Note: Descriptions are shown in the official language in which they were submitted.
t ~ j O
O 96/01268 . ~ '( /a
PURIFICATION OF OLIGODEOXYNIJCLEOTIDE PHOSPHOROTHIOATES
USING DEAE 5PW ANION ION-EXCHANGE CHROMATOGRAPHY AND
HYDROPHOBIC INTERACTION CHROMATOGRAPHY
~ACKGROUNn OF TH~ IIWEI~TION
5 Field of the lnvenrinn
This invention relates to the field of purification of oligode~ yL~_lc~Lides, and in
particular, the purification of oligodco~.ucle~,Lidc phoD~hJluLlliodlf s.
Dec~r~ption of the Prior Art
The use of modified phosphate backbone oligodeoxynucleotides as antisense
10 oli~v...,clcvlides in the field of selective gene regulation for i' , purposes has
received increasing attention over the }ast several years. There are numerous types of
modified phosphate linkages, e.g., u~lllyl~ ' , phoD~,llo~ulllioate,
laulidulc, that have been i..cu.~,ol~ d into antisense oli~ f s and studied.
E.g., Erickson and Izant (Eds.), Gene Regulation: Biology of Antisense RNA and DNA
15 (Raven Press, New York, 1992). Oligodeo~y,il,uuu,,l_otidc pho~hololllioat~, for
example, have been found to inhibit ~ y virus (Agrawal et al., Proc. Narl.
Acad. Sci. Usa 8!;, 7079 (1988); Agrawal et al., Proc. Natl. Acad. Sci. USA 86, 7790
(1989); Agrawal et al., in Advanced Dr~lg Deliver~ Reviews 6, 251 (R. Juliano, Ed.,
Elsevier, A ~hl.l, 1991); Agrawal et al. in Prospects for Amisense Nucleic Acid
20 Thef-apy of Cancer and AlDS, 143 (E. Wickstrom, Ed., WileylLiss, New York, 1991); and
Zamecnik and Agrawal in Annual Review of AIDS Research, 301 (Koff et al., Eds.,
Dekker, New York, 1991)), and influenza virus (Leiter et al., Proc. Natl. Acad. Sci. USA
87, 3420-3434 (1990)) in tissue culture. In addition, oli6~,dco~y.il,u -leotidei ' , ' uihioates have been the focus of a wide variety of basic research (e.g., Agrawal
25 et al., Proc. Natl. Acad. Sci. USA 87, 1401 (1990) and Eckstein and Gish, Trends
Biochem. Sci. 14, 97 (1989)), enzyme inhibition studies (Mujumdar et al., Biocf~emistry
28, 1340 (1989)), regulation of oncogene expression (Reed et a]., Cancer Res. 50, fi565
(1990~) and IL-1 e~ Di~u (Manson et al., L~.. r.'~!f~;re Res. 9, 35 (1990)) in tissue
culture.
Automated syllth~ have proven an invaluable tool for obtaining
oliyo. ~If c~ - c Oligo ~ 1f ol;Af~s are produced stepwise, with the addition of one
monomer at a time to the nascent olign ~ IP~.l;'l' chain. 2-3f~ of the reactions fail during
each cycle in which a nucleotide monomer is to be added however. COnDC~l ly, the
wos6/0126s 2 1 9 ~ 3 ~ . /a
resulting products are generally a ' u~, . - mixture of oli~n~ oti~ of varying
length. For example, in a typical 20mer synthesis, the 20mer product represents only 50-
60% of the recovered oligJ '~ product.
~ i' c, I~l.tJ4l.~iiU.. of oligvdcvAy -lecti~lPc on a solid phase support
5 requires that the oligvdcoAy ' '- be cleaved from the support. Cleavage of the oligo
from the support is typically accomplished by treating the solid phase with conc.,LIllaled
nliu~ll hydroxide. The a~.lll.o hydroxide is cv..~. 71y removed under
reduced pressure using, for example, a rotary .,~4lJu.4tOl . This method for removing the
ammonium hydroxide, however, is not ideal for use in large scale isolation of
10 olig;)dcvAy '- " '
For most purposes ~e.g., th~ r~pell~ir or diagnostic) the purity of the ~ ,v~n~ is
extremely important. Consc~ ly, there has been an interest in developing
4~.'dc i ' . for purifying olig.. koti~1~ s Because of their ~
potential, much of the focus has been on purifying olig~ P pL(,D~,l.b.u~hioates.
I5 Conventional methods for purifying oligodeu~y ~ Ps employ reverse-phaseliquid chromatography. Such methods require explosion-proof el . because
acetonitrile is typically used in the elution buffer.
Methods of oligodeoxynucleotide phJD~holvLLiv~lle ~ r~ n have been
published. Agrawal et al., J. Chroma~ograpky 509, 396 (I990), reported the analysis of
20 o1ic. lrv~ JLuD,uhulvLllio~t~D using high-p~,.ru.lll~-l.c liquid cL.~ a~hy with a
reverse-phase column. In that study, Agrawal et al. converted the olig '-~ '
phoD~,h~,.ulL.o4t~ to its phoi, ' ~ ~41L and then carried vut HPLC analysis.
Using this method they were able to analyze nlignmlrlPoti~P !)hOD~hJ~ containing
10 or fewer ~-alid.,s on a strong anion-exchange column (P~IL;"~}.~.~; SAX column).
25 01il,.,.- 1~ol;-1~r' ,' uLhiu&t~Dhavingmorethanlo ' :' couldnotbeanalyzed,
however, because of the strong interaction with the SAX medium.
Metelev and Agrawal, Anal. Biochem. 200, 342 (1992), reported the ion-exchange
HPLC analysis of oligodeoxy-l jl .. .~ l, v~ r' . ' utLio.lt.,D on a weak anion-exchange
column (Pa. ,' ~ WAX) in which the weak anion exchanger ntilizes a
30 di-~.~,.Lyl4~lhlv~u~yl functional group bonded to P4-liD~,hele silica. This medium, with
an ion-exchange capacity of 0.l8 meqlg, exhibits an h..~,l4~1ioll with anions weaker than
those observed with strong anion-exchange media. The authors of this study found that
~WO96/01268 21 ~ 50 r~ "
s~lJaldlioll was length dependent for oligu,. koti.~F phoD~JhuluLllhJdt~a up to 25
:id in length. ru.Lll~.l..u.~, n-l peaks were well separated from the parent peak.
They also found that 30-mer and 35-mer ~ r~ ulhiOat~,O were separable
with the same gradient, although better separation could be obtained with a shallower
5 gradient.
Metelev et al., Ann. N.Y. Acad. Sci. 660, 321-323 (1992), reported the analysis of
oli~,~,-il,, '- :' andchimericoligoribo-oli6ud~ y~ ro~ ,usingion-exchange
HPLC. They found that the retention time of the oli~ s studied depended on the
number of . il" '~otide moieties in the oligonucleotide . In addition, the retention time
10 of oligo.il o-lu-leotides was found to be length ~F ' . The authors noted that
oliG~,Hl,ul-u~lc~,~idcs of length up to 25 r ~rl~oti~ C could be purified and analyzed.
Bigelow et al., J Chromarography 533, 131 (1900), reported the use of ion-pair
HPLC to analyze nliG~ f oLide r~ . ' uLIliuat a. Stec. et al., J. Chromarography 326,
263 (1985), and Agrawal and 7.: ~ ' Nuclefc Acids Res. 19, 5419 (1990), reported15 HPLC analysis of oligodeo~ylil,o-lu"h,uLidcs ~ ;,.e one or two phoD~ o,uLllioàt~
of i~ linkages using a reversed-phase column.
These methods of olig. ' : h phoa~slJluLllioDle p. ~ use HPLC. While
this technique is useful for small scale operations, it is unsuitable for large, cuu....~.~idl
scale use. COua~ , improved methods of oli~u~u~lcoLidc IJ~-ir~dtion suitable for
20 use in large scale oliE,~ preparations is desirable.
~UMMARY OF T~F. ~NV~,NTION
The present invention provides improved methods for purifying
oli odc~,Ay - ' . :- ' pl.oD~hu. vi' . In particular, the invention p}ovides p. ., i r;. -f ion
suitable for large scale separation of oligonucleotide phoaf,holuLl.ioates~ The
25 ~n i r;, -~ ;nn methods of the invention do not require the use of reduced pressure to remove
a.. .oniuul hydroxide or the use of conventional C-18 silica gel reverse-phase liquid
clllullldtc~ld~S~y. The inventive methods replace these procedures with Lydll,' ,hic
d~Lion cL., -~,ld~Ly or DEAE-5PW anion ion-exchange chromatography.
In one aspect of the present invention, the oligodeoAy . ~ D are purified using
30 hydluphOhiC t~ r~ tinn Chrulllàt(~ hy~ .UI hydroxide is used to cleave
oligonuc-leotides from the solid support on which they were ayllLL~Di~ Typically, roto-
SlleSTlTUTE SHEET (RULE 26)
W0~6/01268 2 ~ q ~ 3 5 ~ P~I/L~
(.v~l~u~aliOII under reduced pressure has been used to remove most of the _hydroxide. This is then typically followed with reversed phase .,LI~ ~ a~hy to
separate the DMT-on oliboll__lc.JLidcs from the DMT-off olig.~ I~o~ c~ These
n ~ c, however, are unsuitable for large scale use. In this aspect of the invention,
5 hydlu~ obic h.t.la.~ ch.. O ph~ is used in place of roto-cvalJulaLio.l alone or both
roto-evaporation and reversed phase cl~ ty. HIC is preferable to roto
evaporation because it simplifies and acccll.ldt~,s ~1,.. ll.;ll.. hydroxide removal and can
be used for large scale purification. If it is to be followed by RPLC, HIC increases the
purity of the olig~l.,_clcvlide relative to roto-e~a~JolaLion, resulting in less potential for
10 fouling the RPLC column and reducing the purification challenge presented to the RPLC
column.
When used in place of roto-cval,or.ltion and RPLC, HIC provides the benefit of
accomplishing two tasks (removal of A~ l hydroxide and separation of DMT-onoligos from DMT-off oligos) at once. S bstihltinn of RPLC with HIC also reduces the
15 resin cost, eliminates the need for organic solvents (which require more stringent
handling, including special disposal, explosion-proof euvi., t, and ~vap(JIalivec~ ), provides for more rapid elimination of co..~ a~ (e.g., unreacted
monomers and failure sequences), and increases ~hll ,h~ This increase in ~hlUI ~h~ '
is made possible by use of short columns and high linear velocities.
HIC also reduces both the expense (in terms of column packing and equipment) andpotential problems that can arise with HPLC, e.g., dirri~ul,;~ in pack.ing and mRirn.ining
HPLC columns. Suitable HIC colurrms that can be used in the present invention include,
but are not limited to, phenyl-sepharose fast flow (high ~ ' ) and TSK-gel phenyl-
5PW.
JlL~ iLIo]y~ although HIC and DEAE-5PW cLIu~ LGolaplly are m~c~4nic~ir.4.11y
quite different, they can be used h~ cllànO~,ably to serve the same purpose. They both can
be used to purify DMT-off oligo- ~rl~oti~l~c~ although, as dCI~lOll~Llatud below, DEAE-
5PW results in better yields when purifying 25mers. Olig. '-~tidc mixtures having
purities of about 98~o can regularly be obtained using these LL I ~ ,
The use of DEAE-5PW column to purify DMT-off ol gc ~'-: ' , like HIC
columns, does not require HPLC and, therefore, offers the same advantages âS described
SUBSTiTUTE SHEET (RULE 26)
WO96/01268 2 1 '~ 4 3 f) O r~.,.,... /a
'~
on a relatively short column, which increases lhl., h, . and eases packing,
and requires simple step gradients for elution, which simplifies t~ c~hu~ ;5
and the chance for error.
In yet another aspect of the present invention, ion-e~change is ~rc~lmpli~h~d with
~ 5 a DEAE-5PW column. When oli ull.. ~ Wes intended for i' ~ ~ _ - use, it is essential
that all allll~lOIliulll cations be replaced with, for example sodium cations. This can be
accomplished with a Dowex cation ion exchange column followed by desalting with
sephadex gel filtration. In this aspect of the invention, standard ion-exchange methods are
replaced by the DEAE-SPW column. The resin is relatively cheap compared to the more
10 recently introduced anion ion-exchange styrene divinylbenzene polymer supports (e.g.,
PerSeptive Biosytems, Polymer Labs), yet is sturdy enough (in terms of particle size and
resistance to currcntly used cleaning l"u~,c.ll".,s) for production use.
When the oligodeu~y.,u.,l~utides a}e purified using a DEAE-5PW resin, the eluategenerally has a very high salt ~ullu~l.L.~tion, rendering typical sephadex gel filtration
15 desalting inpractical or h."l,ci~iblc. [n place of sephadex gel filtration, other salt removal
techniques, e.g., RPLC and tangential flow filtration (TFF), should be used. In the
preferred ernho~l , the DEAE-5PW oli~,u..u~lcotidc mixture is desalted using
tangential flow filtration.
The foregoing merely ~u~ a~ certain aspects of the present invention and is not
20 intended, nor should it be construed to limit the invention in any way.
All patents and publication cited in this ~l,ccir~ ,n are hereby iu~o.~,ul~.tud by
reference in their entirety.
DFTAlT F.l) DE.CCRlPTION OF l'T~F INVFNTION
The present invention provides improved methods for purifying
25 oligodeoAy '- ' phc,~l,l.ulùLllioates Inparticular,theinventionprovides~,u,iri,,~Lju,
~r~ 1 r5 suitable for large scale separation of oli~u..ucl~otide r~ ' -,I.,tl.ioat,,s. The
purification methods of the invention do not require the use of reduced pressure to remove
Ulil,~lU hydroxide or the use of cu..~ .iul.~l C-18 silica gel reverse-phase liquid
cLl~",latot.~,l,y. The inventive methods replace these plucclulcs with h.ydlu~ ic
30 interaction chromatography or DEAE-5PW anion ion-exchange chl~,,,latocl~t.y.
SUBSTITUTE ~HEET ~RULE 26)
W0 96~01268 P~ l.,9!' /a
2~ 943~
Following solid phase synthesis, oligon ~l~oti~iPs are cleaved from the solid support
by incubating the support in: hydroxide. Not only are the desired oligonucleo-
tides cleaved from the support, but so too are failure seq~n~ 5, i.e., olj~ cleolidc
sequences being fewer than the desired number of ~' ,lid~,s in length. Such failure
5 sequences arise from less than complete coupling of ~ .. h oC;~I~ s to the growing
oligo. r~Ul;~ir chain and less than complete capping of unreacted functional sites. The
desired oligonucleotide must be separated from failure sequences if it is to be used
effectively for IL.la~ tic or other purposes.
Cu~ lh~ lly, the bulk of the ~ .. hydroxide is driven off by roto-
10 e~ olatiull This is then followed by separating the desired DMT-on oligos (i.e.,
oligo l~leo~i~lf s having a 5' dimethoxytrityl protecting group) from undesired DMT-off
oligos (i.e., olig. -'~fi~C not having a 5' protecting group) by reversed phase liquid
~Llulllaluola~Jlly~ The I~IT-on oligos, being more Lydlu,ullOlJ;c, bind to the reverse phase
column more tightly than DMT-off oligos.
If the desired oligonllrlpoti~ll is to be used for Ll ~ purposes, any ,.. -
cations complexed with the oliO~ 5~ ~Lide must be exchanged with, for example, sodium
cations. This is generally accomplished by ion exchange chromatography followed by
desalting using sephadex gel.
The present invention comprises improved methods for separating or purifying
20 oligomlrleoti~ pho~l,ho.u~hioates. As used herein, the terms "separatingN and"purifyingR are intended to be used h~ lldl~o-ably and mean a process by which
oligu _t _" 1~ having a particular molecular structure are physically segregated from
olig~ . c have a different molecular structure and by which .u.. ~.. ;..... hydroxide
and/orothersaltsareremovedfromtheoliL...~ 1~oli~l~mixture. Inparticular,thepresent
Z5 invention comprises the use of hydlu~ Ob;c h.t.,l~-,,i Chl~ o a~lly (HIC) and DEAE-
5PW ion exchange l,Llulllatc~Ola~/lly to separate and purify oligonucleotide
phJsl~huluiLioatu~
In a first: bodill,.,.,l of the present invention, the desired oliOu ~lrul;l1r is
separated from excess amml hydroxide and prepared for treatment by reverse phase30 liquid cL.. -- -:u~ by subjecting the niig.,.. l~u~ containing ,.. ~ .... hydroxide
solution to h~.llu~llobic interaction chromatography. Preferably, the EIIC comprises
SURSTITUTE SHEET (RUL~ 26)
W0 96/01268 J ~
. ~ 2~9'~5~
passing the oli~,u l-u~ containing ~u~uu~ ' hydroxide through a phenyl s~La.usc
fast flow chr~,matogr~phy resin or a phemyl-5PW chromatography resin.
In a second e.l.bodi~ ..l of the present invention, the desired oli~c ~ is
separated from excess a,ul.mlliu,.l hydroxide and DMT-off oligos by ~uh;cclil~,5 the
~ 5 olig. -'- ~ c~ aining . hydroxide solution to hydluullol;ic i,..u.~uliuu
clllulllAlo~;ldph~. Preferably the HIC comprises passing the oligonculcotide-contAining
a""..o"iu"- hydroxide through a phenyl-sepharose fast flow chromatography resin or a
phenyl-SPW chromatography resin.
In a third ~l.lbodh.lcl.; of the present invention, ion exchange (wherein a~ uolliu-l-
10 ions are eYrhangPd for sodium ions) and ~ iri~alioll are arComrlichpd with a DEAE-5PW
column. In this (..llbC ' it is found that salt . Alions of the eluant are generally
too high for desalting with sephadex-gel, as is traditionally done after standard
~~tl i~gl.... I< ~llid~ ion-exchange is con~l ~t~ ~1 In this c...l,c " t, desalting is accomplished
by any suitable technique capable of effectively handling high salt collc~llll~liu,.~.
15 Desalting is preferably conducted by tangential flow filtration (TFF). TFF provides the
advantage of being able to a~co~ t~. and desalt large volumes of solutions having very
high salt cùuc. ullAAliul~ (e.g., 2 M NaCI).
The invention also encoll",dsscs a process for preparing purified mixtures of
olig~ l.~oti~Ps~ This aspect of the invention comprises:
(a) HIC to remove - .l.. :.. hydroxide and to remove oligos not having DMT
protecting groups;
(b~ detritylation to remove the DMT protecting groups;
(c) anion exchange on DEAE-5PW;
(d) tangential flow filtration; and
(e) Iyophilization.
The methods of the invention may be used to purify crude mixtures COlllailli
olig.,.- Iro~i~lr pho~,hu,ulhio~;~i, that have just been cleaved from a solid phase synlhesis
support using Al~l~l..lll;..lll hydroxide or mixtures that have been previously subJected to a
l,ulifiuaLioll procedure but that are not of acceptable purity.
We have found that oli~ull.,~lcolidc pho~Jllolulllioates can be purified on DEAE-
5PW ion-exchange columns, phenyl-5PW and phenyl s ~ u~e columns using aqueous
SUBSTITUTE SHEET (RULE 26)
WOg6f01268 21 ~435~ r~
SPW results in effective separation of oli~u.. I uli<l< phor~ ulutl~iu~t~.D fromrl~n~ in high yields. The protocols presented herein can be employed in large
scale purification of oligo~ lcvLide ph~l.lloLutl.ioates to replace the COll~
IUIU~,~a~JVldliUII and C~8 silica gel protocols required during the p,J. iri ~ n process. This
S results in fewer steps in the II~AI' r~- u h~e process and allows for better purity and
recovery of product.
According to the methods oF the present invention, oligonucleotide phoDI,hulu-
thioates from about 10 to about 35, 1l ,"i.1. c in length can be separated on a DEAE-SPW
ion exchange column, a phenyl sepharose column or phenyl-5PW column. In a preferred
10 .,lllb- ~ ~, ol ;~ .. . rlr o~ pLu~f~hu~ùlllioates having a length of from about 20 to about
3'i and more preferably 25-30. can be separated using the present methods.
According to the invention, the olig~,. Ir "J _ separable by the present method
may have as few as one ~nd as many as all ph- 1 ~ ollliùat~ eoli~lP linkages. As
used herein, the term "oligonucleotide pho~ uluLllioate~ is used to describe such an
15 oligonucleotide. Oligonucleotide pho~haln~l;(l.i(. ~, of the same size as theoligu ' :' phosphorothioates described can also be separated by the inventive
methods .
The oli~o., r 1~ c are placed on the column and eluted at or near ambient (room)t~ f/~,~alul~i with either a gradient or a non-gradient buffer. For purificationlpreparation
of ammoniacal solutions by HIC, A.l.l.l.~l.;.. acetate is used in - . IihrAlion buffers at
cv~lc~ t~a~ions ranging from about 0.5 to about 2.0 M, preferably 0.75 M. Although
addition of A~...._ ~I~;IIn~ acetate may reduce the pH somewhat, the pH should typically be
high to minimize the loss of the trityl group. pH values of 7.5 - ll.0 have been used
s~ ~ceccfi-lly. A typical value of 10.0 is preferred since this more alkaline pH minimizes
25 loss of the trityl group from the olitJI~ Ir vl ;~1 . Elution is most effectively ~ Ch~d
with water, although other schemes employing buffers may prove useful in particular
Aprlir~firn~ Organic solvents are not required for separation, although their inclusion
may be desirable when preparing oligon~r!P<!til1f~s by HIC for ~b~ ..1 RPLC
separation. While taller columns can be used (see the height diameter ratios in the
30 Examples, infra), shorter columns work very well and are preferred for large-scale work.
Colurnn loads can go as high as 650 OD unitslml packing, depending on elution conditions
SUBSTITUTE SHEET (RULE 26)
W0 96~01268 1~ l /a
~ ) 5 0
and column geometry. Preferably the column load is about 200 OD units/ml packing.
Linear velocities can range as high as 300 cm/hr, but are preferably about 150 cm/hr.
When HIC is used to purify DMT-off oliL.... lrori~ , a variety of buffers can beused at a pH ranging from about 7.2 to about 8.5. In the preferred c..lbr '- Tris-HCI
5 is used at a pH of about 7.5. Salts such as sodium chloride must be added to the load and
to the equilibration and wash buffers in concentrations rapging from about 1 M to about
3 M, preferably about 3 M. Elution is Prcn~plich~d by either linear or step gradients via
salt reduction. Phenyl sepharose columns are preferably loaded at high flow rates. Rates
exceeding about 250 cm/hr are found to work eYrel~ingly well. Elution at a rate of about
10 150-200 cmlhr is desirable.
When DEAE-SPW is used to purify DMT-off oli~;~ '- lc , buffers such as Tris-
HCI (preferably in a ~ - dLiu.l of about 10 to about 50 mM) are to be used. Salts such
as sodium chloride are used for equilibration and elution. Preferably, sodium chloride
gradients of 0 to 2 M are used. For shorter columns a range of sodium chloride
15 collcl .1 between 0.85 and 2 M are preferred. The pH of the oligi ' ' solution
that is loadel may be between about 7.0 and about 10.5~ preferably about 7.2. The pH of
eq~ilihr~tion, wash and elution buffers can range from about 7.2 to about 8.5, and is
preferably about 7.2. Chelating agents such as 1 mM EDTA and organic solvents such as
acetonitrile or ethanol can be added to the buffers in some applications. While linear
20 gradients can be run, excellent reco~eries and purities are obtained using simple step
gradients. In many cases, better recoveries can be obtained using step gradients -- linear
gradients appear to cause a slow "bleed-off" of bound product from the column that often
works against effective sepa~ation and high recovery. While taller columns can be used,
relatively short columns provide excellent results and are preferable for large-scale work.
25 Column capacity for olig ~'~ ' is somewhat lower that that seen for the HIC resins,
with optimum column loads ranging up to about 200 OD units/ml resin, preferably about
- 150 OD units/ml packing. Optimum linear velocities range between about 50 and about
150 cm/hr and are preferably about 70 cm/hr.
Column geometry has a cigniFi~ rt effect on the l,, tuil.,.l..,ut~ for equilibration
3û buffer salt condiliuns -- taller columns (larger height:diameter ratio) require higher salt
as compared to shorter columns (lower height:diameter ratio). When a mixture of
SUSSTITUTE S! IEET ~RULE 26)
W096/01268 .~,IIU~
2' ~'~3~0
the rel~r~ hir between the column height:diameter ratio and sodium chloride
c~ . ~ . . ,.1;.... is relevant to product yield. As the I - i~,' di~ ~t~,l ratio of the DEAE-5PW
columm is increased, the sodium chloridc ~ ~ion of the eq~ilihr:ltinn buffer should
preferably be increased to obtain a greater recovery of purified product.
The separation method of the present invention is essentially i, l ~ . a of column
particle size. Sizes ranging from 25 to 90 ~m can be used s~l~~cescfnlly. In preferred
emhorii~nt~, the particle size is 25 ~m - 40 llm or 45 llm - 165 ~m. A particularly
preferred hylllvpllobh, interaction chromatography resin has a particle size of about 90
llm.
One skilled in the art will recognize that ~ ~lifir~inns may be made in the present
invention without deviating from the spirit or scope of the invention. The invention is
illustrated further by the following examples which are not to be construed as limiting the
invention or scope of the specific p-v.c.lu-~,s described herein.
EXAMPLES
In each of the following Examples, equipment and materials were obtained as
indicated from Tosollaas (Mv..l~ ville, PA). Amicon (Beverly, MA) Waters
(Milford, MA), Hewlett-Packard (HP) (Palo Alto, CA), Perkin-Elmer (Norwalk, CT~,ISCO (I.incoln NB), Rainin (Woburn, MA), Filtron (Nollihl,(,.~ ll, MA), Pharmacia
(Piscll~dw~y, NJ). and Biorad (Hercules, CA).
1~. '~ 1
Purification witi2 DEAE-SPW
A 25-mer oligonucleotide having the sequence
CTCTCGCACCCATCTCTCTCCTTCT ~GEM 91) was purified using a 0.51 column of
30 ~um TSK DEAE SPW (~vll....~l.,i~lly available from TosoHaas). ï'he resin was pacl~ed
25 into a Pharmacia 5.0 cm diameter glass colurnn. Purification was monitored using an
ISCO UA-5 detector equipped with a 254 nm filter and a Rainin rabbit pump.
Two sources of GEM 91 were used in this e~ample:
(1) GEM 91 fractions from a Sephadex G-15 column, subsr ~ '~/ desalted and
~,oncc~ .tcd on Filtron tangential flow filtration (TFF) ~ al~s prior to
cl.. , 'b~'llJ~I~Y;
SUBSTITUTE Sl EET (RULE 26)
WO 96/01268 ~ I /a
21 ~35~
11
(2) two lots of Iyophilized GEM 91 powder: one having low ion exchange-HPLC
purity, the other having low CGE (capillary gel elc~ u~,ho.~i.is) purity.
The GEM 91 olib~ was loaded onto the column aL 150 A76~ O~D~ (optica]
density) units/ml packing~
25 mM Tris-HCI, pH 7.2 room t~ cldlult having from 0.85 to 1.0 M NaCI was
used as the buffer with, in some t~ liUI.,I..~ mM EDTA~ Elution was p~lrullllcd with
25 mM Tris-HCI, pH 7~2 (RT) containing 2 M NaCI and, in certain ~ .Hlll~,l,L~, 1 mM
EDTA~ The column was washed with 6 column volumes (4 column volumes in run no~ 4)
of the appropriate equilibration buffer after application of sample~ Fractions were
10 collected, O~D~ recoveries rlPtl-rmil~rrl and aliquots analyzed for purity by ion exchange
HPLC (IEX) and, in some cases, by CGE analysis~
ln addition, to runs done on the 0~5 I column, three e~ ,.hl.t.ll~ were done at a 9
ml column-scale in order to evaluate the effect of column collri~sulalioll on p.~l rol.~ ce
Results are shown in Table I Percent purity was analyzed by ion exchange (IEX)
HPLC (% purity = % of phc,~l o-ull.ioate, including N and N-1 sequences); CGE data
is also presented for run #5 (% purity = % of N oligonucleotide, including both
pho~pl.o.ù~l~io~ and ph~ ~p~o~ ster oligos)~
Three ~ i" ~ - ~ were conducted on a 9 ml (2.2 cm diam~ x 2.4 cm ht ) column
The results are shown in Table 2~
The results rh.n~m 11~ that crude olig~ -I-.t ' can be purified with DEAE
5PW with high purity and excellent yield~
SUBSTITUTE SHEET (RULE 26
WOg6~01268 2 ~ r:~ 12
Table 1
Pllrifi~fi pool
Run % recovery% purity% recovery
No. Conditions O.D. I product~
1 ~411: 0.85M NaCI + EDTA 54.1 99.6 59.7
15~: 90% IEX purity3
2 ~lU: 1.0M NaCI/+ EDTA 85.8 98.6 94.0
l~ai: 89~o IEX purity
3 _41U: l.OM NaCI 86.0 99.2 96.2
lQla 89 % IEX purity
4 _41U: l.OM NaCI 58.7 98.0 64.7
~lai: 89% IEX purity
~4lU: l.OM NaCl 87.0 97.8 84.94
Llal: 89% IEX purity, 77% CE 84.7
purity
on retests
6 ~4lU: 1.0M NaCI 78.3 98.2 90.1
15laLl: 877~o IEX purity on retest
7 e~uil: 1.0M NaCI 78.7 98.6 89.3
1~1: 87% IEX purity on retest
8 _41U: 1.0M NaCI 83.5 98.4 93.4
ll~i: 88~o IEX purity on retest
' percent recovery based on total O.D. (optional density) units initially loaded on column.
2 percent recovery of product in terms of "purity units" initially loaded on column.
3 IEX purity units = (total O.D. units) x ~percent purity by IEX-HPLC analysis); CEi
15 purity units = (total O.D. units) x (percent purity by CE analysis). CE purity units
indicate the recovery efflciency of the column.
4 based on CE data.
SUBSTITUTE SHEET (RULE 2
WO9~i/01268 ~ ~ 9 '11 3 ~ ~ r ~ /a
Table 2
Purified Pool
percent
[NaCI] irl recovery Y0 recovery % purity (IEX- % recovery of
equil. buffer O.D. ~total~ O.D. HPLC) product
90.85 M 109.0 87.0 97.0 94.0
100.85 M 101.0 80.0 97.0 86.0
111.0 M 113.0 67.0 99.0 73.0
r ~-2
Purificafion by HydropholJic Inreraction Chromatography
A GEM 91 sample was subjected to hydrophobic interaction ~hl. .~ phy. This
sample had the composition shown in Table 3.
Table 3
percent percent PO oli~ulluclcuLides by CGE analysis
DMT-on n n- ] n-2 n-3
62 0.4 55.2 11.4 2.8 3.5
This olign~ Poti~P was purified on Phenyl-sepharose fast flow (high sllhslitl-tinn~ using
an Amicon 2.2 cm glass column, a Waters 650 solvent delivery system, an HP integrator,
15 a Rainin Dynama~ UV-C absulballue detector and a Perkin-Elmer s~ lu~hui : .
Theresinwaspackedintooneoftwocolumncù--riL,.--,'iu~~c ahPigh~ pterratio
of 2.2:1 or 2.5:1.
The following elution protocols were employed:
(A) the sample load was adjusted to 1.7 M .IUIIIIU ' acetate, pH 10.7; the
20 column was equilibrated and washed with 2.5 M ~l...,.ù~ ... acetate; sample was eluted
using 0.1 M acetate, pH 8.5, followed by water;
(B~ the sample load was adjusted to 1.0 M ammonium acetate, pH 10.7; the
column was equilibrated and washed with 1.0 M i~ acetate, pH 8.5; the sample
was eluted with water;
SUBSTITUTE SHEEI ~RULE 26)
WO96101268 21 ~35~
14
(C) the sample load was adjusted to 1.0 M .,, ~'~'!';""' aceLate, pH 10.7; the
column was equilibrated and washed with 1.0 M: acetate, pH 8.S; the sample
was eluted using 0.01 M NaOH, pH 11.5, followed by water;
(D) the sample load was adjusted to 0.75 M .tllllllOlliUIII acetate pH, 11.0: the
S column was e.~ aL~d and washed with 0.75 M: acetate, pH 8.5; the sample was eluted using 0.01 N NaOH, pH 11.5, followed by water.
Ali4uots of wash and elution fractions were analyzed by ;on exchange and reversephase (RP) HPLC. The results are shown in Table 4.
Table 4
10 Sample Height: Load' Velocity Elution Percent Percent recovery Elution
diameter (cm/hr)2 protocol recovery OD3 product4 purity5
ratio Elution Total Elution Total
2.2:1 252 316 A 53.3 99.2 64.0 86.4 97.8
2 2.2:1 252 316 B 66.1 101.7 96.0 102.0 80.5
3 2.2:1 252316/158 C 50.9 98.8 80.8 93.8 88.0
4 2.5:1 2003161158 C 55.8 100.8 80.5 85.7 89.8
2.5:1 200316/158 C 62.7 99.2 84.8 91.1 83.8
6 2.5:1 200316/158 C 63.9 100.0 79.7 88.2 77.3
7 2.5:1 200316/158 D 59.1 98.3 82.3 86.3 86.3
8 2.5:1 1783161158 D 66.1 95.3 83.9 85.9 87.7
9 2.5:1 1783161158 D 57.9 89.8 76.6 80.1 91.1
20 ' Total OD units loaded/ml packing.
2 Samples 3-9 were loaded at 316 cm/hr and eluted at 158 cm/hour.
3 % Recovery OD = % recovery of oligon-lcl~oo~ s loaded on column
Elution = % recovery in elution pool
Total = % recovery in all f}actions
25 4 ~a Recovery product = % recovery of "DMT-on' product loaded on column
Elution = % recovery "DMT-on" product in elution pool
Total = % recovery ~DMT-on" product in ull fractions
SUBSTITUTE SHEET (RUL~ 26)
WO96/01268 2 l ~ ~ ~ 5 0 r~L~ 11 .
5 Indicated in terms of percentage of DMT contAining material (5fo DMT-on); calculated
as (% DMT-on by ion exchange HPLC + % DMT-on by reverse phase HPLC)/2.
An i~ ,onis Al solution of GEM 91 having an avarage of 62% DMT-on was
purified on phenyl sepharose fast flow (high ~llbs~hl~tion) using a Waters 650 solvent
delivery system, a HP integrator, a Rainin Dynamax UV-C absoll,al.~c detector, 1 cm
(Biorad) and 2.2 cm (Amicon) diameter glass columns and a Perkin-Elmer
spectrophotometer
The deblock solutions were adjusted to 1.7 M~ ~m. acetate pH 10 3 before
loading onto phenyl sepharose The phenyl sepharose was previously equilibrated with
0 _UllllVlliUIll acetate. Elution of the ol4,u..uclcvtide was accomplished with reverse
gradients of Allm~ll;lllll acetate, pH 7.35, followed by water The following runce~itinnq were used:
(A~ the column was equilibrated and washed with 2.5 M al~lnlui~u acetate, pH
8.5; oligonucleotide eluted with step gradient of8% ACN in 0.1 M Al l l~.ni
acetate pH7.85, followed by water;
(B) the column was equilibrated and washed in 1 7 M A~ niul~l acetate, pH8.5;
oligonucleotide was eluted with consecutive linear gradients of1.5 M-1.0 M and
1 0 M-0 M al..l..vniu u acetate, followed by water;
(C) the column was equilibrated and washed with 7 M A~ acetate, p~ 8.5;
olig,~ otiAf was eluted with step gradient of 0.5 M al~, .nl,; ", acetate, followed
by linear gradient of 0.5-0 M Ammrlnillm acetate, followed by water
(D) the column was eqnilihratP(l and washed with 2.6 M a""~n; " acetate, pH
8.5; oligonucleotide was eluted with step gradient of 0.5 M a,.. oniu.. acetate,
followed by a linear gradient of 0 5-0 M ..iu... acetate, followed by water
Aliquots of wash and elution fractions were analyzed by ion exchange and reversephase HPLC. The results are shown in Table 5.
SUBS~ITUTE SHEET (R~!LE 26)
~ 9435~ 16 r~
Table 5
height: Load' linear elution recovery of OD recoYery of elution
diameter velocity protocol (~c) pr~ purity2
ratio (cm/hr) (%)
Elution Total Elution Total
S.4:1 192 462 A 46.9 106.4 75.4 78.1 9g.7
2.2:1 385 315 B 42.4 97.7 66.7 83.6 97.8
2.2:1 385 315 C 18.9 101.5 30.2 93.1 99.0
2.2:1 252 315 D 32.3 96.6 52.5 85.3 97.5
' OD units/ml packing.
10 ~ percent DMT-on: (% DMT-on by ion exchange by ion exchange HPLC + % DMT-on
by reverse phase HPLC)/2.
These results d~ ..D~ DMT-on oligon~cll otifl~os can be purified from
ammoniacal solutions and DMT-off oli~ k~ os.
Example 3
HIC as a Su'ostitutefor Rotoe~!aporation and RPLC
Two GEM 91 samples were subjected to L,ydluyLOb;~, interaction ~,Llullld~o~l~yh ~ .
These samples had the ~v~yv~;lioll shown in Table 6. Sample 3-1 had been stored at 4~C
as an ammoniacal solution; sample 3-2 had been recently prepared.
Table 6
CGE
Sample %DMT- %PO
on n n- 1 n-2 n-3
63 0.4 56.5 10.3 2.~ 2.9
2 69 0.4 57.6 6.5 2.0 3.7
These oligom~rl~ oii~es were purified on Phenyl sepL.~.u3c fast flow (high
~ ) using Amicon 2.2 cm glass columns, a Waters 650 solvent delivery systcm,
25 an HP integrator, a Rainin Dynamax UV-C abbo~ cc detector, and a Perkin-Elmer
., ' tVIII~t~..
SUBSTITUTE SHEET IRULE ~6
wos6/~l26s ~ 1 9 -1 3 5 0 P~
17
,~mmo~:or:ll solutions of GEM 91 were adjusted to 0.75 M a acetate and
loaded at 75 cmlhr onto the column which was previously equilibrated with 0.75 Mall~ OIliull~ acetate, pH 10.2. After loading, the column was washed at 317 cm/hr with
(0.75 M) a7nn oni~ acetate and DMT-on product was eluted by washing the column at
5 159 cmlhr with water.
Aliquots of wash and elution fractions were analyzed using ion exchange and
reverse phase HPLC. The results are shown in Table 7. The purity and yield of the eluted
product is equivalent to that achieved using reverse phase liquid ~hlul~ 6~h2hy. This
dcl..u.~ , that by proper ~dj~lc~m~nt of load and elution conditions, high purity DMT-
10 on product can be obtained with excellent yield on a relatively short (low height:diameter
ratio) chl~ u~ hy columns with HIC.
Table 7
Crude GEM 91 % recovery of OD% recovery of purity of
product elution
oligo Loadl Elution Total Elution Total % D
sample
200 55.8 102.1 85.0 85.3 97.5
300 48.5 104.1 73.5 88.8 98.5
200 56.6 96.1 83.5 88.9 96.0
2 200 64.0 1~0.4 89.0 94.3 96.0
20 I total OD units loadedlml packing.
2 Purity calculated as: (% DMT-on by IEX-HPLC + % DMT-on by RP-HPLC)12.
Example 4
Olig~ o~ Purification by ~IC
GEM 91 was purified on TSK-gel phenyl-SPW (TosoHaas) and an agarose based
25 phenyl sepharose fast flow (high ,~I,s~ ioll) (Pharmacia) using a Biorad glass column
(1.5 cm diam.), an ISCO UA-5 detector equipped with a 254 nm filter, a Rainin rabbit
pump, and a Perkin-Elmer ~ tlu~ uLulll,,ter.
The GEM 91 (sodium salt form) was recovered from desalt column (gel filtration)
side fractions from production lots via tangential flow filtration (TFF) on 2,000 MWCO
30 modified polyether sulfone filters (Filtron. Inc.l. The oli~odeoxvnucleotide s~ tionc
SUBSTITUTE SHEET (RULE 26~
W096/01268 ~ F~ a
J _J
18
were adjusted to 3M NaCI in 25 mM Tris-HCI, pH 7.4-7.5, prior to oppliro~inn to the
columns.
E~,.. I Conditions:
A. Phenyl 5PW; oli~ ~ loaded in 3M NaCI/25 mh~ Tris-HCI pH 7.5,
5 oligo eluted using linear gradient of 3M - OM NaCI.
B. Phenyl sepharose; same oli~ l Ioading and elution conditions as in
A.
C. Phenyl sepharose; olig. ' :ide loaded in 3M NaC1125 mM Tris-HCI, pH
7.4; oligolluclc~lidc eluted using step gradient of 2M and lM NaCI followed by linear
10 gradient of lM - OM NaCI.
D. Phenyl sepharose; oli~5~,. -' '- loaded in 3M NaCl/25 mM Tris-HCI pH
7.4; olig ~ eluted using step gradient of lM NaCI foliowed by linear gradient of
lM - 0.7M NaCI, followed by step gradients of 0.7M and OM NaCI;
The results for each column are shown in Table 8. These results dc.l.ol.~ e that15 the sodium salt form of an olig~ ti~iP can be purified using either phenyl sepharose
fast fiow resin or phenyl 5PW resin.
Table 8
OD units Prodl~r~ Rrr.-very. IF.X (%)1 Product Rec.overy. C&E (%)2
elution recovery: elution elution recovery: elution elution recovery:
recovery combincd recovery purity combined recovery purity combined
(%)fractions (%) (%)fractions (%) (%)fractions
(%) (%) (%)
A 53.694.5 59.2 95.991.7 nd nd nd
B 74.8100.8 77.0 95.695.8 77.3 91.8 96.3
C 62.5g3.0 65.2 94.990.1 63.9 91.7 96.7
D 66.0113.9 68.3 96.1112.2 67.1 92.0 110.1
' [(number of CGE units in fraction)/~number of CGE units in load)] x 100.
2 [(number of IEX units in fraction)/(number of IEX units in load)] x 100.
SUBSTITUTE SHEET (RULE 26)
_ .
WO 96ro1268 r~ /a
2~ ~~35()
19
Example 5
An a~l solution of GEM 91 having an avarage of 62% DMT-on was
purified on phenyl sepharose fast flow (high ~Uh~ Ulivll) using a Waters 650 solvent
delivery system, a HP integrator, 1 cm (Biorad) and 2.2 cm (Amicon) diameter glass
S columns and a Perkin-Elmer s,u~ u,uh~
Phenyl sepllalua,, was packed into one of three column cl-nfig-lr~rions at
height:diameter ratios of: 5.4:1, 2.2:1, and l:1.
Phenyl sepharose columns were eq~ in .I. ~ acetate pH 8.5. The run
cr~hi~ln~ were as follows:
(A) no adj.. ~h~ to the GEM 91 _ I solution; the column was equilibrated
with 2.5 M ,.,... ~ ... acetate, pH 8.5; oligu ' .i ~- was eluted with step
gradient of 8% acetonitrile (ACN) in water pll 7.85, followed by water; or
(B) the GEM 91 .. ~llul~ia~,al solution was adjusted to 1.7 M a.. w.. :~.l.. acetate, pH
7.'a'S before loading; the column was equilibrated with l.S M ,~ .. nlli...,. acetate,
IS pH 7.85; oligon~lrleotPIe was eluted with step gradient of 8% ACN in water,
followed by water.
Aliquots of wash and elution fractions were analyzed by ion exchange and reversephase HPLC. The results are shown in Table 9. These results demonstrate that phenyl
sepharose fast flow resin can be used to prepare llmmnni~ l solutions of crude DMT-on
20 olig~ '-ulidcs for sl-hseqllent RPLC.
SUBSTITUTE SHEET (RULE 26)
wo 9~012 ~82 1 9 ~r ~ S 3 1~ ~ la
Table 9
height: Load'linear elution recovery of ODrecovery of e]ution
diametervelocity protoco1 (%) p-rQsl~(7o) purity'
ratio (cmlhr) Elution TotalElution Total
5.4:1 192 462 A 92.1 103.594.4 94.4 68-
5.~:1 450 462 A 80.8 104.497.8 97.8 75
5.4:1 642 462 A 54.8 100.188.5 91.9 80
5.4:1 1285 462 A 24.5 105.5 nd3 nd nd
l:l 400 396 A 62.7 101.181.4 91.~f 81
1:1 305 396 A 53.5 98.3 86.3 89.2 72
1:1 305 IS~ A 72.4 100.087.6 92.4 75
1:1 250 157 A 80.4 102.090.8 91.0 70
2.2:1 355 396 A 42.1 103.555.1 97.2 81
2.2:1 400 317 B 85.9 105.393.5 93.5 68
15 ' Load: OD unitslml packing
2 Elution purity: as fl. l,-....;. ~d by IEX-HPLC
F. '-6
GEM 91 is syufll~ D;L. .i using a CPG support and an automated synthesizer. The
olic. '- ' iscleaved'rromthesupportusing c~' ~..f ~ff-d A~"~r'-' hydroxide. The
20 oli~ hydroxidemixtureisthen-.l..l c aphedusingllydlupLobic
interaction chromatography using phenyl-sepharose or phenyl-5PW resins essentially
according to the ~lucedulr s set forth above in Examples 1-6. The resulting solutions
containing GE~M 91 are pooled and optionally .,1,-~ ' using preparative reverse-
phase liquid chlu laLucla~rlly. The combined GEM 91 fractions are acidified to remove
25 the dimethoxytrityl (DMT~ protecting group. The dcLLiLylut~ d GEM 91 is suspended in
water and ~hll cl ,'-~ over a DEAE-SPW ion-exchange column essentially as
described above in Example 1 to purify it and convert it to the sodium salt form of GEM
91. The olic. ' : ~fP is then "desalted" via tangential flow filtration (TFF) to remove
salt and any remaining small molecule impurities, and then dr~.yluO_~laLr d on a membrane
SUBSTITUTE SHEET (hLlLE 26)
. .. . . . .. . _ _ _ . _ _ _ _ _
W0~1268 2 1 ;' ~ 3 5 a ~ /a
referred to as purified BDS. The overall recovery of product after these steps is about
70% with a purity of about 98%. Thus, using this technique, HIC and DEAE 5PW
chrom~L~ pl.i~s can be effectively combined to obtain good recovery of high purity
oligo ~ oti~1~ s from ~ul....Juiacdl solutions.
~ 5 From the foregoing, it will be .~ .cci.lt.d that although specific e hoJ;~ of
the invention have been described herein for purposes of il~ ~Lh~ll, various
may be made without deviating from the spirit or scope of the invention.
SUESTITUTE SHEET ~RULE 26~
W09~0~2(i8 ~ 3 ~"~
22
SEQUENCX LISTING
(l) GENERAL INFORMAT}ON:
~i) APPLICANT: Puma Ph.D., Patricia
lii) TITLE OF INVENTION: Purific~tion of Oligodeoxynucleotide
Phosphorothioates Using DEAE 5P~ Anion Exchange Chromatography
and Hydrophobic Interaction Chromatography
~iii) NUMBER OF sE~u~c~b: 1
(iv) CORRXSPONDENCX ADDRXSS:
0 (A) ADDRXSSXE: Allegretti ~ Witco~f, Ltd.
(B) STRXET: 10 South W~cker Drive, Suite 3000
(C) CITY: Chicago
(DJ STATE: IL
(X) COUNTRY: USA
(F) ZIP: 60606
(~) COMPUTXR RXADABLE FORM:
(A) MXDIUM TYPE: Floppy disk
(B) COMPUTXR: IBM PC compatible
(C) OPERATI~G SYSTEM: PC-DOS~MS-DOS
(D) SOFTWARX: PatentIn Release #1.0, Version ~1.25
(vi) CURRENT APPLICATION DATA:
(A) APPLICATION NUM3ER:
(B) FILING DATX:
(C) CLASSIFICATION:
~iii; ATTORNEY/AGXNT INFORMATION:
(A1 NAME: Greenfield Ph.D., Michael 5.
(B) REGISTRATION NUMBER: 97,142
(C) RXFXRXNCXJDOC~ET NUMBER: 94,444
(ix) TELECOMMUNICATION INFORMATION:
(AJ TELEP~ONE: (312J715-1000
(B) TELEFAX: (312)715-1234
(2) INFORMATION FOR SEO ID NO:1:
(i; SEQUENCE CHARACTERISTICS:
(A) LENGTH: 25 base pair~
(B) TYPE: nucleic acid
SUBSTITUTE SHEET ~RULE 2&)
WO 96101268 r 1 ~,~ru~ -h la
2 1 9 4 3 ~ ~
23
(C) STR~Nn~n~'~qS: single
(D) TOPOLOGY: linear
(iii) HYPOTHETICAL: NO
(ix) FEATURE:
(A) NAME~EY: misc ~eature
(B) BOCATION: 1..25
(D) OTHER INFORMATION: /note= "All internucleotide
linkages are phosphorothioates~
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:1:
10 CTCTCGCACC CATCTCTCTC CTTCT 25
SUBSTITUTE SHEET (RULE Z6)
