A Amicon ultrafiltration membranes allows a simultaneous and total recovery of the compounds of interest. Analysis is performed by reverse-phase highperformance liquid chromatography on a 5-pm Supelcosil LC-18 column and uv detection. Extraction and analysis require about 30 min. Levels of adenine and pyridine nucleotides in normal adults are also presented. 0 1985 Academic PESS, IX. KEY WORDS: reverse-phase high-pressure liquid chromatography; alkaline extraction; nucleotide separation; erythrocytes.
The adenine ribonucleotides (ATP, ADP, AMP) and pyridine nucleotides play a vital role in cellular metabolism and their involvement in differentiation and dedifferentiation processes has been reported (1). These compounds participate in biochemical reactions which are responsible for the energetic and redox state of the cell and are important cofactors in erythrocyte glycolysis. The body and accuracy of the information which can be obtained are obviously related to the precision, accuracy, and sensitivity of the analytical method employed. Speed of analysis is an important factor since routine determination is generally required and inconveniences due to possible decomposition can be minimized. However, quantitative analysis of erythrocyte nucleotides is a difficult and tedious task because of problems encountered during extraction and quantitation. Problems are caused in extraction by the
presence of hemoglobin (2), and oxidized and reduced pyridine nucleotides show different stabilities in acids and alkali, respectively (3). For this reason, a double-step procedure is generally used (4), although a single-step extraction has been proposed (5). Both methods are lengthy and require considerable analytical work. Furthermore, recoveries lower than 70% are obtained for the single coenzymes. As far as quantitative analysis is concerned, high-performance liquid chromatography is a more powerful tool than are spectrophotometric, fluorometric, or enzymatic cycling methods. Radioimmunoassay is more sensitive but more complicated (6). Ion-exchange HPLC is widely used for the separation and determination of nucleotides (7-9). However, the compounds of interest either are only partially separated (9) or require exceedingly long analysis times (N 1 hr). Moreover, it is necessary to use highconcentration buffers and/or a fast gradient elution. Reverse-phase liquid chromatography has been shown to be a simpler and more
’ This work was partially supported by the Cassa di Risparmio di Pesaro and Minister0 della Pubblica Istruzione. 0003-2697/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproductmn in any form reserved.
118
ALKALINE
EXTRACTION
AND
CHROMATOGRAPHY
OF
NUCLEOTIDES
119
Extraction. Perchloric acid (PCA)2 extracts of erythrocytes were obtained following the method reported by Beutler (15). slightly modified. Two milliliters of 7.5% (v/v) HClO4 was added to I ml of freshly drawn whole blood. After centrifugation at 3000 r-pm X 10 min, the supernatant was separated and neutralized with K2C03. The solution was again centrifuged at 3000 rpm X 10 min and the superna~nt was removed and brought to a final volume of 4 ml (pH 6.5). Alkaline extraction was performed as follows: 1 ml of freshly drawn blood was added to 1 ml of ice-cold 0.5 M KOH solution and immediately deproteinized under vigorous shaking on Vortex. After 3 min standing in ice, 2 ml MATERIALS AND METHODS of ice-cold HZ0 was added and the resulting solution was placed on a CF 50A Amicon C~~~~~~~s. Nucieotide standards of the membrane and centifuged at 2500 rpm X 10 highest grade available were purchased from min. One milliliter of clear ultrafiltered soSigma (St. Louis, MO.). The coenzymes lution was obtained and the pH was adjusted NAD’, NADH, NADP’, and NADPH of to 6.5 by adding 0.1 ml of a 1 M KH2P04 the highest purity were obtained from Boeh- solution. This volume was largely sufficient ringer (Mannheim, West Germany). for the subsequent HPLC determination. Analytical-grade potassium dihydrogen Nevertheless, a greater volume can be recovfrom Merck phosphate was obtained ered and used without any change in the (Darmstadt, W. Germany), HPLC-grade concentration of the compounds. A micromethanol was from Hoecst (Frankfu~, W. method employing 75 X 1-mm-i.d. capillary Germany) and Fluka (Buchs, Switzerland), tubes was used for hematocrit estimation. while double-distilled water was prepared in Chrmnatographic apparatus and condithe laboratory. CF 50A and PM 30 (4 = 22 tions. Two liquid chromatographic systems mm) ultrafiltration Amicon membranes were were used throughout this work. One was a purchased from Amicon (Lexington, Mass.) Varian 5000 liquid chromatograph (Varian, and Millipore filters (0.22 pm) were from Palo Alto, Calif.) equipped with a singleMillipore (Bedford. Mass.). Stock solutions piston reciprocating pump and a variable(1 mM) of ATP, ADP, AMP, NAD’, and wavelength uv detector (Model UV-SO) emNADP’, as well as of the other nuc~eotides ploying a lo-p1 flow cell. Samples were inand nucleosides, were prepared by dissolving jected by means of a Valco injector valve a weighed mass of the dried material in 0.1 and retention times and peak areas were M KH2P04 buffer solution (pH 6.5) and were obtained using a Shimadzu Chromatopack stored at -20°C. The only exception was R-IB electronic integrator. The other one xanthine, which was prepared as a 0.1 mM was a Beckman system (Beckman, Berkeley, standard solution. Standards of various con- Calif.) which consisted of two Model 1 12 centrations were prepared by proper dilution pumps, a 420 solvent programmer. a 2 10 of the stock solutions. Standard solutions of sample injection valve, and a 160 fixedNADH and NADPH were prepared before wavelength (254 nm) uv detector equipped use in KH2P04 buffer solution (pH 6.5) kept * Abbreviation used: PCA, perchloric acid. at 0°C.
versatile technique and has in most cases replaced ion-exchange HPLC in the analysis of compounds such as nucleotides, nucleosides, and related bases ( 10-l 3). We recently reported the separation of selected nucleotides present in the PCA extracts of human erythrocytes, utilizing a Supelcosil LC-18 reversephase column (14). In this paper we describe a simple and rapid (30-min) analytica method, based on alkaline extraction and RPLC quantitation, which allows a simultaneous and accurate determination of ATP, ADP, AMP, NAD+, NADP+, NADH, and NADPH in human red blood cells.
120
STOCCHI
with a 18,5-&l flow cell. Integration of peak areas was obtained by means of a 3390A HP (Hewlett-Packard, Avendale, Pa.) electronic integrator. In both cases a 5-pm Supelcosil LC-18 [25 cm X 4.6 mm i.d. (Supelco, Bellefonte, Pa.)] was the analytical column and a stainless-steel guard column (2 cm X 4.6 mm i.d.) packed with pellicular reversedphase material was used. The mobile phase consisted of two eluants: 0.1 M KH2P04 solution, pH 6, (Buffer A), and a 0.1 M KH2P04 solution, pH 6, containing 10% (v/v) of CH30H (Buffer B). All buffer solutions, after preparation and pH adjustment, as well as standards and sample solution, were filtered through a 0.22~pm Millipore filter. The chromatographic conditions used to obtain the chromatograms reported in Figs. 1 and 2 were the following: 9 min at 100% of Buffer A, 6 min at up to 25% of Buffer B, 2.5 min at up to 90% of Buffer B, 2.0 min at up to 100% of Buffer B, and hold for 6 min. The gradient was then returned to the first buffer and the initial conditions were restored in 5 min. The flow rate was 1.3 ml/
I 0
, 5
ET AL.
min and detection was performed at 254 nm. The chromatogram of Fig. 3 was obtained isocratically by using Buffer B as mobile phase. The flow rate was again 1.3 ml/min while detection was pe~o~ed at 340 nm. The temperature was 18-22°C in both cases. Peaks identities were confirmed by coelution with standards, enzymatic peak-shift, and detection at different wavelengths. Quantitative measurements were carried out by injection of standard solutions of known concentration. The following millimolar extinction values were used for standards calibration: 13.8 for ATP, ADP, and AMP; 16.6 for NAD+ and NADP+; and 13.0 for NADH and NADPH at 254 nm. The value of 6.2 was used for NADH and NADPH at 340 nm. RESULTS AND DISCUSSION
HPLC analysis. Figure 1 shows the separation of a standard mixture of 20 nucleotides and nucleosides where the compounds of interest are completely separated in 22 min. This is a noteworthy improvement with re-
10
15
20
28
min
FIG. 1. Reverse-phase HPLC on a S-pm (25-cm X 4.6mm-i.d.) Supelcosil LC-18 column of OS-2 nmol of a standard mixture of nucleotides and nucleosides using the chromato~phi~ conditions reported under Materials and Methods. The compounds were monitored at 254 nm.
ALKALINE
0
I 5
, 10
EXTRACTION
15
I 20
AND
25
CHROMATOGRAPHY
min
FIG. 2. Reverse-phase determination of adenine nucleotides and oxidized and reduced pyridine coenzymes in human erythrocytes of normal subjects. Fifty microliters of the alkaline extract obtained as described under Materials and Methods was injected. Chromatographic conditions were as in Fig. 1.
spect to reported separations, obtained in either the ion-exchange or the reverse-phase mode, which require 45-60 min. The choice of mobile phase parameters (buffer concentrations and pH, percentage of organic modifier), limits of detection, and precision of the method, together with the experimental procedure which should be followed to obtain reproducible and accurate measurements and to avoid routine problems (rapid column deterioration, lines and loops clogging, etc.), are described in detail elsewhere (14). We have shown that if one needs to determine, ATP, ADP, AMP, NAD, and NADP+ only, as in the case of PCA extracts, analysis can be completed in 13 min by simply changing the gradient program (14). In Fig. 2, the chromatogram of an erythrocyte alkaline extract obtained as described under “Materials and Methods” is reported. It should be pointed out that ATP, ADP, AMP, NAD+, NADP+, NADH, and NADPH are completely separated from interferences and the gradient used does not cause an excessive baseline drift. These factors, coupled to the baseline separation of the seven compounds, allows an easy and accurate quantitation of AMP and NADH, which are present at con-
OF
121
NUCLEOTIDES
centrations much lower than those of the other nucleotides. The chromatogram also shows that other important metabolites, i.e., xanthine and hypoxanthine, can be determined. NADPH and NADH can be separately determined at 340 nm (Fig. 3) in about 6 min by using the same column and isocratic conditions. Because of the lower extinction coefficients, larger volumes of sample have to be injected. The column shows a quite long lifetime since 3000 sample injections have been made up to now without our observing any irreversible deterioration effects. The guard column needs to be changed every 200-300 injections when analysing 20111aliquots. Extraction qf adenine and p.vridine nucleotides from human red blood cells. Preliminary experiments were carried out to find a method that would allow the simultaneous extraction of ATP, ADP, AMP, NAD+, NADH, and NADPH and yield extracts which could be readily analyzed by HPLC. Different extraction mediums were tried: (i) Hydroalcoholic KOH solution containing various amounts, ranging from 10 to 80%, of CH30H or C2HSOH; (ii) KOH containing O.l-0.2% Tri-
0
2.5
5
min.
FIG. 3. Reverse-phase HPLC separation of NADPH and NADH in extracts of human red blood cells. A 500~1 aliquot of a potassium hydroxide extract of erythrocytes, obtained as reported under Materials and Methods, was injected. Mobile phase: 0.1 M KH2P04, pH 6.0, containing 10% (v/v) methanol. The compounds were monitored at 340 nm.
122
STOCCHI
ET AL.
TABLE 1 RECOVERY
TABLE 3
OF ADENINE AND P~RIDINE NUCLEOTIDES FROM HUMAN ERYTHROCYTES”
Compound
ADENINE AND P~RIDINE HUMAN ERYTHR~CYTES
Recovery (W)
ATP ADP AMP NADP+ NADPH NAD+ NADH
96 97 94 94 92 101 92
+ f k t k +
NUCLEOTIDE LEVELS IN OF 10 NORMAL ADULTS
Mean ct SD (nmoI/ml RBC)
1.0 3.2 2.9 2.1 3.4 3.6
ATP ADP AMP NADP+ NADPH NAD+ NADH
+ 3.1
1488 1: 90 83.0 rt 9.0 10.6 + 2.2 26.0 xk 3.7 16.0 f. 3.5 48.0 rfr 6.4 1.40 It; 0.57
’ Mean and standard deviation of five experiments.
ton X- 100; (iii) KOH added with 0. l- 10 mM solutions of s~bilizing agents such as cysteine and dithiothreitol; (iv) KOH solutions of different concentrations (0. l- 1.O M). Different ultrafiltration membranes, i.e., PM 30 and CF 50A Amicon, were tested in each case. To evaluate the percentage of recovery under physiological conditions, a sample of fresh whole blood was divided into four aliquots and, to three of these, standard solutions of nucleotides at various concentrations were added. The four aliquots were immediately extracted and analyzed by HPLC. The method employing 0.5 M KOH and CF 50A Amicon membranes yielded the best results, as shown by Table 1, where recovery data are reported for the single compounds. Average recoveries better than 95% were ob-
tained for ATP, ADP, AMP, NAD+, and NADP+, while a slightly lower recovery (~92%) was obtained for NADH and NADPH. Reproducibility was tested by preparing four separate extracts from the same blood sample, and a coefficient of variation of 1% was found for all compounds. Stability was also checked by analyzing the same extract, maintained in ice at pH 6.5, over a period of 12 h. No significant variation was observed with the exception that NADPH and NADH showed an oxidation rate of about 3%/h. However, this did not affect the accuracy of their determination, since the overall analysis time was about 30 min. The advantages of the present method are seen in Table 2, where concentration levels,
TABLE 2 COMPARISON
BETWEEN THE CONCENTRATION OF ADENINE AND PVRIDINE NUCLEOTIDES IN HUMAN CELLS OBTAINED FROM PERCHLORIC ACID AND ALKALINE EXTRACTS’
PCA extract A+B+C+Db Total PCA’ extract Present method (alkaline extract) ’ pmol/ml
red blood
RED BLEND
ATP
ADP
AMP
NADP+
NADPH
NAD+
NADH
1.013
0.067 0.074 0.141
0.0076 0.0032
0.0283
0.0101
0.0398
-
0.0347
0.317 1.331
-
1.497
0.091
0.0101
0.0252
0.0174
0.0415
cell; N.D.
0.0115
0.0091 0.0440
0.002 1
= not detectable.
’ A, B: first and second wash, respectively, of the protein precipitate; C, D: first and second wash, respectively, of the precipitate after neutmli~tion. ’ PCA and alkaline extractions were carried out utilizing the same blood sample.
XTRACTION
AND
CHROMATOGRAPHY
I
I I I
I
I I I
I
I I I
I
I I I
I
I I I
OF
NUCLEOTIDES
123
124
STOCCHI ET AL.
obtained by using PCA extraction and our alkaline extraction, are compared. First, the PCA extraction obviously does not allow determination of the reduced coenzymes NADPH and NADH, while their conversion to the oxidized forms affects the actual determination of NAD+ and NADP+ (2,5). Second, appreciable amounts of all compounds remain in the protein and perchloric precipitates, so that operative conditions must be carefully followed each time to obtain reproducible results. Third, the PCA method is quite time consuming (~45 min) compared to our method, which requires only LO-12 min. Finally, it should be pointed out that the agreement on the ATP, ADP, and AMP levels is a further proof that good recoveries can be obtained with the alkaline extraction. In Table 3 the concentrations of adenine and pyridine nucleotides in human erythrocytes of 10 normal adults are reported. These results are then compared with those obtained from other authors (Table 4) by using different methods of extraction and analysis. It is known that oxidized coenzymes are unstable in alkali (2,5,6). However, this was not observed in our experiment, probably due to the fact that during extraction the compounds remained in the alkaline medium for a short time. Also, the highest values of the ATP/ ADP and ATPfAMP ratios, 18 and 135, respectively, show that no degradation occurs and are a good indication of the reliability of the method. As far as adenine nucleotides levels are concerned, good agreement is found with the data of Dean et al. (2 1), while large discrepancies are observed, mainly in ADP concentration, with respect to other authors. In the case of pyridine nucleotides, it should be noted that only the data reported by Reinauer and Bruns ( 19) and by Sander r?tal. (5), with the exception of NADP+, show good correlation with our results. In conclusion, the method described in this paper based on alkaline extraction and reverse-phase liquid chromatography allows a simuItaneous and accurate determination
of adenine and pyridine nucleotides in human red blood cells. The short analysis time makes it a very useful tool for the study of cellular metabolic pathways. Finally, the results obtained indicate that the method can be extended to determine other metabolites, i.e., xanthine and hypoxanthine, while the applicability to different biological systems needs to be tested. REFERENCES 1. Clark, J. B.. Greenbaum. A. L.. and McLean. P. (1966) B&hem. J. 98;546-5;6. Burch, H. B., Bradley, M. E., and Lowry, 0. H. (1967) .I Biol. Chem. 246,4546-4554. Lowry, 0. H., Passoneau. J. V., and Rock, M. K. (1961) .I. Biol. Chem. 236,2756-2759. Klingenberg, M. (1974) in Methods of Enzymatic Analysis (Bergmeyer, H. V., ed.), Vol. 4, pp. 2046-2077, Academic Press, New York/London. 5, Sander, B. J., Oelshlegel, F. J., Jr., and Brewer, G. J. (1976) Anal. Biochem. 71, 29-36. 6. Bredehorst, R., Lengeyel, H., and Hilz, H. (1979) Eur. J. Biochem. 99,40 1-4 11. 7. Hartwick, R. A., and Brown, P. R. (1975) J. Chromatogr. 112, 651-662. 8. Brown, E. G., Newton, P. R., and Shaw, N. M. (1982) Anal. B~ochem. 123, 378-388. 9. Perret, D. (1982) Chromatographia 16, 2 1 l-2 13. IO. Anderson, F. S., and Murphy, R. C. (1976) J. Chramatogr. 121, 25 I-262. Il. Hoffman, N. E., and Liao, T. C. (1977) Anal. Chem. 49,223 I-2234. 12. Hartwick, R. A., Assenza, S. P., and Brown, P. R. (1979) J. Chromatagr. 186, 647-659. 13. Brown, P. R., and Krstulovich, A. M. (1979) Anal. Biochem. 99, l-12. 14. Crescentini, G., and Stocchi. V. (1984) J. Ch~amatagr. 290, 393-399. 15. Beutler, E. (1975) Red Cell Metabolism, pp. 8-18, Grune & Stratton, New York. 16. Grass, R. T., Schroeder, E. A,, and Gabrio, 8. W. (1966) J. Clin. Invest. 45, 249-255. 17. Omachi, A.. Scott, C. B., and Hegarty, H. (1969) Bighorn. Siuphys. Acta 184, 139- 147. ts. Marshall, W. E., and Omachi, A. (1974) Biochim. Biophys. Acta 354, l- 10. 19. Reinauer, H., and Brims, F. H. (1964) Z. Physio!. Chem. 337,93- 100, 20. Scholar, E. M., Brown, P. R., Parks, R. E., Jr., and Calabresi, P. (1973) Blood 41, 927-936. 21. Dean, B. M., Perrett, D., and Sensi, M. (1978) Biochem. Biophys. Res. Commun. 80, 147-154. 22. Ericson, A., Niklasson, F., and de Verdier, C, H. (1983) Cl&. Chim. Acta 127, 47-59.