Journal of Chromatography, 260 (1983) 419-427
`Elsevier Science Publishers B.V., Amsterdam — Printed in the Netherlands
`
`CHROM. 15,668
`
`COMPARISON OF SEVERAL HIGH-PERFORMANCE LIQUID CHROMA-
`TOGRAPHY TECHNIQUES FOR THE SEPARATION OF OLIGODEOXY-
`NUCLEOTIDES ACCORDING TO THEIR CHAIN LENGTHS
`
`WOLFGANG HAUPT and ALFRED PINGOUD*
`Abteilung Biophysikalische Chemie, Zentrum Biochemie, Medizinische Hochschule Hannover, Konstanty-
`Gutschow-Strasse, D-3000 Hannover 61 (G.F.R.)
`
`(Received December 10th, 1982)
`
`
`
`SUMMARY
`
`The use of reversed-phase, reversed-phase-ion-pair and anion-exchange high-
`performance liquid chromatography (HPLC) wasinvestigated for the analytical and
`preparative separation of oligodeoxynucleotides according to their chain length. The
`data obtained with homooligonucleotides and oligonucleotides of defined sequence
`show that reversed-phase-ion-pair and anion-exchange, but not reversed-phase,
`HPLC canbe used reliably to separate oligodeoxynucleotides accordingto their chain
`length, largely irrespective of their base composition. The chain length limits for
`complete separation within 30 min by reversed-phase, reversed-phase-ion-pair and
`anion-exchange HPLCare approximately 10, 15 and 20 nucleotides, respectively.
`
`INTRODUCTION
`
`High-performance liquid chromatography (HPLC), because of its high reso-
`lution, reproducibility and ease of operation, has become an indispensible method
`for the separation of oligodeoxynucleotides, for both analytical and preparative pur-
`poses. A variety of different techniques have been developed, based on reversed-phase
`HPLC' >, reversed-phase-ion-pair HPLC*-5, stronganion-exchange HPLC®° andsize
`exclusion HPLC!°, and have been used so far mainly for the separation of short
`oligonucleotides. The separation of oligodeoxynucleotides by these techniquesis de-
`termined by the base composition and chain length of the oligodeoxynucleotides to
`be separated, which in turn determine their polarity and size. These properties, albeit
`to different extents in the various chromatographic systems, control the degree of
`retention of a particular oligodeoxynucleotide. The retention mechanisms involved
`are not fully understood.
`Reversed-phase HPLCofoligodeoxynucleotides is dominated by hydrophobic
`interactions between the solute and the bonded phase!; it is more sensitive to the
`base composition than reversed-phase-ion-pair, anion-exchange and size exclusion
`HPLC. Nevertheless the retention sequence of oligonucleotides with similar base
`composition is controlled by the charge of the oligonucleotide.
`In reversed-phase-ion-pair HPLC,negative charges on the oligodeoxynucleo-
`
`0021-9673/83/$03.00
`
`© 1983 Elsevier Science Publishers B.V.
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 1
`Page 1
`
`

`

`420
`
`W. HAUPT, A. PINGOUD
`
`tide are neutralized by positively charged alkylammonium ions.It is not clear, how-
`ever, whether this process occurs in the mobile phase!! or in the stationary phase??,
`leading to dynamic ion exchange. The retention is determined mainly by the charge
`of the oligodeoxynucleotide; in addition to these dominating electrostatic interac-
`tions, hydrophobic interactions between the oligodeoxynucleotides and the reversed
`phase also play a role in the chromatographic process‘,
`The separation of oligodeoxynucleotides by anion-exchange HPLC is depen-
`dent on differences in charge; as currently used anion exchangers are based onasilica
`gel matrix, in which the tertiary ammonium moiety is connected to the resin via an
`aliphatic spacer, there are hydrophobic interactions in addition to the ionic interac-
`tions. The inclusion of non-polar solvents in the mobile phase diminishesthe effect
`of the reversed phase on the chromatography!?-1*.
`Recently, silica gel bonded polyol phases have become commercially available
`for the separation of oligodeoxynucleotides by size exclusion HPLC. The separation
`can be carried out with a variety of mobile phases. The resolution depends on the
`mobile phase, indicating that size is not the only determinant for the chromatogra-
`phy. For small and medium sized oligodeoxynucleotides the resolution is not as good
`as with the above-mentioned chromatographic systems!°.
`Onecharacteristic feature of a particular oligodeoxynucleotide, and one that
`can be used conveniently for the purpose of analysis or preparative purification, is
`its size. It is therefore desirable to have chromatographic systems that allow sepa-
`rations accordingto size, i.e., chain length. A separation in such a chromatographic
`system ideally should be unaffected by base composition and should allow separation
`with high resolution over a wide range of chain lengths. For this purpose we inves-
`tigated the suitability of two standard and one new chromatographic system for chain
`length-dependent separations of oligodeoxynucleotides, namely (1) reversed-phase
`HPLC on Zorbax ODScolumns, (2) reversed-phase-ion-pair HPLC on LiChrosorb
`RP-8 columns and (3) anion-exchange HPLC on Partisil SAX columns. We used
`commercial homooligodeoxynucleotides of specified chain lengths as well as chemi-
`cally or enzymatically synthesized oligodeoxynucleotides ofdefined sequence to study
`the limits of resolution of these chromatographic systems.
`The results show that both reversed-phase-ion-pair HPLC and anion-exchange
`HPLC,but not reversed-phase HPLC,can be used reliably to separate oligodeoxy-
`nucleotides according to their chain length, with only slight interference by base
`composition. Whereas, reversed-phase-ion-pair HPLC hasa cut-off for an acceptable
`resolution at a chain length of around 15-20 nucleotides, anion-exchange HPLC
`allows even longer oligodeoxynucleotides to be separated.
`
`EXPERIMENTAL
`
`Instrumentation
`Two chromatographswereused: (1) a DuPont 850 liquid chromatograph,con-
`sisting of a gradient controller, a three-head pump,a temperature-controlled column
`compartment, a UV spectrophotometer and a Rheodyne 7125 injection valve, and
`(2) a Pye Unicam liquid chromatograph,consisting of an LC-XP gradient program-
`mer, an Altex Model 100A pump, an LC-UV detector and a Rheodyne 7125 injection
`valve, to which a Merck LiChrocart-Autofix pneumatic mounting device was at-
`tached.
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 2
`Page 2
`
`

`

`HPLC OF OLIGODEOXYNUCLEOTIDES
`
`421
`
`Chemicals and oligonucleotides
`All homodeoxyoligonucleotides were obtained from P.L. Biochemicals (St.
`Goar, G.F.R.).(d(GGAATTCC) and d(CGAATTCG} were kindly given by Dr. H.
`Bloecker and Dr. R. Frank (GBF, Stéckheim, G.F.R.) and by Dr. M. Zabeau
`(EMBL,Heidelberg, G.F.R.), respectively. Tetrabutylammonium hydrogen sulphate
`(TBAHSO,),
`triethylamine and formamide were products of Fluka (Neu-Ulm,
`G.F.R.) and were used without further purification. HPLC-grade acetonitrile was a
`product of Burdick & Jackson Labs. (Muskegon, MI, U.S.A.) and was provided by
`Macherey, Nagel & Co, (Diiren, G.F.R.). T4 DNA ligase was obtained from BRL
`(Neu-Isenberg, G.F.R.). All other chemicals used were of pro analysi grade from
`Merck (Darmstadt, G.F.R.).
`
`Columns and chromatographic conditions
`All eluents were filtered through a Whatman GF/Fglass-microfibre-filter (Vet-
`ter, Wiesloch, G.F.R.) and degassed prior to chromatography.
`Forreversed-phase separations Zorbax ODS columns (25cm x 4.6mm I.D., 5
`pum) from DuPont (Bad Nauheim, G.F.R.) were used. The triethylammonium ace-
`tate (TEAA)buffer was prepared as described elsewhere’. Solvent A was 0.1 4’ TEAA
`(pH 7.0)-1% acetonitrile and solvent B was 0.1 M TEAA(pH 7.0)-50% acetonitrile.
`If not stated otherwise, a linear gradient from 15 to 30% B in 30 min was applied.
`The fiow-rate was 1.0 ml/min. Chromatography wascarried out at ambient temper-
`ature.
`
`Reversed-phase-ion-pair separations were performed on Hibar LiChrocart
`cartridges (25cm x 4mm [I.D.) packed with LiChrosorb RP-8, 10 um (Merck). Sol-
`vent A was 50 mM potassium phosphate (pH 5.9)-2 mM TBAHSO,andsolvent B
`was 50 mM potassium phosphate (pH 6.5)}-2 mM TBAHSO,-60%acetonitrile. The
`standard linear gradient used was from 15 to 80% B in 60 min. The flow-rate was
`1.0 ml/min. Chromatography wascarried out at ambient temperature.
`WhatmanPartisil 10 SAX columns (25 x 4.6mm I.D., 10 pm) obtained from
`IC-Chemikalien (Munich, G.F.R.) were used for anion-exchange HPLC.Solvent A
`was 1 mM potassium phosphate (pH 6.3) in formamide-water (6:4) and solvent B
`was 0.3 M potassium phosphate (pH 6.3) in formamide-water (6:4). The linear gra-
`dient used was from 0 to 80%B in 60 min. The flow-rate was 1.0 ml/min. Chro-
`matography was carried out at 45°C.
`
`RESULTS
`
`Figs. 1 and 2 show chromatogramsof mixtures of oligo-(dA)s and oligo-(dT)s
`separated by reversed-phase chromatography. Oligonucleotides of different chain
`length are well resolved within the oligo-(dA) and oligo-(dT) series up to a chain
`length of at least ten nucleotides under the conditions given.
`Oligo-(dT)s are considerably more retarded than oligo-(dA)s of the same chain
`length. Thus, for example, p(dA),9 is eluted before p(dT)¢. This demonstrates that
`the base composition ofthe oligonucleotide has a considerable effect on the retention
`of the oligonucleotides. This effect, of course, can be of advantage when oligomers
`of similar chain length or even sequence isomers have to be separated. Thisis illus-
`trated in Fig. 3, where a separation of d(GGAATTCC)from its sequence isomer
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 3
`Page 3
`
`

`

`W. HAUPT,A. PINGOUD
`
`260
`
`422
`
`260
`
`0.02
`
`0.02
`
`0
`
`10
`
`20
`
`min
`
`0
`
`10
`
`20
`
`30
`
`min
`
`Fig. 1. Reversed-phase HPLC ofoligo-(dA)s. Column: Zorbax ODS (25 cm = 3.6 mm I.D.). Eluents: A,
`0.1 M TEAA (pH) 7.0-1% acetonitrile; B, 0.1 / TEAA (pH 7.0)}-50% acetonitrile; linear gradient from
`15 to 30% B in 30 min at ambient temperature; flow-rate, 1.0 ml/min. Unidentified peaks are due to
`contaminants in the individual samples.
`
`Fig. 2. Reversed-phase HPLC of oligo-(dT)s. Chromatographic conditionsas in Fig. 1.
`
`d(CGAATTCG)is shown.In fact, our analysis of commercially available oligo-(dT)s
`andin particular of oligo-(dA)s has shownthat these preparations contain consider-
`able amounts of impurities (¢f., Figs. 1 and 2), which are not detected by polyacryl-
`amide gel electrophoresis under denaturing conditions, by homochromatography
`after labelling the 5’-end with [3?P]phosphate or by anion-exchange HPLC(see be-
`low), suggesting that they consist of chemical derivatives of these oligonucleotides
`with the same chain length or charge.
`Figs. 4 and 5 show separations of mixtures of oligo-(dA)s and oligo-(dT)s in
`a reversed-phase-ion-pair system. The individual oligonucleotides are extremely well
`resolved up to a chain length ofat least 16 nucleotides, the peaks being sharper and
`more symmetrical than in the reversed-phase system (compare Figs. 4 and | or Figs.
`5 and 2). Oligo-(dA)s are slightly more retained than oligo-(dT)s, e.g., p(dA)x is
`eluted after p(dT),9. The influence of base composition is less pronounced in the
`reversed-phase system. Accordingly, reversed-phase-ion-pair HPLCis not as useful
`as reversed-phase HPLCforthe separation ofoligonucleotides of different base com-
`position but similar chain length. Thisis also apparent from Fig. 6, which showsthe
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 4
`Page 4
`
`

`

`HPLC OF OLIGODEOXYNUCLEOTIDES
`
`423
`
`pidAlg
`
`~—pidalyp(dAly2“~~pidAg
`
`0.02 10
`
`hmno 3a 3
`
`>
`
`pidAls,
`
`ata
`3a
`
`0
`
`20
`
`min
`
`0
`
`10
`
`20
`
`30
`
`min
`
`Fig. 3. Reversed-phase HPLC of the sequence isomers d(GGAATTCC) and d(CGAATTCG). Column:
`Zorbax ODS (25cm x 4.6mm I.D.). Eluents: A, 0.1 4f TEEA (pH 7.0)-1% acetonitrile; B, 0.1 44 TEAA
`(pH 7.0}-50% acetonitrile; linear gradient from 20 to 25% B in 20 min at ambient temperature;flow-rate,
`1 ml/min.
`
`Fig. 4. Reversed-phase-ion-pair HPLC ofoligo-(dA)s. Column: LiChrocart cartridge (25cm = 4.6 mm
`LD.) filled with LiChrosorb RP-8. Eluents: A, 50 mM potassium phosphate (pH 5.9)-2 mM TBAHSO,;
`B, 50 mM potassium phosphate (pH 6.5}-2 mM TBAHSO,-60%acetonitrile; linear gradient from 15 to
`80% B in 60 min at ambient temperature; flow-rate, 1.0 ml/min.
`
`reversed-phase-ion-pair chromatogram of a mixture of d(GGAATTCC) and
`d(CGAATTCG). Although the retention time is longer than in the reversed-phase
`chromatography shownin Fig. 3, the separation is not as good.
`Figs. 7 and 8 show chromatogramsof mixtures of oligo-(dA)s and oligo-(dT)s
`separated by strong anion-exchange HPLC. Theretention times under the chosen
`conditions are nearly constant with respect to base composition, e.g., p(dA)1o0 is
`eluted at the same position as p(dT),9. The resolution of oligonucleotidesof increas-
`ing chain length is excellent up to a chain length of about 20. The high resolution
`is also demonstrated in Fig. 9, which shows the chromatogram ofthe separation of
`commercial oligo-(dA),,. More than 30 discrete peaks can be seen. The usefulness of
`ion-exchange HPLC for the separation of longer oligonucleotides is particularly
`apparent from Fig. 10. In the HPLC run shown the mixture of products resulting
`from the T4 DNAligase catalysed oligomerization of pd(GGAATTCC) was separ-
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 5
`Page 5
`
`

`

`0.02
`
`30
`20
`10
`0
`min
`30
`20
`10
`0
`Fig. 5. Reversed-phase-ion-pair HPLCofoligo-(dT)s. Chromatographic conditions as in Fig. 4.
`Fig. 6. Reversed-phase-ion-pair HPLC of the sequence isomers d((GGAATTCC)and d(CGAATTCG).
`Chromatographic conditions as in Fig. 4.
`
`min
`
`ated by anion-exchange HPLC. Resolution was achieved up to the “octamer’’,i.e.,
`up to a chain length of 64 nucleotides.
`We have also used commercial preparations of oligodeoxynucleotides of the
`(dG) and (dC)series in reversed-phase, reversed-phase-ion-pair and anion-exchange
`HPLC.In general, the oligonucleotides of defined chain length as stated on the label
`proved to be much less homogeneous than the corresponding (dA)- and (dT)-oli-
`godeoxynucleotides. Nevertheless, similar results were also obtained with the oligo-
`nucleotides of the (dG) and (dC) series compared with those of the (dA) and (dT)
`series (data not shown).
`
`DISCUSSION
`
`We have compared the utility of three different chromatographic systems
`for the separation of oligodeoxynucleotides by HPLC according to chain length.It
`was not our aim to improve previously developed chromatographic system, but rath-
`er to use them, with minor modifications, in order to establish their limits of reso-
`lution with respect to size and base composition. Ourresults show that withinaseries
`of homooligonucleotides all three systems reliably lead to separation according to
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 6
`Page 6
`
`424
`
`260
`
`W. HAUPT, A. PINGOUD
`
`3 P
`
`e3i
`
`d(GGAATTCC)
`
`

`

`HPLC OF OLIGODEOXYNUCLEOTIDES
`
`425
`
`
`
`270
`
`0.02
`
`0.02
`
`30
`20
`10
`0
`min
`30
`20
`10
`0
`Fig. 7. Anion-exchange HPLC ofoligo-(dA)s. Column: Whatman 1 SAX (25cm x 4.6mm I.D.). Eluents:
`A, | mM potassium phosphate (Ph 6.3) in formamide—water(6:4); B, .3 Mf potassium phosphate (Ph 6.3)
`in formamide-water (6:4); linear gradient from to 8% in 6 min at 45°c; flow-rate,
`|. ml/min.
`
`min
`
`Fig. 8. Anion-exchange HPLC ofoligo-(dT)s. Chromatographic conditions as in Fig. 7.
`
`chain length. The chain length limits for separation by reversed-phase, reversed-
`phase-ion-pair and strong anion-exchange HPLCare ca. 10, 15 and 20 nucleotides,
`respectively. These limitis apply for the complete separation of homologous oligo-
`deoxynucleotides within ca. 30 min. With suitable modifications, e.g., by using dif-
`ferent gradients, these limits may well be extended towards longer chain lengths.It
`is known from the separation of short oligonucleotides that the retention of oligo-
`deoxynucleotides in reversed-phase HPLCis very sensitive to the base composition.
`Ourresults show that oligo-(dT)s are much more strongly retained on Zorbax ODS
`(this work) and LiChrosorb RP-8 columns (data not shown) than oligo-(dA)s. The
`retention of oligodeoxynucleotides in reversed-phase-ion-pair HPLC, on the other
`hand,is only slightly influenced by the base composition. Interestingly, our results
`indicate that with this chromatographic system oligo-(dA)s are more strongly re-
`tained than oligo-(dT)s, probably a consequence of the different base stacking',
`which might affect the accessibility of the phosphate residues of the oligodeoxynu-
`cleotides for the ion-pairing agent.
`Under the conditions applied the retention of oligodeoxynucleotidesis least
`affected by the base composition of the sample in anion-exchange HPLC.It has been
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 7
`Page 7
`
`

`

`426
`
`0
`
`20
`
`40
`
`min
`
`0.03
`
`W. HAUPT, A. PINGOUD
`0.01
`
`
`
`Fig. 9. Anion-exchange HPLC of a material that was designated by the supplier as p(dA)25-30. Column
`and eluent as in Fig. 7; linear gradient from 0 to 100% B in 75 min at 45°C; flow-rate, 1 ml/min.
`
`0
`
`20
`
`40
`
`min
`
`Fig. 10. Anion-exchange HPLCofthe product mixture of the T4 DNAligase catalysed oligomerization
`of pd(GGAATTCC) that was obtained by incubating 6 uM [3??P]pd(GGAATTCC) and 24 pM
`d(GGAATTCC)at 22°C for 16 h with 48 units/ml of T4 DNAligase in a mixture of 40 mM Tris HCl (pH
`7.6), 8 mM magnesium chloride, 1 mM dithioerythritole (DTE) and 470 ym ATP. The separated peaks
`were identified by polyacrylamide gel electrophoresis'’: a = d(GGAATTCC) monomer; b = dimer; c
`= trimer; d = tetramer; e = pentamer; f = hexamer; g = heptamer; a’ = AMP-pd(GGAATTCC)
`ligation intermediate. Analysis of the fractions by Cerenkov radiation in a scintillation counter indicated
`that oligonucleotides beyond the octamer were resolved. The small satellite peaks of peaks a, a’, b, c, d
`nd
`babl:
`due to
`slight hete
`eity
`in the starting
`material.
`CUREVAC EX2029
`and
`€ pro
`y are due
`to
`slight
`heterogeneity 1
`starting materi
`CUREVAC EX2029
`Page 8
`Page 8
`
`

`

`HPLC OF OLIGODEOXYNUCLEOTIDES
`
`427
`
`reported, however, that under different conditions, in particular in the absence of
`organic solvents in the mobile phase, anion-exchange HPLC canby used to separate
`sequence isomers of small oligonucleotides*:17. This demonstrates thatthe inclusion
`of organic solvents in the mobile phase is necessary to suppress hydrophobicinter-
`actions between the sample and the stationary phase!?-'+,
`Taken together, our results suggest that for short to medium sized oligonu-
`cleotides reversed-phase-ion-pair and anion-exchange HPLCare equally suitable for
`separations according to chain length (cf. also ref. 18), whereas for longer oligonu-
`cleotides anion-exchange HPLCis superior to reversed-phase-ion-pair HPLC.It
`must be mentioned in this context, however, that anion-exchange HPLC columns
`seem to suffer from shorterlifetimes* than reversed-phase HPLC columns. Thus, for
`analytical or preparative work with small
`to medium sized oligonucleotides re-
`versed-phase and reversed-phase-ion-pair HPLCis the combination of choice, one
`system separating according to polarity and size, the other predominantly according
`to size. For separations of larger oligonucleotides by HPLC there is as yet no alter-
`native to anion-exchange or size exclusion HPLC.
`
`ACKNOWLEDGEMENTS
`
`Wethank Dr. W. Block for his active support during the initial phase of this
`work. The expert technical assistance of Mr. R. Mull is gratefully acknowledged.
`This work was supported by grants from the Deutsche Forschungsgemeinschaft and
`the Fonds der Chemischen Industrie.
`
`REFERENCES
`
`1 H.-J. Fritz, R. Belagaje, E. L. Brown, R. H. Fritz, R. A. Jones, R. G. Lees and H. R. Khorana,
`Biochemistry, 17 (1978) 1257.
`R. A. Jones, H. J. Fritz and H. G. Khorana, Biochemistry, 17 (1978) 1268.
`G. D. McFarland and P. N. Borer, Nucl. Acid Res., 7 (1979) 1067.
`J. B. Crowther, R. Jones and R. A. Hartwick, J. Chromatogr., 217 (1981) 479.
`W
`. Jost, K. Unger and G. Schill, Anal. Biochem., 119 (1982) 214.
`6M
`. J. Gait and R. C. Sheppard, Nucl. Acid Res., 4 (1977) 1135.
`7M
`. J. Gait and R. C. Sheppard, Nucl. Acid Res., 4 (1977) 4411.
`8M
`. Dizdaroglu and W. Hermes, J. Chromatogr., 171 (1979) 321.
`9
`M. J. Gait, H. W. D. Matthes, M. Singh, B. S. Sproat and R. C. Titmas, Nucl. Acid Res., 10 (1982)
`6243.
`10 D. Molko, R. Derbyshire, A. Guy, A. Roget, R. Teoule and A. Boucherle, J. Chromatogr., 206 (1981)
`493,
`
`2 34S
`
`11 Cs, Horvath, W. Melander, I. Molnar and P. Molnar, Anal. Chem., 49 (1979) 2295,
`12 R. P. W. Scott and P. Kucera, J. Chromatogr., 175 (1979) 51.
`13 P. J. M. van Haastert, J. Chromatogr., 210 (1981) 229.
`14 P. J. M. van Haastert, J. Chromatogr., 210 (1981) 229.
`15 T. Maniatis, A. Jeffrey and H. van de Sande, Biochemistry, 14 (1975) 3787.
`16 T. N. Solie and J. Schellmann, J. Mol. Biol., 33 (1968) 61.
`17 M. Dizdaroglu, W. Hermes, C. von Sonntag and H. Schott, J. Chromatogr.,, 169 (1979) 429.
`18 H. J. Fritz, D. Eick and W, Werr, in H. G. Gassen and A. Lang (Editors), Chemical and Enzymatic
`Synthesis of Gene Fragments, Verlag Chemie, Weinheim, 1982, p. 199.
`
`CUREVAC EX2029
`CUREVAC EX2029
`Page 9
`Page 9
`
`