Separation of DNA fragments by high-resolution ion-exchange chromatography on a nonporous QA column

ANALYTICAL

BIOCHEMISTRY

l&%9,126-130

(1990)

Separation of DNA Fragments by High-Resolution lonExchange Chromatography on a Nonporous QA Column Kondo,

Yoshihiro

Ohmiya,

Yoichi

Department

of Physical

Biochemistry,

Received

February

and Toshihiko

Institute

Kondo’

of Endocrinology,

University,

Maebashi

371, Japan

13, 1990

A nonporous QA column (a strong anion exchanger) was used for HPLC of DNA fragments. This column was successfully employed to separate small (ca. 10 bp) and intermediate size (ca. 10 kbp) DNA fragments from each other. The column also separated double-stranded DNA from its single-stranded form, and circular DNA molecules from linear ones. The entire separation process was completed within 60 min. The recovery of DNA fragments in each run was above 95%. High resolution was obtained both at an analytical level (microgram scale) and at a preparative level (100 pg scale). In view of time efficiency, recovery, and resolution, the nonporous QA column is superior to other porous ionexchange columns and expected to be a very useful tool in molecular biological studies. o Isso Academic PMS, IW.

Separation of DNA fragments is currently being performed by preparative gel electrophoresis of the fragments and their extraction from the gel. The recovery of DNA fragments throughout this process is generally low, probably due to various technical difficulties during extraction. Recently, various types of liquid chromatographies have been employed for DNA fragment separation. Macroporous ion-exchange columns have been used for the separation of DNA fragments smaller than 5 kbp (l-3). Reverse-phase HPLC has proved useful in the separation of small DNA fragments (~500 bp) (4), and, especially, in the purification of synthesized singlestranded DNA fragments. Size-exclusion HPLC has been employed for separating DNA fragments of intermediate sizes (800-5000 bp) (5) and also fragments of sizes larger than 10 kbp (6,7). These methods, however, are not satisfactory because they are time consuming

1 To whom

Gunma

correspondence

should

be addressed.

and result in poor resolution. In addition, the capacities of the columns are not large enough for preparative purposes. This situation led us to search for a chromatography column with qualities better than those of the columns previously used, for both analytical and preparative purposes. In this paper, we describe the application of a nonporous QA-type anion-exchange column, which enables the fine separation of various types of DNA molecules and their fragments of various sizes within only 60 min. MATERIALS

AND

METHODS

DNA samples. A plasmid DNA (pGEM-7Zf(+), 3 kbp) was purchased from Promega (Madison, WI). Linear DNA was prepared by digesting the plasmid DNA with EcoRI, and the three DNA fragments (1.8, 0.82, and 0.38 kbp were prepared by digesting with ScaI, NueI, and HindIII, respectively. Double-stranded and Single-stranded preparations of bacteriophage DNA (M13mp18, 7 kbp) were obtained from Pharmacia (Uppsala, Sweden) and DNA size markers (1-kbp DNA ladder and 123-bp DNA ladder) from BRL (Bethesda, MD). Concentrations of double-stranded DNA were calculated by assuming 1 A,,, unit to be 50 pg/ml, and those of single-stranded DNA by assuming 1 A,,, unit to be 36 pg/ml. Chromatography. We used a Pharmacia DfB HPLC pump equipped with a nonporous hydrophilic polymethacrylate gel-packed column, Shodex IEC QA-620N (trimethyl ammonium groups are introduced on the surface of spherical gel particles; particle size, 2.5 pm; column size, 6 X 75 mm; Showa Denko, Tokyo, Japan). Samples were applied via a Rheodyne 7125 injector (Cotati, CA), monitored by a Tosoh UV-8 uv absorbance detector (Tokyo, Japan), and collected on a Pharmacia Frac-100 fraction collector. An appropriate volume of a sample dissolved in TE buffer (10 mM Tris/HCl con-

126 All

Copyright 0 1990 rights of reproduction

0003-2697190 $3.00 by Academic Press, Inc. in any form reserved.

ION-EXCHANGE

CHROMATOGRAPHY

taining 1 mM EDTA) was injected and eluted with a linear gradient of NaCl in 20 mM Tris/HCl, pH 8.0, containing 1 mM EDTA. Absorbance was monitored at 260 nm. The retention time (ts) and the width at half-height (zu~,~)of each DNA peak were determined on the basis of the chromatographic pattern. The resolution of two components (R,) and the number of theoretical plates (N) were calculated according to the standard method (8). Electrophoresis. Electrophoresis of DNA samples was carried out on a 1% SEAKEM agarose plate in a Tris/borate buffer (TBE; 89 mM Tris and 89 mM boric acid containing 2 mM EDTA, pH 8.3) using the mini-gel electrophoresis system Mupid-2 (Advance, Tokyo, Japan). The gel was stained with ethidium bromide and photographed under a uv illuminator using Type 667 Polaroid film.

OF

DNA

127

FRAGMENTS

0.03

0

RESULTS

AND

DISCUSSION

10

20 Retemtim

Since materials packed in QA columns are strongly basic, a higher concentration of NaCl is needed for the elution of adsorbed substances than in the case of DEAE columns. In fact, most of the DNA fragments were eluted with between 0.68 and 0.90 M NaCl in the elution buffer described under Material and Methods. In contrast to the buffers generally used for DEAE column elution, the elution buffer for the QA column contained EDTA. EDTA is known to stabilize DNA molecules in terms of their physical properties, which may influence their chromatographic behavior. Nevertheless, EDTA had not been added in the DNA elution from DEAE columns, since EDTA is adsorbed to the anion exchangers when NaCl is below 0.3 M. The high salt condition for the QA column enabled the use of EDTA as a constituent of the elution buffer to avoid DNA degeneration. Figure 1 shows a chromatogram of a 123-bp DNA ladder which consists of 34 DNA fragments from 123 to 4182 bp. We determined the size of DNA fragments contained in each peak on the basis of their mobilities in 1% agarose gel electrophoresis. The results showed that the DNA fragments were eluted according to their size. Resolution decreased with an increase in fragment size. Fragments larger than 984 bp did not show separate peaks, probably because the steps of the 123-bp ladder were too narrow to be distinguished by this column, especially with larger fragments. In fact, when we applied a 1-kbp DNA ladder preparation to this column, the ladder steps from 3054 to 12,166 kbp appeared as distinct peaks (Fig. 2A). DNA fragments of each l-min fraction were analyzed by 1% agarose gel electrophore-

30

40

ttme (m&t)

FIG. 1. Elution pattern of a 123-bp DNA ladder. 4 pg of the DNA ladder was applied to a QA-620N column and eluted for 40 min with a linear gradient of 0.68 to 0.84 M NaCl in 20 mM Tris/HCl, pH 8.0, containing 1 mM EDTA at a flow rate of 0.8 ml/min. The lengths of the DNA fragments were: 123 bp (1); 246 bp (2); 369 bp (3); 492 bp (4); 615 bp (5); 738 bp (6); 861 bp (7); 984-4182 bp (8-34).

sis (Fig. 2B). The results demonstrated that smaller fragments tend to elute faster. This particular commercial preparation of a DNA ladder also contained HinfI fragments of pBR.322 DNA, composed of 11 fragments with sizes from 75 to 1636 bp. As shown in Fig. 2A, all of these small fragments were well separated by the QA column in a roughly size-dependent manner. The best resolution was obtained with sizes between 506 and 517 bp, which was not expected in view of agarose gel electrophoresis. Such a high resolution between small fragments and the efficient applicability for lo-kbp fragments have not been reported previously (l-3). For further characterization of the nonporous QA HPLC of DNA, we determined the number of theoretical plates and the change in resolution of the column when the three restriction digests of DNA (0.38, 0.82, and 1.8 kbp) were used as model samples. Generally the number of theoretical plates is estimated on the basis of the chromatographic behavior of low-molecular-weight substances such as acetone. However, in contrast to ordinary gels, small molecules like acetone were trapped by minute pores on the surface of the nonporous gel particles, resulting in an incorrect estimation of theoretical plates. Therefore, we used relatively large molecules

128

OHMIYA,

KONDO,

AND

KONDO

FIG. 2. (A) Chromatography of 15 pg of a 1-kbp DNA ladder. Chromatographic conditions were the same as described in Fig. 1, except that the gradient slope was 0.16 ~/60 min. The size of DNA in each peak was estimated by analyzing the peak fraction by 1% agarose gel electrophoresis. (B) An agarose gel electrophoretogram of DNA fragments obtained from the chromatographic fractions. Lanes a to j were obtained from the respective chromatographic fractions in (A). 0 is the original 1-kbp DNA ladder preparation. The lengths of DNA fragments were: 75 bp (1); 134 bp (2); 154 bp (3); 220 bp (4); 201 bp (5); 298 bp (6); 344 bp (7); 396 bp (8); 506 bp (9); 517 bp (10); 1018 bp (11); 1636 bp (12); 2036 bp (13); 3054 bp (14); 4072 bp (15); 5090 bp (16); 6108 bp (17); 7126 bp (18); 8144 bp (19); 9164 bp (20); 10,180 bp (21); 11,198bp (22); 12,216 bp (23).

like DNA fragments to obtain reliable data of theoretical plates. Figure 3 shows the representative chromatogram of the three differently sized DNA fragments at a flow rate of 0.8 ml/min and a gradient slope of 4.2 mM/

ml. The number of theoretical plates of all three peaks was about 7000, indicating that the number of theoretical plates of the column is independent of DNA size under the present conditions. Figure 4 shows the resolution of fragments between 0.38 and 0.82 kbp and between 0.82 and 1.8 kbp as a function of flow rate at a gradient slope of 4.2 mM/ml. When the flow rate was increased from 0.4 to 1.2 ml/min, the resolution of two pairs of fragments did not appreciably change. This shows that the resolution does not depend on the flow rate of the elution buffer under these conditions. On the other hand, in HPLC using a Mono Q column (a porous

10 (0.33-0.32

kbp)

0 0

Retention

time

(mln)

0.4

0.6 Flow

FIG. 3. Elution pattern of three DNA fragments (0.38,0.82, and 1.8 kbp). Chromatographic conditions were the same as described in Fig. 1, except that the concentration gradient from 0.7 to 0.8 M was accomplished within 30 min. Three DNA fragments were obtained from plasmid DNA (pGEM-7Zf) digested with HindIII, NaeI, and ScaI restriction enzymes, respectively.

0.6 rate

1.0

1.2

(mllmln)

FIG. 4. Resolutions of DNA fragments between 0.38 and 0.82 kbp (0) and 0.82 and 1.8 kbp (w) at a gradient slope of 4.2 mM/ml, as a function of flow rate. Experimental conditions were the same as described in Fig. 3, except for flow rate. The shaded area indicates resolution where theoretically two peaks could not be distinguished.

-b Flow

ION-EXCHANGE

rate

CHROMATOGRAPHY

OF

DNA

; 0.0 ml/mln

Gradient

slope

(mM/ml) Retention

FIG. 5. Resolutions of DNA fragments between 0.38 and 0.82 kbp (0) and 0.82 and 1.8 kbp (m) at a flow rate of 0.8 ml/min, as a function of gradient slope. Experimental conditions were the same as described in Fig. 3, except for gradient slopes. The shaded area indicates resolution where theoretically two peaks could not be distinguished.

QA column) the resolution either between two relatively small or between two large fragments has been reported to decrease when the flow rate is increased (3). Figure 5 shows the change in resolution between the same two pairs of DNA fragments as shown in Figure 4 as a function of gradient slope at a flow rate of 0.8 ml/ min, demonstrating the dependency of resolution on the gradient slope. Nevertheless, the resolution of the fragments is hardly satisfactory when the gradient slope is below 8.3 mM/ml and the fragments are smaller than 1.8 kbp. Hydroxylapatite columns have usually been used for the separation of single-stranded from double-stranded DNA (9). However, caution is necessary because the packed material is very sensitive and damage could result from drastic changes in pressure. With the QA column this disadvantage can be avoided. Figure 6 shows

Reterltioll

129

FRAGMENTS

tlrne

(m&l)

FIG. 6. Separation of single-stranded and double-stranded DNA of M13mp18. The amounts of single-stranded and double-stranded DNA were 7 and 10 pg, respectively. Experimental conditions were the same as described in Fig. 1, except that the concentration gradient from 0.7 to 0.9 M NaCl was accomplished within 30 min.

the

(min)

FIG. 7. Separation of supercoiled circular and linear DNA of pGEM-7Zf. The amounts of supercoiled circular and linear DNA were 8 and 10 pg, respectively. Experimental conditions were the same as described in Fig. 1, except that the concentration gradient from 0.8 to 0.9 M NaCl was accomplished within 30 min.

the chromatogram of a mixture of a single-stranded and a double-stranded DNA (M13mp18) on the QA column. Both DNAs were completely separated within 30 min. It is known that in 1% agarose gel electrophoresis, supercoiled circular, linear, and relaxed circular molecules derived from the same species of DNA move in that order. In contrast, in the QA column, a supercoiled circular plasmid DNA (pGEM7) eluted later than its linear form, as shown in Fig. 7. The relaxed circular form of this DNA eluted faster than the supercoiled circular one and later than the linear one (data not shown). Thus, the behavior of circular DNA in agarose gel electrophoresis and in QA HPLC was almost opposite to that observed with the DNA ladders, in which the mobility order of DNA fragments in agarose gel electrophoresis was basically similar to the elution order in the QA HPLC. These results suggest that the separation mechanism of DNA fragments by the QA column is not very simple and is related partly to size differences among the fragments and also to their differences in stereoscopic structures. Reproducibility of the DNA separation procedure by the QA column is very high; repeated applications of the same linear DNA samples provided very similar elution patterns, and the differences in retention times of each peak in repeated experiments were within kO.12 min. Sample recovery estimated from peak area was above 95% in each run (data not shown). The separation characteristics of the QA column seem to be very stable. We have employed the same column continuously for over 5 months with no apparent change in chromatographic characteristics. High resolution was obtained both at an analytical level (microgram scale) and at a preparative level (100 pg scale).

130

OHMIYA,

KONDO,

The nonporous QA column would be a very useful tool in various aspects of molecular biological studies.

We are grateful Ms. M. Kimura for her excellent secretarial assistance. Y.O. thanks Mr. H. Hasegawa for his valuable technical suggestions on the HPLC system used. This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan.

REFERENCES Y., Sasaki,

265,342-346.

and Hashimoto,

T. (1983)

J. Chromatogr.

S. E. (1987)

Anal.

&o&em.

166,

158-171.

4. Usher, D. A. (1979) Nucleic Acids. Res. 6, 2289-2306. 5. Kato, Y., Yamasaki, Y., and Hashimoto, T. (1985) J. Chromatogr. 320,440-444. 6. Boyes, B. E., Walker, D. G., and McGeer, P. L. (1988) And. Bio&em. 170, 127-134. 7. Hirabayashi, J., and Kasai, K. (1989) Anal. Biochem. 178, 336341.

a. M.,

KONDO

2. Hecker, R., Colpan, M., and Riesner, D. (1985) J. Chromatogr. 326,251-261. 3. Westman, E., Eriksson, S., Laas, T., Pernemalm, P. A., and Skold,

ACKNOWLEDGMENTS

1. Kato,

AND

Fallon, A., Booth, R. F. G., and Bell, L. D. (1987) Applications of HPLC in Biochemistry, pp. 8-22, Elsevier, Amsterdam. S. W., and Engelhorn, S. C. (1983) LC Mag. 1, 294. 9. Compton,