Novel amino linkers enabling efficient labeling and convenient purification of amino-modified oligonucleotides

Article abstract of DOI:10.1016/j.bmc.2007.10.011

We developed new amino linker reagents for an oligonucleotide (ONT) terminus. These reagents consist of an aminoethyl carbamate main linkage and a side-chain residue, which was a naphthylmethoxymethyl, methoxymethyl, or methyl group or a hydrogen atom. Th

Full text of DOI:10.1016/j.bmc.2007.10.011

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 941–949
Novel amino linkers enabling efficient labeling
and convenient purification of amino-modified oligonucleotides
Yasuo Komatsu,a,* Naoshi Kojima,a Maiko Sugino,a Akiko Mikami,a Ken Nonaka,b
Yumi Fujinawa,b Takashi Sugimoto,b Kousuke Sato,a
Kenichi Matsubarab and Eiko Ohtsukaa
aResearch Institute of Genome-based Biofactory, National Institute of Advanced Industrial Science & Technology
(AIST Hokkaido), 2-17-2-1 Tsukisamu-Higashi, Toyohira-ku, Sapporo 062-8517, Japan
bDNA Chip Research Inc., 1-1-43, Suehirocho, Tsurumi-ku, Yokohama 230-0045, Japan
Received 6 September 2007; revised 1 October 2007; accepted 4 October 2007
Available online 10 October 2007
Abstract—We developed new amino linker reagents for an oligonucleotide (ONT) terminus. These reagents consist of an aminoethyl
carbamate main linkage and a side-chain residue, which was a naphthylmethoxymethyl, methoxymethyl, or methyl group or a
hydrogen atom. The primary amine was protected with a monomethoxytrityl (MMT) group. The chemical properties of ONTs con-
taining these amino-modifications were investigated. The MMT group of these amino-modifications could be quite rapidly removed
from the amine under very mild acidic conditions, which are not strong enough for the deprotection of a conventional aliphatic
amine. This significant feature enabled the amino-modified ONTs to be conveniently purified with a reverse phase column. Further-
more, the amino-modifications efficiently reacted to active esters, as compared with other amino-modifications. We also found that
the pKa values of the amino-modifications were lower than that of the aliphatic amine. All of the experimental results showed that
these chemical properties are closely related to their structures. We report here the chemical properties and the availability of the
new amino linker reagents.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
are also chemically immobilized on micro beads by
immersing the beads in a reaction solution. The probe
immobilization requires functional groups, such as an
aliphatic amine or a sulfhydryl group at the ONT termi-
nus, and the modified ONT probes are used after purifi-
cation. This purification process is not carried out for
in situ synthetic method.
DNA and RNA detection techniques have been remark-
ably advanced by including physicochemical techniques.
Oligonucleotide array and beads-based systems are rep-
resentative technologies, which utilize chemically syn-
thesized oligonucleotides (ONTs). There are in situ
synthesis and deposition methods for the preparation
of oligonucleotide arrays. In the in situ synthetic meth-
od, ONT probes are directly synthesized on a solid sur-
face in combination with the standard amidite
chemistry,1,2 photolithography3,4 or electrochemical
techniques.5 In the deposition method, synthesized
ONT probes are spotted on the array surface by an
ink-jet6 or thin needles,7,8 and then become immobilized
on the surface by forming a covalent bond. ONT probes
A sulfhydryl group is frequently used for ONT-immobi-
lization on a gold surface.9–11 However, a sulfhydryl
group requires thiol-protection, to prevent intermolecu-
lar dimerization, and this protecting group should re-
main attached to the sulfhydryl group until just before
the ONTs are used. In addition, the sulfhydryl modifica-
tion is expensive. Although there are other terminal
modifications such as hydrazide,12 anthraquinone,13
and acrylamide,14 these modifications yield side prod-
ucts under the standard deprotecting conditions for syn-
thesized ONTs. On the other hand, an aliphatic amine is
cost-effective and is very easy to handle in comparison
with other modifications. From this point of view,
Keywords: Amino linker; Nucleic acids; Oligonucleotides; Labeling;
Microarray.
*
Corresponding author. Tel.: +81 11 857 8437; fax: +81 11 857
8954; e-mail: komatsu-yasuo@aist.go.jp
0968-0896/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.10.011

942
Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
amino-modification is frequently used for the prepara-
tion of diverse ONT probes.15,16
those amino linker reagents, one amino linker reagent,
ssN, which consists of a naphthylmethoxymethyl group
at an aminoethyl carbamate linkage displayed the most
efficient reactivity. To examine the high reactivity of the
ssN-modification, we synthesized three amino linker
amidite units, with methoxymethyl (ssMeO), methyl
(ssMe), and hydrogen (ssH) in place of the naphthyl-
methoxymethyl group of the ssN (Fig. 1). All of these
amino linker reagents (ssR; R = N, MeO, Me, H) had
an aminoethyl carbamate structure in the main chain.
In the first report describing the ssN-modification, the
primary amine of the reagent was protected by a triflu-
oroacetyl group (TFA). Since the TFA was removed
by the standard alkaline treatment, the hydrophilic ami-
no group was exposed. The hydrophobicity of only the
naphthalene residue was insufficient to retain the ami-
no-modified ONTs within a reverse phase column. Thus,
there were still some difficulties in purifying the ssN-
modified ONTs with a reverse phase column. To achieve
easier purification by using a reverse phase column, all
of the primary amines of the ssR-amidite units were pre-
pared with monomethoxytrityl (MMT) protection,
which is highly hydrophobic. However, the removal of
an MMT group from a primary amine requires a strin-
gent acid treatment, and the MMT group thus has a
great disadvantage in terms of high throughput and con-
venient purification of diverse amino-modified ONTs.
Therefore, various deprotecting conditions for the
MMT group were first examined.
An amino hexyl amidite reagent is a typical aliphatic
amino-modification,17,18 and the amino group is intro-
duced to the 50-end of the ONT at the final step of the
chemical synthesis. There are mainly two kinds of com-
mercially available reagents, which differ in the protect-
ing groups for the primary amine. One is
a
monomethoxytrityl (MMT) group, and the other bears
a trifluoroacetyl group (TFA) as the amino-protection.
The MMT group assists in the purification of the ami-
no-modified ONTs by reverse phase column chromato-
graphy, due to its strong hydrophobicity; however, the
removal of the MMT requires stringent acid treatment.
The stringent acidic treatment is unfavorable for deoxy-
ribonucleotides, because it leads to the generation of
apurinic sites. Furthermore, the long deprotection time
is disadvantageous in terms of the high throughput
preparation of diverse ONT probes. On the other hand,
the TFA group is easily removed during the standard
alkaline deprotection step, but it is difficult to purify
the amino-modified ONTs from impurities, such as
defective ONTs, because there is little difference in the
hydrophobicity between the amino-modified ONTs
and the defective ONTs, which fail to combine with
the amino linker reagent. Therefore, some problems still
exist in the current procedure to achieve the high
throughput preparation of an amino-modified ONT li-
brary. In addition, the reaction efficiency of the ami-
no-modification should be increased in the labeling
reaction in aqueous solution.
2.2. MMT-deprotection reaction
We synthesized amino-modified ONTs of 25-bases,
which had the MMT group at the 50-terminus (MMT-
X-25; Fig. 2a). The commercially available C5 reagent,
which bears an aminoethoxyethyl structure, and a con-
ventional aliphatic amino reagent (Con) were used for
the preparation of control ONTs (Fig. 2b). After aque-
ous ammonia treatment of these ONTs, we checked
We previously constructed new types of amino linker
reagents with aromatic residues.19,20 The introduction
of the aromatic residue resulted in an increase of the
labeled products in the coupling reactions with active
esters, due to the enhanced hydrophobic interaction be-
tween the aromatic residue and the target molecule.
However, these amino linker reagents require many syn-
thetic steps, and the high throughput purification of the
ONTs was insufficient.
a
ssR-amidite unit
R
O
In our further development of amino linker reagents, we
found that a simple aminoethyl carbamate structure can
improve the reaction efficiencies and the purification
process, in comparison with the conventional amino-
modification. This amino modification is very useful
for not only labeling of ONTs but also preparation of
diverse probes for gene expression analysis. We report
here the properties and the application of this amino-
modification for the ONT-terminus.
N(
i
-Pr)2
O
MMT-NH
O
N
P
H
CN
O
ssN
ssMeO ssMe ssH
CH3OCH2
H
CH3
R=
CH2OCH2
b
ssPro-amidite unit
2. Results and discussion
O
N(i
-Pr)2
O
2.1. Synthesis of amino linker reagents containing an
aminoethyl carbamate linkage
MMT-NH
O
N
P
H
CN
O
We previously reported amino linker reagents, which
were connected to an aromatic residue by various alkyl
linkers.19,20 The amino-modified ONTs were efficiently
labeled with a fluorophore in aqueous solution. Among
Figure 1. Structures of amino linker amidite reagents. (a) ssR-amidite
units. (b) ssPro-amidite unit. R indicates a naphthylmethoxymethyl
(N), methoxymethyl (MeO), or methyl (Me) group or a hydrogen
atom (H).

Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
943
2a), and the MMT deprotection rate was measured.
ssPro showed a much slower deprotection rate than
ssH, even though the structural difference is only one
methylene between these molecules. The results obtained
from the ssR, ssPro, and C5-modifications indicated
that the aminoethyl carbamate linkage is an adequate
structure for the fast removal of the MMT group from
the primary amine under mild acidic conditions.
a
X-25
5’
5’
3’
X-T5
X=ssR, ssPro, Con, C5
b
X-
2.3. Cartridge column purification
R
O
-
H2N
H2N
OPO3
O
N
ssR
Trityl groups can facilitate the purification of trityl-pro-
tected (‘Tr on’) ONTs with the use of a reverse phase
column, because the trityl groups have high affinity to
the reverse phase resin. Especially, the dimethoxytrityl
(DMT) group, which is a protecting group for a primary
alcohol of the 50-terminus, enables ONTs to be conve-
niently purified, because the DMT can be rapidly re-
moved within the resin, by passing an acidic solution
through the column. On the other hand, the MMT
group protecting an amine cannot be deprotected com-
pletely within the column resin, even by adding an acidic
solution.
H
O
-
OPO3
O
N
ssPro
Con
C5
H
-
OPO3
H2N
H2N
O
-
OPO3
Figure 2. Amino-modified ONTs (X-25 and X-T5). (a) Sequences of
X-25 and X-T5. X indicates ssR, ssPro, Con or C5. (b) 50-Terminal
structures of amino-modified ONTs. Con and C5 were used as control
modifications.
Since it was found that the MMT deprotection rate of
the ssR was very fast under mild acidic conditions, we
investigated whether the ssR-modified ONTs would be
rapidly purified by using a cartridge column filled with
reverse resin. ssH, which showed the slowest deprotec-
tion rate among the ssR-modifications, was tested in
the cartridge purification. After ssH- or Con-modified
ONTs (MMT-ssH-25, MMT-Con-25) were subjected
to the standard alkaline treatment, each ONT was
loaded on the column resin. After the resin was washed,
3% aqueous trifluoroacetic acid was passed through the
column to promote the MMT-deprotection. The total
exposure times to the acidic solution for the ssH- and
Con-modified ONTs were 10 or 60 min. After neutral-
ization of the column resin, the amino-modified ONTs
were eluted. HPLC analysis of each eluted fraction indi-
cated that ssH-25 could be purified by the 10-min acid
treatment but Con-25 still retained the MMT group
even after the longer acid treatment (Supplementary
data). The 10-min treatment of Con-25 yielded almost
all of the ONT in the ‘Tr on’ form. We also found that
other ssH-modified ONTs of 40 bases could be purified
with a cartridge column (Supplementary data). These re-
sults indicate that the ssR-modification enables the rapid
and high quality purification of the amino-modified
ONTs with the use of a cartridge column, and this
feature provides a great advantage for the high through-
put purification of diverse ONT probes.
the acid lability of the MMT group of each amino-mod-
ification. First, MMT-X-25 was treated with an 80%
acetic acid solution for 20 min, and the ratios of ‘trityl
off’ ONTs were calculated from the HPLC analysis (Ta-
ble 1). Interestingly, the MMT groups were completely
deprotected from the MMT-ssR-25 under these condi-
tions, whereas MMT-Con-25 gave only 20% ‘trityl off’
products. The same reactions were then performed un-
der milder conditions (1% acetic acid). The MMT
groups of the ssR were still removed very fast, within
10 min, under the mild acidic conditions (Table 1),
whereas MMT-Con-25 gave only a small amount of ‘tri-
tyl off’ ONTs. The observed rate constants in the 1%
acetic acid solution were obtained (Table 1). Although
ssH showed a slightly slower deprotection rate than
the other ssR, the rate of ssH was still much faster than
that of Con-modification. The deprotection rate of the
C5-modification was an intermediate value between
Con and ssH. The fast MMT-deprotection of the ssR-
modification was also observed in other ONTs with dif-
ferent sequences (data not shown).
To investigate the structure–function relationship more
precisely, an amino linker reagent with an aminopropyl
carbamate linkage (ssPro) was synthesized (Figs. 1b and
Table 1. Analysis of MMT-deprotection reactions by treatment with aqueous acetic acid (AcOH)
Con
ssN
ssMeO
ssMe
ssH
C5
ssPro
80% AcOH (20 min)a
20.7
0
100
100
100
100
1.3
100
97.6
100
0.49
100
91.5
45.0
79.4
0.067
82.5
32.1
46.5
0.033
1% AcOH (pH 2.7; 10 min)a
1% AcOH (pH 2.7; 20 min)a
97.3
98.7
0.37
88.5
98.4
0.25
4.5
kobs (minꢀ1 b
N.D., not detected.
a Percentages of ‘Tr off’ amino-modified ONTs (X-25) after 10 or 20-min reaction.
b Observed rate constants were calculated from the percentages of the products, which were generated from 1% acetic acid treatment.
)
N.D.

944
Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
2.4. Labeling reactions with active esters
and phosphate buffers (Fig. 3c). The labeling efficiency
of the ssR-modifications remarkably depended on the
side-chain size. On the other hand, the Con-modification
decreased the reaction products in the phosphate buffer.
Since the pKa value of the aliphatic amine (Con) is about
11 (the pKa is described in Section 2.5), it is generally
thought that the decrease in the labeled product is de-
rived from the increase in the protonated amino form,
by changing the buffer’s pH from pH 9 to 8. Although
the reactions of the ssN, ssMeO, and ssMe-modifica-
tions were not influenced by this buffer change, the
labeled products of ssH were slightly decreased. The
C5- and ssPro-modifications showed similar reactivities
to Con-modification.
Amino-modified ONTs are frequently used as biological
probes by conjugation with various reporter groups,
such as fluorophores or biotin. Therefore, we investi-
gated the labeling efficiency of the ssR-modification,
using succinimidyl (NHS) esters (Cy3-NHS and biotin-
NHS) and isothiocyanate group (FITC). These reac-
tions were carried out under standard labeling condi-
tions using bicarbonate (pH 9) or phosphate (pH 8)
buffers.
First, the labeling reactions to Cy3-NHS and biotin-
NHS were examined. The ssR-modified ONTs (ssN-
25, ssMeO-25, ssMe-25, and ssH-25) were labeled less
efficiently than Con-25 in the bicarbonate buffer
(Fig. 3a and b; white bars). On the other hand, the same
coupling reactions using the phosphate buffer resulted in
a drastic increase in the product for the ssR (Fig. 3a and
b; black bars). Especially, the ssN- and ssMeO-modifica-
tions gave approximately 6-fold higher amounts of the
labeled products by this buffer change. C5 and ssPro
also yielded slightly increased labeled products in the
phosphate buffer. In contrast to the ssR-modifications,
Con-25 did not exhibit any difference in the labeling effi-
ciency between both reaction buffers.
The results of the labeling reactions revealed that the
reaction property of the ssR was different from that of
the standard aliphatic amine, and an apparent struc-
ture–function relationship was observed from the reac-
tions of ssR, ssPro, and C5. Among all of the
reactions, the ssR-NHS ester reactions performed in
the carbonate buffer were quite characteristic. To further
examine the effect of the buffer composition on the ssR
reactivity, the labeling reactions of ssN-25, ssH-25, and
Con-25 to biotin-NHS were carried out with various
concentrations of the three reaction buffers [bicarbonate
(pH 8, 9), phosphate (pH 8) and borate (pH 8)]. These
reactions revealed that the ssN- and ssH-modified ONTs
reacted to the NHS ester more efficiently than the Con-
In contrast to the NHS ester, FITC reacted with the
ssR-modified ONTs efficiently in both the bicarbonate
100
100
a
b
80
60
40
20
0
80
60
40
20
0
Con
N
MeO Me
H
C5 Pro
Con
N
MeO Me
H
C5 Pro
100
80
60
40
20
0
c
100
80
60
40
20
0
d
biotin-NHS/carbonate buffer
ssH
ssN
Con
Con
N
MeO Me
H
C5 Pro
20 250 500 20 250 500 900
carbonate
buffer (mM)
pH9
pH8
Figure 3. Labeling reactions of X-25 (X = Con, ssN, ssMeO, ssMe, ssH, C5, ssPro) with reporter groups. Each ONT was labeled with Cy3-NHS (a)
biotin-NHS (b) or FITC (c). The reactions were carried out in 250 mM carbonate buffer (pH 9; white bars) or 250 mM sodium phosphate buffer (pH
8; black bars) for 30 min. Con-25 (solid diamonds), ssH-25 (solid circles) and ssN-25 (solid triangles) were labeled with biotin-NHS in the presence of
various concentrations of the carbonate buffer (pH 8 and 9) (d). Percentages of the products are plotted versus the buffer concentrations.

Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
945
modified ONT in either the phosphate or the borate buf-
fer, and the percentages of the products were the same at
all concentrations (Supplementary data). On the other
hand, a decrease in the product was observed as the con-
centration of the bicarbonate buffer increased (Fig. 3d).
It is unclear why the ssR reaction to the NHS ester was
inhibited at high bicarbonate ion concentrations. There
might be some interactions between the aminoethyl car-
bamate structure and the bicarbonate ions. However,
the results show that the labeling reaction proceeds effi-
ciently at the low concentrations (about 20 mM) of the
bicarbonate ions.
and the labeling efficiency were also related to the R res-
idues. Although we could not detect this intramolecular
hydrogen bond by the NMR measurement, this struc-
tural effect induced by the side chain is probably related
to the chemical properties of the ssR-modification. ssH,
C5, and ssPro have also the potential to form the con-
formation of five- or six-membered rings by forming
the intramolecular hydrogen bonds; however, these
modifications have no side-chain residues, and the con-
formations might not be stabilized unlike ssN. There-
fore, the reactivities of these modifications would
mainly associate with the electron withdrawing induc-
tive effect of the main chain structures as discussed in
the pKa measurement rather than the hydrogen bond
factor.
2.5. pKa measurements
The pKa is an important factor to characterize the ami-
no group, and thus the pKa value of each amino-modi-
fication was examined. In order to determine the pKa,
each cognate alcohol unit of ssN, ssH, C5, and Con
was prepared. 13C NMR analyses of these monomer
units (ssNOH, ssHOH, C5OH, ConOH) were measured in
various pH buffers, and each pKa value was determined
from the chemical shift changes21,22 (Supplementary
data). The pKa values of ssNOH, ssHOH, C5OH and Con-
OH were found to be 8.9, 9.6, 10.1, and 11.0, respectively.
The ssNOH and ssHOH units containing the aminoethyl
carbamate linkage, showed lower pKa values than the
aliphatic amine, and the pKa values of ssNOH and ssHOH
are probably derived from the electron withdrawing
effect of the carbamate linkage. This effect might be also
responsible for the rapid MMT removal and the efficient
labeling reaction. The difference in the pKa values be-
tween ssNOH and ssHOH is notable. The pKa difference
is probably related to their tertiary structures induced
by the side chains at the aminoethyl carbamate linkage.
2.7. Stability of ssR-modified ONTs
If the ssN adopts the anti-conformation, as predicted in
the computational simulation, then trans-acylation may
occur at the primary amine. Therefore, to examine the
stability of the ssR-modified ONTs, penta thymidylic
acids containing ssR- or Con-modifications (ssR-T5
and Con-T5; Fig. 2) were subjected to the standard
aqueous ammonia treatment without MMT-protection,
and the reactions were analyzed by reverse phase HPLC.
Aqueous ammonia treatment of ssN-T5 provided new
two peaks (P1 and P2), in addition to the starting mate-
rial (ssN-T5) (Fig. 5a). Both products were purified by
HPLC, and their molecular weights were measured. P1
was identified as Con-T5 from its molecular weight
and the retention time of the HPLC analysis. Although
the molecular weight of P2 was the same as that of ssN-
T5, the retention time of P2 was different from that of
ssN-T5. Alkaline treatment of ssR-T5 also yielded the
same decomposition pattern as ssN-T5 (Supplementary
data), but the product yields were different from each
other (Fig. 5b) and depended on the R residue on the
carbamate linkage. ssH-T5 provided the least amount
of the products among the ssR series, and ssPro-T5 gave
much less product than ssH-T5, while no side products
were found with the same treatment of the Con-modifi-
cation. These reactions required heating conditions and
did not proceed under the labeling conditions [at 40 °C
in bicarbonate buffer (pH 9)]. Furthermore, the P1 and
P2 products were not detected after the alkaline treat-
ment of MMT-protected ONTs. This result shows that
2.6. Computational simulations
The ssR-modification exhibited unique chemical proper-
ties, which are expressed by the presence of the amino-
ethyl carbamate structure. In addition, the chemical
properties of the ssR series depended on the side-chain
residues. Especially, ssN showed the highest reactivity
among the ssR series, and therefore, we performed a
computational simulation of the ssN-monomer (ssNOH
)
to elucidate the structural effect based on the main and
side chains. MacroModel version 9.0 was used for the
computational simulation,23 and most stable structure
of ssNOH is shown in Figure 4. The modeling of ssNOH
indicates that the primary amine and the naphthylmeth-
oxymethyl residue are aligned with an anti-conforma-
tion, and that the primary amine is close to the O4
oxygen atom of the carbamate linkage. This conforma-
tion allows a hydrogen bond between the amino proton
and the oxygen atom. If the primary amine forms an
intramolecular hydrogen bond with the O4 atom, then
the nucleophilicity of the amine (N1) might be increased,
leading to reduction in the pKa, which is consistent with
the results of the pKa measurement.
a
b
ssNOH
5’
O
4’
3’
O
7
3
6
4
O
OH
2
N
5
1
H
N
H
H
The stability of the anti-conformation is thought to de-
pend on the bulkiness of the side-chain residue on the
C3 atom of ssR. Actually, the MMT-deprotection rate
Figure 4. Computational analysis of the ssN-monomer (ssNOH).
The most stable anti-conformation structure (a) and its schematic
drawing (b).

946
Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
100
80
60
40
20
0
P2
a
c
b
P1 P2
P1
ssN-T5
N
MeO Me
H
Con Pro
HN
O
Y
R
O
R-c-carbamate form
5
H2N
R
+
path 1
path 2
Con (P1)
O
2
R
H
1
O
O
Y
5
O
N
O
H
N
H
O
Y
5
N
H
N
H
OH
ssR-T5 Y=penta thymidylic acids
ssNOH R=N, Y=H
R-urea form (P2)
Figure 5. Stability of free ssR-modifications under heated alkaline conditions. (a) HPLC analysis of ssN-T5 without trityl-protection after aqueous
ammonia treatment at 60 °C for 16 hr. (b) Products (P1 and P2) of each ONT after alkaline treatment at 60 °C for 16 hr. White and black bars
indicate percentages of P1 and P2. (c) Scheme of the intramolecular reaction of ssR-modified ONT and ssN-monomer (ssNOH). Y indicates ONT or a
hydrogen atom. Aliphatic amine and cyclic-carbamate (c-carbamate) forms were generated by path 1, and the urea-form was generated by path 2.
the presence of a free primary amine of the ssR is asso-
ciated with the decomposition reaction.
consisted of poly(dT)10-linkers and 30-base capture
sequences, which were designed to be complementary
to the anti-sense RNAs (aRNAs) of selected endogenous
mRNAs. These ONTs were chemically synthesized and
purified with a reverse phase column, according to our
purification protocol. All of the ONTs were highly puri-
fied with high quality and then were immobilized on a
plastic slide. The aRNAs were generated from total
RNA preparations of K562 and HepG2 cells, followed
by labeling with Cy3 and Cy5, respectively.
To investigate this reaction mechanism, the ssN-mono-
mer unit (ssNOH, Fig. 5; R = N, Y = a hydrogen atom)
was also subjected to the alkaline treatment. As ex-
pected, two main products with UV-absorbance were
obtained, and their structures were determined to be
cyclic carbamate (c-carbamate; yield 11%) and urea-
ssNOH forms (yield 52%), as shown in Figure 5. These
products were generated from two reaction routes,
which are triggered by the attack of the primary amine.
From this result, it is thought that P1 and P2, generated
from the alkaline treatment of ssR-T5, were Con-T5 and
urea-R-T5 (Fig. 5c; Y = penta thymidylic acid), and this
identification is consistent with the results of the molec-
ular weight and the HPLC analyses. P1 and P2 of each
ssR-T5 were subjected to labeling reactions with FITC,
which revealed that only P1 was labeled with FITC. This
result is consistent with the identification that P1 is Con-
T5 and P2 is urea-R-T5 lacking the primary amine. The
efficiency of the intramolecular trans-acylation de-
pended on the R residue. This result might be derived
from the anti-conformational effect, as shown in Figure
4, which places the primary amine proximal to the car-
boxyl group.
The relative RNA expression levels of these cell lines
were obtained from microarray analyses (Fig. 6). A
quantitative PCR (QPCR) experiment was also per-
formed using the same RNAs. The expression ratios of
genes obtained using microarray were compared to
those obtained using QPCR. For a few genes, namely
1.0E+04
QPCR
microarray
1.0E+02
1.0E+00
1.0E-02
1.0E-04
1.0E-06
1.0E-08
2.8. Application to microarray
It was proven that ssR modification enables high
throughput purification and increases reactivity to active
esters. Then, to investigate their capability as microarray
probes, ssH-modified ONTs were subjected to a micro-
array analysis. The RNA expression of cell lines (K562
and HepG2) was analyzed using microarrays with
immobilized ssH-modified ONTs. All of the ONTs
PON2
LDHB APOH IGFBP1 HBE1
SSR4
Figure 6. Microarray analysis using ssH-modified ONTs as probes.
The gene expression of two cell lines was examined by the microarray
and quantitative PCR (QPCR). Concordance in the profiling of gene
expression is shown.

Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
947
APOH and IGFBP1, the ratios of QPCR were larger
than those of microarray. This is partly due to the differ-
ence in calculation methods. Overall, the direction of the
expression ratios of all genes was concordant to each
other for both systems. This result demonstrates that
ssH-modified ONTs are applicable to microarray
analyses.
search. All amino linker units were coupled for 600 s
at the final step of the ONT synthesis. The coupling effi-
ciencies of ssR and Con were calculated by measuring
the 478 nm absorbance of the monomethoxytrityl
group, which was removed from the termini. Cleavage
from the CPG support and deprotection of the ONTs
were carried out by incubating the CPG in concentrated
ammonium hydroxide (2 mL) at 55 °C for 16 h.
3. Conclusions
4.3. Cartridge column purification of 50-MMT-oligo-
nucleotides
We have developed novel amino linker reagents (ssR),
which have an aminoethyl carbamate structure. The
aminoethyl carbamate structure increased both the
MMT-deprotection rate and reactivity to active esters,
as compared to a standard aliphatic amine. The fast
deprotection of the MMT group from the aminoethyl
carbamate linkage facilitated the rapid and convenient
purification of the amino-modified ONTs with the re-
verse phase cartridge column. The pKa value of the ami-
noethyl carbamate structure became lower than that of
the aliphatic amine. These significant properties were de-
rived from the aminoethyl carbamate structure and the
substituent at the linkage. Computational simulation
of the ssN-monomer, which contains a naphthylmeth-
oxymethyl group on the main chain, revealed the stable
anti-conformation between the primary amine and the
naphthylmethoxymethyl residue, leading to an intramo-
lecular hydrogen bond between the primary amine and
the oxygen atom of the carbamate structure. Due to
the anti-conformation, intramolecular trans-acylation
occurred at the free aminoethyl carbamate structure un-
der heated alkaline conditions, and this intramolecular
reaction depended on the substituent size.
A reverse phase cartridge column (YMC, C18, 500 mg)
was washed with acetonitrile (6 mL) and was equili-
brated with 0.2 M TEAA (6 mL). After the deprotection
of amino-modified oligonucleotides (MMT-ssR-25,
MMT-Con-25, MMT-ssR-T5, and MMT-Con-T5), a
concentrated ammonium hydroxide solution was mixed
with the same volume of 0.2 M TEAA, and then, the
solution was loaded on the cartridge column. The col-
umn was washed with 0.2 M TEAA (4.2 mL), 10% ace-
tonitrile-0.2 M TEAA (4 mL), and distilled water
(3 mL). To remove the MMT group from the primary
amine, 3% aqueous trifluoroacetic acid (TFA; 5 mL)
was slowly passed through the column over 10-min per-
iod (60 min for the Con-modified ONT). The column
was washed with distilled water (3 mL), 0.1 M TEAA
(3 mL), and distilled water (3 mL) to neutralize and de-
salt the column resin. Finally, the amino-modified
ONTs were eluted from the column with 20% aqueous
acetonitrile (3 mL). The oligonucleotides were further
analyzed by HPLC, using reverse phase column
chromatography.
4.4. Analysis of MMT-deprotection from the synthesized
ONTs
The terminal modification containing an aminoethyl
carbamate linkage was superior to the aliphatic amino-
modification, in terms of the reactivity and the purifica-
tion. Especially, ssH lacking an asymmetric atom was
the most stable and cost-effective modification. It was
proven that ssH-modified ONTs could function as
microarray probes. We have revealed here the unique
chemical properties of the amino-modification contain-
ing an aminoethyl carbamate linkage. The simple ami-
noethyl carbamate structure is a useful modification
for the functionalization of ONT termini.
After the standard deprotection step of the amino-mod-
ified ONTs (MMT-Con-25, MMT-ssR-25, MMT-ssPro-
25, MMT-C5-25; 0.2 lmol), an aliquot (10 lL) was ta-
ken from the concentrated ammonium hydroxide solu-
tion (2 mL), and was immediately transferred into
distilled water (5 lL). The solution was evaporated un-
der reduced pressure. Aqueous acetic or trifluoroacetic
acid solutions of various concentrations (25 lL) were
added to the amino-modified ONTs to deblock the trityl
group. Aliquots (5 lL each) were taken from the reac-
tion solution at time intervals and were neutralized with
a dilute aqueous ammonia solution. The percentages of
‘trityl off’ products were determined by an HPLC anal-
ysis, using a reverse phase column. These HPLC charts
are presented in the Supplementary data. The observed
rate constants of ssN, ssMeO, ssMe, and ssH were calcu-
lated from the reactions using a 1% aqueous acetic acid
solution.
4. Experimental
4.1. Synthesis of phosphoramidite units
The syntheses of the ssN, ssMeO, ssMe, ssH, and ssPro
phosphoramidite units are presented in the Supplemen-
tary data.
4.2. Synthesis and deprotection of oligonucleotides
4.5. Coupling reaction with reporter groups
All oligonucleotides (ONTs) were chemically synthe-
sized by using standard phosphoramidite chemistry.
The amidite units of the canonical bases and the control
amino linker reagents (50-amino-modifier C6-MMT, 50-
amino-modifier C5) were purchased from Glen Re-
The amino-modified ONTs (1 nmol) were reacted with
an active ester group (Cy5-succinimidyl ester, biotin-
succinimidyl ester, fluorescein isothiocyanate (FITC))
in a solution (100 lL) containing 250 mM sodium car-
bonate buffer (Na2CO3–NaHCO3, pH 9) and 10%

948
Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
dimethylformamide. The buffer composition was chan-
ged to 250 mM sodium phosphate (pH 8) or 100 mM so-
dium tetraborate (pH 8), and labeling reactions were
also carried out. The concentrations of the succinimidyl
ester groups (Cy3 and biotin) were 0.3 mM, and the
FITC labeling reactions were carried out at 5 mM. Ali-
quots were taken from the reaction solutions after 30
and 60 min, and were desalted with a cartridge column
(NAP5, Amersham Pharmacia). The products were ana-
lyzed by HPLC, using a reverse phase column and a
photodiode array detector (Waters). The percentages
of the products at each time point were determined from
the HPLC analysis.
(Supplementary data). QPCR was also carried out,
according to the standard method (Supplementary
data).
Acknowledgments
We thank Dr. Lim Chun Ren, Dr. Junya Hashida, and
Youji Ueda for their helpful discussions, and we thank
Naonori Inoue for technical assistance with the MAL-
DI-TOF/MS measurements. This work was carried
out with financial support from the National Institute
of Advanced Industrial Science and Technology
(AIST).
To study the effect of the carbonate ion on the coupling
reaction, 10 lM of the amino-modified ONT (Con-25,
ssN-25) was reacted with 0.3 mM biotin-NHS in the
presence of 250 mM sodium phosphate buffer (pH 8),
containing 0, 50, 100, 250 or 500 mM sodium carbonate
buffer (pH 8). The reactions were analyzed by the same
method as that used for the coupling reaction described
above.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2007.10.011.
4.6. Computational simulation and pKa determination
See the Supplementary data.
References and notes
1. Maskos, U.; Southern, E. M. Nucleic Acids Res. 1992, 20,
1675.
2. Southern, E. M.; Case-Green, S. C.; Elder, J. K.; Johnson,
M.; Mir, K. U.; Wang, L.; Williams, J. C. Nucleic Acids
Res. 1994, 22, 1368.
3. Fodor, S. P.; Read, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A.
T.; Solas, D. Science 1991, 251, 767.
4. Jacobs, J. W.; Fodor, S. P. Trends Biotechnol. 1994, 12, 19.
5. Egeland, R. D.; Marken, F.; Southern, E. M. Anal. Chem.
2002, 74, 1590.
6. Okamoto, T.; Suzuki, T.; Yamamoto, N. Nat. Biotechnol.
2000, 18, 438.
7. Beier, M.; Hoheisel, J. D. Nucleic Acids Res. 1999, 27,
1970.
8. Ramakrishnan, R.; Dorris, D.; Lublinsky, A.; Nguyen, A.;
Domanus, M.; Prokhorova, A.; Gieser, L.; Touma, E.;
Lockner, R.; Tata, M.; Zhu, X.; Patterson, M.; Shippy,
R.; Sendera, T. J.; Mazumder, A. Nucleic Acids Res. 2002,
30, e30.
4.7. Alkaline treatment of amino-modified ONTs
After the amino-modified ONTs (Con-T5, ssR-T5,
ssPro-T5; 50 nmol) were purified on the reverse phase
column, they were dissolved in concentrated aqueous
ammonia (3.3 mL), and heated at 60 °C. After 16 h,
the ammonia was removed under reduced pressure,
and then the residue was dissolved in distilled water
(1100 lL). The products (P1 and P2) were purified
by HPLC, and the molecular weights were determined
by MALDI-TOF/MS measurements (Supplementary
data). P1 and P2 were mixed with Con-T5 and the
starting material (ssR-T5, Con-T5 or ssPro-T5),
respectively, and the HPLC analysis using the reverse
phase column was carried out to compare the retention
times.
9. Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000,
289, 1757.
10. Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R.
L.; Mirkin, C. A. Science 1997, 277, 1078.
11. Pathak, S.; Choi, S. K.; Arnheim, N.; Thompson, M. E.
J. Am. Chem. Soc. 2001, 123, 4103.
12. 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.
13. Koch, T.; Jacobsen, N.; Fensholdt, J.; Boas, U.; Fenger,
M.; Jakobsen, M. H. Bioconjug. Chem. 2000, 11,
474.
14. Rehman, F. N.; Audeh, M.; Abrams, E. S.; Hammond, P.
W.; Kenney, M.; Boles, T. C. Nucleic Acids Res. 1999, 27,
649.
15. Epstein, J. R.; Lee, M.; Walt, D. R. Anal. Chem. 2002, 74,
1836.
16. Epstein, J. R.; Ferguson, J. A.; Lee, K. H.; Walt, D. R.
J. Am. Chem. Soc. 2003, 125, 13753.
17. Agrawal, S.; Christodoulou, C.; Gait, M. J. Nucleic Acids
Res. 1986, 14, 6227.
4.8. Microarray analyses of mRNAs collected from cell
lines
ssH-modified ONTs were designed and synthesized to
detect anti-sense RNAs (aRNAs) of PON2, LDHB,
APOH, IGFBPI, HBE1, GAPDH, and SSR4 (Supple-
mentary data). All ssH-modified ONTs were purified
with cartridge columns, according to our purification
protocol.
Microarrays were fabricated using a Prime Surface kitR
(BS-11607 Sumitomo Bakelite Co., Ltd.). The ssH-mod-
ified ONTs were dissolved in the spotting buffer to a
concentration of 10 lM, and then each ONT was depos-
ited in triplicate on the plastic plate with a spotting ma-
chine (MARKS-I, Hitachi Software Engineering, Co.,
Ltd.). Cy3 or Cy5 labeled anti-sense RNA (aRNA)
was prepared from HepG2 and K562 cells, and then
these aRNAs were subjected to microarray analyses

Y. Komatsu et al. / Bioorg. Med. Chem. 16 (2008) 941–949
949
18. Emson, P. C.; Arai, H.; Agrawal, S.; Christodoulou, C.;
Gait, M. J. Methods Enzymol. 1989, 168, 753.
Komatsu, Y. Bioorg Med. Chem. Lett. 2006, 16,
5118.
19. Komatsu, Y.; Kojima, N.; Nonaka, K.; Fujinawa, Y.;
Sugino, M.; Mikami, A.; Hashida, J.; Ohtsuka, E.;
Matsubara, K. Nucleic Acids Symp. Ser. 2004,
21.
21. Aurup, H.; Tuschl, T.; Benseler, F.; Ludwig, J.; Eckstein,
F. Nucleic Acids Res. 1994, 22, 20.
22. Ogawa, A.; Tanaka, M.; Sasaki, T.; Matsuda, A. J. Med.
Chem. 1998, 41, 5094.
20. Kojima, N.; Sugino, M.; Mikami, A.; Nonaka, K.;
Fujinawa, Y.; Muto, I.; Matsubara, K.; Ohtsuka, E.;
23. Sato, K.; Seio, K.; Sekine, M. J. Am. Chem. Soc. 2002,
124, 12715.