Journal of Pharmaceutical and Biomedical Analysis
34 (2004) 277–283
Separation of purine and pyrimidine bases by capillary
electrophoresis using ␤-cyclodextrin as an additive
Ping Wang1, Jicun Ren∗
College of Chemistry and Chemical Engineering, Shanghai Jiaotong University, 800 Dongchuan Road, Shanghai 200240, PR China
Received 8 May 2003; received in revised form 8 September 2003; accepted 8 September 2003
Abstract
Capillary electrophoresis was applied to separate purine and pyrimidine bases in the basis of their partial ionization in the
alkaline buffer. The effects of buffer pH, buffer and ␤-cylclodextrin concentration were systematically investigated using a
commercial capillary electrophoresis instrument with UV detector at 254 nm. We found that the resolutions of bases (especially
for adenine and thymine) were significantly improved in the presence of ␤-cylclodextrin. The satisfactory separation of five
bases such as cytosine, thymine, adenine, guanine and uracil were achieved by capillary electrophoresis using ␤-cylclodextrin
as an additive. Under the optimal conditions, the linear range was from 2 to 200 ␮g/ml for bases (R = 0.991–0.997) and the
detection limits were from 0.8 to 1.8 ␮g/ml (S/N = 2). The detection limit of 0.05 ␮g/ml (S/N = 2) for uracil was obtained
by stacking injection mode. The assay was used to determine the deamination of cytosine to uracil by heating in the presence
of sodium hydroxide. Our primarily results show that capillary electrophoresis is a very useful tool for determination of purine
and pyrimidine bases and study on nucleic acids.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Capillary electrophoresis; Purine; Pyrimidine; Deamination; ␤-Cylclodextrin
1. Introduction
Purine and pyrimidine bases are the building blocks
in both DNA and RNA that play important roles in
cell metabolism. Base changes in DNA may affect seriously
the structure and function of products of gene
expression—protein, which is considered to be the
main causes of inherited diseases and most human can∗
Corresponding author. Tel.: +86-2154746001;
fax: +86-2154741297.
E-mail address: jicunren@mail.sjtu.edu.cn (J. Ren).
1 Present address: Nanhua University, Hengyang, Hunan
Province, PR China.
cers [1–3]. Many areas such as pharmacological studies,
clinical diagnosis and DNA damage assay, need
a quick, inexpensive and accurate method for determination
of purine and pyrimidine bases [4–9]. High
performance liquid chromatography (HPLC) is a commonly
used method for analysis of purine and pyrimidine
bases [10–12].
Capillary electrophoresis (CE) is a powerful alternative
to HPLC for the separation of charged and
polar compounds [13–16]. CE has been successfully
used for analysis of nucleic acids and nucleotides because
they are negatively charged in neutral pH buffer
[17–20]. But only a few reports involved the separation
of purine and pyrimidine bases [21–24]. Differing
0731-7085/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0731-7085(03)00502-8
278 P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283
from nucleic acids, purines and pyrimidines are only
partially ionized in the solution. On the other hand,
due to their similar pKa values, it is very difficult
to achieve baseline separation of purines and pyrimidines
by CE in free solution, especially for thymine
and adenine. Certain efforts such as changes of pH
and buffer types could not improve the resolution between
thymine and adenine. Cyclodextrin (CD) has
been extensively used as a chiral selector for separation
of enantiomers and as an additive for improving
separation of some cyclic compounds [25–29]. CDs
are cyclic oligosaccharides with truncated cylindrical
molecular shapes. They have particular names:
␣-CD, ␤-CD and ␥-CD for those containing six,
seven and eight glucopyranose units, respectively.
Since the inside surfaces of the CD’s cavity are hydrophobic,
CDs tend to form inclusion complexes
with certain compounds whose molecular size and
structure match with CD’s cavities by hydrophobic
interaction. Since the cyclic size of thymine is smaller
than that of adenine, we predict that there is certain
difference in their interactions with CD. In this paper
we will explore the possibility of improving the
separation of purines and pyrimidines using ␤-CD
as an additive. Furthermore, we try to develop a CE
method to determine the deamination of thymine to
uracil.
2. Experimental
2.1. Reagents and materials
Adenine (A), guanine (G), thymine (T), cytosine
(C), uracil (U) and ␤-cyclodextrin were purchased
from Sigma (St. Louis, MO, USA). Sodium tetraborate
and sodium hydroxide were provided by Shanghai
Reagents Co.(Shanghai, China). Double-distilled water
was used for preparation of all aqueous solutions.
The pHs of various borate solutions were adjusted by
1 M sodium hydroxide using a pH meter (Lezi Instrument
Factory, Shanghai, China). Buffers were filtered
with a membrane filter of 0.45 ␮m pore size prior to
use.
Fused silica capillaries with 75 ␮m internal diameter
(i.d.) and 365 ␮m outer diameter (o.d.) were from
Yongnian Optical Fiber Factory (Yongnian, Hebei,
China).
2.2. Deamination of cytosine into uracil
Several hundreds ␮l of 1 mg/ml cytosine sample
solutions were heated at 100 ◦C from 5 to 60 min in
the presence or absence of 6 mM sodium hydroxide.
One hundred ␮l of the reaction products were taken
out and analyzed by CE using 2 ␮g/ml thymine as an
internal standard.
2.3. Capillary electrophoresis procedure
A Waters Quanta 4000E Capillary Electrophoresis
system with a UV detector (Milfod, MA, USA) was
used in the experiment. Caesar software (version 4.0,
Princes Technologies, Emmn, The Netherlands) was
used for data collection and processing.
A new capillary (43 cm total length, 36 cm effective
length) was rinsed successively with 0.1 M sodium
hydroxide, water and borate buffer. The temperature
was set at 25 ◦C and the detection wavelength was at
254 nm. The samples were introduced by hydrostatic
injection, and electrophoresis was performed at positive
polarity under the conditions specified in the figure
legends. The capillary was rinsed with fresh buffer
between each run.
3. Results and discussion
The structures of five purine and pyrimidine bases
to study are displayed in Fig. 1. At first, we wanted to
develop a CE method for separation of bases by optimizing
buffer pH and concentration and electrophoresis
conditions. However, the separation of thymine
and adenine was not achieved (partial data shown in
Fig. 5A).
3.1. Buffer pH
Since purine and pyrimidine bases are partially ionized
to negative charged ions in the alkaline solution,
the buffer pH is an important factor affecting their
separation by CE. Fig. 2 shows the effects of buffer
pH on the migration times of purine and pyrimidine
bases. Twenty mM sodium tetraborate solution containing
25 mM ␤-CD was used as running buffer, and
the influences of pH range from 9.2 to 11.0 were investigated.
In the Fig. 2, we examined that the migration
P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283 279
Pyrimidine Bases
N
H
N
O
NH2
N
H
NH
O
O
H3C
N
H
NH
O
O
N
N
N
N
H
NH2
N
NH
N
N
H
O
NH2
Cytosine Thymine Uracil
Adenine Guanine
Purine Bases
Fig. 1. Structures of pyrimidine and purine bases.
times of bases dramatically increased with buffer pH.
This result was mainly attributed to an increase in the
dissociation degrees of bases with an increase at buffer
pH. The resolutions between bases were also considerably
improved with buffer pH. However, when buffer
pH was over 10.5, the resolutions were not further im-
9.0 9.5 10.0 10.5 11.0
8
12
16
20
24
C
A
T
G
U
Migrationtime(min)
Buffer pH
Fig. 2. Effects of buffer pH on the migration time of bases. The solutions of 20 mM sodium tetraborate including 25 mM ␤-CD were used
as running buffer and their pH range was from 9.2 to 11.0. All sample concentrations were 50 ␮g/ml. Hydrostatic injections for 10 s were
used, the temperature was 25 ◦C, and the applied voltage was 8 kV.
proved, and the poor reproducibility of migration time
was examined. This phenomenon was due to the deterioration
of the inside surface of the silica capillary
in the high pH buffer.
3.2. β-CD concentration
Fig. 3 displays the effects of ␤-CD concentration
on the separation of purine and pyrimidine bases.
We found that the differences of migration time of
thymine and adenine gradually increased with an increase
in ␤-CD concentration. This result likely owed
to the different interaction between thymine -␤-CD
and adenine-␤-CD. The cyclic size of adenine possibly
matched well with the cavity of ␤-CD and led
to formation of the adenine-␤-CD inclusion complex.
From the results in Fig. 3, we expected that high concentration
of ␤-CD was benefited for improvement
of the separation of bases. However, the dissolution
degree of ␤-CD limited further increasing in the
concentration of ␤-CD.
3.3. Buffer concentration
Buffer concentration effects on the separation of
purine and thymine bases are shown in Fig. 4. The
280 P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283
0 5 10 15 20 25
8
10
12
14
16
18
Migrationtime(min)
β-CD Concentration (mM)
C
A
T
G
U
Fig. 3. Effects of ␤-CD concentration on the migration time of bases. The solutions of 15 mM sodium tetraborate, pH 10.5, were used as
running buffer. Other conditions were as described in the caption to Fig. 2.
migration times of bases were gradually increased and
the resolutions were improved with buffer concentrations.
These effects were mainly attributed to the reduction
of the electro-osmotic flow with the buffer
concentration. Moreover, as the buffer concentration
10 15 20 25
8
12
16
20
24
C
A
T
G
U
Migrationtime(min)
Buffer concentration (mM)
Fig. 4. Effects of buffer concentration on the migration time of bases. The solutions of sodium tetraborate including 25 mM ␤-CD (pH 10.5)
were used as running buffer. The concentration of sodium tetraborate ranged from 5 mM to 25 mM. Other conditions were as described
in the caption to Fig. 2.
increased, the bases were stacking into narrower zones
and formed sharper peaks, because of a marked difference
in the effective electric field between the buffer
and sample zone. But, the current in electrophoresis
rapidly increased with buffer concentration, which
P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283 281
5 10 15 20
0.014
0.016
0.018
0.020
0.022
0.024
U
G
A+T
C
AU
Migration time (min)
0 5 10 15 20
0.014
0.016
0.018
0.020
0.022
U
G
T
A
C
AU
Migration time (min)(A) (B)
Fig. 5. Separation of purine and pyrimidine bases by CE. (A) 20 mM sodium tetraborate (pH 10.5) was used as running buffer; (B)
20 mM sodium tetraborate including 25 mM ␤-CD (pH 10.5) was used as running buffer. All sample concentrations were 100 ␮g/ml. Other
conditions were same as described in the caption to Fig. 2.
resulted in Joule heating effect. When the buffer concentration
was more than 25 mM, it was observed that
Joule heating effect led to substantial broadening of
the peaks and a decrease in the resolutions of bases
(data not shown).
3.4. Separation of bases
Fig. 5 shows the separation of purine and pyrimidine
base in the borate buffer with or without ␤-CD.
In the borate buffer without ␤-CD, five bases were
not completely separated, and the co-migration of adenine
and thymine was observed (as shown in Fig. 5A).
Notably, five bases were well separated in the presence
of 25 mM ␤-CD (as shown in Fig. 5B). The
data demonstrated that use of ␤-CD as an additive
was of paramount importance to improve the separaTable
1
Bases Equations for calibration curves Linear range (␮g/ml) R Detection limit (␮g/ml)
Adenine A = 0.055 + 7.53 × 10−4C 2–200 0.996 0.8
Gaunine A = 0.038 + 5.60 × 10−4C 2–200 0.996 1.2
Cytosin A = 0.067 + 4.16 × 10−4C 2–200 0.991 1.5
Thymine A = 0.036 + 4.25 × 10−4C 2–200 0.995 1.3
Uracil A = 0.031 + 5.01 × 10−4C 2–200 0.997 1.8
tion of purine and pyrimidine bases by CE. This result
likely was attributed to the interaction between
thymine-␤-CD and adenine-␤-CD.
3.5. Detection limits, linearity and reproducibility
We measured the detection limits and linearity for
purine and pyrimidine bases and the results obtained
were shown in Table 1. These data illustrated that the
wide linear ranges and low-detection limits for purine
and pyrimidine bases were obtained by CE with UV
detection. The relative standard deviations (R.S.D.) of
migration times in one day and between days were less
than 1.5% (n = 5) and 4.3% (n = 5), respectively, and
the R.S.D. of peak areas in one day and between days
were less than 2.4% (n = 5) and 5.2% (n = 5), respectively.
These data showed that CE with UV detector
282 P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283
N
H
N
O
NH2
NaOH/H2O
Heating
N
H
NH
O
O
Fig. 6. Reaction of cytosine into uracil by heating in the presence
and absence of sodium hydroxide.
had a good reproducibility. Stacking injection [30,31]
of a large volume of samples markedly improved the
concentration detection limits of bases. The concentration
detection limit of 0.05 ␮g/ml (S/N = 2) for
uracil was obtained at 50 s injection time when sample
dissolved water.
3.6. Deamination of cytosine by heating and
sodium hydroxide
The deamination of cytosine into uracil in DNA
may occur in response to exposure to chemical mutagens
such as hydrogen peroxide [32–34]. In addition,
it also occurs spontaneously when DNA is exposed
to high temperature and sodium hydroxide. Deamination
reaction of cytosine into uracil was expressed in
10 20 30 40 50 60 70
0
1
2
3
4 (NaOH)
(no NaOH)
Signalratio(uraciltointernalstandard)
Time of heating (min)
Fig. 7. Progressing of deamination of cytosine into uracil by heating in the presence and absence of sodium hydroxide. The reaction
temperature was set in 100 ◦C. The solutions of 20 mM sodium tetraborate (pH 10.5) were used as running buffer. Two ␮g/ml of thymine
was used as an internal standard. Other conditions were same as described in the caption to Fig. 2.
Fig. 6. In present study, we wanted to use CE technique
to examine the temperature-dependence of the
cytosine deamination in the presence or the absence
of sodium hydroxide. In order to improve the precise
of determination of uracil from deamination of cytosine,
we employed thymine as an internal standard in
CE analysis. The progressing of deamination of cytosine
by heating is shown in Fig. 7 in the presence and
absence of sodium hydroxide. The data demonstrated
that sodium hydroxide significantly speeded up the
deamination of cytosine into uracil. This preliminary
result indicated that CE was a useful tool for study on
the chemistry of nucleic acids.
4. Conclusions
We systematically investigated the effects of buffer
pH, buffer concentration and ␤-CD on the CE separations
of purine and pyrimidine bases, and demonstrated
that separations of thymine and adenine was
significantly improved by use of ␤-CD as an additive.
Our data also illustrated that the CE with UV
detection possessed wide linear ranges and low detection
limits for purine and pyrimidine bases. Furthermore,
we first demonstrated that CE technique was
P. Wang, J. Ren / Journal of Pharmaceutical and Biomedical Analysis 34 (2004) 277–283 283
successfully applied to study on the deamination of
cytosine.
Acknowledgements
This work was supported by The National Natural
Science Foundation of China (Grant No. 20271033),
The Natural Science Foundation of Shanghai
(02ZA14057), The Foundation from Personnel Ministry
of China, and The Foundation from Educational
Ministry of China.
References
[1] C.L. Vnencak-Jones, Am. J. Clin. Pathol. 112 (1999) S19–
S32.
[2] D.P. Lane, Br. J. Cancer 80 (1999) 1–5.
[3] A. Lindblom, M. Nordenskjold, Acta Oncol. 38 (1999) 439.
[4] J. Cadet, M. Weinfeld, Anal. Chem. 65 (1993) 675A–682A.
[5] H.C. Tseng, R. Dadoo, R.N. Zare, Anal. Biochem. 222 (1994)
55–58.
[6] H. Lin, D.K. Xu, H.Y. Chen, J. Chromatogr. A 760 (1997)
461–466.
[7] B.N. Ames, Toxicol. Lett. 102/103 (1998) 5–18.
[8] H. Atamna, I. Cheung, B.N. Ames, Proc. Natl. Acad. Sci.
U.S.A. 97 (2000) 686–691.
[9] B.N. Ames, N. Y. Acad. Sci. 889 (1999) 87–106.
[10] D. Perrett, H.A. Simmonds, Biomed. Chromatogr. 4 (1990)
267–272.
[11] T. Grune, G.A. Ross, H. Schmidt, W. Siems, D. Perrett, J.
Chromatogr. 636 (1993) 105–111.
[12] G. Minniti, U. Caruso, R. Cerone, E. de Toni, Adv. Exp. Med.
Biol. 431 (1998) 843–848.
[13] Q. Yang, K. Hidajat, S.F.Y. Li, J. Chromatogr. Sci. 35 (1997)
358–373.
[14] K.D. Altria, J. Chromatogr. A 856 (1999) 443–463.
[15] J.S. Fritz, J. Chromatogr. A 884 (2000) 261–275.
[16] S.N. Krylov, N.J. Dovichi, Anal. Chem. 72 (2000) 111R–
128R.
[17] M.C. Boyce, Electrophoresis 22 (2001) 1447–1459.
[18] A.S. Cohen, S. Terabe, J.A. Smith, B.L. Karger, Anal. Chem.
59 (1987) 1021–1027.
[19] D.L. Deforce, F.P.K. Ryniers, E. Van den Eeckhout, L.
Lemiere, E.L. Esmans, Anal. Chem. 68 (1996) 3575–3584.
[20] S.E. Geldart, P.R. Brown, J. Chromatogr. A 828 (1998) 317–
336.
[21] S.S. Zhang, H.X. Liu, Y. Chen, Z.B. Yuan, Biomed.
Chromatogr. 10 (1996) 256–257.
[22] S.E. Geldart, P.R. Brown, J. Chromatogr. A 831 (1999) 123–
129.
[23] T. Wessel, C. Lanvers, S. Fruend, G. Hempel, J. Chromatogr.
A 894 (2000) 157–164.
[24] M.K. Grob, K. O’Brien, J.J. Chu, D.D.Y. Chen, J. Chromatogr.
B 788 (2003) 103–111.
[25] J. Ren, H. Liu, J. Chromatogr. A 732 (1996) 175–181.
[26] G. Vigh, A.D. Sokolowski, Electrophoresis 18 (1997) 2305–
2310.
[27] Z. Aturki, C. Desiderio, L. Mannina, S. Fanali, J. Chromatogr.
A 817 (1998) 91–104.
[28] G. Gubitz, M.G. Schmid, Electrophoresis 21 (2000) 4112–
4135.
[29] E. Schneiderman, A.M. Stalcup, J. Chromatogr. B 745 (2000)
83–102.
[30] J.P. Qurino, S. Terabe, J. Chromatogr. A 902 (2000) 119–135.
[31] Z.K. Shihabi, J. Chromatogr. A 902 (2000) 107–117.
[32] L.A. Federio, T.A. Kunkel, B.R. Shaw, Biochemistry 29
(1990) 2532–2537.
[33] M. Ehrlich, X.Y. Zhang, N.M. Inamdar, Mutat. Res. 238
(1990) 277–286.
[34] S. Burney, J.L. Caulfierld, J.C. Niles, J.S. Wishnok, S.R.
Tannenbaum, Mutat. Res. 424 (1999) 37–49.