Iowan! of Chromatography. 389 ([987) [65—116
`Lilscvicr Science Publishers EAL, Amsterdam — Printed in The Netherlands
`
`(2H ROM. [9 176
`
`COLUMN CHROMATOGRAPHIC PURIFICATION OF GUANYLATE-RICH
`SYNTHETIC OLIGODEOXYRIBONUCLEOTIDES
`
`HERBERT SCHOTT*, ROLF SEMMLER and IIEINER ECKSTEIN
`
`Jasmin .fiiir Organise-he Chcmire, Unilwxmit Tiibingeii, An]! der Momensmfle i8, 0—?400 Trilyingan—i
`(ERG)
`(Received October ZOth. 1986)
`
`SUMMARY
`
`The purification of the guanylate—rich DNA fragments d(T4G4]. d(G4'I‘4),
`d(G4T4G4), d(T4G4T4) and d(T4G4T4(]4) using column chromatography on a pre-
`parative scale is described. The crude oligonucleotides were obtained after deprotcc—
`tion ofthe chemically synthesized compounds. The separation can be performed with
`commonly used sorbents (DEAE—cellulose, QAE—Sephadex, Nucleosil C13. Partisil
`lO-SAX), however with high losses during the chromatography. Guanylate-rich oli—
`gonucleotides ol‘ different chain lengths associate with each other, thus causing iden-
`tical compounds to be contained within different peaks. At the same time. part of
`the product remains irreversibly adsorbed on the sorbent. The recoveries could be
`improved by application of ion-pair reversed-phase high-performance liquid chro—
`matography. The oligonucleotidcs were Fractionated with linear increasing gradients
`using acetonitrile as the organic modifier and tetrabutylammonium hydrogensulphate
`as the ion-pair reagent.
`
`INTRODUCTION
`
`The Synthesis of guanylate—rich DNA fragments is much more laborious than
`the Synthesis ol‘comparable oligonucleotides, which contain only few or no guanylate
`monomer units in their sequences. Besides distinctly lower yields obtained in the
`condensation reaction, there are additional difficulties in separation, isolation and
`identification, which have not been solved.
`We have synthesized the guanylate-rich oligonucleotides d(T4G4), (1(G4T4),
`d(G¢T4G4), d(T4G4T4) and d(T4G4T4G4) in preparative amounts. These oligonu-
`cleotides correspond to fragments of the terminus of the macronuclear DNA of hy-
`potrichous ciliates'_fi. The syntheses were carried out in solution according to the
`phosphotriester method and will be published elsewherei. The same oligonucleotides
`were prepared in three ways, applying differently protected guanylate monomer units.
`The results ol‘ thirty difTercnt condensation reactions, several of which have been
`performed repeatedly, may be summarized as follows: the synthesis and isolation of
`the protected guanylate-rich oligonucleotides can be achieved in gram amounts. The
`
`UGZ]—9673;’BT,E'S(]3_50
`
`at}
`
`[987 Elsewer SelenCe Publishers BM.
`
`CUREVAC EX2041
`CUREVAC EX2041
`Page 1
`Page 1
`
`

`

`”,6
`
`H. scnorr, R. SEMMLER, II. ECKSTEIN
`
`condensation reactions result in good yields, which are essentially independent of the
`kind of protecting groups and of the choice of the agents used For the condensation.
`However, serious difficulties arise during the chromatographic purification of the
`deblockcd oligonueleotides as will be reported in this paper.
`
`EXPERIMENTAL
`
`Materials
`
`Sephadex G-lS-and QAE—Sephadex A—Zfi were obtained from Pharmacia
`(Uppsala. Sweden), DEAE-cellulose from W. R. Balstonc (Maidstone, U.K.), Dowex
`50W—X8 from Serva (Heidelberg, F.R.G.) and Nucleosil 7 C13 from Macherey &
`Nagel (Dtiren, F.R.G.). Tetrabulylammonium hydrogensulphate (TBA) for ion-pair
`chromatography was from Merck (Darmstadt, F.R.G.).
`
`Fractionation
`
`Coiamn chromatography of the demarcated oiigoaacieotides. The deprotected
`oligonueleotides were fractionated on DEAE—cellulose or QAE~Sephadex at a flow-
`rate of 200 mlih, according to the conditions listed in Table l'. Fractions of about 20
`ml were collected. The absorbance of every fifth fraction was measured at 250, 260
`and 280 nm. The values measured at 260 nm were plotted versus the clution volume
`(Fig. l). Fractions were collected within the vertical dotted lines of Fig. 1. On re-
`peated addition of pyridine,
`the volatile triethylammonium hydrogencarbonate
`(TEAB) was removed in vacao. The pyridine was removed by co-evaporation with
`3% aqueous ammonia. Finally the remaining solution was lyophilizcd. The sodium
`chloride—Tris—HCI buffer was removed by gel chromatography on a Sephadex G-15
`column. In order to remove the urea, the combined peak fractions were diluted to
`1:25 in water and pumped on a DEAE—Sephadex column (40 cm X 2 cm) previously
`equilibrated with water. The column was washed with water until free of chloride,
`and was then eluted with a l M sodium chloride solution. The oligonucleotides eluted
`with the salt were desalted by gel chromatography and lyophilized.
`High—performance liquid chromatography {HPLC} of the deprorecied oiigoaa—
`decades. HPLC was performed according to the conditions summarized in Table II
`on an analytical column (250 mm X 4.6 mmm LB.) and a preparative column (250
`mm X 8 mm LD.) equipped with a precolumn (30 mm x 8 mm l.D.) packed with
`Nucleosil 7 C13. Experiments 1 3 of Table II were carried out at room temperature,
`445 at 50°C. One A260 unit of the oligonucleotide was dissolved in 1—10 at water and
`applied to the column. The combined fractions were desalted as follows: the TBA
`solution obtained was added to 50 ml dichloromethanc. A saturated aqueous picric
`acid solution was added dropwise to the stirred mixture until the aqueous layer had
`become slightly yellow. After separation of the layers, the aqueous phase was treated
`with Dowex SOW-XS (H+) and ehromatographed on a Sephadex G-lS column (40
`cm X 4 cm). The fractions containing product were combined, evaporated to dryness
`in racuo and lyophilized.
`HPLC ofdi T4G4T‘4G4) after i01alh_t’dr0t‘ysis byformir acid. One A360 unit of
`d(T4G4_T4G4) was treated with 500 pl of 90% formic acid at 170°C during 45 min.
`The reaction mixture was lyophilizcd and dissolved in about 200 pl of 50 mM
`aqueous ammonium acetate (pH 6.8). About 0.10 A260 units of this solution were
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 2
`Page 2
`
`

`

`PURIFICATION OF OLIGOI) EOXYRIBONUCLEOTIDES
`
`[57
`
`TABLE I
`
`CIIROMATOGRAPHIC PURIFICATION OF PROBES (DISSOLVED IN WATER) OBTAINED AFTER THE
`DEPRO'I‘ECTION OF THE PROTECTED DODECAMERS AND IIEXADECAMERS USING DIEAE-CEL—
`LUI.OSE (EXPERIMENTS 1. 2, 3:1, 4) OR QABSEPHAIH‘ZX (EXPERIMENT 3)
`The columns [diameter 2 cm) were eluted with increasing salt concentration using triethylamrnoni um hydrogencar—
`bonate (pl-I 18) (A) or sodium chloride—0.05 M Tris—HCI (pH 7.6) + T M urea (B).
`
`Expert?»
`men:
`No.
`
`Dept-steered
`oligonu—
`deonkte
`
`Applied
`probe
`(/1260
`unite-"mu
`
`Column
`length
`(cm)
`
`Volume (U and
`salt concentration {Mf’
`—_..
`Mixing
`Reser-
`vesse!
`1-0:}-
`
`
`Elation conditions
`
`Temper
`were
`("Cj
`
`Eluem
`
`Step
`No.
`
`1“
`
`T4G4T4
`
`10 200300
`
`50
`
`2
`
`3
`
`G4T4G4
`
`45003150
`
`TaGaTaGit
`
`[4 500.850
`
`25
`
`25
`
`3a
`
`T4G4T4Gf“
`
`39003100
`
`25
`
`4
`
`Lorrie.
`
`6700;200
`
`25
`
`50
`
`25
`
`25
`
`50
`
`so
`
`l
`2
`3
`4
`
`l
`2
`3
`
`I
`2
`3
`4
`
`[
`2
`3
`4
`
`I
`2
`3
`4
`
`1.0, 0.05
`B
`2.0. 0.05
`B
`2.0. 0.15
`B
`1 M NaCl 1.0
`
`A
`A
`A
`
`1.0. 0.10
`2.0, 0.10
`0.5,
`|.0
`
`1.0. 0.05
`B
`2.0. 0.05
`B
`1.0, 0.5
`I3
`l M NaCl 0.?
`
`1.0. 0.05
`B
`2.0. 0.05
`B
`1.5. 0.20
`B
`I M NaCl I0
`
`1.0, 0.05
`B
`2.0, 0.05
`B
`1.5, 0.20
`B
`1 M NaCl |.0
`
`—
`2.0. 0.15
`2.0, 0.30
`—
`
`2.0, 0.40
`—
`
`—
`2.0, 0.50
`
`—
`2.0, 0.20
`1.5, 0.35
`
`,
`2.0. 0.20
`1.5, 0.35
`
`" When B is used as the eluent M refers only to the sodium chloride concentration.
`1" The total amounls ol‘ deprotected oligonucleotides were chromatographed in three experiments.
`“* Rechromalography or the mixture of d(T4G+) and d(T4G.,T4G4) isolated from experiment 3.
`
`fractionated on a Nucleosi] 7 (713 column (250 mm x 4.6 mm ID.) with 50 mM
`ammonium acetate (pl-I 6.8) as the eluent [see Fig. 5}.
`
`RESULTS AND DISCUSSION
`
`The DNA fragments d(G4T4G4), d(T4G4T4) and d(T4G.._T4G 4) were obtained
`from the corresponding fully protected oligonucleotides after cleavage of the pro—
`tecting groups and chromatographic separation of' the oligonucleotides. The crude
`product d(T4G4T4), obtained after deproteetion of 1.05 g dodecamer. was fraction-
`ated on DEAE-cellulose in three portions, employing a linear increasing gradient of
`sodium chloride containing 7 M urea at 50°C. as indicated in Table l (experiment I).
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 3
`Page 3
`
`

`

`H. senor]: R. SEMMLER, II. [ECKSTEIN
`
`
`I E
`
`2
`:3—
`
`050
`
`can
`
`(no
`
`020
`mo
`ans
`
`
`
`0.1.0
`
`
`
`
`
`
`
`.._
`{130
`a;
`'6
`£120 5\
`
`L3
`010
`(105 E
`
`—-
`'r_
`75
`E\
`
`‘6Z
`
`163
`
`6a
`
`it
`
`3
`4
`
`DDN
`4
`
`Eluiion vol. inl
`
`
`
`Fig. l. Chromatographic purification ofdt'I‘4G4T4G4) resulting after deprolcction of the prolecled hexa~
`dccanucleolidcs, which were synthesized using dificrcnt strategies. (21] Fractionation (experiment A Tables
`I, III) ofthe first hexadocanucieotidc d('I‘4G.,T4G4) on a QAE-Sephadcx column at 25"C with an increasing
`sodium chloride gradient. buffered to pIi 7.6 by (1.05 M Tris—HG]. (h) Rcchromillography of the mixture
`corresponding to peak I] resulting from (a) on a DEAE-oellulosc column (experiment 3a, Tables I. III}
`at 50°C with an increasing sodium chloride gradient in 7 M urea, buHereti to pH 7.6 by 0.05 M Trin—HCI.
`(c) Fractionation (experiment 4. Tnblesi, III) of the second hexadwanucleotidc d[‘l'4G‘T,,G‘) using the
`same conditions as in (b). Column: 25 cm X 2 cm. Flow—rate: 200 mlls'h. Within the dotted lines, fractions
`of peaks 1 and I] were pooled, desalted and lyophilizcd.
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 4
`Page 4
`
`

`

`PURIFICATION OF OLIGUDEOXYRIBONUCLEOTIDES
`
`169
`
`TABLE II
`
`CONDITIONS AND RESULTS OF ION—PAIR REVERSED—PHASE HPI.C OF (JLIGONUCLIEOTIDES ON
`A NUCLEOSIL ? C”, COLUMN (LENGTH 250 mm)
`
`Elucnls: A == 7.5 mM tetrabutylammonium hydrogcnsulphate (TBA) pH 7.0; B = 15 mM TBA pH 7.0 in 35%
`aqueous acetonitrile; C = 5 mM TBA pH 6.8; D = 5 mM TBA pH 6.8 in 70% aq. acetonitrile,
`
`
`Chromamgmphed oiigonucieotider Fig. Column Elation Retention
`
`
`
`--
`diameter
`conditions
`time
`from J
`(min)
`
`Designation
`
`Experi—
`mam
`No.
`
`Yield
`Amount
`{A zoo
`("4)
`units}
`
`
`1
`2
`3
`
`4
`S
`
`6
`
`d(G4T4)
`dt'T4G4'F4)
`d(T..,G4T.,)
`
`d(T4G.¢)
`d{G.{l}G..)
`
`d(T.tG4T4G4)
`
`60.0
`2.5
`50.0
`
`0.5
`0.3
`
`0.3
`
`9]
`98
`94
`
`92
`60
`
`9]
`
`8.0
`4.6
`8.0
`
`4.6
`4.6
`
`4.6
`
`?0% A, 30% B
`50% A, 50% B
`50% A, 50% B
`
`60% C. 40% D
`changed within
`48 min to
`20% C, 80% D
`
`9.23
`12.92
`12.22
`
`26.22
`31.78
`
`36.23
`
`2a
`--
`2b
`
`33
`3h
`
`3c
`
`Of the A250 units applied to the column, 39% were due to d(T4G4'f4) and 18% to
`d(G4T4) . The remaining 43% consisted of removed protecting groups and of several
`oligonuclcotides of shorter chain length. Fractions which contained d(T4G4T4) or
`d(G4T4) from three experiments were pooled and worked up, thus giving 35! mg of
`dodecamers and 150 mg ol‘octamer. On the basis of the fully protected dodecamer,
`the yield ol‘d(T4G4T4) was 53%.
`The crude product of the second dodecamer d(G4T4G4) was also fractionated
`by means of DEAR—cellulose. When a small quantity of dodecamer d(G4T4G4) was
`chromatographed on a DEAE-cellulose column, no clear peaks could be detected in
`the region of the dodecamer. This result was quite a surprise, especially since the
`formation of a fully protected condensation product had been confirmed by thin
`layer chromatography. After the deprotection of 600 mg of the corresponding pro-
`tected d{G4T4G4), preparative chromatography (experiment 2, Table I) could be
`performed only with a considerable loss of oligonucleotides. Contrary to the pre-
`viously described fractionation of d(T4_G4T4), the column was eluted with an increas—
`ing concentration of triethylammonium hydrogencarbonate buffer (TEAB) at 25”C.
`The oligonuclcotide leaving the column at a salt concentration of 0.35—0.39 M was
`identified as d(T4G4}. The required dodecamer d(G4T4G4) was finally eluted by l M
`TEAB (see Table III). 15% of the applied A260 units were due to d(T4(}4) and 20%
`to d(G4T4G4). On working up the pooled fractions, 20 mg d(T4G4) and 30 mg
`d(G4T4G4) were obtained corresponding to a yield of only 8% in relation to the fully
`protected dodecamer.
`This rather low yield might be explained by assuming that part of the dode-
`eamer was degradatcd during the cleavage ofthe protecting groups. This explains the
`elution of numerous short—chain oligonucleolidcs. Another reason for the low yield
`lies in the fact that substantial losses of guanylate-rich dodecamers occur during
`chromatography. Chromatography at elevated temperature (45°C) did not increase
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 5
`Page 5
`
`

`

`no
`
`TABLE Ill
`
`H. scnorr, R. SEMMLL-‘R. H. ec-xsram
`
`RESULTS OF THF. CHROMATOGRAPHIC PURIFICATION (SEE TABLE I) OF THE DEPRUTECTED
`()LIGONUCLEUTIDES
`
`Oit'gonncfeoride eluted
`Expert?
`n18”! — --
`-- -
`
`—-—-
`
`-
`
`- —
`
`- --
`
`--—
`
`
`maimed oiigodmxynudemide
`-
`
`.
`
`. _—
`
`No.‘
`
`Peak
`Amount
`Salt
`concentration — (Fig. l)
`{M}
`(/1269
`1%)“
`units)
`
`
`Designation
`
`Weight
`(mg)
`
`l
`
`2
`
`3
`
`3a
`
`4
`
`0.13—0.11r
`0.21—0.25
`0.35 0.39
`1.00
`0.23-0.30
`0.42 0.50
`
`0.15—0.18
`0.27 0.30
`0.14 0.16
`0.30—0.33
`
`1880§
`3980§
`670
`920
`[350
`4630
`
`18.4
`39.0
`14.9
`20.4
`9.3
`32.2
`
`Not shown
`Not shown
`Not shown
`Not shown
`1(a)
`"(3)
`
`d(G4T4)
`d('I'4G4T4)
`d(T4G4)
`d(G..T..G4]
`d(T4Cu)
`d(T4(i4]
`+
`d(T4641434)
`dCT4G4)
`1(b]
`41.5
`1620
`dCT‘ci‘T‘G‘)
`11(b)
`21.0
`820
`2520
`37.6
`1(c)
`d(T4G4)
`
` 1800 26.9 I l(c) d(T..G4T4G4}
`
`
`
`
`Stti
`I 17‘j
`20
`30
`40
`
`170
`
`50
`30
`1'0
`55
`
`l'ield'“
`{94)
`
`52.":r§
`
`"L9
`
`4.?
`
`22. l
`
`" See Table I.
`" Based on the total amount of the probe applied.
`“* Based on the protected oligonucleotidcs.
`§ Average of three experiments.
`
`the recovery of guanylate-rich oligonucleotides. In this case, extended elution with
`strong buffer solution resulted in a broad second peak of nucleotide material. During
`rechromatography a part of this material was eluted with the normal retention time.
`Similar problems have been reported” when purifying guanylate—rich oligonucleotides
`on Partisil
`IO—SAX. Even when chromatographing small amounts, other authors”
`have reported unusually low recoveries {40%) from a PE! column in the case of
`oligonucleotidcs containing three or more consecutive deoxyguanosine monomer
`units.
`
`the deprotected hexadecamcr
`Chromatography of small quantities of
`d(T4G4T4G4) on a DEAR—cellulose column yielded no clear peaks in the region
`where octamcrs and longer-chain oligonucleotides are clutcd. The chromatographic
`purification of larger amounts of a hexadecamer was performed as follows. A 700—
`mg amount of the first hexadccamer, which was synthesized using only one nucleo-
`base protecting group, was dcprotected and the d(T¢G4T4(}4) obtained was frac-
`tionated on QAE—Scphadex with an increasing gradient of sodium chloride at 25°C,
`according to the conditions given in Tables 1, III (experiment 3). The elution profile,
`shown in Fig. la, exhibited two main peaks: 9.3% of the applied A260 units were
`contained in peak I and amounted to 40 mg d(T4G4) after isolation, peak 11 contained
`32% of the applied A260 units and 170 mg ofa mixture ofd(T4G4) and d(T4G4T4G4)
`(see Table 111). Furthermore, the elution profile indicates two different shorter—chain
`oligonucleotides eluted previous to the octamer. These oligomeric units might be the
`result of chain degradation, occurring during cleavage of the protecting groups. For
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 6
`Page 6
`
`

`

`PU R lFlCATlUN 0F ULIGODEOXYRIBONUCLEOTIDES
`
`l'r'l
`
`the isolation of d(T4G¢T4G4), the mixture corresponding to peak II was rechro—
`matographed on a DEAE—cellulose column, see Table III (experiment 3a). with an
`increasing sodium chloride gradient, containing 7 M urea, at 50°C. The elution profile
`(Fig. 1b) again showed two main peaks, with numerous smaller side peaks forming
`a high baseline. It is remarkable that oligonucleotides were still eluted from the
`DEAF-cellulose column at 50°C with l M sodium chloride, although the same mix—
`ture was eluted from the more basic anion exchanger QAE—Sephadex during the
`separation with 0.42 0.50 M sodium chloride at 25°C (see Fig. la). From the fractions
`corresponding to peak 1, containing 42% of the applied A260 units, 50 mg d(T4Gx)
`were isolated. The work-up of peak ll, containing 21% of the A260 units applied,
`resulted in 30 mg d(T4G4T4(34), which is only 4.7% of the fully protected hexade-
`camer.
`
`The greatest portion of the deprotected guanylate-rieh oligonuclcotides was
`lost during the preparative fractionation on the ion exchangers QAE-Sephadex and
`DEAF-cellulose. As clearly demonstrated in Fig. 1a, quite a large part of d(T4G4)
`is associated with d(T4G4T4G4). Therefore both oligonucleotides are eluted together
`within peak ll, although the octamer differs significantly from the hexadccamer in
`its negative charge. By rcchromatography (experiment 3a, Fig. lb), using 7 M urea
`and a temperature of 50°C, however, d(T4G4) and d(T4G4T4Gx] were separated.
`Both 7 M urea and the increased temperature during the elution counteracted the
`formation of aggregates. The high baseline in the elution profile (Fig. lb) also indi-
`cated that the mixture corresponding to peak [I in Fig. la. besides both main prod—
`ucts, contained additional oligonucleotides of various chain lengths, which are as-
`sociated with the main products. On the other hand, it cannot be excluded that octa-
`and hexadeeanucleotides were also eluted in the background. Despite the drastic
`elution conditions, part of the applied mixtures was retarded to such an extent that
`it was not eluted without the use of l M sodium chloride. According to our experi-
`ence, mixtures of corresponding guanylate-poor oligonuclcotides do not exhibit such
`difficulties. For example, d(G4T4) could be separated from d(T4G4T4) (see exper-
`iment l in Tables I, Ill), although these oligonucleotides differ less in their negative
`charges in comparison to d(T4G4) and d(T4G.tT4G4],
`The purification of a second hexadecamer, which was synthesized using an—
`other strategy, resulted in comparable results. After deproteeting 350 mg of fully
`protected hexadecamer, the solution containing d(T4G4T4G4) was directly fraction-
`ated on a DEAE-cellulose column at 50"C with an increasing sodium chloride gra—
`dient, containing 7 M urea without any previous separation (see experiment 4, Tables
`I, III). The elution profile (Fig. 1c} generally corresponds to that in Fig. lb, except
`that the bulk of the shorter—chain oligonucleotides was eluted prior to peak I,
`d(T4G4). Although the first hexadecamer, in contrast to the second one, was syn-
`thesized using guanylate monomer units with doubly protected guanine residues,
`both solutions exhibited similar percentages of short-chain oligonucleotides, after the
`protecting group had been cleaved. Because most of these side products had been
`removed during the preseparation of the first hexadeeamer (experiment 3, Tables I,
`11]), they are lacking in the elution profile (experiment 3a, Fig. lb) upon rechroma-
`tography. The fractions corresponding to peak I (Fig. 1c), which contained 316%
`of the applied A260 units, amounted to 70 mg d(T4G4). The work-up of peak [1,
`corresponding to 27% of the A260 units, resulted in 55 mg d(T4G4T4(}4). On the
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 7
`Page 7
`
`

`

`I'll
`
`n. senorr, R. SEMMLER, ll. ECKSTEIN
`
`basis of the fully protected component, d(T4G4T4G¢) was obtained in 22% yield,
`whereas the yield of the hexadecarner synthesimd according to the other strategy was
`only 4.1%. It cannot be excluded that the low yields result chiefly from the chro—
`matography and not from insufficient protection of the bases during the synthesis.
`The non-specific and irreversible adsorption, which caused the heavy losses of
`the deprotected oligonucleotides on the ion exchangers, occurred also on Nucleosil
`C”, which is commonly used for reversed—phase HPLC of oligonucleotides. in re-
`versed—phase HPLC on Nucleosil 7 C13 recoveries > 90% could be achieved only
`when analytical amounts ( < 3 Am, units) were applied. On applying 30 A250 units,
`only 50% were eluted within the expected region. The other part appeared in sub-
`sequent peaks or even at the end of the gradient. Especially when fractionating
`d(G4T4G4), identical compounds had different retention times. Finally, we found
`that the totally deprotected oligonucleotides could be separated satisfactorily by ion-
`pair reversed-phase HPLC10_13, as described below.
`In order to remove any contamination, HPLC was performed at room tem-
`perature with a preparative Nucleosil 7 C13 column (250 mm x 8 mm l.f).] permit—
`ting up to 60 A250 units to be fractionated. Elution of the columns was achieved by
`a two-component system (see Table II). The elution was monitored at 260 nm and
`resulted in the elution profiles shown in Figs. 2 and 3. The oligonucleotides d((}4T4]
`and d(T4G4T4] were eluted under isocratic conditions (experiments I 3) using 7.5
`mM tetrabutylammonium hydrogensulphate (TBA) pH 7.0 as eluent A and 15 mM
`TBA pH 7.0 in 75% aqueous acclonitrilc as eluent B. The oligonucleotides d(T4G4).
`d(G4T4(]4,) and d(T4G4T4G4_) were fractionated with a linear increasing gradient
`(experiments 4 6}, the concentration of eluent D (5.0 mM TBA pH 6.8 in 70%
`aqueous acetonitrile) increasing from 40 to 80% within 48 min. Fluent C was 5.0
`mM TBA pH 6.8. The integration of the elution profiles (Figs. 2 and 3) showed that
`the oligonucleotides, except d(G4T4G4), were contaminated to an extent of less than
`10%, demonstrating that
`the previous column chromatographic separations on
`DEAF—cellulose or QAE-Sephadex led to oligonucleotides of sufficient purity. The
`fractions corresponding to the main peaks within the vertical dotted lines were
`pooled, desalted and lyophilized. As a test of purity, the oligonucleotides were re—
`chromatographed in amounts of lO—IS ltg at 50°C on an analytical Nucleosil "i C13
`column (250 mm x 4.6 mm l.D.), cg, experiment 2 (Table II) demonstrates that
`the purity of the isolated oligonucleotides exceed 98%.
`The guanylate-rich d(G4T4(34), however, could not be purified to an extent
`beyond 60%, by ion exchange or by ion-pair reversed-phase HPLC, as could be
`concluded from the integration of the elution profile (Fig. 3b). It is possible that the
`dodecamer was of much higher purity judging from the elution profile. In support
`ofthis is the elution ofidentieal guanylate-rieh oligonucleotides at different retentiOn
`times. Furthermore, the dodecarner could be used successfully for enzymatic ligation,
`as will be described elsewhere? Therefore contaminations up to the presumed
`amount can certainly be excluded.
`Both the purity and sequence of d(G4T4), d(T4G4) and d(T4G4T¢) were con—
`firmcd by sequencing the oligonucleotides, carried out according to the well known
`two-dimensional fingerprint method 14—13. Contrary to our expectation, the fingerprint
`method could not be used for the sequencing of d(G4T4G4) and d(T4G4T4G4). Be-
`cause of the strong adsorption of the guanylatc—rich oligonucleotides on the poly-
`
`CUREVAC EX2041
`CUREVAC EX2041
`Page 8
`Page 8
`
`

`

`PURIFICATION OF OLIGOD‘EOXY RIBON UCLEOTIDES
`
`173
`
`
`
`2.0
`
`1,0
`
`““250
`
`0
`
`5
`
`—'—:l
`1O
`
`-l—
`15
`
`0
`
`5
`
`TE m e in m i n
`
`Hg. 3. Ion—pair reversed-phase HPLC of d[G4T..) and (HT4641);) on a Nuclcosi] 7 Cu; column (250 mm
`X 8 mm |.I}.] at room temperalure under isocratic conditions (see Table II} with a flow—rate ol‘ 2 nth-"min.
`(a) d(G4'I'4) eluted with a mixture of 70% A and 30% B; d(T4GAT4) chromatographed with 50% A and
`50% B. A = 15 mM TBA, pH 7.0; B = 7.5 mM TBA. pH 7.0 in 75% aqueous acetonitrile.
`
`saccharide matrix, a significant separation ol‘ the partial hydrolysates of these oli—
`gonucleotides by means of two-dimensional chromatography failed, thus a finger-
`print could not he obtained. Therefore, the partial hydrolysates of the radioactively
`labelled dodecamer d([32P]G4T4G4) and hexadeeamer d([-‘2P]T4G4T4G4) were sep-
`arated only one-dimensionally on a polyaerylamide gel under denaturing condi—
`tions“) by means of electrophoresis. The separation of the twelve or sixteen spots
`confirmed that the oligonucleotides synthesized indeed correspond to dodecamers
`and hexadecamers, respectively.
`dCT4G¢T4G4) was also sequenced according to the method of Maxam and
`Gilbert20 (see Fig. 4). This method is based on a specific chemical modification of
`Cyt, Cyt + Thy, Ade + Gua and Gua in four parallel reactions. During the partial
`
`Ici
`
`'
`
`
`
`Time in min
`
`Fig. 3. Ion—pair reversed—phase HPLC of [a] d(T4G4). (b) d(G.iT.iG..) and (c) d('I'4G4T.¢U4) on a Nucleosil
`T Cm column (250 mm X 4.6 mm ID.) at 50°C with a gradient of 40 to 80% D over {I to 48 min (see
`Table II); fiow~rate:
`l ml_.-'min. C = 5 mM TBA, pH 6.8: D = 5 mM TBA in 70% aqueous acetonilrile.
`
`CUREVAC EX2041
`CUREVAC EX2041
`Page 9
`Page 9
`
`

`

`W4
`
`[1. SCHOT’I', R. SEMMLER, H. ECKSTEIN
`
`hydrolysis only the modified nucleobases are supposed to be eliminated and the po-
`lynucleotide chain should be cleaved at the point where these nucleobases are missing.
`The partial hydrolysate obtained is separated into fragments ofdiil'ercnl chain lengths
`by gel electrophoresis, resulting in the autoradiogram of Fig. 4. The sequence of the
`hexadccamer from the 5’- to the 3'-terminal is obtained by following the most black-
`ened bands in the four lanes from the top (hexadecamer) to the bottom (monomer
`unit). The interpretation oi“ the autoradiogram is given in the right part of Fig. 4.
`The degradation pattern confirms the sequence of d(T4G4T4G4). Possible failure
`sequences of synthetic oligonucleotides cannot be detected conclusively and can
`therefore not be excluded. For example, the “C lane", to which the “Cyt degrada-
`tion” was added for control purposes, contains strong bands in its upper part, which
`might be correlated with C eontaminations. However, this is to be excluded in this
`case, because only T- and G-monomer units have been employed in the synthesis.
`The presence of other nucleobases was independently excluded by totally de-
`grading the hexadecamer chemically. Using formic acid. the oligonueleotide was de—
`graded, according to well known methodszmz, to its nucleobases. The total hydrol-
`
`C CIT #66
`
`C
`
`CIT MG
`
`6
`
`—
`.......
`
`--
`
`$749474 ‘31.
`03
`
`G
`
`“—
`
`We 547:.
`
`m'pTz. 9:.
`
`‘—“*
`
`all5‘
`
`G3
`(3CDN
`
`llll
`
`U"!
`
`a E
`
`g
`35
`E
`LLI
`._J
`LLI
`
`...._
`
`
`
`t
`
`o
`
`the Ituclcotide‘spcetific degraded
`Fig. 4. Left part: autoradiogram after gel electrophoresis of
`d[[32P]T.G4T4G4} using the Maxam and Gilbert method. C, CIT. A_.-"G. G denote C-specific.
`(T + T
`cleavage, A + G cleavage and G—specifie cleavage of the oligonucleotide. The chemically degraded oli«
`gonuclcotide is fractionated on a "20% polyaerylamidc gel“ [0.025 cm X 20 cm X 40 cm) with 50 mM
`Tris-borate 1 mM EDTA butler. Electrophoresis proceeded at 2.5 kWh mA for 2 h. Right part: inter—
`pretation of the sequence patterns. *1) denotes {52 P].
`
`CUREVAC EX2041
`CUREVAC EX2041
`Page 10
`Page 10
`
`

`

`PURIFICATION OF OLIGODEOXY R] BON UCLEOTIDES
`
`]',r'5
`
`ysate was fractionated on a Nucleosil 7 C13 column by means of reversed-phase
`HPLC under isocratic conditions. Only two peaks were obtained (see Fig. 5b). As is
`seen from the clution profile (Fig. 5a) obtained by chromatography of the four nu~
`cleobases under identical conditions, the retention times (9.76 and 11.69 min) match
`those of Gua (9.77 min) and Thy (l 1.67r min), respectively. Having found only the
`two nucleobases expected in the total hydrolysate, the presence of other nucleobases
`within the hexadecamer synthesized can be excluded. From the integration of the
`peak areas of Fig. 5b and in view of the molar absorption coefficients at 260 nm
`(Thy, 96001 mol‘l cm'l; Gua, 13 700 I mol‘l cm”), a molar ratio of Thy: Gua
`of 1.02:] was calculated. The nucleobase composition of d(T4G4T4G4) determined
`was very close to that expected (1.00:1).
`
`
`
` 0
` Ade29,16
`
`10-20-30
`
`1.0
`Time in min
`
`Fig. 5. Reversed-phase IIPLC on a Nucleosil 7 C13 column (250 mm x 4.6 min 1.0.) at room temperature.
`Elucnt: 5t} rnM ammonium acetate, pH 6.8; flow—rate, l mlj'min. Applied probe: (a) the test mixture of the
`four nucleobases Cyt, Gua, Thy and Ade; (b) about 0.1 .426" units of the totally hydrolysed d[T4G4T4G4].
`
`CONCLUSIONS
`
`The preparative chromatography of guanylate-rich oligonuclcotides, employ-
`ing different separation materials (DEAE-eellulose, QAE-Sephadex, Partisil lO-SAX
`and Nucleosil C18), can be performed only with considerable loss of oligonuclcotides.
`Therefore, in the oligonucleotide synthesis there is only a limited possibility of sep—
`arating impurities using chromatography. Guanylate—rich oligonucleotidcs of dif-
`ferent chain lengths associate with each other, thus causing identical compounds to
`be contained within different peaks and be eluted from the column at different times.
`At the same time, part of the product remains irreversibly adsorbed on the ion-
`exchanger matrix. The formation of aggregates between both the oligonucleotides
`andfor their derivatives and between the oligonucleotidcs and the polymer matrix is
`the reason why the desired oligonucleotide cannot be obtained when small quantities
`of condensation product are worked up by column chromatography.
`Remarkably, the dodecamer d((34T4G4), which could be purified only partially
`and characterized not unequivocally, however, resulted in the 36mer and other po—
`
`CUREVAC EX2041
`CUREVAC EX2041
`Page 11
`Page 11
`
`

`

`i'tfi
`
`H. SCHOTT, R. SEMMLER. H. ECKSTEIN
`
`lynucleotidcs upon enzymatic ligation with the dodecamer d(T4G4T4)7. This result
`demonstrates that the oligonucleotides can be used in enzymatic reactions directly
`after their synthesis, without tedious final purification. Also that the common en-
`zymatic reactions employing oligonueleotides do not require an high standard of
`purity, because the enzymes are able to select the “fitting compound" from the mul-
`titude offered.
`
`The increasing demand for oligonueleotides necessitates their preparation in
`large amounts. Therefore, preparative synthesis on the largest scale possible is an
`urgent objective. In our opinion there is no need for considerable improvements in
`the strategy of the oligonueleotide synthesis, but there is a great demand for more
`efficient separation methods for purifying the oligonucleotides after the deprotection
`without major losses.
`
`ACKNOWLEDGEMENT
`
`This work was suported by the Deutsche Forschungsgemeinsehaft.
`
`REFERENCES
`
`wa—tOmJ—‘nww—
`
`.l. Lipps and P. Erhardt, FEBS Lem, 126 ([980 219.
`11.
`t.-".S.A.. 79 (1982) 2495.
`H. .l. Lipps. W. Gruissem and D. M. Prescott, Pror. Natl. Aired. Stsi.
`R. F. Boswell, L. A. Klobutcller and D. M. Prescott. Free. Nat). Amd. Sci. L-".S.A., 79 (1982) 3255.
`D. Dawson and G. Herrick. Coil, 36 {1984) 1T1.
`D. Dawson and G. Herrick, Moi. Ce”. Biol. 4 (1984] 2661.
`E. l-Ielflenbein, Nae-fete Acids Res, 13 (1985] 415.
`ll. Schott. R. Semmler. K. Closs and H. Eckstcin. Makmmot‘. Chem. in press.
`M. I). Edge, A. R. (.lreene, G. R. Heathclifie. PA. Meacock, W. Sehuch. D. B. Scanlon,T.C. Atkinson.
`(LR. Newton and A. F. Markham, Nature (London), 292 “98” 756.
`9 T. G. Lawson, F. E. chnier and 11. L. Weith, Anni. Biochem., 133 (1983) 85.
`ll] 1. M. Johansson, K.-G. Wahlund and G. Schill. J. C‘hramatagr._ 149(19781281.
`11 J. H. Knox and J. Jurand. J. Citrottmtogr.. 149 (1978) 297.
`12 A. Tilly Melin. M. Ljungcrantz and G.Schill, J. Chromatogr, 185 (1979': 225.
`13 M.Kwiaikowski. A. Sandstrom. N. Balgobin and .l. Chattopadhyaya. Nucleic Acids Res. Syn-3p. See.
`No. 14 (I934).
`14 E. Jay, R. Bambara, R.Padmanabhan and R. Wu, Nair-[etc Aerials Rain,
`15 C.D.Tu, E. Jay. (T.P.Bt|ltl and R.Wu, Anal. Biochem, 3‘4 [1926) i3.
`16 R. Frank and H. Bliicker. in H.G. Gassen and A. Lang (Editors), C{remind} and En:_|.‘mrttic Synthesis
`ofGettt-t Fragments, Verlag Chemie, Weinheim, 1982, p. 225.
`17 H. Schott and H. Schrade, J.C'itromtttogr., 284 (1984] 331.
`18 H. Schott, H. Schrade and H.Watzlawick, J. Chrotttatogr., 285 (1984) 343.
`19 T. Maniatis, A. Jeffrey and H. van dc Sande, Biochemistry. 14 (1975) 328?.
`20 A. M. Maxam and W. Gilbert. Methods Enzymai, 65 (1980) 499.
`21 C. Y. Ko, J. L. Johnson, L. B. Harnell. H. M. MeNair and J. R. Vercelotti, Anni. Bfm'ht’flt. 80(1977)
`183.
`
`1 (19M) 331.
`
`22 H. J. Fritz, 1). Rick and W. Werr, in H. G. Gassen and A. Lang (Editors). Chemical and Enzymatic
`Synthesis of Gene Fragtttents, Verlag Chemie, Weinheim, 1982, p. [99.
`
`CU REVAC EX2041
`CUREVAC EX2041
`Page 12
`Page 12
`
`