Total stepwise solid-phase synthesis of oligonucleotide-(3'-->N)-peptide conjugates.

Article abstract of DOI:10.1021/ol026502u

[structure: see text] An efficient total stepwise solid-phase synthesis of oligonucleotide-(3'-->N)-peptide conjugates is described that makes use of either a controlled pore glass support or macroporous polystyrene beads. Extending our previous homoserin

Full text of DOI:10.1021/ol026502u

ORGANIC
LETTERS
2002
Vol. 4, No. 19
3259-3262
Total Stepwise Solid-Phase Synthesis of
Oligonucleotide-(3′fN)-Peptide
Conjugates
Dmitry A. Stetsenko, Andrey D. Malakhov, and Michael J. Gait*
Medical Research Council, Laboratory of Molecular Biology, Hills Road,
Cambridge CB2 2QH, UK
mgait@mrc-lmb.cam.ac.uk
Received July 11, 2002
ABSTRACT
An efficient total stepwise solid-phase synthesis of oligonucleotide-(3′fN)-peptide conjugates is described that makes use of either a controlled
pore glass support or macroporous polystyrene beads. Extending our previous homoserine linker approach, we prepared a range of conjugates
containing one of four different cell or nuclear penetration peptides together with oligonucleotides containing 2′-deoxynucleoside or 2′-O-
methylribonucleoside phosphodiesters, or gapmers containing 2′-deoxyphosphorothioates. The route also allows incorporation of a fluorescent
label within the conjugate for cell uptake studies.
Oligonucleotides and their analogues have been studied as
specific inhibitors of gene expression in cells for almost a
quarter of a century.1 However, their use as gene regulation
agents within cultured cells has been much hampered by poor
cellular uptake, which has led to investigations of many
different transfection agents.2 One promising approach is to
conjugate the oligonucleotide or analogue to a peptide that
possesses cell penetration properties to facilitate cellular
delivery and/or to alter internal localization.3 Among alterna-
tive methods for the preparation of peptide-oligonucleotide
conjugates, total stepwise solid-phase synthesis seems to be
the most direct and a number of routes have been suggested.4
Of particular significance have been methods based on the
use of branched linkers that allow successive peptide and
oligonucleotide assembly from the branch point.5 We de-
scribed recently a novel and facile method of synthesis of
oligonucleotide 3′-conjugates, including some short peptides,
using an ω-aminoalkyl succinate/L-homoserine combination
linker.6 Now we report the modification and extension of
this method toward the total stepwise solid-phase synthesis
of longer cell penetration peptides 3′fN conjugated to
antisense oligonucleotides and their 2′-O-methyl analogues
(see Scheme 1).
As antisense oligonucleotide models (Table 1) we have
used a 15-mer 2′-deoxyoligonucleotide and 12-mer and 16-
mer 2′-O-methyloligoribonucleotide (OMe) sequences. We
have shown recently that OMe oligonucleotides comple-
mentary to the apical stem-loop of the trans-activation
(1) Stein, C. A.; Krieg, A. M. Applied Antisense Oligonucleotide
Technology; Wiley-Liss: New York, 1998.
(2) Reviewed in: Hughes, M. D.; Hussain, M.; Nawaz, Q.; Sayyed, P.;
Akhtar, S. Drug DiscoVery Today 2001, 6, 303
(3) Reviewed in: Stetsenko, D. A.; Arzumanov, A. A.; Korshun, V. A.;
Gait, M. J. Mol. Biol (Russ.) 2000, 34, 852. Tung, C.-H.; Stein, S. Bioconj.
Chem. 2000, 11, 605.
(5) (a) Antopolsky, M.; Azhayev, A. HelV. Chim. Acta 1999, 82, 2130.
(b) Antopolsky, M.; Azhayev, A. Tetrahedron Lett. 2000, 41, 9113. (c)
Antopolsky, M.; Azhayev, A. Tetrahedron Lett. 2002, 43, 527.
(6) Stetsenko, D. A.; Gait, M. J. Bioconj. Chem. 2001, 12, 576.
(4) Reviewed in: Zubin, E. M.; Romanova, E. A.; Oretskaya, T. S. Russ.
Chem. ReV. 2002, 71, 239.
10.1021/ol026502u CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/28/2002

Scheme 1
responsive element TAR RNA inhibit in vitro transcription
membrane penetration activity: a 16-mer fragment of the
of HIV-1 by steric competition with HIV Tat-mediated viral
transcription to decrease the amount of full-length transcript.7
3′-Fluorescein-labeled oligonucleotide analogues delivered
by a cationic lipid were also tested in a HeLa cell reporter
assay, where cellular activity was found for chimeric
oligonucleotides containing both OMe and locked nucleic
acid (LNA) units.8 As further models (Table 1), we chose
HIV-1 gp41 fusion sequence (1),10 15-mer Kaposi fibroblast
Table 2. Sequences of Cell-Penetrating Peptidesa
sequence
no.
AVGIGALFLGFLGAAG
AVALLPAVLLALLAP
AGYLLGK(Ac)INLKALAALAKKIL
AKKKKLDK
1
2
3
4
Table 1. Sequences of Antisense Oligonucleotidesa
a All peptides are C-terminal 6-hydroxyhexylamides and contain L-
homoserine residue at the N-terminus, optionally N-acylated with 6-car-
boxyfluorescein.
sequence
no.
cucccaggcuca
I
CTCCCAGGCTCAAAT
GCTCCCAGGCTCAAA
AGCTCCCAGGCTCAA
ugugcTATTCTGTgaauu
uaagcTGTTCTATguguu
cucccaggcucagauc
II
III
IV
V
VI
VII
VIII
growth factor peptide (K-FGF) (2),11 and a 21-mer short
version of Transportan (a chimeric galanin/mastoparan
peptide) (3).12 An 8-mer peptide 4 that constitutes part of
the nucleoplasmin nuclear localization signal (NLS)13 was
also included. The first two peptides do not require amino
acid side-chain protection, By contrast, Transportan includes
asparagine and tyrosine residues as well as lysines, the
ꢀ-amino groups of which are normally protected as trifluo-
roacetamides.6 Since satisfactory base-labile protection for
the tyrosine phenol group compatible with Fmoc/tert-butyl
chemistry14 is lacking, we opted for the mild acid-labile
agcucccaggcucaga
a 2′-Deoxynucleotides are in capitals, 2′-deoxy phosphorothioates in bold
capitals, and 2′-O-methyls in lower case.
antisense and mismatch gapmers directed against survivin
mRNA9 consisting of a 2′-deoxy phosphorothioate octamer
sequence flanked 3′ and 5′ by OMe pentanucleotide wings.
For initial conjugate synthesis, we have chosen three
peptides (Table 2) that were reported to have cellular
(10) Soukchareun, S.; Tregear, G. W.; Haralambidis, J. Bioconj. Chem.
1995, 6, 43.
(11) Dokka, S.; Toledo-Velasquez, D.; Shi, X.; Wang, L.; Rojanasakul,
Y. Pharm. Res. 1997, 14, 1759.
(7) Arzumanov, A.; Walsh, A. P.; Liu, X.; Rajwanshi, V. K.; Wengel,
J.; Gait, M. J. Nucleosides, Nucleotides Nucl. Acids 2001, 20, 471.
(8) Arzumanov, A.; Walsh, A. P.; Rajwanshi, V. K.; Kumar, R.; Wengel,
J.; Gait, M. J. Biochemistry 2001, 40, 14645.
(9) Chen, J.; Wu, W.; Tahir, S. K.; Kroeger, P. E.; Rosenberg, S. H.;
Cowsert, L. M.; Bennett, F.; Krajewski, S.; Krajewski, M.; Welsh, K.; Reed,
J. C.; Ng, S.-C. Neoplasia 2000, 2, 235.
(12) Soomets, U.; Kilk, K.; Land, T.; Langel, U. In InnoVations and
PerspectiVes in Solid-Phase Synthesis and Combinatorial Libraries 2000;
Epton, R., Ed.; Mayflower Press: Kingswinford, UK, 2001.
(13) de la Torre, B.; Albericio, F.; Saison-Behmoaras, E.; Bachi, A.;
Eritja, R. Bioconj. Chem. 1999, 10, 1005.
(14) Chan, W. C.; White, P. D. Fmoc Solid-Phase Peptide Synthesis:
A Practical Approach; Oxford University Press: Oxford, UK, 2000.
3260
Org. Lett., Vol. 4, No. 19, 2002

2-chlorotrityl protecting group instead.15 The appropriate
monomer, Fmoc-Tyr(2-ClTrt)-OH, is available from Nova-
biochem. The asparagine residue was incorporated as its
amide-unprotected form via use of Fmoc-asparagine pen-
tafluorophenyl ester to avoid side reactions resulting from
carboxylic activation. The rest of the amino acids were
coupled via the HATU/DIEA/DMF in situ activation protocol
and Fmoc groups were removed throughout by use of 20%
piperidine/DMF.14 The NLS sequence contains aspartic acid
of which the â-carboxy group requires protection. Here, we
found the recently described base-labile Dmab ester16 to be
a good solution to this problem.
complished in a machine-assisted way by using the condi-
tions for normal Fmoc-amino acid coupling, as was dem-
onstrated by the synthesis of peptide 2 on PS200 support.
Finally, 2% CF3CO2H-CH2Cl2 (v/v) was used to remove
the trityl group.
Then, peptide-loaded resins were subjected to standard
oligonucleotide chain assembly on a ABI 394 DNA/RNA
Synthesizer with either 2′-deoxyribonucleoside or 2′-O-
methyl ribonucleoside phosphoramidites and manufacturer’s
protocols. Beaucage’s reagent (Glen Research) was used for
phosphorothioate synthesis. Conjugates were cleaved from
the resin and deprotected by concentrated aqueous ammonia
treatment at 55 °C overnight, conditions we have found to
be safe for a range of peptides.6,17 The resulting products
were isolated in good to moderate yield and analyzed by
reversed-phase HPLC, and their respective molecular masses
checked by MALDI-TOF mass spectrometry (Table 3).
For initial trials, manual peptide assemblies were carried
out on 500 Å long-chain alkylamine controlled pore glass
(LCAA-CPG) support functionalized by sarcosine followed
by an aminohexyl succinate linker as described previously.6
Peptide 1 was assembled by fragment coupling with tripep-
tides Fmoc-AAG-OH and Fmoc-FLG-OH, and dipeptides
Fmoc-IG-OH and Fmoc-VG-OH. For peptide 4, all monomer
couplings were used. For both peptides, coupling steps were
followed by acetic anhydride/N-methylimidazole/2,4,6-col-
lidine capping. Synthesis efficiency was monitored by
measurement of N-(9-fluorenylmethyl)piperidine absorbance
values14 and double couplings were used when yields
dropped below 95%. The average stepwise yield was >96%.
Then, we explored two alternative polymer supports that are
based on macroporous polystyrene beads: PS200 resin
(Amersham Biosciences), 30 µm particle size, and ArgoPore
low loading resin obtained from Aldrich, >100 µm bead size.
To our knowledge, the latter has not been tested previously
in oligonucleotide synthesis. Peptide sequence 2 was as-
sembled first on ArgoPore resin manually to check its
performance, using two dipeptide blocks Fmoc-AP-OH and
Fmoc-LP-OH, the rest being monomer couplings. Again,
Fmoc deprotection yields were carefully checked, and double
couplings were used when appropriate. PS200 resin was used
for Transportan (3) synthesis. The first two amino acids were
introduced manually to test the support under peptide
synthesis conditions. Then, the resin was subjected to
automated peptide assembly on a Pioneer Peptide Synthesizer
(Applied Biosystems), except for the asparagine residue
which was coupled manually. Double couplings with Fmoc-
amino acid monomers and manufacturer’s standard protocols
were used throughout the automated syntheses. Acetic
anhydride capping steps were omitted, except before and after
all manual coupling steps. In the case of peptide 3, after
assembly the 2-chlorotrityl group on the tyrosine residue was
removed by treatment with 2% CF3CO2H-CH2Cl2 (v/v) for
2 min, and the resin was capped, but using isobutyric
anhydride instead of acetic anhydride, to protect the phenolic
group of tyrosine with a base-labile isobutyryl group.
After all peptide assemblies, Fmoc deprotection was
followed by Fmoc-Hse(Trt)-OH manual coupling and an
optional fluorescein label could then be introduced by further
Fmoc deprotection and coupling with 6-carboxyfluorescein
diacetate. Homoserine incorporation also could be ac-
Table 3. Properties of Oligonucleotide-(3′fN)-peptide
Conjugates
MALDI-TOF, RP-HPLC retention purity,
c
no.a
I.1
FAMb
calcd/found
time, min
%
no
5591.5/5591.3
5961.8/5958.7
6741.3/6739.3
5117.5/5118.6
6548.8/6545.8
6201.9/6198.6
6981.3/6984.2
5715.6/5711.5
6073.6/6073.4
6226.9/6225.3
7006.3/7002.5
7006.3/7003.6
7944.9/7943.7
8724.4/8720.8
7944.9/7944.4
8724.4/8722.6
7303.0/7297.7
8082.5/8078.7
6996.9/6992.9
25.2
57.3
56.3
16.5
27.7
54.3
57.8
11.4
18.2
30.7
58.0
33.0
30.7
31.0
30.6
31.0
30.8
31.2
27.5
69.4
22.5
26.9
79.0
54.7
90.6
17.7
73.9
70.4
54.7
11.6
20.0
50.5
19.2
48.9
18.5
59.4
35.6
45.8
I.2
I.3
I.4
yes
yes
no
yes
no
no
no
yes
no
no
II.1
II.2
II.3
II.4
II.4
III.2
III.3
IV.3
V.2
no
yes
yes
yes
yes
yes
yes
yes
V.3
VI.2
VI.3
VII.2
VII.3
VIII.1
a Oligonucleotides are designated by Roman numbers, peptides by Arabic.
b 6-Carboxyfluorescein label. c Integrated from HPLC traces of crude
conjugates.
Examples of typical RP-HPLC traces are shown in Figure
1.
Comparison of the results of conjugate synthesis on the
different supports showed unequal performance in peptide
and oligonucleotide assemblies. The PS200 support (Amer-
sham Biosciences) gave the best overall purity/yield balance
in both oligonucleotide and peptide synthesis. CPG-500
(conjugates I.1, I.4, II.1, II.4, and VIII.1) showed compa-
rable or slightly better results in oligonucleotide synthesis,
(16) Chan, W. C.; Bycroft, B. W.; Evans, D. J.; White, P. D. J. Chem.
Soc., Chem. Commun. 1995, 2209.
(15) Athanassopoulos, P.; Barlos, K.; Gatos, D.; Hatzi, O.; Tsavara, C.
Tetrahedron Lett. 1995, 36, 5645.
(17) Stetsenko, D. A.; Williams, D.; Gait, M. J. Nucl. Acids Res. Suppl.
2001, 1, 153.
Org. Lett., Vol. 4, No. 19, 2002
3261

nucleotide conjugates based on our previously reported
homoserine linker. This approach is an effective and expedi-
tious way to obtain cell-penetrating peptide-oligonucleotide
conjugates necessary for antisense inhibition and cell delivery
studies, or for other applications such as improving hybrid-
ization properties of oligonucleotides in vitro.18 Several
amino acids that are common in such peptides and which
require side-chain protection were successfully incorporated
by using ammonia-labile protecting groups (lysine, aspartic
acid) or transient protection (tyrosine). The incorporation of
arginine remains problematic, but one possible solution
through use of protected ornithine has been published.5b,c
Cellular studies with these antisense peptide-2′-O-methyl-
oligonucleotide and gapmer conjugates are in progress and
will be reported elsewhere.
Acknowledgment. One of us (A.D.M.) thanks Astra-
Zeneca Ltd for financial support. The authors thank Per
Denker (Amersham Biosciences) for the generous gift of
PS200 support, and Donna Williams for advice and for
carrying out some oligonucleotide assemblies.
Figure 1. RP-HPLC traces of crude oligonucleotide-(3′fN)-
peptide conjugates: I.2 (a) and II.4 (b).
but was noticeably inferior in peptide synthesis. ArgoPore
resin (conjugate II.2) gave excellent product purity, but low
isolated yield (10% calculated from initial peptide loading),
which may indicate that it is more suitable for peptide
synthesis than for oligonucleotide synthesis.
In conclusion, we have presented a new approach toward
the total stepwise solid-phase synthesis of peptide-oligo-
Supporting Information Available: Synthesis details,
HPLC traces, and MALDI-TOF spectra for the conjugates.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL026502U
(18) Corey, D. J. Am. Chem. Soc. 1995, 117, 9373.
3262
Org. Lett., Vol. 4, No. 19, 2002