Oligonucleotide analysis by gel capillary electrophoresis

METHODS: A Companion to Methods in Enzymology 4, 213-226 (1992)

Oligonucleotide Analysis Capillary Electrophoresis

Alex Andrus

by Gel

Applied Biosystems, Inc., 850 Lincoln Centre Driue, Foster City, California 94404

Several million oligonucleotides are synthesized each year for a broad variety of molecular biology applications. Steady improvements in the synthesis chemistry efficiency and the automated DNA synthesizers have made production of oli- gonucleotides routine and reliable. Many applications, such as PCR and sequencing, are often successful when the prim- ers have not been rigorously purified. To ensure an adequate level of quality and purity, rapid and convenient analytical methods are necessary for the dozens of oligonucleotides produced each day by a DNA synthesis laboratory. Traditional methods of analysis have been HPLC and polyacrylamide slab gel electrophoresis (PAGE). Gel capillary electrophoresis is a new option, combining the advantages of the HPLC and PAGE, with unprecedented resolution and speed. D 1992 Academic

Press, Inc.

Gel capillary electrophoresis has emerged as a pow- erful new tool for the analysis of synthetic oligonucle- otides (l-5). The advantages are dramatically de- creased analysis time, excellent resolution, in-capillary detection, reduced sample quantities, and automation. Capillary electrophoresis (CE) methods are established for other biomolecules, such as proteins, peptides, and high-molecular-weight, double-stranded nucleic acids (6,7). Now the CE method has been extended to single- stranded oligonucleotides using polymeric gel-filled capillaries. In a gel matrix capillary, as with polyacryl- amide slab gel electrophoresis (PAGE), DNA separates primarily, and predictably, by increasing mass (length) under the influence of an electric field. The elution

1046.2023/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

pattern of small oligonucleotides followed by the larg- est, usually the product, is almost always obtained. Single base resolution beyond 100 bases can often be attained. The electropherogram can be displayed, stored, integrated, and printed like HPLC chromato- grams. Slab gels (PAGE), on the other hand, are vi- sualized at a single time point in the analysis and do not easily yield to integrative quantitation. The reso- lution of gel CE surpasses that of the current traditional techniques, HPLC and PAGE, for oligonucleotide analysis. The combination of the gel materials, acryl- amide/urea and others, and heating of the capillary confer a significant denaturing effect for predictable elution patterns devoid of secondary structure artifacts. The gel capillaries sustain multiple injections depend- ing on the gel matrix and storage and handling con- ditions. The parameters affecting oligonucleotide analysis, resolution limits, time, sample preparation and concentration, reproducibility, electrophoresis settings, and adaptibility to automation are examined here. Electropherogram examples of a variety of oli- gonucleotides are presented.

MATERIALS AND METHODS

There are a variety of commercial electrophoresis instruments available, with a range of features, spec- ifications, and degree of automation. The system used to generate the data shown here is the Model 270A/ HT capillary electrophoresis system equipped with MicroGel capillaries. The generic instrument config- uration, shown in Fig. 1, consists of a high-voltage power supply with electrodes immersed in static buffer

213

214 ALEX ANDRUS

reservoirs which are spanned by the fused-silica, gel- filled capillary. At a fixed point, the capillary is mounted at the detector site. A variable-wavelength, uv/vis detector collects light passed directly through the capillary, giving the typical HPLC type analog voltage signal to be processed by an A/D converter.

The typical gel capillary is a fused-silica tube (Fig. 2), about 50 cm long, coated on the outside with poly- imide. The small internal diameter of the capillary, 25-100 pm, allows fast and efficient heat dissipation, permitting the use of high field strengths to achieve rapid analysis times. At the detection point, the poly- imide coating is removed to give a clear window loca- tion, about 5 mm long. The gel matrix must be loaded inside the capillary under carefully controlled condi- tions. The gel concentration must be constant from one end to the other and identical for all capillaries upon which data are to be compared. In addition, a void free gel must be produced, otherwise, capillary failure, erratic current, and spurious peaks may occur. Small bubbles may even travel through a capillary, causing a prolonged period of absorbance that oblit- erates all sample detection.

The gel materials useful for CE are crosslinked or chain-entangled elastic polymers, with hydrophilic and

Electrode - MicroGel Capillary

500 pl Sample Tube

Oligonucleotide Solution .l to .2 odu/ml 20 ul minimum volume

FIG. 2. Sample tube with capillary and electrode. Detector- wavelength, 260 nm; rise time, 1 s; range, 0.5; zero, yes (auto zero at the beginning of each run). Sample-loading time, 5 s; voltage, -5 kV, temp, 4O’C; Time-run time, 22 min; voltage, -15 kV; temp, 40°C. Acquisition time-15-50 min. Sampling rate-5 points per second. Signal-O to 1 V. Integration format-baseline to baseline or peak to peak. Area reject-50,000.

denaturing properties. Polyacrylamide, the traditional polymer used for slab electrophoretic analysis of nucleic acids, has been extensively investigated. The optimum acrylamide, bisacrylamide, buffer, and denaturant for- mula has been intensively sought (1,2). The commer- cial development of polyacrylamide capillaries and their

A/D Data SF-

7 -BufferReservoir

Sample/Buffer Reservoir v Sample Cam&/Auto-Sampler

FIG. 1. Schematic diagram of a capillary electrophoresis instrument.

OLIGONUCLEOTIDE ANALYSIS BY GEL CE 215

use even by experts have been hampered by instability problems (8). It is clear that when urea is present, cap- illary stability is compromised and the capillary will have a short shelf life. Acrylamide/urea capillaries also must be refrigerated off the instrument when not in use. MicroGel capillaries do not contain acrylamide or urea, but are filled with a viscous-elastic, high-molec- ular-weight polymer (3-5). Although nucleobases do not contain traditional chemical denaturants, such as urea or formamide, hydrogen bonding between nucleo- bases is almost totally disrupted upon interaction with the MicroGel matrix and the elevated capillary tem- perature. Given proper handling and analysis condi- tions, typically 50-100 analyses per MicroGel capillary are attained. The buffer chambers and the MicroGel capillaries are filled with 75 mM Tris-phosphate buffer/ 10% methanol, pH 7.5, and are changed every few hours of electrophoresis time, especially on the sample side. On most systems, this is the smaller volume buffer chamber, which undergoes ion depletion during elec-

50.0

45.0

40.0

z 35.0

trophoresis. Changes in ionic strength or pH of the buffer or sample solutions have large effects on elution time, resolution, and peak size. Because there may be some variations in the absolute elution times between runs, exact identification may be facilitated by using relative migration times. When a reference marker is included in the same run as the unknown sample, a relative elution time may be calculated by dividing the elution time of the unknown oligonucleotide peaks by the time of the reference peak. This relative elution time is very reproducible and may be used to predict the elution time of the same oligonucleotide in sub- sequent runs.

Elution time variability of oligonucleotides on MicroGel capillaries is about 1.0% when an internal standard is included with the sample in a run. Accurate peak area quantitation also requires an internal stan- dard because of possible differences in molar extinction coefficients. Injection anomalies can be caused by sam- ple conductivity differences or changes in electrode

FIG. 3. 18mer: 5’ TCA CAG TCT GAT CTC GAT 3’.

216 ALEX ANDRUS

alignment. MicroGel capillaries should be held at con- stant temperature between 30 and 5O’C. Temperature cycling or temperatures higher than 50°C can slowly degrade the gel capillary matrix. All examples shown here were conducted at 40°C. Higher temperatures disrupt secondary structures of hydrogen bonding, which can lead to extra or broadened peaks. Together with a data system, parameters such as resolution, speed, and sample effects can be optimized. The injec- tion and elution of the oligonucleotide occur under the influence of a precisely controlled electric field. The capillary establishes an electrical circuit at high voltage and low current. The voltage is held constant and the current may change as a function of the resistance.

Introduction of an oligonucleotide onto the capillary occurs by electrokinetic injection, whereby the highly charged sample is induced to migrate into the capillary inlet under the influence of an electric field. The amount of oligonucleotide sample that enters the cap- illary increases, but not linearly, with higher concen-

tration, longer injection duration, or higher voltage. It can be highly influenced, however, by the presence of other salts that alter the conductivity of the sample solution.

The crude sample is usually sufficiently free of salts to allow the proper conductance for adequate loading. Samples containing inhibiting levels of salt and other contaminants should undergo purification (e.g., oli- gonucleotide purification cartridge, OPC (9)) or de- salting operations, such as precipitation, prior to gel CE analysis. The presence of small amounts of organic solvents does not affect injection. Samples that have been purified by HPLC and eluted in salt-containing media can be analyzed directly as long as they do not contain excessive amounts of salts, which will limit the electrokinetic loading. Oligonucleotide concentrations of about 0.1-2.0 Az6,, nm units (3-66 pg)/ml in deionized water are appropriate. There is a wide range in sample concentrations due to the nature of electrokinetic in- jection, which is sensitive to conductance and the pres-

FIG. 4. 29mer: 5’ CCA TGA AGC TTT GAC CAT GAA AAT GGA GA 3’.

OLIGONUCLEOTIDE ANALYSIS BY GEL CE 217

ence of competing salts in the sample. Analysis can be done directly from crude reaction mixtures containing ammonium hydroxide or from lyophilized (dried in a vacuum centrifuge) samples dissolved in water. No spurious peaks on the electropherogram will be caused by the ammonium hydroxide, but changes in the sample pH may inhibit sample introduction. At least 20 ~1 of sample solution (about 0.004 Azeonm unit of DNA) in a 500-~1 microcentrifuge (Eppendorf style) tube is re- quired to ensure adequate immersion of the capillary tip and electrode during injection. Several repetitive injections can be made from this amount of sample. However, after a few injections, the sample introduced onto the capillary will diminish, resulting in smaller peaks, due to the carryover of salts from the buffer by the capillary and electrode, and depletion of the sample.

Electropherograms generated from gel-filled capil- laries present graphic and highly resolved represen- tations of oligonucleotide purity. With a data system,

22.0

1 21.0-

20.0-

19.0-

lB.O-

2 17.0-

2 2 16.0-

z B 15.0-

d 2 14.0-

12.0

11.0

10.0

9.0

quantitative results can be obtained from peak areas. The data acquisition system used to process the data shown here is the Model 1020s PE/Nelson personal integrator with a Panasonic KX-P1080i dot-matrix printer and a Hewlett Packard Model 7550A graphics plotter. As with any chromatographic data, reliable conventions must be developed to obtain consistent interpretations. Variables such as attenuation, rise time, and peak width on a data acquisition system can dramatically affect purity analysis. Once data acqui- sition parameters are set, consistent sample concen- tration, injection duration, and injection voltage should also be set. Generally, synthetic yields are consistent enough that analysis conditions can be held constant. Long oligonucleotides often require higher sample concentrations or increased injection times due to re- duced product yields, and therefore reduced purity. Run times also need to be increased for longer oligonucle- otides.

minutes

FIG. 5. 72mer: 5’ AGG GCC GAG CGC AGA AGT GGT CCT GCA ACT TTA TCC GCC TCC ATC CAG TCT ATT AAT TGT TGC CGG GAA GCT 3’.

218 ALEX ANDRUS

FIG. 6. 120mer: 5’ CAA CAG GGG ATT TGC TGC TTT CCA TTG AGC CTG TTT CTC TGC GCG AGG TTC GCG GCG GCG TGT TTGTGCATCCATCTGGATTCTCCTGTCAGTTAGCTTTTCACAGTCTGATCTCGATAT3’.

EXAMPLES including phosphorothioates, can also be evaluated. Figure 3 shows an electropherogram of a typical, crude

Single-stranded, synthetic oligonucleotides, DNA or oligonucleotide, an 18mer made under standard syn- RNA, from 4 to about 150 bases, can be effectively thesis conditions. Using the sample parameters given analyzed by gel CE. Various oligonucleotide analogs, above, the electropherogram and integration report

FIG. 7. Mixture of 12Omer and 119mer.

OLIGONUCLEOTIDE ANALYSIS BY GEL CE 219

show the product, the largest peak eluting at 12 min, mixture have decreasing velocities and increasing res- to constitute 84% of the total integrated area. The in- idence times at the detector point (3). When uncor- tegrated percentage of the product peak is often cited rected, this effect exaggerates the purity of the product as a measure of purity by gel CE and by HPLC. But oligonucleotide, typically the slowest eluting compo- several factors affecting integration should be noted: nent in the electropherogram. In practical terms, the A single wavelength, typically 260 nm, is monitored. purity shown in Fig. 3 would predict excellent perfor- Since the relative extinction coefficients at a particular mance for this oligonucleotide in most, if not all, ap- wavelength of all the species in the sample, nucleotidic plications, such as PCR, sequencing, and probe exper- and not, are uncorrected, the integration has no molar iments. Most of the sample components that elute prior or stoichiometric significance. Without knowledge of to the product are shorter, failure sequences, sometimes the respective extinction coefficients of the migrating referred to as “N-” species, whereas the desired, full- species, the eluting materials cannot be accurately length product is called the “N” peak or band. Later- quantitated. However, the integration values are of eluting impurities arise from other synthesis imper- value, at least in a relative sense, in comparison with fections, composed of either higher molecular weight other samples. The integration instrument settings, or less charged species. These are termed “N+“. They such as threshold reject values and peak identification may be due to incomplete deprotection or branching format, also influence the quantitation. Finally, the of the oligonucleotide during synthesis. This elution peak areas must be corrected for relative peak velocities pattern is, of course, consistent with slab gel electro- since the components of an eluting electrophoretic phoresis in which the components, N-, N, and N+ are

16.0

17.0

16.0

15.0

14.0

13.0

12.0

11.0

10.0

minutes

FIG. 8. Mixture of two 18mers: shorter retention time, 5’ Qs’J CAG TCT GAT CTC GAT 3’; longer retention time, 5’ 3 CAG TCT GAT CTC GAT 3’.

220 ALEX ANDRUS

viewed in a vertical lane format of bands. The sample in Fig. 3 shows several discrete N- peaks and virtually no N+ peaks. Since the detectability limit by uv shadow or staining detection with PAGE is about 0.05 Azcon,,, unit, these low-level impurities may not be apparent by those methods. In addition, the high resolution of gel CE is sometimes required to separate some impu- rities from the product. Figure 4 shows a 29mer with a relatively high amount of N+ impurities.

As the length of the oligonucleotides increases, res- olution decreases. The size of the product peak (N) relative to the amount of impurities decreases also. This is simply a function of the imperfect synthesis chem- istry. For example, at the typical 98% stepwise effi- ciency, the yield of a 20mer will be about 68%. At this same efficiency, the yield of a 40mer will be 45%. As a consequence, longer oligonucleotides are less pure and the product is less distinct, by any analytical method. By injecting more sample, either with a higher con- centration, longer injection time, or higher injection

35.0

30.0

> E 25.0

15.0

voltage, the capillary may be overloaded and loss of resolution may result. The absorbance of the 72mer in Fig. 5 returns nearly to baseline during elution of all the components of the sample.

Even longer oligonucleotides will show single base resolution when critical parameters, such as concen- tration, injection time, and injection voltage, are op- timized, often empirically. The 120mer shown in Fig. 6 gives a quantitative assessment of purity, unavailable by any other method, for such a long oligonucleotide. The integrated product purity of 10% is consistent with a stepwise efficiency of about 98%, probably the max- imum that can realistically be attained. Higher trityl measurements can sometimes be obtained, but they are augmented by side reactions that contribute to the trityl release, but detract from product purity.

As an illustration of the resolving power of gel CE, the 120mer in Fig. 6 and a 119mer (same sequence as the 120mer minus the third base from the 3’ terminus) were mixed and analyzed. Figure 7 is the electropher-

FIG. 9. Mixture of 5’ OH and 5’ aminohexyl phosphate: 5’ TGT AAA ACG ACG GCC AGT 3’.

OLIGONUCLEOTIDE ANALYSIS BY GEL CE 221

ogram showing partial separation of the two long oli- gonucleotides, differing in length by only one base. Considering that both oligonucleotides were in a crude state of approximate 10% purity, the method demon- strates excellent resolution.

While separations of oligonucleotides by gel CE occur primarily according to the sieving efficiency, and therefore length, there is also a hydrophobic component to the elution order. The MicroGel matrix is an en- tangled polymer in which the large oligonucleotide molecules must migrate through pores in response to the electric field. Polyacrylamide also provides a sieving type matrix through which charged molecules pass. The shape and/or hydrophobicity influences elution pat- terns. As an example, a mixture of 18mers sharing the same base composition (A4G3C5T6) separates in Fig. 8. Even with oligonucleotides of the same molecular weight and net charge, a different sequence order of the three bases at the 5’ end allows separation. Achiev- ing separation between oligonucleotides of the same

75.0-

70.0-

65.0-

length, let alone the same base composition, is not a typical occurrence. The effect on mobility of a partic- ular combination of bases suggests that higher order structure is exerting some unique, unpredictable influ- ence. This does not necessarily mean the formation of hairpins, but may result from an alteration in pK of certain functional groups due to the proximity or in- teraction with other nucleotides or an alteration in the shape of the oligonucleotide.

More and more frequently, oligonucleotides are de- rivatized postsynthesis and labeled by covalent attach- ments with other molecules to serve a growing variety of novel applications (10, 11). The most common site of attachment is the 5’ end. A protected amino phos- phoramidite reagent, such as AminoLink 2 (6-N-tri- fluoroacetylaminohexyl-1-diisopropylaminomethyl phosphoramidite), can be used as the final base on the DNA synthesizer, giving a nucleophilic amine group after cleavage and deprotection (12). The 5’ amino- linked oligonucleotide can be analyzed and purified by

FIG. 10. Crude 5’ biotin: TCA CAG TCT GAT CTC GAT 3’.

222 ALEX ANDRUS

the same methods as underivatized, 5’ OH oligonucle- can be made by reaction of the amino-link oligonucle- otides. The aminohexyl phosphate moiety exerts a otide with a large excess of the NHS-biotin (e.g., Sulfo- slight retarding effect during electrophoresis, relative Biotin, Pierce Chemical Co.) reagent. Careful purifi- to the 5’ hydroxyl, due to the slight increase in mass cation is necessary to remove the unreacted biotin and and net zero charge difference. Gel CE is a convenient the organic solvents, base, and salts present in the cou- method for assessing the efficiency of the amino phos- pling reaction. Alternatively, biotinylated oligonucle- phoramidite coupling and for gauging the purity of the otides may be conveniently prepared on the DNA crude amino-link oligonucleotide product. Figure 9 synthesizer with biotin phosphoramidite reagents, shows a mixture of 5’ hydroxyl and 5’ amino-link oli- available from many sources. Figure 10 shows a crude, gonucleotides. Their identities were established by in- biotinylated 18mer. The corresponding unlabeled, 5’ jections of the purified oligonucleotides and mixtures OH 18mer elutes 1 min earlier (data not shown). Some at other ratios (data not shown). The amino-link oli- versions of biotin phosphoramidites allow multiple in- gonucleotides consistently elute later than their 5’ hy- corporation of biotin throughout the oligonucleotide. droxyl counterpart sequences. Further purification at Biotin-oligonucleotides elute significantly later than this stage may not be necessary. their 5’ hydroxy counterparts, due to the increased hy-

Amino-linked oligonucleotides are then reacted fur- drophobicity and mass addition of the biotin and the ther, in solution, usually with an active ester molecule, linker moieties. such as NHS-biotin or NHS-fluorescent dyes, to make By similar procedures, a variety of active-ester flu- biotinylated and fluorescent dye oligonucleotides, re- orescent dyes are available to construct fluorescent dye- spectively. For example, biotinylated oligonucleotides labeled oligonucleotides. A crude 5’, 5-carboxyfluores-

30.0

25.0

> E

2 s 20.0

0, 0

5

15.0

FIG. 11. Fluorescent dye-labeled lamer 5’ (5-FAM-linker): 5’ TGT AAA ACG ACG GCC AGT 3’.

OLIGONUCLEOTIDE ANALYSIS BY GEL CE 223

cein-labeled 18mer is shown in Fig. 11. Like biotin, conjugated fluorescent dyes retard the electrophoretic velocity of oligonucleotides, but otherwise behave nor- mally and exhibit well-formed peaks. The tallest peak in Fig. 11, at 8.0 min, is the dye-labeled product 18mer. The unlabeled 5’ OH 18mer, which failed to couple with the fluorescent dye compound, elutes earlier at 7.7 min.

Oligoribonucleotides, RNA, are excellent samples for gel CE. Being more hydrophilic, they elute slightly ear- lier than their DNA counterparts. RNase degradation does not seem to be significant when the RNA sample is dissolved in purified, filtered water. All the same considerations and separation parameters pertaining to DNA are relevant for RNA. Considering the lower synthesis efficiency, higher costs, and more stringent applications for RNA, it is all the more important to conduct the high-resolution, quantitative analysis that gel CE provides. Figure 12 shows an electropherogram of a crude 22mer RNA oligonucleotide (13).

Internucleotide phosphate analogs are being inten- sively studied for their inhibition of gene expression

with nuclease protection. These experiments, usually referred to as the antisense effect, are most frequently conducted with phosphorothioate oligonucleotides, in which one of the nonbridging oxygen atoms of the in- ternucleotide phosphate has been replaced with a sulfur atom. The synthesis of the phosphorothioate analogs is very efficiently conducted on the DNA synthesizer with the usual phosphoramidite chemistry, using tet- raethylthiuram disulfide (TETD) as a sulfurizing agent (14), instead of the iodine oxidizing reagent. They are significantly more hydrophobic than their phospho- diester, oxygen-containing counterparts. Also, the sul- furizing reaction is not stereospecific at the chiral phosphorous center, yielding a large number of chem- ically distinct diastereomeric products. The net result upon analysis, by any of the common methods, is a slight broadening of the product peak or band. Gel CE is again a very useful technique for assessing the purity of phosphorothioate oligonucleotides, either in the crude or purified state. Oligonucleotides containing both phosphodiester and phosphorothioate linkages

120.0-

llO.O-

lOO.O-

90.0-

so.o- P

=; 70.0-

x ifi 60.0-

f ‘”

50.0-

40.0 1

FIG. 12. Crude RNA 22mer: 5’ AUA AUG GUU UGU UUG UCU UCG U 3’.

224 ALEX ANDRUS

50.0-

45.0-

40.0-

35.0-

25.0-

20.0-

15.0-

FIG. 13a. 1 P-S, 23 P-O: 5’ AGT CAG TCA GTC AsGT CAG TCA GTC T 3’.

may be easily prepared in a single synthesis operation. Figures 13a, 13b, and 13c show three electropherograms of a 25mer sequence consisting of phosphodiester and phosphorothioate in three different arrays. The se- quence with a single phosphorothioate linkage, Fig. 13a, shows the narrowest peaks and the earliest elution of the product. The other two electropherograms, Figs. 13b and 13c, show the effects of the presence of many phosphorothioate linkages: broadened peaks, lower resolution, and slightly increased elution times.

CONCLUSION

Gel capillary electrophoresis offers a direct method of evaluating the purity of synthetic oligonucleotides. The denaturing gel matrix, under high-voltage condi- tions, gives the predictable, familiar elution patterns for oligonucleotides, as well as DNA analogs, RNA, and their labeled conjugates. Coupled with an appro- priate data system, integrator, or chart recorder, gel

CE thus attains most of the virtues of PAGE and HPLC. With the attributes of high resolution, speed, and ease of automation, gel CE is well suited to oli- gonucleotide analysis.

Further investigations are needed to understand the factors controlling resolution. It is clear that electro- phoretic mobility of an oligonucleotide in a gel-filled capillary is influenced by more than mass and charge. Molecular size and shape, charge distribution, hydro- phobic elements, hydrogen bonding, and interactions of all types with the gel matrix seem to have more influence in gel CE than in PAGE. The highly desirable denaturing ability of the MicroGel matrix is consid- erable. The gel materials are still in an early develop- ment phase and there remain opportunities to further optimize resolution of oligonucleotides. Practical fea- tures such as reliability, durability, and shelf life of the gel capillaries must also be improved. The variables influencing the mass of oligonucleotide sample entering a capillary during the electrokinetic injection period is not yet well understood. Advances here will better en- sure reproducible peak heights, accurate integrations,

1o.c

35.0-

30.0-

> E 25.0-

zi B

9 d 20.0- ‘:

z

15.0-

FIG. 13b. 12 P-S, 12 P-O: 5’A G T C A G T C A G T C A G T C A G T C A G T C T 3’. oooossssoooossssoooossss

30.0

25.0

> E

zi 20.0 B

2 d I: 8

15.0

10.0

minutes

FIG. 13~. 23 P-S, 1 P-O: 5’ AGT CAG TCA GTC A,,GT CAG TCA GTC T 3’.

225

226 ALEX ANDRUS

increased capillary lifetimes, and optimal resolution. 2. Coupled with second-generation capillary electropho- 3. resis instruments, advances in gel matrices and method developments will establish gel capillary electropho- 4’ resis as a powerful technique for oligonucleotide anal- ysis. 5.

ACKNOWLEDGMENTS

7. The author thanks his colleagues, Christie McCollum, Dr. Pete

Theisen, Peter Wright, Jay Kaufman, Dr. Ravi Vinayak, Larry DeDionisio, Bob Dubrow, and Junko Stevens, for their technical assistance, and Jenny Andrus and Beth Sanchez for their editorial assistance.

8.

9.

10.

11. REFERENCES 12.

13.

1. Cohen, A. S., Najarian, D. R., Paulus, A., Guttman, A., Smith, J. A., and Karger, B. L. (1988) Proc. N&l. Acad. Sci. USA 85, 14. 9660-9663.

Paulus, A., and Ohms, J. I. (1990) J. Chromatogr. 50’7,113-123.

Demorest, D., and Dubrow, R. (1991) J. Chromatogr. 559, 43- 56.

Dubrow, B., DeDionisio, L., and Andrus, A. (1991) International Meeting of the Electrophoresis Society, Washington, DC, March 19-21. Andrus, A. (1992) Evaluating and Isolating Synthetic Oligonu- cleotides, Applied Biosystems, 6-l-6-16; Dubrow, R. S. (1991) Am. Lab. March. Grossman, P. D., Lauer, H. H., Moring, S. E., Mead, D. E., Old- ham, M. F., Nickel, J. H., Goudberg, J. R. P., Krever, A., Ransom, D. H., and Colburn, J. C. (1989) Am. Biotechnol. Lab.

Moring, S. E., Colburn, J. C., Grossman, P. D., and Lauer, H. H. (1990) LC GC 8,34. Karger, A. E., Harris, J. M., and Gesteleand, R. F. (1991) Nucleic AcidsRes. 19,4955-4962. User Bulletin Number 59, Applied Biosystems (March 1991) New Applications for the Oligonucleotide Purification Cartridge.

Keller, G. H., and Manak, M. M. (1989) DNA Probes, pp. 105- 148, Stockton Press, New York.

Goodchild, J. (1990) Bioconjugare Chem. 1, 165-187. Applied Biosystems User Bulletin No. 49, Aminolink 2.

Vinayak, R., Anderson, P., McCollum, C., and Hampel, A. (1992) Nucleic Acids Res. 20, 1265-1269. Vu, H,, and Hirschhein, B. L. (1991) Tetrahedron Lett. 32,3005- 3008.