Kinetic model studies on the chemical ligation of oligonucleotides via hydrazone formation

Article abstract of DOI:10.1016/j.bmcl.2004.11.068

We report on the suitability of hydrazone formation for activator-free ligation of oligonucleotides. 5′-Acyl hydrazides were synthesized using a previously described phosphoramidite modifier, whereas 3′-hydrazides resulted from a hydrazinolysis of an ester group serving as a linker to the solid support. Aromatic aldehydes could be directly introduced on the 5′-terminus via the respective phosphoramidates. Aliphatic aldehydes were generated by periodate cleavage of the corresponding 3′- and 5′-modified diol precursors. Ligation of a 3′-hydrazide-modified oligonucleotide with oligonucleotides bearing an aromatic aldehyde in 5′-position showed a fast reaction kinetics (k₁ about 10 ^-1) and irreversible hydrazone formation. The ligation of a 5′-hydrazide-modified oligonucleotide and a 3′-ribobisaldehyde appeared to proceed reversibly at the beginning, but became irreversible with increasing reaction time. Hydrazide-modified oligonucleotides were found to be somewhat unstable in aqueous solutions.

Full text of DOI:10.1016/j.bmcl.2004.11.068

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1229–1233
Kinetic model studies on the chemical ligation
of oligonucleotides via hydrazone formation
K. Achilles and G. v. Kiedrowski*
Lehrstuhl fu¨r Organische Chemie I, Ruhr-Universita¨t-Bochum, D-44780 Bochum, Germany
Received 19 October 2004; revised 23 November 2004; accepted 25 November 2004
Available online 19 January 2005
Abstract—We report on the suitability of hydrazone formation for activator-free ligation of oligonucleotides. 50-Acyl hydrazides were
synthesized using a previously described phosphoramidite modifier, whereas 30-hydrazides resulted froma hydrazinolysis of an ester
group serving as a linker to the solid support. Aromatic aldehydes could be directly introduced on the 50-terminus via the respective
phosphoramidates. Aliphatic aldehydes were generated by periodate cleavage of the corresponding 30- and 50-modified diol
precursors. Ligation of a 30-hydrazide-modified oligonucleotide with oligonucleotides bearing an aromatic aldehyde in 50-
position showed
a ) and irreversible hydrazone formation. The ligation of a
fast reaction kinetics (k1 about 10À1
50-hydrazide-modified oligonucleotide and a 30-ribobisaldehyde appeared to proceed reversibly at the beginning, but became irrevers-
ible with increasing reaction time. Hydrazide-modified oligonucleotides were found to be somewhat unstable in aqueous solutions.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
aldehyde-substituted nucleoside under reductive amina-
tion conditions.6
In the context of our search for oligonucleotide based
chemical self-replicating systems we were interested in
finding ligation methods that are not based on hydrolyz-
able condensing agents. Former approaches, which
made use of the formation of phosphodiesters,
pyrophosphates, and phosphoramidates to ligate
oligonucleotides always necessitated the addition of
activating reagents such as water soluble carbodiimide
or cyanogen bromide.1 Recent publications reported
on a number of alternative chemical ligation reactions
proceeding without activating reagents using so-called
Ôpre-activatedÕ building blocks. Examples include the
reaction of 30-phosphorothioate- or 30-phosphorosele-
noate components with 50-iodine-substituted com-
pounds,2 the ligation of 30-phosphorothioates with
50-mercapto oligonucleotides under formation of
monophosphoryldisulfides,3 the formation of metallosa-
len-DNA-complexes,4 the photochemical ligation of
oligonucleotides using 5-vinyldesoxyuridine5 or the
template-directed oligomerization of a 50-amino-30-
However, the pre-activated building blocks described so
far are either unstable, difficult to synthesize, or lack
broad applicability. A very promising ligation reaction
in this context is the formation of hydrazones using
aldehyde-modified and hydrazine or hydrazide-modified
oligonucleotides. This chemistry is claimed to be stable,
irreversible, easy to use, highly selective, and not suscep-
tible to non-specific binding.7 It is widely used for the
immobilization of oligonucleotides on surfaces or for
the conjugation of oligonucleotides with, for example,
proteins or small organic compounds.8 During the
course of studies on chemical copying of trisoligonucleo-
tides the formation of hydrazones proved to be particu-
larly suitable.9 Herein we report on the evaluation of
suitable hydrazide- and aldehyde-modified oligonucleo-
tides and on kinetic studies of their ligation reaction.
2. Results and discussion
It is highly desirable to introduce any new functionality
at the 30- and 50-ends of oligonucleotides as modular
building blocks compatible with automated oligonucleo-
tide synthesis. This also holds for aldehyde and hydr-
azide modifiers needed for hydrazone-based ligation.
For both, the 30-modification of oligonucleotides with
Keywords: Kinetic; Hydrazone; Hydrazide; Aldehyde; Modified
oligonucleotide.
*
Corresponding author. Tel.: +49 234 32 23218; fax: +49 234 32
14355; e-mail: kiedro@rub.de
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.11.068

1230
K. Achilles, G. v. Kiedrowski / Bioorg. Med. Chem. Lett. 15 (2005) 1229–1233
a hydrazide group and with an aldehyde function, the
modification of the solid phase used for automated oli-
gonucleotide synthesis seemed most suitable. Solid
phase 2 was obtained over several steps: Starting from
glycolic acid the hydroxyl group was protected with 4-
methoxytrityl chloride (MMT-Cl)/pyridine and the carb-
oxylic group was subsequently reacted with aminoprop-
yl-controlled-pore-glass (CPG) using [2-(1H-benzo-
triazol-1-yl)-1,2,3,3-tetramethyluroniumhexafluoropho-
sphate] (HBTU) as coupling reagent. Cleavage of the
MMT-group with trifluoroacetic acid and repeated cou-
pling of 1 gave 2 (Scheme 1). Automated oligonucleotide
synthesis was carried out and deprotection of the oligo-
nucleotide (1. NHEt2, dichloromethane, 18 h, rt; 2. 24%
NH2NH2, 18 h, 4 ꢁC),8a removal of excess hydrazine via
SepPakꢂ-cartridges and HPLC-purification resulted in
50-(4,40-dimethoxytrityl) (DMT)-protected-30-hydrazide-
modified oligonucleotides. 30-Diol-modified oligonucleo-
tides were employed as precursors for 30-aldehydes. So-
lid phases 6 and 9 were used for the synthesis of 30-diol-
modified oligonucleotides. Support 6 was synthesized
starting frompropane-1,2,3-triol ( Scheme 1). Since the
synthesis previously described did not result in good
yields in our hands,10 we made use of an alternative syn-
thesis strategy. In the first step, one of the primary
hydroxyl groups was protected with DMT-Cl/pyridine.
In the second step, the remaining primary hydroxyl
group was esterified with benzoyl chloride, and last,
the secondary hydroxyl group was acylated with succi-
nic acid anhydride. The fully protected compound was
subsequently attached to aminopropyl-CPG using
HBTU as coupling reagent. Starting compound for solid
phase 9 was 5-hexen-1-ol, which was protected with
DMT-Cl/pyridine (Scheme 1). Following oxidation with
osmium tetroxide/N-methylmorpholine-N-oxide to give
the vicinal diol, the introduction of the other protective
groups and the attachment of the protected compound
to the solid phase was performed as described for
2. 50-DMT-protected-30-(vicinal-dihydroxyalkyl)phos-
phate-modified oligonucleotides were obtained via auto-
mated oligonucleotide synthesis on supports 6 and 9,
deprotection (25% ammonia, 6–8 h, 65 ꢁC), and
HPLC-purification. These compounds were converted
to the respective aldehyde-modified oligonucleotides 20
and 21 (Scheme 4) using sodiumperiodate (1. 4-fold ex-
cess, 75 min, rt; 2. 20-fold excess of sodium thiosulfate,
lyophilization). 30-Ribobisaldehyde 22 was obtained
starting froma solid support preloaded with the corre-
sponding ribonucleoside, automated synthesis, cleavage
(1. 25% ammonia, 16 h, 55 ꢁC; 2. HF*triethylamine/
DMF (3/1), 1.5 h, 55 ꢁC), and glycol oxidation using
O
O
O
O
O
O
i
N
OH
10
O
O
O
ii
N
H
OH
11
O
O
O
O
O
O
O
iii
CH3
CH3
12
iv
O
OH
13
Scheme 2. Synthesis of aromatic alcohols 11 and 13. Reagents and
conditions: (i) N-hydroxysuccimide, dicyclohexyl carbodiimide, diox-
ane/THF, 0 ꢁC; (ii) 3-aminopropane-1-ol, dioxane/water; (iii) ethyl-
eneglycol, toluol, 4-toluenesulfonic acid (cat.); (iv) lithiumaluminium
hydride, THF, acidic work-up.
O
i
ii
HO
HO
OH
O
OMMT
*NEt3
N
H
OMMT
O
O
O
2
1
iii
i
i
OH
HO
OMMT
OMMT
HO
OH
ODMT
OH
OH
7
3
iii, iv, ii
OH
v
ii
HO
iv
ODMT
BzO
BzO
OMMT
OSuc
OH
8
*NEt3
5
4
ODMT
O
H
N
BzO
OMMT
Bz = benzoyl
Suc = succinyl
O
O
H
N
OBz
O
O
= aminopropyl-CPG
9
O
6
Scheme 1. Synthesis of solid phases 2, 6, and 9 for 30-modification of oligonucleotides. Reagents and conditions: (i) 1.1 equiv MMT-Cl or DMT-Cl,
pyridine, 0 ꢁC, 16 h; (ii) HBTU, diisopropylamine, DMF, 1 h; (iii) benzoylchloride, pyridine, 0 ꢁC, 16 h; (iv) succinic acid anhydride, pyridine, 16 h;
(v) OsO4 (cat.), N-methylmorpholine-N-oxide, acetone/water (8/1), 24 h, 0 ꢁC.

K. Achilles, G. v. Kiedrowski / Bioorg. Med. Chem. Lett. 15 (2005) 1229–1233
1231
sodiumperiodate (1. 4-fold excess, 75 min, rt; 2. 20-fold
11
excess sodiumthiosulfate).
and the required alcohol for the synthesis of 16 could
be obtained by reacting hexan-1,2,6-triol with carbon-
yldiimidazole analogously to a previously described
method.13 Modified oligonucleotides derived from 15a
and 15b could be cleaved under standard conditions.
Following HPLC-purification, the ketal was cleaved
with 80% acetic acid within 1 h.14 The cyclic carbonate
in the oligonucleotide derived from 16 was directly
deprotected using lithiumhydroxide/triethylamine in
methanol as cleaving reagent.15 The 50-diol modified oli-
gonucleotides obtained from 15b and 16 were equiva-
lent. One argument favoring the synthesis via the ketal
is that the purification of the ketal by HPLC is improved
compared to the less lipophilic diol. Diols could be con-
verted to aldehydes following the above-mentioned con-
ditions. Aromatic alcohol 11 (Scheme 2) needed for the
synthesis of phosphoramidite 17 was synthesized analo-
gously to literature.16 Phosphoramidite 18 was obtained
starting from methyl 4-formylbenzoate by protecting the
aldehyde function with ethylene glycol and reducing the
ester to the respective alcohol 13 using lithiumalumi-
numhydride followed by acidic work-up ( Scheme 2).
All phosphoramidites applied for 50-labeling are sum-
marized in Scheme 3. Scheme 4 shows the 30- and 50-
modified trinucleotides 19–26, whose sequence in the
case of 50-modified oligonucleotides was d(CGG),
whereas it was d(CCG) in the case of 30-modified tri-
mers. Oligonucleotides were stored as aqueous stock
solutions (6.25 mM).
For 50-labeling, phosphoramidites derived from the
reaction of the appropriate alcohol and N,N-diisoprop-
yl-chloro-phosphoramidite as phosphitylating reagent
were used.12 50-Hydrazide-modified oligonucleotides
were synthesized fromthe phosphoramidite 14 (Scheme
3) as previously described.8a For the synthesis of 50-alde-
hyde-modified oligonucleotides two different approaches
were taken: Amidite building blocks were either pre-
0
pared from5 -acetal or 50-carbonyl-protected diols as
precursors or in the case of aromatic and non-enolizable
aliphatic aldehydes without protection. Quantitative
activation of various 50-acetal-modified oligonucleotides
under strongly acidic conditions failed, presumably due
to depurination. Phosphoramidites 15a, 15b, and 16
(Scheme 3) were used for 50-diol modification, whereas
the corresponding alcohols are commercially available
N(iPr)2
H
P
Trityl
N
OCE
O
N
H
O
14
N(iPr)2
P
O
N(iPr)2
P
n
O
O
OCE
O
O
OCE
O
O
15a: n = 1
15b: n = 4
16
Hydrazone formation kinetics were performed in MES-
buffer (pH 6.1, 0.1 M) at 30 ꢁC, final concentration of
oligonucleotides was about 1 mM, all experiments were
run at least in duplicate. At various time-points samples
were withdrawn and diluted hundredfold with water.
The compositions of the reaction mixtures were ana-
lyzed by RP-HPLC (ammonium bicarbonate (0.1 M)/
acetonitrile 95/5 ! 80/20 in 15 min, detection: 254 nm,
O
O
H
N
N(iPr)2
OCE
O
N(iPr)2
OCE
O
P
O
18
17
Scheme 3. Phosphoramidites 14–18 for 50-modification.
R1
R1
O
R1
O
O
G
G
G
O
O
O
R1
O
G
O
O
P
O
HO
C
O
P
O
O
P
O
O
O
O
O
O
H
N
NH2
O
O
O
O
R1=
O
O
O
P
O
20
22
21
19
O
O
C
O
O
O
O
O
P
O
O
P
O
O
C
N
O
O
O
P
C
H
O
O
O
O
O
O
O
O
R2
24
O
R2
23
O
P
O
R2=
O
O
G
O
H
N
O
P
O
O
H2N
O
O
O P O
O
C
C
n
O
O
O
O
O
O
P
O
O
G
O
O
R2
O
R2
26
25a: n = 1
25b: n = 4
OH
Scheme 4. 30- and 50-Aldehyde and hydrazide-modified oligonucleotides 19–26 applied for investigations.

1232
K. Achilles, G. v. Kiedrowski / Bioorg. Med. Chem. Lett. 15 (2005) 1229–1233
k
flow: 1 mL/min). The reaction between 20 and 26 yielded
not only the expected hydrazone, but also a number of
by-products. The analysis of the reaction mixture by
MALDI-TOF-MS showed two main by-products,
which could be explained by an oxidation of the 30-phos-
phoglycaldehyde 20 to glyoxal under elimination of the
30-phosphate oligonucleotide.17 Glyoxal itself was
reported to react with amino-group containing nucleo-
bases under formation of a 1-N2-glyoxal-adducts and
thus should exhibit this reaction also with 25a. Due to
this side reaction aldehyde 20 and its corresponding 50-
modified oligonucleotide 25a are unsuitable for quanti-
tative ligation experiments. To circumvent this side reac-
tion, the next experiments were carried out with 21,
which differs from 20 by a longer spacer between the
aldehyde function and the phosphate group. No quanti-
tative product formation was observed in this case too,
since 21 was consumed in a side reaction leaving 50-
hydrazide 26 unattached. The concentration of 21 also
decreased when it was incubated with the unmodified
trimer d(CGG). A MALDI-TOF analysis of the reac-
tion mixture showed small amounts of the expected
hydrazone and the remaining hydrazide—but no other
products. Due to these results, experiments with com-
pound 25b were also not continued. The ligation reac-
tion of the bisaldehyde 22 and hydrazide 26 appeared
to be too fast under the above-mentioned conditions.
The reaction was decelerated by increasing the pH and
lowering the final concentration of oligonucleotides
(HEPES-buffer, pH 7.5, oligonucleotide concentration:
0.79 mM). Three products having the same mass and
close retention times were found and tentatively
assigned to be the three major diastereomers of the
dihydroxymorpholine unit (Fig. 1) as reported for simi-
lar cases.18 Further experiments were carried out with
the 30-hydrazide 19 and the aromatic aldehydes 23 or
24. The time course of product formation was moni-
tored by HPLC and analyzed using our SimFit program
(Figs. 2 and 3).19 During fitting it became obvious that
in all cases the concentration of the hydrazide-modified
oligonucleotides decreased more rapidly than the prod-
uct was formed. In independent studies we found evi-
dence for the formation of dimeric and even
1
19 + 23 or 24
hydrazone
by-product
k
3
19
1.2
1
19
23
hydrazone
0.8
0.6
0.4
0.2
0
0
2
4
6
(A)
time [h]
19
1
0.8
0.6
0.4
0.2
24
hydrazone
0
0
2
4
6
(B)
time [h]
Figure 2. Reaction model and concentration versus time plot. (A) 19
and 23; (B) 19 and 24.
k
k
1
3
morpholino-
product
22 + 26
intermediate
k
-1
k
2
by-product
26
1
0.8
0.6
0.4
0.2
hydrazone
0
0
OH
B
OH
B
N
O
B
N
4
8
12
time [h]
16
20
O
O
OH
OH
N
HO
27a
HO
27c
Figure 3. Reaction model and concentration versus time plot for the
reaction of 22 and 26.
27b
22
26
oligomeric products of 50- and 30-hydrazide-modified
oligonucleotides.20 Reactions of the nucleobases with
nucleophiles such as bisulfite, hydroxylamine, or hydr-
azine and O-alkyl or N-alkyl derivatives depending on
pH are well known side reactions.21 Up to now such side
reactions have not been described for reactions with acyl
hydrazides. Variation of the reaction model (Figs. 2 and
3) by taking into account a side reaction to an undetect-
able (possibly oligomeric) product, which lowers the
hydrazide concentration during the reaction, resulted
in acceptable fits. Rate constants are summarized in
Table 1. In cases where the oligonucleotides were
27
8
10
12
14
16
retention time [min]
Figure 1. HPLC-chromatogram and possible isomers of the formed
morpholine 27.

K. Achilles, G. v. Kiedrowski / Bioorg. Med. Chem. Lett. 15 (2005) 1229–1233
Table 1. Kinetic parameters for the ligation of 19 and 23 or 24 and 22 and 26
1233
Rate constants
19 and 23
19 and 24
22 and 26
k1/MÀ1sÀ1
6.15 0.01·10À1
—
2.85 0.02·10À5
—
4.7
1.31 0.06·10À1
—
1.08 0.12·0À5
—
2.9
2.42 0.13·10À1
1.57 0.36·10À4
5.73 3.34·10À6
1.18 0.40·10À4
2.6
k
À1/sÀ1
k2/sÀ1
k3/sÀ1
RMSa/%
a RMS = root-mean square.
modified with aromatic aldehydes in the 50-position hav-
ing a 30-hydrazide as the reaction partner, the reaction
was fast and irreversible exhibiting comparable rate0con-
stants. The kinetic data fromthe reaction of the 3 -bis-
aldehyde 22 and the 50-hydrazide 26 could be
explained by the assumption of a consecutive mecha-
nism: The reaction seems to proceed reversibly in the
beginning, but becomes irreversible with increasing reac-
tion time. As a reason for this finding we presume a two-
step model: first the hydrazide reacts reversibly with one
aldehyde function of the bisaldehyde, before the forma-
tion of the morpholine ring renders the reaction irrevers-
ible. The reversibility of morpholine formation from
ribobisaldehydes and amino group containing com-
pounds has been described, but for morpholines derived
fromacyl hydrazides reversibility could not implicitly be
suspected.8d,22
2. Xu, Y.; Nilesh, B.; Kool, E. T. Nat. Biotechnol. 2001, 19,
148–152.
3. Metelev, V. G.; Borisova, O. A.; Volkov, E. M.; Orets-
kaya, T. S.; Dolinnaya, N. G. Nucl. Acid Res. 2001, 29,
4062–4069.
4. Czlapinski, J. L.; Sheppard, T. L. J. Am. Chem. Soc. 2001,
123, 8618–8619.
5. Fujimoto, K.; Matsuda, S.; Takahashi, N.; Saito, I. J. Am.
Chem. Soc. 2000, 122, 5646–5647.
6. Li, X.; Zhang, Z.-Y. J.; Knipe, R.; Lynn, D. G. J. Am.
Chem. Soc. 2002, 124, 746–747.
7. For further information see: www.solulink.com.
8. (a) Raddatz, S.; Mueller-Ibeler, J.; Kluge, J.; Wass, L.;
Burdinski, G.; Havens, J. R.; Onofrey, T. J.; Wang, D.;
Schweitzer, M. Nucleic Acids Res. 2002, 30, 4793–4802; (b)
Timofeev, E. N.; Kochetkova, S. V.; Mirzabekov, A. D.;
Florentiev, V. L. Nucleic Acids Res. 1996, 24, 4793–4802;
(c) Grimm, G. N.; Boutorine, A. S.; Helene, C. Nucleos.,
Nucleot. Nucleic Acids 2000, 19, 1943–1965; (d) Ghosh, S.
S.; Kao, P. M.; Kwoh, D. H. Anal. Biochem. 1989, 178,
43–51.
9. Eckardt, L. H.; Naumann, K.; Pankau, M. W.; Rein, M.;
Schweitzer, M.; Windhab, N.; von Kiedrowski, G. Nature
2002, 429, 286.
10. Urata, H.; Akagi, M. Tetrahedron Lett. 1993, 34, 4015–
4018.
11. Langland, J. O.; Pettiford, S. M.; Jacobs, B. L. Protein
Express. Purif. 1995, 6, 25–32.
12. Beaucage, S. L.; Radhakrishnan, P. I. Tetrahedron 1993,
49, 1925–2138.
13. Kang, S.-K.; Jeon, J.-H.; Nam, K.-S.; Park, C.-H.; Lee,
H.-W. Synth. Commun. 1994, 24, 305–312.
14. Lewbart, M. L.; Schneider, J. J. J. Org. Chem. 1969, 34,
3505.
15. Surzhikov, S. A.; Timofeev, E. N.; Chernov, B. K.;
Golova, J. B.; Mirzabekov, A. D. Nucleic Acids Res. 2000,
28, E29.
We have described the synthesis and utilization of a
number of versatile building blocks for the 30- and 50-
labeling of oligonucleotides with aldehyde and hydr-
azide functionalities. The reactions of hydrazide- and
aldehyde-modified oligonucleotides as a means to
achieve activator-free ligation were found to be fast, effi-
cient, and irreversible for the case of aromatic alde-
hydes. Hydrazone ligation using ribobisaldehydes
appeared to be reversible at the stage of open-chained
hydrazones or another intermediate and irreversible at
the stage of morpholines. Side reactions of the hydrazide
building blocks raised difficulties for quantitative liga-
tion studies. The nature of this side reaction has to be
ascertained in further studies. Nevertheless, the hydr-
azone formation fulfills most requirements to be em-
ployed as an efficient activator-free ligation chemistry.
16. Schwartz, D. A.; Hogrefe, R. I. PCT Int. Appl. WO
2002010431, 2002.
17. Awada, M.; Dedon, P. C. Chem. Res. Toxicol. 2001, 14,
1247.
Acknowledgements
18. Summerton, J.; Weller, D. Antisense Nucl. Acid Drug Dev.
1997, 7, 187–195.
19. Sievers, D.; von Kiedrowski, G. Chem. Eur. J. 1998, 4,
629–641.
This research was supported by the Deutsche Fors-
chungsgemeinschaft (DFG), the BMBF, and Cost D27.
20. By-products could not be observed by RP-HPLC and in
MALDI-TOF-MS, but by gel electrophoresis and by ion
exchange chromatography. Further experiments regarding
the nature of this side reaction are currently being
pursued.
21. Shabarova, Z.; Bogdanov, A. In Advanced Organic
Chemistry of Nucleic Acids; VCH: Weinheim, 1994; Vol.
1, p 400f.
References and notes
1. (a) von Kiedrowski, G.; Wlotzka, B.; Helbing, J.; Matzen,
M.; Jordan, S. Angew. Chem. 1991, 103, 456–459; (b) von
Kiedrowski, G.; Wlotzka, B.; Helbing, J. Angew. Chem.
1989, 101, 1259–1261; (c) Naylor, R.; Gilham, P. T.
Biochemistry 1966, 5, 2722–2728; (d) Sokolova, N. I.;
Ashirbekova, D. T.; Dolinnaya, N. G.; Shabarova, Z. A.
FEBS Lett. 1988, 232, 153–155.
22. Gritsenko, O. M.; Gromova, E. S. Russ. Chem. Rev. 1999,
68, 241–251.