Oligodeoxynucleotides containing 2′-deoxy-1-methyladenosine and Dimroth rearrangement

Article abstract of DOI:10.1002/hlca.200790093

2′-Deoxy-1-methyladenosine was incorporated into synthetic oligonucleotides by phosphoramidite chemistry. Chloroacetyl protecting group and controlled anhydrous deprotection conditions were used to avoid Dimroth rearrangement. Hybridization studies of int

Full text of DOI:10.1002/hlca.200790093

928
Helvetica Chimica Acta – Vol. 90 (2007)
Oligodeoxynucleotides Containing 2’-Deoxy-1-methyladenosine and Dimroth
Rearrangement
by Edward N. Timofeeva), Sergey N. Mikhailov*a), Andrei N. Zueva), Ekaterina V. Efimtsevaa),
Piet Herdewijn*b), Robert L. Somersc), and Marc M. Lemaitrec)
a) Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Vavilov str. 32, Moscow,
119991 Russia (phone: þ7-495-1359733; fax: þ7-495-1351405; e-mail: smikh@eimb.ru)
b) Katholieke Universiteit Leuven, Rega Institute for Medical Research, Laboratory of Medicinal
Chemistry, Minderbroedersstraat 10, B-3000 Leuven
(phone: þ32-16-337387; fax: þ32-16337340; e-mail: Piet.Herdewijn@rega.kuleuven.be)
c) Glen Research Corporation, 22825 Davis Drive, Sterling, VA, 20164, USA
2’-Deoxy-1-methyladenosine was incorporated into synthetic oligonucleotides by phosphoramidite
chemistry. Chloroacetyl protecting group and controlled anhydrous deprotection conditions were used to
avoid Dimroth rearrangement. Hybridization studies of intramolecular duplexes showed that introduc-
tion of a modified residue into the loop region of the oligonucleotide hairpin increases the melting
temperature. It was shown that modified oligonucleotides may be easily transformed into oligonucleo-
tides containing 2’-deoxy-N6-methyladenosine.
1. Introduction. – Naturally occurring 1-methyladenosine (m1A) attracted much
attention due to its role in building of tRNA secondary structure [1 – 5]. This modified
nucleoside has shown to be responsible for cloverleaf conformation of tRNA [3]. Post-
transcriptional methylation of adenosine in tRNA with methyl-1-adenosine transferase
dramatically changes charge and hydrophobic properties of this heterocyclic base. The
consequences of such a modification influence both stacking and pairing properties of
m1A. Inability of the base to form a Watson – Crick (WC) duplex favors the loop
formation which may be additionally stabilized by the positive charge of m1A. Indeed,
some stabilization effect was observed in synthetic oligonucleotide hairpins due to the
presence of m1A in the loop [6].
Recently, the synthesis of protected phosphoramidite of m1A and hybridization
properties of short modified synthetic oligoribonucleotides have been reported [6][7].
The selection of appropriate N6-protecting group and mild ammonolysis conditions
allowed us to avoid complications associated with incomplete deprotection and
Dimroth rearrangement.
Oligodeoxynucleotides containing 2’-deoxy-1-methyladenosine (m1dA) possess a
great potential in studies of DNA alkylation and repair mechanism [8 – 10] and in
construction of deoxyribozymes [11][12]. Substitution of adenosine (pK 3.6 – 3.8) with
1-methyladenosine (pK 8.2 – 8.8) [13][14] may have considerable effect on the catalytic
properties of deoxyribozymes. Particular interest may arise with respect to non-
canonical DNA studies due to the feature of m1A to form reverse-Hoogsteen base pairs
[15].
ꢀ 2007 Verlag Helvetica Chimica Acta AG, Zürich

Helvetica Chimica Acta – Vol. 90 (2007)
929
Here, we report the preparation of a m1dA synthon suitable for standard
oligonucleotide synthesis, its incorporation into oligodeoxynucleotides, and their
hybridization behavior in short DNA hairpins. Dimroth rearrangement in concentrated
aqueous ammonia solutions have been also studied in detail.
2. Results and Discussion. – 2.1. Synthesis of m1dA Phosphoramidite and Modified
Oligonucleotides. The synthetic route to the protected m1dA cyanoethyl phosphor-
amidite (4) is shown in the Scheme. We followed the synthetic strategy that was
successfully implemented for the preparation of m1A phosphoramidite [6]. However,
some modifications were introduced in the present strategy. Prior the methylation of 2’-
deoxyadenosine, we introduced 5’-O-MMTr (MMTr¼ monomethoxytrityl) protecting
group. This derivative, 1, was further reacted with MeI in N,N-dimethylacetamide
according to the procedure described for adenosine [16] to give 2 in 72% yield. If the
methylation was performed first, as described for m1A [6], introduction of MMTr took
up to 7 d, and the yield of 2 was not more than 50% due to the low solubility of m1dA in
pyridine.
An increase of the acid stability of the 5’-O-protecting group (MMTr vs. DMTr¼
dimethoxytrityl) is preferred in view of the formation of hydroiodide salt of 2 during
alkylation with MeI. We have found that such a replacement did not affect the
efficiency of the deblocking step in oligonucleotide synthesis.
Scheme
a) MeI. b) (ClCH2CO)2O/Py, NH3/MeOH. c) (iPr2N)2P(OCH2CH2CN), 1H-tetrazole, Py. d) Automated
oligonucleotide synthesis.

930
Helvetica Chimica Acta – Vol. 90 (2007)
Reaction of MMTr derivative 2 with chloroacetic anhydride provided protected 3 in
78% yield. Preparation of phosphoramidite 4 was performed with 2-cyanoethyl
N,N,N’,N’-tetraisopropylphosphorodiamidite. Unlike the adenosine derivative bearing
a bulky 2’-O-tBuMe2Si protecting group and requiring the use of 2-cyanoethyl chloro-
N,N-diisopropylphosphoramidite, 2’-deoxynucleoside 3 easily reacted with bis-amidite
in the presence of pyridinium tetrazolide. We found that purification of m1dA
phosphoramidite 4 may be carried out with high efficiency in AcOEt/hexane or hexane/
toluene mixtures, allowing easy separation of the product.
Incorporation of modified deoxyadenosines during automated oligonucleotide
synthesis was carried out without modifications of the standard protocol. The m1dA
amidite 4 coupled very well using the standard DNA coupling time (30 s) and with
0.45m 1H-tetrazole as activator. Efficiency of coupling for 4 was found to be at least
99% as estimated by comparing DMTr cation absorbance at 495 nm for n ꢀ 1 and n þ 1
base. Complete removal of the MMTr protecting group was achieved when using
deblock parameters for thymidine. Reverse-phase (RP) HPLC of crude oligodeoxy-
nucleotides was also used to determine the percentage of full-length product and
deduce overall coupling efficiency specific monomers. This technique gave consistent
results.
It was demonstrated for m1A-modified oligoribonucleotides [6][7] that cleavage
from solid support and further deprotection with 2m NH3 in MeOH for 60 h at room
temperature does not induce Dimroth rearrangement. According to this procedure, we
synthesized six oligodeoxynucleotide hairpins, ON1 – ON7, modified with m1dA
(Table). In a separate study, we found that the use of labile protecting groups for dA
(PAC ¼ phenylacetyl), dC (Ac), and dG (iPr-PAC) may reduce deprotection time in 2m
NH3 in MeOH or in 0.05m K2CO3 in MeOH to 24 h at room temperature.
Table. Thermal Stability of DNA Hairpin Oligodeoxynucleotides and Mass-Spectrometric Results
No.
Sequencea)
Mcalc
Mexper
Tm [8]
ON1
ON2
ON35
ON4
ON5
ON6
ON7
ON8
ON9
5’-GGACGTAATAGTCC
5’-GGACGTAATAGTCC
’-GGACGTAATAGTCC
5’-GGACGTAATAGTCC
5’-GGACGTAATAGTCC
5’-GGACGTAAATAGTCC
5’-GGACGTAAAATAGTCC
5’-GGACGTAATAGTCC
5’-GGACGTAATAGTCC
4287.9
4301.9
4301.9
4301.9
4301.9
4615.1
4942.34942.2
4301.9
4286.8
4302.2
4304.7
4301.4
4303.2
4613.8
45.7
4304.2
4304.5
53.3
53.8
55.6
56.9
n/d
48.5
46.6
56.7
4301.9
a) m1dA Residues are shown in bold, m6dA residues are underlined.
MALDI Mass spectra of synthesized oligonucleotides confirmed incorporation of
modified residues. A representative spectrum of the oligodeoxynucleotide ON6 is
shown in Fig. 1.
Convincing proofs of the presence of m1dA and the absence of a rearrangement
product were obtained from the HPLC analysis of nucleoside mixtures resulting from
venom phosphodiesterase and phosphatase digest of modified oligodeoxynucleotides.

Helvetica Chimica Acta – Vol. 90 (2007)
931
Fig. 1. MALDI-MS Analysis (in negative-ion mode) of modified oligodeoxynucleotide ON6
We observed formation of m1dA along with four natural 2’-deoxynucleosides (Fig. 2).
Due to the polar nature of this nucleoside, its retention time in RP-HPLC was shorter
than for natural deoxynucleosides. In contrast, the product of Dimroth rearrangement,
2’-deoxy-N6-methyladenosine (m6dA), is more hydrophobic as compared with dA, dC,
dG, and T, and has a considerably increased retention time. It should be mentioned,
however, that minor amount of m6dA was formed when using prolonged incubation
time (2 h) at pH 8.5 during phosphatase treatment. This product did not appear when a
shorter incubation time (30 min) was used.
The later circumstance clearly indicates the possibility of Dimroth rearrangement at
higher pH in aqueous solutions [13]. For example, in concentrated aqueous NH3, the
half-time of Dimroth rearrangement of m1A to m6A was 36 h at room temperature [7].
Under these conditions, nearly quantitative conversion of m1dA into m6dA was
observed after 7 d at room temperature (cf. Exper. Part). In spite of these results, in two
recent studies, for the deprotection of oligodeoxynucleotides containing m1dA,
concentrated aqueous NH3 at elevated temperatures was used [9][10]. It was reported
that deprotection of m1dA-modified oligodeoxynucleotides with aqueous NH3 at 378
for 4 h resulted in 8% Dimroth rearrangement [10]. We verified the compatibility of
the standard deprotection protocol (conc. aq. NH3, 558) with the synthesis of m1dA-

932
Helvetica Chimica Acta – Vol. 90 (2007)
Fig. 2. Enzymatic digest of m1dA-modified oligonucleotide ON4. Retention times for nucleosides: 6.3
(m1dA), 7.4 (dC), 10.2 (dG), 12.4 (T), and 14.8 min (dA). Analysis was performed on a Hypersil ODS
column (5 mm), 4.6 ꢁ 250 mm using 0.05m triethylammonium acetate (TEAA; pH 7.0) and linear
gradient of MeCN (0 – 25%) for 30 min. Flow rate was 1 ml/min.
modified oligodeoxynucleotides. Treatment of m1dA-modified oligonucleotides with
concentrated NH3 at 558 for 10 h resulted in ca. 80% conversion to oligodeoxynucleo-
tide containing 2’-deoxy-N6-methyladenosine. When the treatment time was increased
to 16 h, virtually complete Dimroth rearrangement was observed (Fig. 3). Enzymatic
hydrolysis of the new oligonucleotide product gave a mixture of natural nucleosides
and m6dA. There was no m1dA found in the digest (Fig. 4). As can be seen from Fig. 3,
Dimroth rearrangement proceeded cleanly without any visible degradation. Thus,
synthon 4 may be used successfully for the preparation of oligodeoxynucleotides
containing either m1dA or m6dA, depending on deprotection conditions.
Modified m1dA residues are expected to affect total charge of oligonucleotides. We
expected to find differences in mobility between all-natural and modified hairpins in
non-denaturing gel electrophoresis. Examples of reduced migration rate in polyacryl-
amide for amine-modified oligonucleotides have been reported [17]. Given this
consideration, we failed to observe the effect of m1dA addition. Instead, only minor
irregular variations of mobility were observed in denaturing gel electrophoresis at
pH 8.5 or 7.0 (data not shown).
2.2. Hybridization Properties of Modified Hairpins. We have studied thermal
denaturation of short oligodeoxynucleotide hairpins modified with m1dA (Table). Due
to the inability of m1dA to participate in Watson – Crick base pairing, its incorporation
into oligodeoxynucleotide forces the formation of unpaired, single-stranded fragments.
When introduced into the loop of a hairpin, m1dA is expected to provide some
stabilization effect as was observed for the m1A analog [7]. To compare the effect of
modification with previously reported data, we used the same set of oligonucleotides
used in that study. Two hairpins with enlarged loop were added to this set to follow the
influence of loop size and multiple m1dA residues on the hairpin stability.
As shown in the Table, the Tm values of modified hairpins with a single m1dA in the
loop are higher than that for the unmodified oligonucleotide. The stabilization effect

Helvetica Chimica Acta – Vol. 90 (2007)
933
Fig. 3. a) Oligonucleotide ON9 (m6dA) derived from oligonucleotide ON4 (m1dA) by ammonia
treatment for 16 h at 558. b) Coinjection of oligonucleotides ON4 and ON9. Percentages of MeCN in the
mobile phase are shown for each peak. Analysis was performed on a Hypersil ODS column (5 mm; 4.6 ꢁ
250 mm) using 0.05m TEAA (pH 7.0) and linear gradient of MeCN (0 – 25%) for 30 min. Flow rate was
1 ml/min.
Fig. 4. Enzymatic digest of m6dA modified oligonucleotide ON9. Retention times for nucleosides: 7.4
(dC), 10.2 (dG), 12.4 (T), 14.8 (dA), and 19.6 min (m6dA). Analysis was performed on a Hypersil ODS
column (5 mm; 4.6 ꢁ 250 mm) using 0.05m TEAA (pH 7.0) and linear gradient of MeCN (0 – 25%) for
30 min. Flow rate was 1 ml/min.
observed is very similar to that found for the m1A analog [6] and shows the same order
of stabilities for selected positions of modification. When placed in the duplex region,
m1dA prevented base pairing, and the Tm value for this hairpin could not be determined
as expected (ON5 in the Table).

934
Helvetica Chimica Acta – Vol. 90 (2007)
Increasing the loop size and introduction of an extra m1dA residue into the loop had
a negative effect on the duplex stability. A hairpin with a one-base (dA) larger loop
(ON6) decreased the Tm value by 4.58, while introduction of two bases (dA and m1dA)
induced destabilization by 7.68 (ON7). Thus, the negative effect associated with larger
loop size considerably exceeded stabilization induced by introduction of m1dA.
Transformation of 1-methyl oligodeoxynucleotides to N6-methyl oligodeoxynu-
cleotides via Dimroth rearrangement induced different changes in hybridization
behavior of hairpins, depending on the location of modified base. Hairpin ON8, derived
from oligonucleotide ON5, contained a m6dA residue in the stem region and was able
to form an intramolecular duplex with a Tm value of 46.68. A similar destabilization
effect was observed for DNA duplexes containing m6dA [18]. When present in the loop
(ON9, originally ON4), the m6dA residue had a minor effect on hairpin stability (Tm
56.78 vs. 56.98 in ON4; Table).
3. Conclusions. – We have developed an efficient method to introduce the m1dA
residue into synthetic oligodeoxynucleotides via the corresponding protected phos-
phoramidite 4. Anhydrous deprotection conditions should be used for m1dA-modified
oligonucleotides to avoid Dimroth rearrangement. The synthon 4 may be used
successfully for the preparation of oligodeoxynucleotides containing m1dA or m6dA.
Experimental Part
General. Column chromatography (CC): silica gel (0.06 – 0.20 mm). TLC: Kieselgel 260 F (Merck);
eluents: CH2Cl2/MeOH 9 :1 (A); CH2Cl2/MeOH 1:1 (B); CH2Cl2/MeOH/hexane 9 :1:10 (C); AcOEt/
hexane 2 :1 (D); detection by UV light. UV measurements and thermal denaturation studies were
conducted with a Shimazu UV-160A spectrophotometer. NMR Spectra: Bruker AMX-400 spectrometer;
at 300 K; chemical shifts d in ppm were recorded relative to the solvent signals (1H and 13C) and relative
external ref. ¼ H4PO3 (capil.) (31P); coupling constants J in Hz. The signals were assigned using double-
resonance techniques. MALDI-MS of oligonucleotides were acquired on a Bruker Reflex IV mass
spectrometer in negative-ion mode with hydroxypicolinic acid as a matrix.
Hydroiodide Salt of 2’-Deoxy-1-methyladenosine (m1dA) was obtained as monohydrate [16].
1H-NMR (D2O): 8.47 (s, HꢀC(2,8)); 6.49 (t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.7, HꢀC(1’)); 4.64 (ddd, J(3’,2’a) ¼
6.3, J(3’,2’b) ¼ 4.2, J(3’,4’) ¼ 3.5, HꢀC(3’)); 4.13( ddd, J(4’,5’b) ¼ 5.0, J(4’,5’a) ¼ 3.7, J(4’,3’) ¼ 3.5,
HꢀC(4’)); 3.89 (s, MeN); 3.88 (dd, J(5’a,5’b) ¼ ꢀ12.5, J(5’a,4’) ¼ 3.7, HꢀC(5’a)); 3.75 (dd, J(5’b,5’a) ¼
ꢀ12.5, J(5’b,4’) ¼ 5.0, HꢀC(5’b)); 2.85 (ddd, J(2’a,2’b) ¼ ꢀ14.0, J(2’a,1’) ¼ 6.7, J(2’a,3’) ¼ 6.3, HꢀC(2’a));
2.60 (ddd, J(2’b,2’a) ¼ ꢀ14.0, J(2’b,1’) ¼ 6.7, J(2’b,3’) ¼ 4.2, HꢀC(2’b)). 13C-NMR (D2O): 151.77 (C(6));
148.22 (C(2)); 147.17 (C(4)); 143.54 (C(8)); 123.02 (C(5)); 88.17 (C(1’)); 85.46 (C(4’)); 71.45 (C(3’));
62.02 (C(5’)); 39.77 (C(2’)); 38.21 (MeN).
2’-Deoxy-N6-methyladenosine (m6dA). A soln. of 411 mg (1 mmol) of HI salt of m1dA in 10 ml of
conc. aq. NH3 was kept at 208. According to TLC in system B, the reaction was complete after 7 d. The
starting compound disappeared for 10 or 16 h at 558. The product in system B moved faster (Rf 0.65) than
the starting compound (Rf 0.06). The solvent was evaporated, and the residue was purified on a column
with 100 ml of Dowex-1 (OHꢀ form). The column was washed consequently with 200 ml of H2O and
200 ml of 10% aq. EtOH, and eluted with 20% of aq. EtOH. Product-containing fractions were
combined, evaporated to dryness, and the residue was lyophilized from H2O to give the product as a
powder. Yield: 180 mg (68%). UV Spectra at different pH values were identical with those published in
[16]. 1H-NMR (D2O): 8.22 (s, HꢀC(8)); 8.10 (s, HꢀC(2)); 6.42 (t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.7, HꢀC(1’));
4.72 (ddd, J(3’,2’a) ¼ 6.7, J(3’,2’b) ¼ 3.4, J(3’,4’) ¼ 2.0, HꢀC(3’)); 4.27 (ddd, J(4’,5’b) ¼ 4.4, J(4’,5’a) ¼ 3.1,
J(4’,3’) ¼ 2.0, HꢀC(4’)); 3.95 (dd, J(5’a,5’b) ¼ ꢀ12.6, J(5’a,4’) ¼ 3.1, HꢀC(5’a)); 3.88 (dd, J(5’b,5’a) ¼

Helvetica Chimica Acta – Vol. 90 (2007)
935
ꢀ12.6, J(5’b,4’) ¼ 4.4, HꢀC(5’b)); 3.08 (s, MeN); 2.85 (ddd, J(2’a,2’b) ¼ ꢀ14.0, J(2’a,3’) ¼ 6.7, J(2’a,1’) ¼
6.7, HꢀC(2’a)); 2.64 (ddd, J(2’b,2’a) ¼ ꢀ14.0, J(2’b,1’) ¼ 6.7, J(2’b,3’) ¼ 3.4, HꢀC(2’b)). 13C-NMR (D2O):
155.28 (C(6)); 152.77 (C(2)); 147.73(C(4)); 139.82 (C(8)); 119.53(C(5)); 88.00 (C(1 ’)); 85.12 (C(4’));
71.84 (C(3’)); 62.32 (C(5’)); 39.80 (C(2’)); 27.89 (MeN).
2’-Deoxy-5-O-[(4-methoxyphenyl)(diphenyl)methyl]adenosine (1). To 2’-deoxyadenosine (2.5 g,
10 mmol), dried two times by evaporation with anh. pyridine (2 ꢁ 20 ml) and suspended in dry pyridine
(20 ml), monomethoxytrityl (MMTr) chloride (3.7 g, 12 mmol) was added, and the mixture was stirred
for 16 h at 208. Then, CH2Cl2 (100 ml) and H2O (50 ml) were added, and the org. layer was washed
consequently with 10% aq. NaHCO3 (30 ml) and H2O (20 ml). The org. layer was dried (Na2SO4),
evaporated in vacuo and co-evaporated with toluene (3 ꢁ 30 ml). The residue was purified by CC on
silica gel (50 g). The column was washed with CH2Cl2 (200 ml) and CH2Cl2/MeOH 97:3(200 ml), and
then eluted with CH2Cl2/MeOH 95 :5. The corresponding fractions were evaporated to dryness in vacuo
to give 6 (3.5 g, 67%). Foam. Rf 0.35 (A). 1H-NMR (400 MHz, CDCl3): 8.25 (s, HꢀC(8)); 7.94 (s,
HꢀC(2)); 7.40 – 7.19 (m, 12 arom. H); 6.79 (d, J ¼ 8.8, 2 H, PhOMe); 6.43( t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.3,
HꢀC(1’)); 6.02 (br. s, NH2); 4.66 (ddd, J(3’,2’a) ¼ 6.5, J(3’,2’b) ¼ 4.0, J(3’,4’) ¼ 2.9, HꢀC(3’)); 4.17 (ddd,
J(4’,5’b) ¼ 5.0, J(4’,5’a) ¼ 4.7, J(4’,3’) ¼ 2.9, HꢀC(4’)); 3.76 (s, MeO); 3.41 (dd, J(5’a,5’b) ¼ ꢀ10.3,
J(5’a,4’) ¼ 4.7, HꢀC(5’a)); 3.38 (dd, J(5’b,5’a) ¼ ꢀ10.3, J(5’b,4’) ¼ 5.0, HꢀC(5’b)); 2.78 (ddd, J(2’a,2’b) ¼
ꢀ13.4, J(2’a,3’) ¼ 6.5, J(2’a,1’) ¼ 6.3, HꢀC(2’a)); 2.52 (ddd, J(2’b,2’a) ¼ ꢀ13.4, J(2’b,1’) ¼ 6.3, J(2’b,3’) ¼
4.0, HꢀC(2’b)). 13C-NMR (CDCl3): 158.85 (Ph); 155.63(C(2)); 153.03(C(6)); 149.62 (C(4)); 144.20
(Ph); 139.00 (C(8)); 135.33, 130.45, 128.50, 128.02, 127.22 (Ph); 120.08 (C(5)); 113.37 (Ph); 87.04 (CꢀO);
86.25 (C(1’)); 84.51 (C(4’)); 72.47 (C(3’)); 63.98 (C(5’)); 55.35 (OMe); 40.51 (C(2’)).
2’-Deoxy-5-O-[(4-methoxyphenyl)(diphenyl)methyl]-1-methyladenosine (2). A mixture of 2.62 g
(5 mmol) of 1 and 1.25 ml (20 mmol) of MeI in N,N-dimethylacetamide (8 ml) was stirred in the dark for
16 h at 208. A slightly yellow soln. was diluted with 100 ml of CH2Cl2 and was washed consequently with
H2O (20 ml), 10% aq. NaHCO3 (20 ml), 5% aq. Na2S2O3 (30 ml), and H2O (2 ꢁ 10 ml). The org. layer
was dried (Na2SO4), evaporated in vacuo (bath temp. < 308) to a volume of ca. 5 ml, and powdered into
hexane/Et2O 2 :1 (200 ml), and the mixture was allowed to stand at 08 overnight. The hygroscopic
precipitate was filtered, washed with the same mixture (10 ml), and dried in the vacuum to give 2 (1.94 g,
72%). Slightly yellow powder. Rf 0.20 (B). 1H-NMR (400 MHz, CDCl3): 7.70 (s, HꢀC(8)); 7.56 (s,
HꢀC(2)); 7.41 – 7.21 (m, 12 arom. H); 6.81 (d, J ¼ 8.7, 2 H, PhOMe); 6.26 (t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.5,
HꢀC(1’)); 4.61 (ddd, J(3’,2’a) ¼ 6.5, J(3’,2’b) ¼ 4.3, J(3’,4’) ¼ 2.8, HꢀC(3’)); 4.11 (ddd, J(4’,5’b) ¼ 5.3,
J(4’,5’a) ¼ 5.0, J(3’,4’) ¼ 2.8, HꢀC(4’)); 3.78 (s, MeO); 3.54 (s, MeN); 3.42 (dd, J(5’a,5’b) ¼ ꢀ10.0,
J(5’a,4’) ¼ 5.0, HꢀC(5’a)); 3.32 (dd, J(5’b,5’a) ¼ ꢀ10.0, J(5’b,4’) ¼ 5.3, HꢀC(5’b)); 2.72 (ddd,
J(2’a,2’b) ¼ ꢀ13.4, J(2’a,3’) ¼ 6.5, J(2’a,1’) ¼ 6.5, HꢀC(2’a)); 2.46 (ddd, J(2’b,2’a) ¼ ꢀ13.4, J(2’b,1’) ¼
6.5, J(2’b,3’) ¼ 4.3, HꢀC(2’b)). 13C-NMR (CDCl3): 158.71 (Ph); 155.28 (C(6)); 147.08 (C(2)); 144.12
(Ph); 144.00 (C(4)); 141.72 (C(8)); 136.68, 135.17, 130.36, 128.34, 127.89, 127.07 (Ph); 123.68 (C(5));
113.21 (Ph); 86.87 (CꢀO); 85.96 (C(1’)); 84.15 (C(4’)); 72.53(C(3 ’)); 63.92 (C(5’)); 55.24 (MeO); 40.21
(C(2’)); 35.53 (MeN).
N6-(Chloroacetyl)-2’-deoxy-5-O-[(4-methoxyphenyl)(diphenyl)methyl]-1-methyladenosine (3). To a
cold (08) soln. of 1.61 g (3mmol) of 2 in a mixture of 3ml of dry pyridine and 30 ml of dry CH 2Cl2, 2.05 g
(12 mmol) of (ClCH2CO)2O was added, and the soln. was kept at 08 for 1 h. The soln. was diluted with
CH2Cl2 (50 ml), and H2O (30 ml) was added. The org. layer was washed consequently with 10% aq.
NaHCO3 (20 ml) and H2O (20 ml), dried (Na2SO4), and evaporated in vacuo (bath temp. < 308) to a
volume of ca. 5 ml. To the residue, 2m NH3 in MeOH (15 ml) was added, and the brown soln. was kept at
08 for 20 min. The mixture was evaporated in vacuo (bath temp. < 308) and co-evaporated with toluene
(2 ꢁ 30 ml). The residue was purified by CC on silica gel (100 g). The column was washed with CH2Cl2
(200 ml), CH2Cl2/MeOH 99 :1 (200 ml) and 98 :2 (200 ml), and then eluted with CH2Cl2/MeOH 97:3.
The corresponding fractions were evaporated to dryness in vacuo to give 3 (1.44 g, 78%). Foam. Rf 0.60
(A). 1H-NMR (CDCl3): 7.82 (s, HꢀC(8)); 7.81 (s, HꢀC(2)); 7.40 – 7.21 (m, 12 arom. H); 6.81 (d, J ¼ 8.7,
2 H, PhOMe); 6.29 (t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.4, HꢀC(1’)); 4.60 (ddd, J(3’,2’a) ¼ 6.3, J(3’,2’b) ¼ 4.2,
J(3’,4’) ¼ 3.0, HꢀC(3’)); 4.40 (s, ClCH2); 4.12 (ddd, J(4’,5’b) ¼ 5.3, J(4’,5’a) ¼ 5.0, J(4’,3’) ¼ 3.0,
HꢀC(4’)); 3.79 (s, MeO); 3.60 (s, MeN); 3.40 (dd, J(5’a,5’b) ¼ ꢀ10.0, J(5’a,4’) ¼ 5.0, HꢀC(5’a)); 3.31
(dd, J(5’b,5’a) ¼ ꢀ10.0, J(5’b,4’) ¼ 5.3, HꢀC(5’b)); 2.65 (ddd, J(2’a,2’b) ¼ ꢀ13.2, J(2’a,1’) ¼ 6.4,

936
Helvetica Chimica Acta – Vol. 90 (2007)
J(2’a,3’) ¼ 6.3, HꢀC(2’a)); 2.47 (ddd, J(2’b,2’a) ¼ ꢀ13.2, J(2’b,1’) ¼ 6.4, J(2’b,3’) ¼ 4.2, HꢀC(2’b)).
13C-NMR (CDCl3): 178.15 (C¼O); 158.73(Ph); 147.29 (C(2)); 146.52 (C(6)); 145.13(C(4)); 143.98
(Ph); 138.74 (C(8)); 135.08, 130.36, 128.30, 127.91, 127.09 (Ph); 122.32 (C(5)); 113.24 (Ph); 86.89 (CꢀO);
86.02 (C(1’)); 84.21 (C(4’)); 72.41 (C(3’)); 63.80 (C(5’)); 55.28 (MeO); 45.98 (ClCH2); 40.18 (C(2’));
36.68 (MeN).
N6-(Chloroacetyl)-2’-deoxy-5-O-[(4-methoxyphenyl)(diphenyl)methyl]-1-methyladenosine-3’-(2-cyano-
ethyl N,N-diisopropylphosphoramidite) (4). Compound 3 (0.5 g, 0.81 mmol) was dissolved in 2 ml of anh.
MeCN. 1H-Tetrazole (0.06 g, 0.85 mmol), dry pyridine, (70 ml, 0.85 mmol), and molecular sieves (4 Š;
ca. 5% by volume) were added. After 30 min, 2-cyanoethyl N,N,N’,N’-tetraisopropylphosphoramidite
(0.25 ml, 0.83mmol) was added to the mixture under intensive stirring. Phosphitylation was completed in
15 min as evidenced by TLC (system C). The reaction was quenched with cold sat. NaHCO3 (100 ml),
and the mixture was extracted with AcOEt (100 ml). The org. layer was filtered through Na2SO4, and the
solvent was removed in vacuo to give a white foam. Phosphoramidite 4 was purified by CC on silica gel
(100 g) in hexane/AcOEt mixture (500 ml; gradient from 1:1 to 1:3in the presence of 3% Et 3N) to yield
a mixture of diastereoisomers (520 mg, 79%). Rf 0.50 (C), 0.21, and 0.32 (D). 1H-NMR (CDCl3):
selected signals: 7.85 (s), 7.88 (s) (HꢀC(8)); 7.78 (s, HꢀC(2)); 7.20 – 7.40 (m, 12 arom. H); 6.80 (m, 2 H,
PhOMe); 6.29 (t, J(1’,2’a) ¼ J(1’,2’b) ¼ 6.8, HꢀC(1’)); 4.39 (s, CH2Cl); 3.79 (br. s, MeO); 3.60 (br. s,
MeN). 31P-NMR (CDCl3): 150.65; 150.10.
Oligonucleotide Synthesis and Purification. Oligodeoxynucleotides were synthesized either with an
ASM-102U DNA synthesizer (Biosset Ltd., Russia) or on an ABI 394 (Applied Biosystems, Forster City,
CA, USA). Oligodeoxynucleotides were synthesized in DMTr-ON mode by phosphoramidite chemistry
with standard protecting groups at the 0.2- or 1.0-mmol scales. We used Glen Research reagents (Sterling,
VA) for these synthesis.
For estimation of m1dA average stepwise-coupling yield (ASY), two methods have been used. First,
trityl solns. for preceding and following base couplings were collected, and their absorbance at 495 nm
was measured. Second, we added m1dA to a T6 oligonucleotide, or synthesized [T-T-m1dA]3-T6 and
analyzed the resulting oligomers by RP-HPLC on Spherisorb ODS2 column (5 mm; 4.6 ꢁ 150 mm) using
0.1m triethylammonium acetate (TEAA; pH 7.0) and a linear gradient of MeCN (3– 40%, 30 min for
ꢁtrityl-onꢂ oligonucleotides and 3– 25%, 30 min for ꢁtrityl-offꢂ oligonucleotides). Flow rate was 1 ml/min.
Elution was monitored using a PDA detector, and UV spectra of the eluted peaks were recorded.
Modified oligodeoxynucleotides were deprotected in 2m methanolic NH3 for 60 h at 258. When labile
protecting groups were used (PAC-dA, Ac-dC, and iPr-PAC-dG), deprotection time could be reduced to
24 h. Alternatively, 50 mm K2CO3 in MeOH for 24 h was tested with success on oligomers containing T
and m1dA. This last deprotection protocol must also be used in combination with PAC-dA, Ac-dC, and
iPr-PAC-dG.
For deprotection experiments, we synthesized the oligo m1dA-T6 ꢁtrityl-offꢂ using dT-Q support [19]
(Glen Research – 21-2030-10) which can be cleaved in 60 s. Zero-time sample was cleaved by adding
0.5 ml of 50 mm K2CO3/MeOH (Glen Research – 60-4600-30) to CPG-bound oligomer. The soln. was
then neutralized by adding 0.5 ml of 2m TEAA (Glen Research – 60-4110). Deprotection rates were
determined by removing aliquots at various times, evaporating NH3/MeOH, resuspending in 0.1m TEAA,
and analyzing by RP-HPLC as described above. Progression of the hydrolysis of N6-chloroacetyl
protection of m1dA resulted in a change in retention time and was easy to follow.
Purification of oligodeoxynucleotides was carried out on a Hypersil ODS column (5 mm; 4.6 ꢁ
250 mm) using 0.05m TEAA (pH 7.0) and linear gradient of MeCN (10 – 50%, 30 min for DMTr-
protected oligodeoxynucleotides and 0 – 25%, 30 min for fully deblocked oligodeoxynucleotides). Flow
rate was 1 ml/min. Removal of the 5’-O-DMTr group was achieved by treatment with 2% aq. TFA for
1 min, followed by Et3N neutralization and precipitation with 2% LiClO4 in acetone. Typically, 10 –
15 o.u. (260) of oligodeoxynucleotide was obtained in 0.2-mmol synthesis.
Enzymatic Digest of Modified Oligonucleotides. Oligodeoxynucleotide (0.2 o.u.) was dissolved in
100 ml of buffer containing 50 mm Tris · HCl (pH 7.5), 50 mm NaCl, and 7 mm MgCl2. The soln. was
treated with 1 unit of venom phosphodiesterase for 16 h at 378. Then, 5 ml of 0.5m Tris · HCl (pH 9.0) was
added. The resulting soln. (pH 8.5) was treated with 1 unit of shrimp alkaline phosphatase for 30 min at
378. The mixture was then diluted to 1 ml, filtered, and analyzed by RP-HPLC. Analysis was performed

Helvetica Chimica Acta – Vol. 90 (2007)
937
on a Hypersil ODS column (5 mm; 4.6 ꢁ 250 mm) using 0.05m TEAA (pH 7.0) and linear gradient of
MeCN (0 – 25%) for 30 min. Flow rate was 1 ml/min. Quantification of the constituents was achieved on
the basis of peak areas, which were divided by the extinction coefficients of the nucleoside (260-nm
values: m1dA, 12000; dA, 15000; dC, 7500; dG, 12500; dT, 8500). Retention times for m1dA, dC, dG, T,
dA, and m6dA were 6.3, 7.4, 10.2, 12.4, 14.8, and 19.6 min, resp.
Oligonucleotide Dimroth Rearrangement. Modified oligodeoxynucleotide (0.2 o.u.) was heated in
1 ml of conc. aq. NH3 for 10 or 16 h at 558. NH3 soln. was then evaporated in vacuo, the residue was
dissolved in H2O and analyzed by RP-HPLC as described above for fully deprotected oligodeoxynu-
cleotides. Retention times for dC, dG, T, dA, and m6dA were 7.4, 10.2, 12.4, 14.8, and 19.6 min, resp.
Denaturing Gel Electrophoresis. Denaturing gel electrophoresis was performed in 20% polyacryl-
amide, containing 7m urea. Running buffer was either 0.1m Tris · borate (pH 8.5) or Tris · HCl (pH 7.0).
Thermal Denaturation Profiles. Absorbance vs. temp. profiles were determined with a Shimadzu
UV160A spectrophotometer equipped with a water-jacketed cell holder. Dry N2 gas was flushed through
the cuvette chamber to prevent moisture condensation at low temp. Melting experiments were carried
out at 260 nm and with the oligonucleotide concentration at 1 mm. The solns. were heated from 0 to 908 at
a rate 0.28/min. Melting temps. were calculated from DH and DS values. Thermodynamic parameters
were obtained by fitting of experimental curves using two-state intramolecular model with sloping base
lines.
Financial support of the Russian Foundation for Basic Research, Program of the Russian Academy of
Sciences ꢀMolecular and Cellar Biologyꢁ, and KUL Research Council is acknowledged. The authors would
like to thank Steve Strickland for skillful technical assistance.
REFERENCES
[1] P. F. Agris, Prog. Nucleic Acid Res. Mol. Biol. 1996, 53, 79.
[2] M. Sprinzl, C. Horn, M. Brown, A. Ioudovitch, S. Steinberg, Nucleic Acids Res. 1998, 26, 148.
´
[3] M. Helm, R. Giege, C. Florentz, Biochemistry 1999, 38, 13338.
[4] J. Anderson, L. Phan, A. G. Hinnebusch, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 5173.
[5] H. Sierzputowska-Gracz, H. D. Gopal, P. F. Agris, Nucleic Acids Res. 1986, 14, 7783.
[6] S. N. Mikhailov, J. Rozenski, E. V. Efimtseva, R. Busson, A. Van Aerschot, P. Herdewijn, Nucleic
Acids Res. 2002, 30, 1124.
[7] S. N. Mikhailov, E. V. Efimtseva, A. A. Rodionov, A. A. Shelkunova, J. Rozenski, G. Emmerechts,
G. Schepers, A. Van Aerschot, P. Herdewijn, Chem. Biodiv. 2005, 2, 1153.
[8] B. Sedgwick, Nat. Rev. Mol. Cell. Biol. 2004, 5, 148.
[9] J. Ringvoll, L. M. Nordstrand, C. B. Vagbo, V. Talstad, K. Reite, P. Aas, K. H. Lauritzen, N. B.
Liabakk, A. Bjork, R. W. Doughty, P. O. Falnes, H. E. Krokan, A. Klungland, EMBOJ. 2006, 25, 2189.
[10] J. C. Delaney, J. M. Essigmann, Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14051.
[11] L. Q. Sun, M. J. Cairns, E. G. Saravolac, A. Baker, W. L. Gerlach, Pharmacol. Rev. 2000, 52, 325.
[12] S. K. Silverman, Nucleic Acids Res. 2005, 33, 6151.
[13] J. B. Macon, R. Wolfenden, Biochemistry 1968, 7, 3453.
[14] Y. Takeuchi, I. Tazawa, Y. Inoue, Bull. Chem. Soc. Jpn. 1982, 55, 3598.
[15] P. F. Agris, H. Sierzputowska-Gracz, C. Smith, Biochemistry 1986, 25, 5126.
[16] J. W. Jones, R. K. Robins, J. Am. Chem. Soc. 1963, 85, 193.
[17] S. V. Kochetkova, E. N. Timofeev, E. A. Korobeinikova, N. A. Kolganova, V. L. Florentiev,
Tetrahedron 2001, 57, 10287.
[18] A. Ono, T. Ueda, Nucleic Acids Res. 1987, 15, 219.
[19] R. T. Pon, S. Y. Yu, Tetrahedron Lett. 1997, 38, 3327.
Received February 7, 2007