Technical Notes Anal. Chem. 1994, 66, 2981-2984
Quantitative Analysis of Ribonucleotide Triphosphates in Cell Extracts by High-Performance Liquid Chromatography and Micellar Electrokinetic Capillary Chromatography; A Comparative Study Arianna Loregian,t Carlo Scremln,ti* Mauro Schiavon,s Alessandro Marcello,t and Glorgio PalG'ft Institute of MicrobiologyI University of Padova, via A. Gabelli 63, 35121 Padova, ItalyI and Beckman Analytical S.p.A., Via Roma 108, Palazzo F / l I 20060 Cassina De'Pecchi (MI)I Italy
Two analytical methods have been evaluated for their ability to separate and quantitate the intracellular pool of free ribonucleotide triphosphates in four different cell lines. Micellar electrokinetic capillary chromatography and highperformance liquid chromatography were compared in terms of speed, sensitivity, and efficiency of analysis, with the former being 3 times faster (about 10 vs 30 min) and having an average minimun detectable quantity of 50 fmol vs 30 pmol and approximately 950 000 theorical platedm vs 4500. Different protocols of extraction of nucleotides from whole cells were also evaluated with the two analytical methods. Micellar electrokineticcapillarychromatographyproved to be a powerful tool for the reproducible and reliable quantitation of ribonucleotide triphosphates, allowing the fast processing of extremely small volumes of sample. Detection, identification and quantitation of intracellular unpolymerized free nucleosides and nucleotides in vivo are of importance in studying various aspects of cellular metabolism. The 5'-mono-, di-, and triphosphates of adenosine, guanosine, uridine, and cytidine, the four ribonucleotides found in RNA, represent the bulk of the cellular ribonucleotides. Ribonucleotide triphosphates (rNTPs) are the immediate precursors of polyribonucleotides and, among other functions, intervene in bioenergetic reactions by donation of a phosphate group and in activation reactions by the transfer of nucleoside mono- and diphosphates to acceptor molecules.' Several nucleoside analogues, such as the most powerful antiviral compounds, are activated by viral or cellular kinases to triphosphate + University of Padova. *Present address: Center for Biologics Evaluation and Research, Food and Drug Administration, 8800 Rockville Pike, Bethesda, MD 20892. 8 Bcckman Analytical S.p.A. (1) Adams, R. L. P.; Burdon, R. H.; Campbell, A. M.; Leader, D. P.; Smellie, R. M. S. The Biochemistry of rhe Nucleic Acid, 9th ed.;Fletcher and Son Ltd.: Norwich, UK, 1981. (2) Shaeffer, H. J.; Beauchamp, L.; De Miranda, P.; Elion, G. B. Nature 1978, 272, 583-585. (3) Furman, P. A.; Fyfe, J. A,; St. Clair, M. H.; Weinhold, K.; Ridwut, J. L.; Freeman, G. A.; Nusinoff Lehrman, S.;Bolognesi, D. P.; Brcder, S.;Mitsuya, H.; Barry, D. W. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 8333-8337.
0003-2700/94/0366-2981$04.50/0 0 1994 American Chemical Society
Many techniques have been developed over the past few years to detect and measure cellular nucleotides. Separation and analysis of these molecules have been performed primarly with high-performance liquid chromatography (HPLC),4gas chromat~graphy,~ and enzymatic assays6 These methods, although extensively used, for biological studies, are time consuming, particularly HPLC. In fact, while resolution is generally adequate by this technique, run times are fairly long, e.g., 30-45 min, and in addition, a discrete quantity of biological material is required. Intracellular nucleotideprofiles have been obtained with shorter run time^,^-^ but usually at the expense of separating and identifying fewer nucleotides. As an alternative to HPLC, capillary electrophoresis (CE)-in the capillary zone electrophoresis (CZE) l o or micellar electrokinetic capillary chromatography (MECC) modes-has been proposed. Run times are typically much shorter than those using HPLC (5-20 min), and excellent peak efficiency and resolution are achievable. Here we have performed the extraction of nucleotide pools from four different cellular lines using both acidic and alcoholic solvents. The extracts have been analyzed and ribonucleotide triphosphates have been quantitated with two different separating techniques: HPLC and MECC. MECC has been showed to be a relatively simple technique with a rapid separation time (about 10-12 min), impressive resolving power, small sample size, and high efficiency (hundreds of thousands of theorical plates). Therefore, this technique may offer significant advantages in the separation of nucleic acid precursors with respect to HPLC. (4) Pogolotti, A. L., Jr.; Santi, D. V. Anal. Biochem. 1982,126, 335-345.
( 5 ) Ettre, L. S. In The Practice of Gas Chromatography; Ettre, L. S.,Zlatkis, A., Eds.; Interscience: New York, 1967. (6) Skoog, L. Eur. J . Biochem. 1970, 17, 202-208. (7) Stocchi, V.; Cucchiarini, L.; Canestrari, F.; Piacentini, M. P.; Fornarini, G. Anal. Biochem. 1987, 167, 181-190. (8) Maessen, J. G.; van der Vusse, G. J.; Vork, M.; Kootstra, G. Clin. Chem. 1988, 34, 1087-1090. (9) Hammer, D. F.; Unverferth, D. V.; Kelley, R. E.; Harvan, P. A,; Altschuld, R. A. Anal. Biochem. 1988,169, 300-305. (10) Jorgerson, J. W,; DeArman Lukacs, K. Science 1983, 222, 266-272. (11) Liu, J.;Banks, J. F.,Jr.;Novotny,M.J.MicrocolumnSep. 1989, I, 136-141.
Analytical Chemistry, Vol. 66, No. 18, September 15, 1994 2Q6l
MATERIALS AND METHODS High-Performance Liquid Chromatography. HPLC analyses were performed using a LKB 2150 two-pump apparatus, with a 2158 Uvicord SD absorbance detector (A254 ,,m) and a Shimadzu C-R3A Chromatopac integrator. The samples were injected onto an Ultrasil SAX column (10 pm, 4.6 X 250 mm, Beckman). The column was eluted with 0.4 M NH4H2PO4 pH 3.5 (adjusted with H3P04) at a flow rate of 1.5 mL/ min and at room temperature.12J3 The maximum injection volume of the extract sample was 100 pL. Peaks were identified by coelution with authentic nucleotide samples (Boheringer), and the quantity of each ribonucleotide in the extracts was determined by referring to separate experiments in which the areas of the peaks of known amounts of rNTPs were measured. The relationship between peak area and standard nucleotide concentration was linear between 0.5 and 20 nmol (see also ref 14). To achieve the best performance, the Ultrasil SAX column was washed with 25 mL of 0.4 M H3P04 prior to a series of analyses.12 Micellar Electrokinetic Capillary Chromatography. Capillary electrophoresis was performed using a P/ACE 2100 system (Beckman Instruments). The capillary cartridge contained a 50-pm4.d. untreated fused silica capillary tubing of 67-cm total length, 60-cm effective length to UV detector. Detection was accomplished by an on-column UV-absorbance detector at 254 nm. Runs were performed with a 50 mM sodium phosphate, 100 mM dodecyltrimethylammonium bromide (DTAB), 1 mM EDTA pH 7.0 buffer, under the following conditions: T = 25 OC, V = 15 kV, I = 100 pA. Polarity was also reversed (positive ground). Before use the column was sequentially pretreated with 1 M NaOH for 3 min forward and 3 min reverse, milliQ water for 5 min, and 0.1 M NaOH for 2 min. Finally, capillary was equilibrated with the buffer to be used for the analysis. Samples were diluted 1:3 and injected under pressure for 5 s. Peaks on the electropherogram were identified by spiking with pure samples of standard ribonucleotides. Concentrations of the ribonucleotides in extracts were determined by using the peak areas of standards at a known concentration. The relationship between peak area and standard nucleotide concentration was linear between 0.05 and 2.5 pmol. For each type of extract, at least three runs were performed and standard error was calculated. Preparation of Cell Extracts. African green monkey kidney cells (Vero), mouse embryo fibroblasts (PA3 17), human larynx epidermoid carcinoma cells (HEp-2) and human epitheloid carcinoma cells (HeLa S3)were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heattreated fetal calf serum (FCS). Five to 10 million confluent cells, counted in a hemocytometer after trypsinization of the monolayers, were resuspended with 15 mL of ice-cold phosphate-buffered isotonic saline solution (PBS) and harvested by centrifugation. Pellets were washed with 1 mL of PBS and then treated with 50 pL of either 5% trichloroacetic acid (TCA),l* 10%TCA, or 0.5 M perchloric acid (PCA) in PBS13 on ice for 30 min. After centrifugation, the acidic
Flguro 1. HPLC profile of ribonucleotldetriphosphate pools. Confluent Vero cells were treated with 10% TCA and rNTPs separated onto an Ukrasll SAX column. A volume equivalent to 2 X loe cells was inJected. Separationswere performedas described under Materialsand Methods.
supernatants were separated and neutralized twice by adding a 1.1 vol of cold Freon (1,1,2-trichlorotrifluoroethane) containing 0.5 M tri-n-octy1amine.l2 Alternatively, the cell monolayers were washed twice with ice-cold PBS and treated with 1 mL of cold 60% methanol for 30 min at 4 O C 6 Extracts were desiccated and resuspended with 10 mM TRIS-HCl pH 7.5. Samples were stored at -80 OC until the analysis, which was performed within the next two to three weeks. Three to five extracts for each cell line were analyzed under the different extraction procedures reported above.
(12) Tanaka, K.; Yoshioka, A.; Tanaka,S.; Wataya,Y. Anal. Biochem. 1984,139,
RESULTS HPLC Separation and Quantitation. The procedure used was substantially that of Garret and Santi,13 modified by Tanaka et al.12 This method gives an excellent separation of all four ribonucleotide triphosphates in about 30 min, as shown in Figure 1. The efficiency of the separation can be expressed by the number of theorical plates per meter that in our experimental conditions was on the average 4500 plates/m, with a minimum detectable mass (MDM) of approximately 30 pmol. MECC Separation and Quantitation. Efficient separation of rNTPs, with high theoretical plate numbers (an average of 950 000 plates/m), was obtained with MECC (Figure 2), using an untreated fused silica capillary and including a cationic surfactant (DTAB) at a concentration of 100 mM in the running buffer. It has been alreadyshownl' that DTAB gives a better separation of ribonucleotides (mono-, di-, and triphosphates) compared to the anionic detergent SDS proposed by others.l5. Electroosmotic flow reversal occurs16
35-41. (13) Garret, C.; Santi, D. V. Anal. Biochem. 1979, 99, 268-213. (14) Hartwick, R. A.; Brown, P. R. J. Chromatogr. 1975, 112, 651-662.
(15) Cohen, A. S.;Terabe,S.;Smith, J. A,; Karger, B. L. Anal. Chem. 1987,59, 1021-1 027.
2982
Anaiyticaichemistry, Vol. 66, No. 18, September 15, 1994
coU
r
Table 1. Extractlon of Ribonucleotide Triphosphates from Vero Cells.
extraction solvent
concnb (nmol/ 106 cells) GTP UTP
ATP
TCA, 5% TCA, 10% PCA, 0.5 M MetOH, 60%
1.60f 0.06 1.32f 0.27 2.79f 0.11 2.90 f 0.05 2.33 + 0.29 2.43 f 0.19 5.5 f 0.50 5.8f 0.56
0.48f 0.02 0.41 f 0.05 0.81 f 0.08 0.61 f 0.10 0.58f 0.14 0.39f 0.11 1.31 f 0.12 1.30 0.19
*
0.26 f 0.03 0.35f 0.10 0.49f 0.03 0.40f 0.02 0.52 f 0.10 0.49 f 0.09 1.48 f 0.21 1.55 f 0.15
CTP
0.18f 0.01 0.22 f 0.04 0.31 f 0.01 0.37f 0.07 0.28f 0.08 0.21 f 0.06 0.81 f 0.11 0.53f 0.07
a Confluent Vero cells were treated with various acidic and alcoholic solvents as described under Materials and Methods. b First row of values for each solvent obtained from HPLC analysis. Second row of values for each solvent obtained from MECC analysis. Values are the means f standard errors obtained from three to five separated extractions.
5.50 8.00
8.00
10.00
12.00
Time (min)
Flgure 2. MECC electropherogramof ribonucleotkletriphosphatepools. Cofiuent Vero cells were treated with 10 % TCA to extract nucleotides as described under Materials and Methods. Operating conditions: constant voltage, 15 kV; current, 100 pA; buffer, 50 mM sodium phosphate, 100 mM DTAB, 1 mM EDTA pH 7; temperature, 25 OC; effective capillary length, 80 cm; capillary diameter, 50 pm; pressure injection for 5 s (about 5 nL). See text for details.
Table 2. Ouantltatlve Analyris of Rlbonucleotkk Trlphorphatrr from Four Cell Lines'
concn" (nmol/ 106 cells) cell line Vero HeLa S3
PA317
above the critical micellar concentration (cmc) of DTAB (14 mM), so that the polarity of MECC must be reversed in order to reestablish electroosmotic flow in the original direction. To further increase peak efficiency, 1 mM EDTA was added to the running buffer as suggested by Perret and Ross.17 Between each run the capillary was treated with 0.1 M NaOH for 1 min and reequilibrated first with the running buffer without DTAB for 2 min and then with the running buffer containing the detergent for a further 2 min to increase reproducibility of the analysis. Particularly, the length of the retention times was observed to increase when washes were omitted. This treatment dissociated all bound materials from the capillary wall and provided the same silica surface for repeated runs, resulting in reproducible electropherograms. Under these conditions, run-to-run variation in elution times (as percent relative standard deviation, RSD%) was 0.380.71%. Samples were diluted in milliQ water (optimal dilution was observed at 1:3) prior to injection to decrease the ionic strength of the solution. In fact, samples from biological origin frequently contain significant amounts of "matrix" salts or buffer ions that can dramatically influence resolution with salt-related peak-broadening effects.'* Injection volumes introduced into the capillary by pressure (0.5 psi) were calculated according to the Poiseuille equation. Each injection was performed in triplicate. The minimun detectable concentration (MDC), calculated by 5-nL injections of a known solution of standards, was approximately 10 pM (corresponding to 50 fmol), with a signal-to-noise ratio of 3. This detection limit compares favorably with results reported from other (16) Otsuka, K.; Terabe, S.; Ando. T. J. Chromatogr. 1985, 332, 219-226. (17) Pcrrett, D.; Ross, G. In Purine and Pyrimidine Metabolism in Man VU; Harkncss, R.A., Ed.; Plenum Prcss: New York, 1991; Part B. (18) Nguyen, A.-L.; Luong, J. H. T.; Masson, C. A m / . Chem. 1990,62, 2490-
2493.
HEp2
ATP
GTP
UTP
CTP
2.79 f 0.1 1 2.90 f 0.05 1.79f 0.29 1.96f 0.18 1.35 f 0.09 1.44f0.11 1.96f 0.23 2.01 f 0.14
0.81 f 0.08 0.61 f 0.10 0.60f 0.09 0.41 f 0.07 0.80& 0.05 0.73f 0.07 0.74f 0.20 0.92 f 0.31
0.49f 0.03 0.40f 0.02 0.44 f 0.01 0.43 f 0.04 0.82f 0.06 0.87f 0.10 0.73f 0.09 0.87f 0.15
0.31 f 0.01 0.37f 0.07 0.13f 0.02 0.24f 0.05 0.12 f 0.03 0.14f 0.06 0.21 f 0.02 0.23f 0.09
Nucleotide pools from four different cell lines were extracted with 10% TCA as described under Materials and Methods. Results obtained from both HPLC and MECC analysis are reported as described in the legend to Table 1. @
workers using CZE19 and also MECC.11g20 Extraction and Quantitation of Ribonucleotide Pools. Extraction of nucleotide pools from cells can be achieved by use of either acidic solvents or organic solvents. As shown in Table 1, comparable levels of rNTPs were obtained with 10% TCA and 0.5 M PCA, while extraction with 5% TCA, proposed by others,'* gave lower amounts of rNTPs. Methanolic extraction provided the best yield of rNTPs (2-3 times higher than that obtained with either TCA or PCA). Four cell lines were analyzed (Table 2) for their levels of intracellular rNTPs after extraction with 10% TCA as this procedure provided the lowest amounts of protein carryover in samples (data not shown). All four cell lines grow on monolayers and are among the most commonly used in virological and biological laboratories. As expected, the ribonucleotide with the highest concentration was ATP, whose levels ranged from 1.44 to 2.90 nmol/106 cells. In general there was a good agreement between values obtained by HPLC and MECC. The observed variations in the level of nucleotide pool among the different cell lines were probably due to differences in cell volumes or in the activity of enzymes involved in nucleoside and nucleotide metabolism.
DISCUSSION A good quantitative analysis of intracellular ribonucleotide pools is of relevance in several chemical assays, as ribonucleo~~~
(19) Huang, M.; Liu, S.; Murray, B. K.; Lee, M. L.Anal. Biochem. 1992, 207,
231-235.
(20) Le-, A.-F.; Leuratti, C.; Marafantc, E.; Di Biase, S. J . High. Resolur. Chromatogr. 1991, 14, 667-671.
Analytical Chemistry, Vol. 66,No. 18, September 15, 1994
2983
tides play a vital role in cell bioenergetics and signal transduction pathways. A method that would quickly and easily measure these nucleotides could find many biological and clinical applications. Capillary electrophoresis has been successfully used to analyze a diverse array of molecules, including organic and inorganic ions,21 amino acids and peptides,22923 oligonucleot i d e ~and ,~~ but to date, there are only a few reports on capillary electrophoresis applied to biological samples to separate and quantitate cell nucleotide p001s.'7326-27To our knowledge, MECC studies on nucleotide separation have been conducted only on highly purified materials reconstituted in appropriate buffers. 1,20 Our investigations indicated that MECC was a valid alternative to more traditional liquid chromatography for simple and quantitative analysis of intracellular rNTPs. MECC had proven to be superior to HPLC in terms of speed, small sample size, efficiency, and sensitivity. Run time for HPLC under standard analytical procedure was about 30 min, whereas the same analysis with MECC was about 3 times faster (10-12 min) under our experimental conditions. In addition, the untreated fused silica capillary required less maintainance than a HPLC column and operations such as washing and reequilibrating could be greatly shortened. Over the course of a large number of runs, MECC was a great time-saving method when compared with HPLC. Another unique aspect of MECC was the small injection volume used for each run. This could be extremely important when sample amount was limited, e.g., when only a small number of cells was available for analysis. The possibility of using small samples could be particularly useful when dealing with biological material in short supply and with the need to save reagents. As far as efficiency and sensitivity are (21) Gross, L.; Yeung, E. S.Anal. Chem. 1990,62, 427-431. (22) Cobb, K. A.; Novotny, M. Anal. Chem. 1989, 61, 2226-2231. (23) Grossmann, P. D.; Wilson, K. J.; Petrie, G.; Lauer, H. H . Anal. Biochem. 1988, 173, 265-270. (24) Guttman, A.; Cohen, A. S.;Heiger, D. N.; Karger, B. L. Anal. Chem. 1990, 62, 137-142. (25) Fujiwara, S.; Honda, S.Anal. Chem. 1987, 59, 2773-2778. (26) Takigiku, R.; Schncider, R. E. J . Chromatogr. 1991, 559, 247-256. (27) Ng, M.; Blaschkc, T. F.; Arias, A. A.; Zare, R. N. Anal. Chem. 1992, 64, 1682-1684.
2984
AnaIyticalChemistry, Vol. 66, No. 18, September 15, 1994
concerned, MECC was shown to be far superior to HPLC under our experimental conditions. The number of theorical plates per meter was 950 000 for MECC vs 4500 for HPLC and minimum detectable mass was 50 fmol for MECC vs 30 pmol for HPLC. In addition, for a comparative examination of the characteristics of the two analytical methods, in the present study we also investigated optimal conditions for obtaining cell extracts suitable for nucleotide analysis, involving treatment of cells with acidic or organic solvents. To be effective, acidic extraction of nucleotide pools had to be performed on cells in suspension, after trypsinization of cell monolayers; besides, the solution was neutralized by treatment with Freon and tri-n-octylamine. Methanolic extraction, instead, could be performed directly on cell monolayers and without neutralization of the samples, thus avoiding two laborious and timeconsuming steps in the protocol. In this regard, it is noteworthy that the shorter the time of extraction the higher the yield of ~ N T P sand , ~ this could explain the higher values found for all four rNTPs using this method of extraction. In summary, the presented MECC method appears promising for the study of intracellular rNTPs and has several advantages over HPLC, particularly when a great number of samples is to be processed. There are many interesting areas for future applications of this technique. Automated versions would be useful in routine analyses such as separation of cellfree ribonucleotides in clinical or biological laboratories. Moreover, MECC could be adopted for the measurement of nucleotide analogs used in chemotherapy and their metabolites. Present analysis of these molecules is often conditioned by availability of small-volume samples and by sensitivity limits.
ACKNOWLEDGMENT We thank Dr. S . Di Biase for helpful discussion and Professor Bonora for critically reading the manuscript. Financial support for this work was provided by the Istituto Superiore di Sanitri of Italy, progetto AIDS. Received for review February 21, 1994. 1994."
Accepted May 26,
*Abstract published in Advance ACS Abstracts, July 15, 1994.
