High resolution separation and quantitation of ribonucleotides using capillary electrophoresis

ANALYTICAL

207,231-235

BIOCHEMISTRY

High Resolution Ribonucleotides Mingxian

Huang,

of Chemistry

Received

12, 1992

June

Separation and Quantitation of Using Capillary Electrophoresis

Shaofan

Departments

(19%)

Liu, Byron

and Microbiology,

K. Murray, Brigham

Young University,

A method for the analysis of ribonucleotides using capillary electrophoresis has been developed. A crosslinked polyacrylamide coated column, Tris-HCl, and phosphate-mixed buffer were used, which produced reproducible separations (solute migration time average RSD% of 1.2) of 14 ribonucleotides within 50 min. Linear relationships between peak areas and sample concentrations, an average minimum detectable concentration of 5.4 FM, and an average minimum detectable quantity of 0.08 pmol were obtained. The described method allowed reproducible and reliable qualitative and quantitative analysis of intracellular ribonucleotide pools in HeLa cells. 6) if)92 AC&de& P~~SS, 1~

Ribonucleotides in mammalian cells play key roles in many vital biochemical processes. They are the activated precursors of DNA and RNA, and therefore their levels control the syntheses of DNA and RNA molecules (1). The levels of GTP and GDP pools are closely related to cell signal transduction (2), and the ATP/ (ADP+AMP) ratio has a central role in cell energy metabolism and cellular regulation (3). Moreover, the analysis of intracellular ribonucleotide pools is an important aspect of antitumor and antiviral research (4). These ribonucleotide pools consist mainly of 14 basic ribonucleotides of similar structure which are present in the majority of cells at very low concentrations. A method that can provide both high resolution separation and reliable quantitation of these ribonucleotides is strongly desirable. Several chromatographic and electrophoretic approaches have been used to measure ribonucleotides. Anion-exchange (56) and ion-pairing reversed-phase

’ To whom

correspondence

should

0003-2697192 $5.00 Copyright Q 1992 by Academic Press, AlI rights of reproduction in any form

be addressed.

and Milton

L. Lee’

Provo,

Utah 84602

LC? (7,8) have been found to be suitable for the separations of ribonucleotides. However, these methods require large amounts of biological material and involve time-consuming procedures. Furthermore, several important ribonucleotides cannot be sufficiently resolved. Capillary electrophoresis (CE) has been shown to be very applicable to the analyses of biological samples; its advantages over other separation techniques are that only nanoliters of sample are consumed in each analysis, and that rapid, highly efficient, sensitive, and automated analyses of complex mixtures can be relatively easily achieved (g-12). Recently, CE has been used to determine nucleotides in fish tissues (13), to quantitate ribonucleotides from base-hydrolyzed RNA (14), and to analyze NTP and dNTP pools in mammalian cells (15). However, only several ribonucleotides were reported in these studies. In the present study, we have investigated the separation and quantitation of 14 common ribonucleotides in a HeLa cell extract using CE with a crosslinked polyacrylamide-coated column. MATERIALS

AND

METHODS

Reagents and materials. All ribonucleotide standards and Tris-HCl were obtained from Sigma (St. Louis, MO). Acrylamide was purchased from Aldrich (Milwaukee, WI); methylene bisacrylamide was purchased from Spectrum (Gardena, CA); azobisisobutyronitrile (AIBN) was obtained from Schweizerhall (South Plain Field, NJ); and 7-act-1-enyltrimethoxysilane was obtained from Petrarch (Bristol, PA).

’ Abbreviations used LC, liquid chromatography; CE, capillary electrophoresis; AIBN, azobisisobutyronitrile; EMEM, Eagle’s minimum essential medium; SDS, sodium dodecyl sulfate; MECC, micellar electrokinetic capillary chromatography; RSD, relative standard deviation; MDQ, minimum detectable quantity; DTAB, dodecyltrimethylammonium bromide; HTAB, hexadecyltrimethyIammonium bromide. 231

Inc. reserved.

232

, 0

HUANG

, 20

IO

FIG. 1.

Electropherogram ditions: 50 mM Tris-HCl buffer (pH 5.3); -16 kV electromigration injection given in Table I.

30

40

50

min

of 14 standard ribonucleotides. CE con+ 30 mM sodium and potassium phosphate applied voltage (injection at cathode end); at -15 kV for 5 s. Peak identifications are

HeLa cells were grown in Eagle’s minimum essential medium (EMEM) with nonessential amino acids solutions, supplemented with 10% bovine calf serum, 20 mM Hepes buffer (pH 7.35), and 50 pg/ml of gentamicin reagent. The procedure used for extraction of the ribonucleotides from these cells with an aqueous acid solution was a modified procedure from Ref. (16). The extract was analyzed by CE without any further treatment. Apparatus. A Model CES-1 capillary electrophoresis system (Dionex, Sunnyvale, CA) was used in this study. Sample intruduction was performed by electromigration. Fused silica capillaries (polymicro Technologies, Phoenix, AZ) of 50 pm i.d. (360 pm o.d.) and approximately 80 cm in length were used. The ribonucleotides were detected by on-column uv absorbance at 254 nm. Data were collected using a Model SP4270 integrator (Spectra Physics, San Jose, CA). Capillary columnpreparation. Details of the preparation of cross-linked polyacrylamide coated columns have been reported previously (17). Capillaries were pretreated with 7-act-1-enyltrimethoxysilane and statistically coated with a solution of 50 mg of acrylamide, 10 mg of methylene bisacrylamide, and 10 mg of AIBN in 4 ml of methylene chloride. Each capillary was then sealed at both ends, heated from 50 to 120°C for 4’C mine’, and held at 12O’C for 2 h for polymerization. The capillary was rinsed with 5 ml of methylene chloride and 5 ml of buffer before CE use. RESULTS

AND

DISCUSSION

of ribonucleotides. The selectivities Separation achieved in CE separations are mainly determined by the differences in electrophoretic mobilities of solutes, which are influenced by buffer composition and buffer

ET

AL.

pH. The resolution in CE separations is also significantly affected by electroosmotic flow, which refers to the bulk buffer flow caused by the electrical double layer on the capillary inner surface and the applied voltage. Nucleotides are strong acids with pK= values of approximately 1. When the CE buffer pH is above 1, nucleotides are negatively charged and will move from the cathode to the anode. For an untreated capillary column, a strong electroosmotic flow with direction from the anode to the cathode exists. Therefore, the migration of the ribonucleotides is against the electroosmotic flow which greatly increases the analysis times of some compounds and reverses the direction of others. Micellar electrokinetic capillary chromatography (MECC) (18) has been used to control the electroosmotic flow and enhance the separation of the ribonucleotides by adding surfactants such as SDS, DTAB, and HTAB to the buffer solution (19). Chemically modifying the capillary surface to eliminate the electroosmotic flow (17,2024) is another alternative. In this study, we chose a cross-linked polyacrylamide coated column for the CE separation of ribonucleotides. We found that the electroosmotic flow was eliminated on this column in the buffer pH range from 3.0 to 9.0, and this coating was quite inert to biological samples (17). Therefore, the effect of electroosmotic flow on resolution is not a factor in this study. Ribonucleotides are readily hydrolyzed in high or low pH buffers, but they are relatively stable over pH 5 to 7 (25). We found that Tris-HCl and a sodium and potassium phosphate mixed buffer gave both highly efficient

TABLE CE

Performance

Migration Peak No. 1 2 3 4 5 6

7 i r. 11 12 13

14

Compound UTP CTP ATP GTP UDP CDP ADP GDP XMP UMP CMP IMP AMP GMP

a CE conditions

1

Measurements

Mean

for

time RSD%b

24.13 24.58 25.24

25.93

1.0

1.5

10.0

28.03

1.2

1.5

30.07 30.86 32.16

1.2 1.2 1.3

41.98

1.3

44.48 45.06

1.4 1.2

45.49

1.3

46.84

1.3

same

as those

1 -mu-m.-.

*n=5.

’ Based

on a signal-to-noise

ratio

Minimum detectable cont. (PM)’

1.1 1.1 1.0

29.33

are the

Ribonucleotidesa

of 3.

0.15 0.20 0.08 0.11 0.09 0.08 0.07 0.04 0.05 0.08 0.09 0.05 0.04 0.04

4.3 6.0 5.0 5.0 4.0 2.7 3.5 6.7 8.6 4.3 3.5 4.0

described

Minimum detectable quantity bnolY

in the

legend

to

233

CAPILLARY ELECTROPHORESIS OF RIBONUCLEOTIDES ATP

GTP UMP aP

AMP EP

ADP

XMP

UP

CMP CTP

1000 0 0

100

Concentration FIG.

2.

Plots

of peak

area

as a function

of solute

200

0

(PM) concentration.

and highly resolved separations. When this buffer was adjusted to pH 5.3, all 14 ribonucleotides were completely resolved within 50 min, as shown in Fig. 1, which compares favorably with the typical LC results in which analysis times were more than 80 min and coelution of peaks still occurred (5). We noted that at higher buffer pH (for example, pH 6.3), IMP and AMP coeluted, while at lower buffer pH, solute migration times rapidly increased and the peaks became much broader. The applied voltage was selected based on the desire to minimize the migration times and maximize the peak separations. Migration time reproducibility. Table 1 gives the mean migration times of 14 ribonucleotides and their relative standard deviations (RSD%). In LC, gradient elution is usually required to obtain optimal resolution of the nucleotides. Column reequilibration is necessary between analyses, which lengthens the required analysis times. Using untreated capillary columns in CE, 1 N NaOH was needed to rinse the capillary after each run (13), or a period of several hours was required to condition the capillary with a surfactant containing buffer (14) in order to obtain reproducible electropherograms. In our work, we obtained reproducible electropherograms (migration time RSD%s are close to 1) for approximately 50 injections without additional rinsing of the

100

200

Concentration CE conditions

are the

same

(PM)

as those

described

in the legend

to Fig.

1.

capillary. This was mainly a result of the use of a crosslinked polyacrylamide coated column which is stable, inert, and compatible with biological samples. Quantitation. In CE, the quantity of sample (Q) introduced into the capillary column by electromigration is given by Q = ZAC,

PI

where 1 is the length of the sample zone, A is the crosssectional area of the capillary, and C is the sample concentration. Because the electroosmotic flow was zero for the column used in this study, 1 is determined by the product of the injection time (tinj) and solute ion velocity during injection (ui~j). The value of uti is equal to the ratio of the injection voltage (Vi,,j) to the operational voltage ( VOP) times the solute ion velocity during operation (u&; the latter can be calculated from the column length and solute migration time. Thus, 1 is given by

where L is the column length and tr is the sample migration time under the operational voltage, VOP. Combining Eqs. [1] and [2], we obtain the relation

234

HUANG 3

0

, 10

FIG. 3. Electropherogram HeLa cells. CE conditions legend to Fig. 1.

,

.

20

30

40

of an extract of ribonucleotides are the same as those described

50

min

from in the

L31 In this study, Eq. [3] was used to calculate the minimum detectable quantity (MDQ) by injecting a known solution of solute standards and measuring the signal-tonoise ratio (S/N). The MDQ was defined as the quantity that gave a S/N = 3. The results are given in Table 1. The MDQs are in the low pmol to fmol range, which is much lower than that of conventional LC techniques (20-50 pmol) (26), lower than that of a reported CE method in which surfactants were added to the mobile phase (about 9 pmol) (19), and comparable to that of a microcolumn ,LC method (lo-50 pg at S/N = 2) (27). The smaller i.d. column has superior mass sensitivity due to the high efficiency achieved with the column. The higher detection limits experienced with MECC techniques may result from the effect of the surfactant on the sensitivity of the uv-absorbance detector. We should mention that another advantage of CE over conventional LC for this application is that only 100 ~1 of total sample volume is enough for numerous CE analyses. Furthermore, on-column fluorimetric detection (28,29) on an open tubular column and CE/MS (30,31) can provide much lower minimum detection limits than uv absorbance. Figure 2 shows plots of ribonucleotide peak area vs concentration. Linear relationships between peak area and concentration are found for all 14 ribonucleotides within the concentration range of concern.

ET

AL.

Application. Figure 3 shows an electropherogram of a HeLa cell extract. Table 2 lists the quantitative results. UTP and ATP gave sharper peaks from the cell extract than from the standard mixture. The improvement in peak shapes can be explained by the so-called “sample stacking effect” (32) which occurs when a voltage is applied along a capillary tube containing a sample plug with a lower specific conductivity than that of the surrounding running buffer, resulting in a concentration of the analyte zone. Since the standard mixture was prepared with the running buffer, there was no sample stacking effect. However, the cell extract contained a lower salt concentration, which resulted in lower conductivity and sample stacking to produce sharper peaks. To verify this explanation, we analyzed a standard sample in the concentrated buffer, and then diluted the sample four times with distilled water and analyzed it again without changing any other conditions. We found that the peak heights in the electropherogram of the diluted sample were two-thirds of those of the concentrated sample and the peak efficiencies were twice those of the concentrated sample. The mean migration times listed in Table 1 are a little different than those listed in Table 2, which may result from the difference in solution compositions of the standard mixture and cell sample. Another contributing factor is that there was no temperature control on the column in this work, which could affect the buffer solution viscosity and, hence, the migration time reproducibility. Therefore, standard compounds were added to the cell extract for migration time verification, and the ratios of the intensities of the 280-

TABLE Results

from

the CE Analysis in a HeLa Cell

2 of Ribonucleotides Extract0

Migration Peak No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Compound

Mean

n CE conditions Fig. 1. bn=4.

RSD%b

24.59 24.91 25.42 25.92 27.39 29.96 30.76

UTP CTP ATP GTP UDP GDP ADP GDP XMP UMP CMP IMP AMP GMP

same

Amount in HeLa (rim01 1.5 X 1Or

0.9 0.9 0.9 1.2 1.0 0.6 0.6 0.3 0.8 0.9

41.74 45.25 46.64 are the

time

as those

described

found cells per cells)

15.7 3.6 42.4 14.3 3.3 Trace 27.4 3.7 Trace 3.0 Trace Trace 6.5 0.9 in the

legend

to

CAPILLARY

ELECTROPHORESIS

to 254nm absorption bands were compared to those of known standards. Both types of information were used for compound identification. Since these analyses were carried out under carefully controlled conditions (e.g., the same pH), and most other interfering biological materials were removed from the samples before analysis, few sample matrix problems were encountered. It can be seen from Fig. 3 and Table 2 that high resolution and reproducible separations were achieved for the cell sample ribonucleotides. Using this CE approach, sequential analyses can be rapidly and automatically performed. In most cases, reproducible separations are impossible in CE when using untreated columns for real biological samples. The main reason for this is that biological samples always contain macromolecular compounds which tend to be adsorbed onto the active capillary surface, resulting in changes of electroosmotic flow and solute migration times. The stable and inert crosslinked polyacrylamide coated column played an important role in the cell sample analysis reported in this study. The various unidentified peaks in Fig. 3 may possibly be NAD, NADP, UDPG, or IDP, and current efforts are being directed toward their positive identification.

was supported CA 94088-3603).

by a grant

from

the Dionex

Corporation

I. S., and Chem.

Toguzor,

49,

2231.

9. Jorgenson, J. W., and Lukacs, K. D. (1983) Sc&ce 222,266. 10. Lauer, H. H., and McManigill, D. (1986) An&. Chem. 58,166. 11. Cohen, A. S., Terabe, S., Smith, J. A., and Karger, B. L. (1987) An&. Chem. 59, 1021. 12. Wallingford, R. A., and Ewing, A. G. (1988) Anal. Chem. 60,1972. 13. Nguyen, A., Luong, J. H. T., and Masson, C. (1990) An&. Chem. 62,249O. 14. Huang, X., Shear, J. B., and Zare, R. N. (1990) AnaL Chem. 62, 2049. 15.

Takigiku, 247.

16. Smee, thews,

R.,

and

19. Liu,

Schneider,

R. E. (1991)

D. F., Boehme, R., Chernow, M., T. R. (1985) Biochem. Pharmacol.

17. Huang, M., Vorkink, Sep. 4,233. 18. Terabe, S., Otsuka, (1984) Anal. Chem. J., Banks,

W.

P., and

K., Ichikawa, 56, 113.

Lee,

M.

J. Chromatogr.

559,

Binko, B. P., and 34,1049.

Mat-

L. (1992)

K., Tseuchiya,

J. F., Jr., and Novotny,

M. (1989)

J. Microcol.

A., and Ando, J. Microcol.

T. Sep.

1,136. 20.

McCormick,

21.

Bruin, G. J. M., Chang, J. P., Kuhlman, R. M., Zegers, K., Kraak, J. K., and Poppe, H. (1989) J. Chromatogr. 471,429. Towns, J. K,, and Regnier, F. E. (1990) J. Chromatogr. 516, 69. Cobb, K. A., Dohrik, V., and Novotny, M. (1990) Anal. Chem. 62, 2478. Huang, M., Vorkink, W. P., and Lee, M. L. (1992) J. MicrocoZ. Sep. 4, 135. Perrett, D. (1987) in CRC Handbook of Chromatography: Nucleic Acids and Related Compounds (Krstulovic, A. M., Ed.), Vol. 1, Part B, p. 3, CRC Press, Boca Raton, FL. Dwyer, M. E., and Brown, P. R. (1988) J. High ResoZut. Chromatogr. Chromatogr. Commun. 11, 419. Banks, J. F., Jr., and Novotny, M. (1989) J. Chromatogr. 475,13.

24. 25.

REFERENCES 1. Bray, G., and Brent, T. P. (1972) Biochim. Biophys. Acta 269, 184. 2. Kharbabda, S. M., Sherman, M. L., and Kufe, D. W. (1988) Cancer Res. 46, 5965. 3. Geller, B. L. (1991) Mol. Microbial. 5, 2093. 4. Kirsi, J. J., Mckernan, P. A., Burns III, N. J., North, J. A., Murray, B. K., and Robins, R. K. (1984) Antimicrob. Agents Chemother. 26, 466. 5. Khym, J. X., Bynum, J. W., and Volkin, E. (1977) Anal. Biochem. 77, 446.

235

RIBONUCLEOTIDES

6. Arezzo, F. (1987) Anal. Biochem. 100, 57. 7. Pimenov, A. M., Tikhonov, Y. V., Meisner, R. T. (1986) J. C!rromatogr. 365, 221. 8. Hoffman, N. E., and Liao, J. C. (1977) Anal.

22. 23.

ACKNOWLEDGMENT This work (Sunnyvale,

OF

26. 27. 28. 29. 30.

R. M.

(1988)

Anal.

Chem.

60,2322.

Gross, L., and Yeung, E. S. (1989) J. Chromatogr. 480, 179. Green, J. S., and Jorgenson, J. (1986) J. Chromatogr. 352, 337. Lee, E. D., Muck, W., Henion, J. D., and Covey, T, R. (1989) Biomed. Environ. Mass Spectrom. 18, 844. 31. Smith, R. D., Loo, J. A., Barinaca, C. J., Edmonds, C. G., and Udseth, H. R. (1989) J. Chromatogr. 480, 211. 32. Vinther, A., and Soeberg, H. (1991) J. Chromatogr. 559, 3.