Base-pairing, tautomerism, and mismatch discrimination of 7-halogenated 7-deaza-2-deoxyisoguanosine: Oligonucleotide duplexes with parallel and antiparallel chain orientation

Article abstract of DOI:10.1021/ja0425785

Oligonucleotides containing 2-deoxyisoguanosine (1, iG_d), 7-deaza-2-deoxyisoguanosine (2, c⁷iG_d), and its 7-halogenated derivatives 3 and 4 were synthesized on solid phase using the phosphoramidite building blocks 5-7. The

Full text of DOI:10.1021/ja0425785

Published on Web 05/06/2005
Base-Pairing, Tautomerism, and Mismatch Discrimination of
7-Halogenated 7-Deaza-2′-deoxyisoguanosine:
Oligonucleotide Duplexes with Parallel and Antiparallel Chain
Orientation
Frank Seela,* Xiaohua Peng, and Hong Li
Contribution from the Laboratorium fu¨r Organische und Bioorganische Chemie, Institut fu¨r
Chemie, UniVersita¨t Osnabru¨ck, Barbarastrasse 7, D-49069 Osnabru¨ck, Germany, and
Laboratory of Bioorganic and Biophysical Chemistry, Center for Nanotechnology, GieVenbecker
Weg 11, 48149 Mu¨nster, Germany
Received December 10, 2004; E-mail: Frank.Seela@uni-osnabrueck.de
Abstract: Oligonucleotides containing 2′-deoxyisoguanosine (1, iGd), 7-deaza-2′-deoxyisoguanosine (2,
c7iGd), and its 7-halogenated derivatives 3 and 4 were synthesized on solid phase using the phosphoramidite
building blocks 5-7. The hybridization properties of oligonucleotides were studied on duplexes with parallel
and antiparallel chain orientation. It was found that the 7-halogenated nucleoside analogues 3 and 4 enhance
the duplex stability significantly in both parallel (ps) and antiparallel (aps) DNA. Moreover, the halogenated
nucleosides shift the tautomeric keto-enol equilibrium strongly toward the keto form, with KTAUT [keto]/
[enol] ≈ 104 coming close to that of 2′-deoxyguanosine (104-105), while the nonhalogenated 7-deaza-2′-
deoxyisoguanosine 2 shows a KTAUT of around 2000 and the enol concentration of 1 is 10% in aqueous
solution. Consequently, nucleosides 3 and 4 show a much better mismatch discrimination against dT than
compound 1 or 2 in antiparallel as well as in parallel DNA. 3 and 4 are expected to increase the selectivity
of base incorporation opposite to isoCd in the form of triphosphates or in the polymerase-catalyzed reaction
in comparison to 1 or 2.
Introduction
as inducer of ps-DNA. On the other hand, in aps-DNA, iGd
forms a stable Watson-Crick base pair with 2′-deoxyisocytidine
The polymorphic nature of double-stranded DNA is well-
established.1 The antiparallel strand orientation is the common
feature of the three major families of A-, B-, and Z-DNA. A
recent addition to the well-known DNA families is that of
parallel-stranded (ps) DNAs.2 This type of duplex formation
may offer new opportunities for designing new oligonucleotide
hybridization probes, antisense constructs, or interference RNA
mimics.3 ps-DNA can be constructed from oligonucleotides
containing reverse Watson-Crick adenine-thymine base pair2
and isoguanine-cytosine or guanine-isocytosine pairs instead
of the canonical guanine-cytosine pair.4 The thermodynamic
stability of ps-DNA when formed exclusively by dA-dT base
pairs is lower than that of the corresponding aps-DNA,2 while
2′-deoxyisoguanosine (1, iGd)4f,g can stabilize ps-DNA by
forming three hydrogen bonds with 2′-deoxycytidine (dC).4b,c
Therefore, 2′-deoxyisoguanosine has attracted much attention
(iCd), which can be incorporated into duplex DNA by poly-
merases, thereby expanding the genetic alphabet from four to
six letters.5,6 However, it has been shown that the incorporation
of 2′-deoxyisoguanosine triphosphates opposite to iCd in DNA
causes the formation of iGd-dT mismatches during replica-
tion.5a,7,8 This has been correlated to a particular property of
(4) (a) Seela, F.; Gabler, B.; Kazimierczuk, Z. Collect. Czech. Chem. Commun.
1993, 58, 170-173. (b) Sugiyama, H.; Ikeda, S.; Saito, I. J. Am. Chem.
Soc. 1996, 118, 9994-9995. (c) Seela, F.; Wei, C. HelV. Chim. Acta 1997,
80, 73-85. (d) Seela, F.; He, Y.; Wei, C. Tetrahedron 1999, 55, 9481-
9500. (e) Yang, X.-L.; Sugiyama, H.; Ikeda, S.; Saito, I.; Wang, A. H.-J.
Biophys. J. 1998, 75, 1163-1171. (f) Kazimierczuk, Z.; Mertens, R.;
Kawczynski, W.; Seela, F. HelV. Chim. Acta 1991, 74, 1742-1748. (g)
Seela, F.; Gabler, B. HelV. Chim. Acta 1994, 77, 622-630. (h) Seela, F.;
Fro¨hlich, T. Collect. Czech. Chem. Commun. 1993, 58, 183-186. (i) Seela,
F.; Wei, C. Collect. Czech. Chem. Commun. 1996, 61, 114-115. (j) Li,
H.; Peng, X.; Seela, F. Bioorg. Med. Chem. Lett. 2004, 14, 6031-6034.
(5) (a) Switzer, C.; Moroney, S. E.; Benner, S. A. J. Am. Chem. Soc. 1989,
111, 8322-8323. (b) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner,
S. A. Nature 1990, 343, 33-37. (c) Bain, J. D.; Switzer, C.; Chamberlin,
A. R.; Benner, S. A. Nature 1992, 356, 537-539. (d) Roberts, C.; Bandaru,
R.; Switzer, C. J. Am. Chem. Soc. 1997, 119, 4640-4649. (e) Rice, K. P.;
Chaput, J. C.; Cox, M. M.; Switzer, C. Biochemistry 2000, 39, 10177-
10188. (f) Tor, Y.; Dervan, P. B. J. Am. Chem. Soc. 1993, 115, 4461-
4467.
(1) (a) DNA Topology and Its Biological Effects; Cozzarelli, N. R., Wang, J.
C., Eds.; Cold Spring Harbor Laboratory Press: New York, 1990. (b) DNA
Structure and Function; Sinden, R. R., Ed.; Academic Press: New York,
1994.
(2) van de Sande, J. H.; Ramsing, N. B.; Germann, M. W.; Elhorst, W.; Kalisch,
B. W.; von Kitzing, E.; Pon, R. T.; Clegg, R. C.; Jovin, T. M. Science
1988, 241, 551-557.
(3) (a) Rippe, K.; Fritsch, V.; Westhof, E.; Jovin, T. M. EMBO J. 1992, 11,
3777-3786. (b) Marfurt, J.; Leumann, C. Angew. Chem., Int. Ed. 1998,
37, 175-177.
(6) Johnson, S. C.; Sherrill, C. B.; Marshall, D. J.; Moser, M. J.; Prudent, J.
R. Nucleic Acids Res. 2004, 32, 1937-1941.
(7) Switzer, C. Y.; Moroney, S. E.; Benner, S. A. Biochemistry 1993, 32,
10489-10496.
(8) (a) Kamiya, H.; Kasai, H. FEBS Lett. 1996, 391, 113-116. (b) Kamiya,
H.; Ueda, T.; Ohgi, T.; Matsukage, A.; Kasai, H. Nucleic Acids Res. 1995,
23, 761-766.
9
10.1021/ja0425785 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 7739-7751
7739

A R T I C L E S
Seela et al.
Scheme 1. Structure of Nucleosides and Phosphoramidite Building
Blocks
the isoguanine base that adopts the form of an enol tautomer
much more easily than the canonical DNA nucleobases.9,10
Earlier, Shugar and co-worker examined the properties of iGd
by optical methods and concluded that although iGd favors the
keto (N1-H) form in aqueous solution, the enol tautomer is
present to a significant amount (10%).10 Crystallographic data
of a Watson-Crick iGd-dT pair with isoguanine in the enol
form observed by Wang and co-workers provide the most direct
evidence to date in support of iGd infidelity resulting from the
enol-keto tautomerism.11 This situation makes the fidelity of
replication and translation inherently more difficult to control
for any genetic system bearing iGd. Thus, the ambiguity of base
pairing caused by tautomerism limits the use of iGd as a “letter
in the genetic alphabet” and can cause mutagenic events during
replication.
To find an ideal candidate as a new member of “letters of
the genetic alphabet”, two special targets for a modified
nucleoside should be met: (i) a high thermodynamic stability
of the iGd-dC (or iGd-iCd) base pair and (ii) a neglectible
content of the iGd enol form decreasing the probability of
mutagenesis during replication. Since the keto-enol equilibrium
of iGd depends on the polarity of its microenvironment, one
can expect that the modification of nucleobases would influence
this equilibrium, resulting in an alteration of the coding
properties. Earlier, we reported on the replacement of the
imidazole moiety of 1 by a pyrrole ring leading to 7-deaza-2′-
deoxyisoguanosine (2)12 (purine numbering is used throughout
the discussion and systematic numbering is used in the experi-
ment part). Recently, Benner and co-workers observed that the
keto form in compound 2 is more favored (KTAUT ) [keto]/
[enol] ≈ 103, room temperature, aqueous solution) than that in
2′-deoxyisoguanosine (KTAUT ≈ 10).13 Nevertheless, experi-
mental data obtained from DNA melting curves indicate that
the thermodynamic stability of duplexes containing 2-dC pairs
are indeed somewhat lower than that containing an iGd-dC
pair,12a,14a which in turn is less stable than the dG-dC pair in
aps-DNA. Moreover, compound 2 shows base-pair ambiguity
as it is observed for 1.12a,14b
of 7-halogenated 7-deaza-2′-deoxyisoguanosine analogues 3
(Cl7c7iGd) and 4 (Br7c7iGd) bearing halogen substituents at C-7
with the potential to enhance duplex stability.18 The introduction
of halogen substituents might also affect the tautomerism of 3
and 4. Herein, nucleosides 2-4 have been converted into the
phosphoramidites building blocks 5-7, which were employed
in solid-phase oligonucleotide synthesis (Scheme 1). The
tautomerism of nucleosides 3 and 4 was studied, and base-
pairing properties (stability and selectivity) were investigated
in DNA with parallel and antiparallel chain orientation.
Results and Discussion
Monomers. The 7-deazaisoguanine nucleosides 2-4 were
prepared from the corresponding 7-deazapurin-2,6-diamine
nucleosides 8-10 (Scheme 2).12a,18 The next task was to choose
protecting groups for the nucleobase amino and oxo functions
that could survive in automated oligonucleotide synthesis and
remain labile to postsynthetic treatment with concentrated
aqueous ammonia. As the isobutyryl (iBu) residue had already
been successfully employed in the case of compound 2,12a the
initial choice for the 6-amino protection of nucleosides 3 and 4
was the isobutyryl residue. Unfortunately, during the stage of
removing the transient silyl OH-protecting group by aqueous
ammonia the iBu residue was also lost. Only 10% of the desired
product 12 was obtained, and 77% of starting material 11 was
recovered (Scheme 2). Apparently, the 7-bromo substituent of
4 increases the acidic character of the 6-amino group, thereby
destabilizing the protecting group. This is underlined by the half-
life for deprotection of 12 (τ ) 4 min) compared to 60 min in
the case of compound 212a (Table 1 and Supporting Information).
Earlier, the N,N-dimethylaminoethylidene residue was chosen
for protecting the 6-amino group of 2′-deoxyisoguanosine.4c
Thus, the amino groups of compounds 2-4 were protected with
the N,N-dimethylaminoethylidene residue using dimethylaceta-
mide dimethylacetal/MeOH to yield the amino-protected com-
pounds 13a-c in 89-91% yield (Scheme 3). It has been shown
that in the case of isoguanosine derivatives, the 2-oxo-
unprotected phosphoramidite resulted in inefficient coupling
yields. The protection of the 2-oxo group with a diphenylcar-
bamoyl residue resulted in significantly higher coupling yields
during oligonucleotide synthesis.12a Likewise, the nucleosides
13a-c were treated with diphenylcarbamoyl chloride (dpc-Cl)
in pyridine to furnish the 2-oxo-protected nucleosides 14a-c.
It has already been shown that the introduction of 7-
substituents into 7-deazapurines results in a stabilization of
oligonucleotide duplexes with antiparallel or parallel chain
orientation.15,16 Also, chemical modification of the base, in
particular the addition of an electronegative atom to the
heterocycles, can produce bases with significantly altered
tautomeric ratios.17 Recently, we reported on the synthesis
(9) Seela, F.; Wei, C.; Kazimierczuk, Z. HelV. Chim. Acta 1995, 78, 1843-
1854.
(10) Sepiol, J.; Kazimierczuk, Z.; Shugar, D. Z. Naturforsch. C 1976, 31, 361-
370.
(11) Robinson, H.; Gao, Y.-G.; Bauer, C.; Roberts, C.; Switzer, C.; Wang, A.
H.-J. Biochemistry 1998, 37, 10897-10905.
(12) (a) Seela, F.; Wei, C. HelV. Chim. Acta 1999, 82, 726-745. (b) Seela, F.;
Wei, C.; Reuter, H.; Kastner, G. Acta Crystallogr. 1999, C55, 1335-1337.
(13) Martinot, T. A.; Benner, S. A. J. Org. Chem. 2004, 69, 3972-3975.
(14) (a) Seela, F.; Wei, C.; Melenewski, A.; He, Y.; Kro¨schel, R.; Feiling, E.
Nucleosides Nucleotides 1999, 18, 1543-1548. (b) Kawakami, J.; Kamiya,
H.; Yasuda, K.; Fujiki, H.; Kasai, H.; Sugimoto, N. Nucleic Acids Res.
2001, 29, 3289-3296.
(15) (a) Seela, F.; Thomas, H. HelV. Chim. Acta 1995, 78, 94-108. (b) Seela,
F.; Zulauf, M. Chem.-Eur. J. 1998, 4, 1781-1790.
(16) (a) Ramzaeva, N.; Seela, F. HelV. Chim. Acta 1996, 79, 1549-1558. (b)
Aubert, Y.; Perrouault, L.; He´le`ne, C.; Giovannangeli, C.; Asseline, U.
Bioorg. Med. Chem. 2001, 9, 1617-1624.
(17) (a) Kwiatkowski, J. S.; Pullman, B. In AdVances in Heterocyclic Chemistry;
Katritzky, A. R., Boulton, A. J., Eds.; Academic Press: New York, 1975;
Volume 18, pp 199-335. (b) Hill, F.; Williams, D. M.; Loakes, D.; Brown,
D. M. Nucleic Acids Res. 1998, 26, 1144-1149.
(18) Seela, F.; Peng, X. Synthesis 2004, 1203-1210.
9
7740 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
Scheme 2. Synthetic Route for the Preparation of Compounds 2-4, 11, and 12a
a
(i) Sodium nitrite, AcOH/H2O (v/v 1:5), room temperature, 30 min; (ii) diphenylcarbamoyl chloride, pyridine, N,N-diisopropylethylamine; (iii) Me3SiCl,
iBuCl, pyridine, room temperature, 3 h; aqueous NH3, 0 °C.
Table 1. Half-Life Values (τ) for the Deprotection of
7-Deaza-2′-deoxyisoguanosine Derivatives in 25% Aqueous
Ammonia at 40 °C
nonfunctionalized compound 2, the route discussed above is
superior to that described by Seela et al.,12a which used the iBu
residue for the 6-amino group protection, giving only a total
yield of 36%.
UV
λ
half-life
(min)
τ
UV
λ
half-life
(min)
τ
compd
(nm)
compd
(nm)
The structures of all new compounds were confirmed by 1H-,
13C-, or 31P NMR spectroscopy as well as by elemental analysis
(see Experimental Section). The 13C NMR chemical shift
assignment was made according to gated-decoupled 13C NMR
spectra (Table 2). Compared to the nonhalogenated compounds
2, 13a, 14a, and 15a, the 7-bromo substituent (compounds 4,
13c, 14c, and 15c) leads to a ca. 13 ppm upfield shift of C(7),
while a downfield shift of ca. 3 ppm is observed for the
corresponding chlorinated compounds (3, 13b, 14b, and 15b).
The assignments of C(1′) and C(4′) are based on the difference
11
12
13a
13b
275
303
313
315
66
13c
315
313
315
315
310
92
109
132
4 (iBu), 74 (dpc)a
14ab
14bb
14cb
260
292
a Supporting Information. b Apparent half-lives.
Scheme 3. Synthesis of the 7-Deazaisoguanine Nucleoside
Building Blocks for DNA Synthesisa
1
of the coupling constants J(C, H), which are larger for C(1′)
than for C(4′) (Table 2).19
pKa Values and Tautomerism. The strength of the hydrogen
bonds between nucleobases within a base pair depends on
various parameters, including the pKa values of the nucleobase
moieties. Thus, it was conceivable to determine the pKa values
of compounds 2-4. For this, UV-spectrophotometric titrations20
were performed between pH 1.5-13.5 (Supporting Information).
All three “isoguanine” nucleosides show two pKa values (2: 4.6,
10.5; 3: 3.9, 9.7; and 4: 3.8, 9.8). From the data, it is apparent
that the electron-withdrawing 7-halogeno substituents in 3 and
4 decrease the pKa values significantly compared to that of the
nonhalogenated compound 2. This will affect the base-pair
stability and also the prototropic equilibria.
In aqueous solution, the canonical DNA constituents (e.g.,
2′-deoxyguanosine) exist predominantly in the keto (lactim)
form, which leads to an almost perfect Watson-Crick base
pairing (KTAUT ≈ 104-105).21 As only one of 104-105 of the
dG molecules is enolized, mispairing is low. Also, compound
1 exists mostly in the keto form, which can implement stable
base pairs with dC or isoCd. However, a significant amount of
the enol tautomer is formed in this case (10% in water).10 This
allows the formation of stable mismatches with dT or dU and
decreases the discrimination ability of isoguanine during base
pairing (Figure 1).11 Earlier, 2 was synthesized in our laboratory,
and the N1-H, 2-keto form was identified in the solid state.12b
Recently, it has been reported that this form is also favored in
aqueous solution.13 As the pKa values of 7-halogenated nucleo-
bases 3 and 4 are changing significantly, we assume that this
will also influence the tautomeric equilibrium and might increase
the population of the keto form. Therefore, the base discrimina-
a
(i) Dimethylacetamide dimethylacetal, MeOH, room temperature, 30
min; (ii) diphenylcarbamoyl chloride, anhydrous pyridine, N,N-diisopro-
pylethylamine; (iii) 4,4′-dimethoxytriphenylmethyl chloride, anhydrous
pyridine, room temperature; (iv) 2-cyanoethyl-N,N-diisopropylchlorophos-
phoramidite, N,N-diisopropylethylamine, CH2Cl2.
The stability of the protecting groups (13a-c and 14a-c)
was studied UV-spectrophotometrically in 25% aqueous NH3
at 40 °C (Table 1). The acetamidine derivatives 13a-c (τ: 260-
310 min) are rather stable, whereas the iBu-protected nucleoside
12 (τ: 4 min) is too labile. However, the combination of the
protecting groups on compounds 14a-c was found to be suitable
for further manipulations, including oligonucleotide deprotection
under basic conditions. Subsequently, compounds 14a-c were
converted into the 5′-O-DMT derivatives 15a-c under standard
conditions. Phosphitylation of the DMT derivatives 15a-c was
performed in anhydrous CH2Cl2 in the presence of iPr2EtN and
2-cyanoethyl-N,N-diisopropylchlorophosphoramidite, furnishing
the phosphoramidites 5-7 (Scheme 3). In the case of the
(19) Seela, F.; Bussmann, W. Nucleosides Nucleotides 1985, 4, 391-394.
(20) Albert, A.; Serjeant, E. P. The Determination of Ionization Constants;
Chapman and Hall: London, 1971; pp 44-64.
(21) Topal, M. D.; Fresco, J. R. Nature 1976, 263, 285-289.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7741

A R T I C L E S
Seela et al.
Table 2. 13C NMR Chemical Shifts (δ) of 7-Deaza-2′-deoxyisoguanosine Derivativesa
C(2)d
C(2)
C(4)d
C(6)
C(4a)
C(5)
C(5)
C(7)
C(6)
C(8)
C(7a)d
C(4)
compdb,c
C(1
′
)
C(2
′
)
C(3
′
)
C(4
′
)
C(5
′
)
HC
dN
NMe2
dCMe
212a
318a
418a
11
153.9
153.2
153.6
158.4
154.8
157.1
156.2
156.2
162.5
162.3
162.1
162.6
162.3
162.3
156.3
156.2
156.0
156.1
152.7
156.4
156.0
156.0
155.4
156.0
155.7
155.6
156.2
156.0
92.6
90.2
91.0
99.1
109.0
98.2
100.8
103.7
87.4
87.4
89.1
101.0
104.4
88.6
100.7
104.8
89.0
101.0
105.1
89.4
118.9
116.2
118.8
121.5
126.5
121.7
118.4
120.9
123.4
120.5
122.9
122.9
120.2
122.6
152.6
f
f
83.4
82.6
82.5
82.6
82.9
82.7
82.3
82.3
82.8
82.8
82.7
82.2
82.3
82.4
e
e
e
e
e
e
e
e
e
e
e
e
e
e
71.1
70.9
70.8
70.8
70.8
71.1
71.0
71.0
71.0
71.0
70.9
70.8
70.6
70.7
87.2
87.2
87.2
87.3
87.7
87.1
87.2
87.2
87.3
87.5
87.4
85.5
85.6
85.6
62.0
61.9
61.8
61.8
61.7
62.1
62.0
61.9
62.0
61.9
61.8
64.1
64.2
64.2
151.4
152.4
155.3
155.8
f
151.7
150.6
150.9
151.9
150.7
151.1
12
13a
13b
13c
14a
14b
14c
15a
15b
15c
161.5
161.3
160.9
161.5
161.8
161.5
161.7
162.0
161.7
38.5, 37.8
38.4, 37.9
38.3. 37.8
38.3, 37.8
38.4, 37.8
38.4, 37.8
e
17.4
17.6
17.7
16.5
16.7
16.7
16.5
16.8
16.8
95.7
96.6
109.3
105.7
106.7
109.2
105.7
106.8
e
e
a Measured in DMSO-d6. b Systematic numbering. c Purine numbering. d Tentative. e Superimposed by DMSO. f Not detected.
Scheme 4. Tautomers of 7-Deazaisoguanosines 3 and 4
tion will increase, in particular against dT. For this purpose,
the tautomeric equilibrium between the keto and the enol forms
of 3 and 4 was determined UV-spectrophotometrically in
aqueous and nonaqueous solution.
According to Figure 2A, the UV spectra of compounds 3
and 4, measured in water, show the same maxima (306 and
261 nm), resembling the fixed keto form 4k (Scheme 4).
However, under nonaqueous conditions (100% dioxane), the
UV absorbance changes totally, showing the maxima of the
fixed enol form 4e (265 and 281 nm) (Figure 2B). (The
syntheses and analytical data of compounds 4k and 4e are
described in the Supporting Information.) Therefore, it is
reasonable to assume that in aqueous solution the halogenated
derivative 3 or 4 exists almost entirely in the keto form, while
in dioxane solution the enol form is the predominant species.
Next, the tautomeric equilibria (KTAUT ) [keto]/[enol]) of
compounds 3 and 4 were determined in a dioxane/water mixture
ranging from 98:2 to 2:98. The ratio of the tautomers was
determined by the multiwavelength method of Dewar and
Urch.22a Data were collected at 281 and 306 nm, where the
extinction coefficients of both forms show the highest difference
(Figure 2). When the water/dioxane ratio was increased, the
UV spectra for 3 and 4 shifted gradually, more and more
resembling the UV spectra of the keto form. When the water
content was higher than 25%, the UV spectra became solvent-
Figure 1. Keto and enol tautomers of isoguanine pair with isoC (or C)
and T (or U).
Figure 2. UV spectra of compounds 3, 4, 4k, and 4e in water (A) and in dioxane (B).
9
7742 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
Figure 4. Crystal structure of compound 4: the anti conformation of the
base about the glycosylic bond [torsion angle ø (O4′-C1′-N9-C4) is
-114.9°] and the S conformation of pentofuranose ring (P ) 153.9°, ψm
) 43.6°). The complete data of the structure will be published elsewhere.
Figure 3. Plot of log([enol]/[keto]) versus ET(30) for compounds 2 (4), 3
(O), and 4 (9) in mixtures of dioxane (ET(30) ) 36.0) and water (ET(30)
) 63.1).
the N1-C2 bond length is 0.077 Å longer than that of N3-C2
[1.335 (7) Å], which suggests that the ring proton is localized
at N1. Obviously, in crystalline state the base ring adopts N1-
H, 2-keto-6-amino tautomeric form, which is in line with the
structure of compound 4 in aqueous solution (Figure 4) (Seela
et al., manuscript in preparation). To prove the higher discrimi-
natory potential of compounds 3 and 4 against dT and to
stabilize duplexes by the properties of 7-halogeno substituents,
a series of oligonucleotides were synthesized.
independent (Supporting Information) and the band at 281 nm
nearly disappeared, while the absorbance at 306 nm increased
to the level of the pure keto form. By adding more water, the
whole spectrum is moved to shorter wavelength due to the
increase of the polarity of the solvent mixture. The composition
of keto and enol forms was maintained at the same level. From
these data, the KTAUT values were determined quantitatively by
the methods of Shugar et al.10 and Voegel et al.22b A plot of
the logarithm of the tautomeric equilibrium constant for 3 and
4 versus the polarity parameter ET(30) of a set of the dioxane-
water mixtures is shown in Figure 3. When the values of ET-
(30) are in the range of 38-50, the relationship between the
log [KTAUT] and the ET(30) is linear. This can be used to estimate
the value of KTAUT in aqueous solution. By extrapolating this
linear relationship to the ET(30) value of water (63.1),23 the
tautomeric equilibrium constants for compounds 3 and 4 are
estimated to be about 104.
For comparison, the KTAUT of the parent nucleoside 2 was
measured. The calculated value of KTAUT for compound 2 was
found to be about 2000 (Supporting Information), a value that
agrees with that measured by Benner et al.13 In conclusion, the
population of the keto form of compound 3 or 4 in water is
significantly higher than that of compound 2 and comes close
to the value for naturally occurring 2′-deoxyguanosine (KTAUT
≈ 104-105).21 Consequently, the halogenated nucleosides should
increase the specificity to form a base pair with dC (parallel
DNA) or iCd (antiparallel DNA) over that of dT.
Oligonucleotides. Synthesis. Oligonucleotide synthesis was
performed on solid phase in an ABI 392-08 synthesizer
employing phosphoramidite chemistry.24 Oligonucleotides were
purified by reversed-phase HPLC. The nucleoside composition
of oligomers was determined by MALDI-TOF spectrometry (see
Table 8, experimental part) as well as by enzymatic analysis
using snake venom phosphodiesterase followed by alkaline
phosphatase. The mixture was analyzed on reversed-phase
HPLC (RP-18) (see Figure 10, experimental part). The stability
of oligonucleotide duplexes was determined by temperature-
dependent UV measurements (Cary 1E, Varian) (see Experiment
Section).
Base-Pairing Properties of the 7-Deaza-2′-deoxyisogua-
nosine Analogues 3 and 4. Recently, the base pairing of 1 and
its 7-deazapurine derivatives has been investigated in our
laboratory.4c,12a,25 It was demonstrated that the 7-deazapurine
nucleoside 2 was an effective substitute of ′1 because of the
high stability of the N-glycosylic bond,26 accompanied by similar
base-paring properties. Nevertheless, it was observed that both
nucleosides showed base-pair ambiguity, in particular against
dT.12a,14b Apparently, the higher keto form population of
compound 2 (103) is not sufficient to result in a better
discrimination among the cognate and noncognate bases. Now,
much better base selectivity was expected due to the increased
KTAUT ≈ 104 for the halogenated nucleosides 3 and 4.
Furthermore, the 7-halogeno substituent should be beneficial
to the duplex stability as it was reported for related nucleo-
Next, a single-crystal X-ray analysis of compound 4 was
performed. The crystal structure is characterized by the anti
orientation around the N-glycosylic bond with the O4′-C1′-
N9-C4 torsion angle of ø ) -114.0° and the S conformation
of pentofuranose moiety [the phase angle of pseudorotation: P
) 153.9° (2T1 a 2E) and the puckering amplitude: ψm ) 43.6°].
According to the bond length of C2-O2 [1.261(7) Å], it is
apparent that the molecule adopts the 2-keto form.12b In addition,
(22) (a) Dewar, M. J. S.; Urch, D. S. J. Chem. Soc. 1957, 345-347. (b) Voegel,
J. J.; von Krosigk, U.; Benner, S. A. J. Org. Chem. 1993, 58, 7542-7547.
(23) (a) Reichardt, C.; Harbusch-Go¨rnert, E. Liebigs Ann. Chem. 1983, 721-
743. (b) von Jouanne, J.; Palmer, D. A.; Kelm, H. Bull. Chem. Soc. Jpn.
1978, 51, 463-465. (c) von Dimroth, K.; Reichardt, C.; Siepmann, T.;
Bohlmann, F. Liebigs Ann. Chem. 1963, 661, 1-37.
(24) Users’ Manual of the DNA Synthesizer; Applied Biosystems: Weiterstadt,
Germany, p 392.
(25) Seela, F.; Kro¨schel, R. Nucleic Acids Res. 2003, 31, 7150-7158.
(26) Seela, F.; Menkhoff, S.; Behrendt, S. J. Chem. Soc., Perkin Trans. II 1986,
525-530.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7743

A R T I C L E S
Seela et al.
Table 3. Tm Values and Thermodynamic Data for ps-Oligonucleotides with Alternating Basesa
b
b
duplex
Tm
(
°C)
∆Tm
(°
C)
duplex
Tm
(°
C)
∆Tm
(°
C)
5′-d(1-C-1-C-1-C) (16)c
5′-d(1-C-1-C-1-C) (16)
5′-d(2-C-2-C-2-C) (17)c
5′-d(2-C-2-C-2-C) (17)
5′-d(3-C-3-C-3-C) (18)
5′-d(3-C-3-C-3-C) (18)
5′-d(4-C-4-C-4-C) (19)
5′-d(4-C-4-C-4-C) (19)
33
22
54
56
5′-d(1-T)6 (20)
5′-d(C-A)6 (21)
5′-d(2-T)6 (22)
5′-d(C-A)6 (21)
5′-d(3-T)6 (23)
5′-d(C-A)6 (21)
5′-d(4-T)6 (24)
5′-d(C-A)6 (21)
5′-d(G-T)6 (25)
3′-d(C-A)6 (26)
46
47
59
60
56
-2.2
+4.2
+4.6
+2.2
+2.3
a Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60 mM Na-cacodylate buffer, pH 7.0, with 5 µM + 5 µM single-strand concentration. b Tm increase per
modification. c See Seela et al.14a
23‚21 and 24‚21 showed a somewhat broader melting range, it
is evident that the Tm values of ps-duplex containing 3 or 4 are
significantly higher than those of the corresponding ps-duplexes
20‚21 and 22‚21. The ps-duplexes incorporating the 3-dC or
4-dC base pairs (23‚21: Tm ) 59 °C and 24‚21: Tm ) for 60
°C) were ∼3-4 °C more stable than the corresponding aps-
duplex 25‚26 containing dG-dC base pairs (Tm ) 56 °C) (Table
3). Apparently, the 7-halogeno substituents stabilize the “isoGd-
dC” pair strongly compared to the nonfunctionalized nucleosides
1 or 2 in ps-duplexes 20‚21 (46 °C) or 22‚21 (47 °C) (Figure
5). This stability increase makes parallel DNA as equally stable
as aps-DNA containing the canonical dG-dC base pairs. The
smaller stabilization found in these series of oligonucleotide
duplexes compared to the self-complementary ones results from
the fact that the latter are further stabilized by the base stacking
of the 5′ overhangs.27 As the self-complementary duplexes of
Figure 5. Tm profiles of ps-duplexes containing compounds 1-4 and the
Table 3 are formed by only five base pairs compared to six
corresponding aps-DNA. Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60
pairs in the series of the non-self-complementary duplexes, the
mM Na-cacodylate buffer, pH 7.0, with 5 µM + 5 µM single-strand
contribution of the overhangs is higher than that of one base
pair.
concentration.
sides.16,25 Consequently, compounds 3 and 4 were incorporated
in oligonucleotide duplexes with ps and aps chain orientation,
and their base-pair stability and base selectivity were studied.
Self-Complementary and Non-Self-Complementary ps-
Duplexes with Alternating Base Composition. The self-
complementary hexanucleotide 5′-d(2-C)3 (17) forms a ps-
duplex with two nucleoside overhangs, which was found to
be less stable than the purine counterpart 5′-d(1-C)3 (16).14a
Now, the halogenated duplexes 5′-d(3-C)3 (18‚18 ) and 5′-
d(4-C)3 (19‚19), incorporating 7-chloro- (3) or 7-bromo deriva-
tive (4) of nucleoside 2, were investigated. From the data given
in Table 3, it is apparent that compounds 3 and 4 strongly
increase duplex stability. Their stabilizing effect amounts to
+4.2 °C per modification for compound 3 and +4.6 °C for
Parallel-Stranded Duplexes with Random Base Composi-
tion. Apart from self-complementary and non-self-complemen-
tary ps-duplexes with alternating base composition, ps-duplexes
with random base composition were investigated next (Table
4). To induce parallel chain orientation, particular sequence
motifs were chosen and dG-dC base pairs were replaced by
“iGd”-dC or MeiCd-dG pairs. The role of 7-halogeno substit-
uents was studied on two different duplex series, namely 27‚
28 and 32‚33, both incorporating two modified nucleosides 3
or 4 in place of iGd opposite to dC. In the first series of duplexes,
the modified nucleosides were arranged in a consecutive order
while in the second the modification sites were separated.
As shown in Table 4, both nucleosides 3 and 4 increase the
duplex stability significantly, while nucleoside 2 shows base-
pair stability almost as identical as those containing 1. Although
the stability of these two duplexes are dependent on the position
of incorporation (Tm ) 44 °C for 27‚28 and Tm ) 39 °C for
32‚33), the Tm enhancement of 7-halogenated derivatives
containing duplexes is in the same range with a ∆Tm ) +2.0
°C per modification for 3 (duplexes 30‚28 and 35‚33) and +2.5
°C for 4 (duplexes 31‚28 and 36‚33). These stabilizing effects
are identical to those in non-self-complementary alternating
duplexes, which shows that the nearest-neighbor stacking does
not affect the stabilizing effect of compound 3 or 4 on these
parallel stranded duplexes.
4. Compared to the nonhalogenated compound 2 with ∆Tm
)
-2.2 °C per modification (duplex 17), the incorporation of
nucleosides 3 and 4 results in a Tm increase of 6-7 °C per
modification.
Next, the effect of compounds 3 and 4 on non-self-
complementary duplexes with alternating bases such as 5′-
d(iGT)6‚5′-d(CA)6 was studied. These duplexes do not form
overhangs, and their stability can be directly compared with
their antiparallel stranded counterparts. Figure 5 shows Tm
profiles of ps-duplexes containing compounds 1-4 and the
corresponding aps-DNA. It can be seen that aps-duplex 25‚26
shows high cooperativity during the melting, indicated by a
narrow range of the melting profile. Although the ps-duplexes
(27) Rosemeyer, H.; Seela, F. J. Chem. Soc., Perkin Trans. II 2002, 746-750.
9
7744 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
Table 4. Tm Values and Thermodynamic Data of ps-Duplexes with Random Base Compositiona,b
c
c
duplex
Tm
(°
C)
∆Tm
(°
C)
duplex
Tm
(°
C)
∆Tm
(°
C)
5′-d(TiCATAAiCT11AT) (27)d
5′-d(AGTATTGACCTA) (28)
5′-d(TiCATAAiCT22AT) (29)d
5′-d(AGTATTGACCTA) (28)
5′-d(TiCATAAiCT33AT) (30)
5′-d(AGTATTGACCTA) (28)
5′-d(TiCATAAiCT44AT) (31)
5′-d(AGTATTGACCTA) (28)
44
44
48
49
5′-d(ATiCiCA1TTAT1A) (32)d
5′-d(TAGGTCAATACT) (33)
5′-d(ATiCiCA2TTAT2A) (34)d
5′-d(TAGGTCAATACT) (33)
5′-d(ATiCiCA3TTAT3A) (35)
5′-d(TAGGTCAATACT) (33)
5′-d(ATiCiCA4TTAT4A) (36)
5′-d(TAGGTCAATACT) (33)
39
39
43
44
0
0
+2.0
+2.5
+2.0
+2.5
a Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60 mM Na-cacodylate buffer, pH 7.0, with 5 µM + 5 µM single-strand concentration. b d(iC) ) MeiCd )
5-methyl-2′-deoxyisocytidine. c Tm increase per modification. d See Seela et al.12a
iGd-MeiCd (motif III) or 2-MeiCd pair (motif IVa) is also
suggested for that of the 7-deazapurine nucleosides 3 and 4
(motifs IVb,c) (Figure 7).
Mismatch Discrimination. The hybridization experiments
of iGd toward four canonical nucleosides in vitro have been
investigated by Seela et al.12a and Kawakami et al.14b using iGd-
iCd or dA-dT base pair as the control. It was found that iGd
forms similar stable base pairs with dT, dC, and dG but not
with dA when the base pair was located in the center of the
sequences. If the duplexes contain the mismatch base pairs at
the end, the stability of duplexes was dependent on the
sequences.14b Nevertheless, the thermodynamic stabilities of
mismatch base pairs were in agreement with the tendency of
the mutagenic misincorporation of the nucleosides opposite to
iGd in vitro, suggesting that the thermodynamic stability of the
base pairs may be an important factor for the mutation spectra
of iGd.14b Among the mismatches of iGd toward dC, dG, and
dT during the vitro replication, the most serious problem is the
misincorporation of iGd opposite dT in the DNA template.5a,7,8
This might result from the enol tautomer of isoguanine
nucleosides existing in aqueous solution to a significant amount
(10%).10
Figure 6. “iGd-dC” base-pair motifs in duplexes with parallel chain
orientation.
Figure 6 shows the base-pair motifs of the duplexes with
parallel chain orientation. Analogously to the base-pair motif I
of iGd-dC, motifs IIa-c are suggested for the “c7iGd”-dC pair.
A significant stabilization of the duplex structure by incorpora-
tion of 7-halogenated nucleoside 3 or 4 instead of 2 can be
attributed to (i) a hydrophobization of the small groove, (ii)
increased stacking interaction of the modified base, and (iii)
better proton donor properties of H-N(1) causing stronger
hydrogen bonding within the base pairs. In the case of
compounds 3 and 4, on the one hand, the introduction of
7-halogeno substituents increases the hydrophobic properties of
the modified bases, on the other hand increases the polariz-
abilities (Rm/10-24 cm3) of the nucleobases (14.83 for 2, 16.85
for 3, and 17.55 for 4, calculated by Hyperchem 7.0). Accord-
ingly, the stacking interaction was increased. Furthermore, better
proton donor properties of H-N(1) causing stronger hydrogen
bonding within the base pairs play an important role in the
stability of 3-dC or 4-dC base pairs. This is underlined by
the pKa values of deprotonation (2: 10.5; 3: 9.7; 4: 9.8).
Consequently, compounds 3 and 4 are stronger proton donors
in the base pairs of 3-dC (IIb) or 4-dC (IIc) pairs compared
to that of 2-dC (IIa).
Duplexes with Antiparallel Chain Orientation. As shown
in Table 5, aps-duplexes containing 1-MeiCd base pairs are more
stable than those containing dG-dC pairs (see duplexes 37‚38,
27‚32, 33‚28, and 33a‚28a), an observation that has been already
reported.12a According to Table 5, it is also apparent that the
stability of the iGd-MeiCd base pair is increased when iGd (1)
is replaced by compound 3 or 4. The duplexes with one
incorporation of 3 or 4 opposite to MeiCd (Tm ) 57 °C for both
37‚40 and 37‚41) show Tm values 3 °C higher than that of the
parent duplex 37‚38 (Tm ) 54 °C) or that containing 2 (37‚39,
Tm ) 54 °C) (Table 5). Incorporation of four modified base
pairs showed a Tm increase of 7 °C for duplex 30‚35 and 8 °C
for 31‚36 (Table 5). Both 7-halogenated nucleosides cause very
similar stabilization. A related base-pair motif as reported for
As discussed above, the enol population is significantly
decreased in the case of the halogenated 7-deaza-2′-deoxy-
isoguanosines 3 and 4 compared to their nonhalogenated
counterparts. Now, we are investigating oligonucleotide duplexes
containing the nucleosides 1-4 located opposite the four
canonical nucleosides. These studies include duplexes with
parallel as well as with antiparallel chain orientation. Table 6
summarizes the data for ps-oligonucleotides containing base
pairs with 1, 2, as well as those incorporating compound 3
or 4 in a position opposite the four canonical nucleosides.
From the data, it is apparent that the most stable base pairs
formed in all cases are those with dC, while those incorporating
mismatches show comparable Tm decreases. The two series
of oligonucleotide duplexes containing the modifications at
different positions show that the mismatch discrimination
is position-dependent. As expected, a stronger discriminatory
effect occurs when the modification is in the middle of the
duplexes (35‚37 and 36‚37), while it is less pronounced near
the termini (duplexes 30‚45 and 31‚45), which was also
observed in the case of 1 and 2 (Table 6). The most striking
result of this investigation is the finding that the base discrimi-
nation of the halogenated nucleosides 3 and 4 is superior
to that of the nonfunctionalized nucleosides 1 and 2. In
particular, the base pair of compounds 3 and 4 with dT shows
a significantly enhanced base discrimination (higher ∆Tm
values).
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7745

A R T I C L E S
Seela et al.
Table 5. Tm Values and Thermodynamic Data of aps-Duplexes with Random Base Compositiona,b
c
c
duplex
Tm
(
°
C)
∆Tm
(
°C)
duplex
Tm
(°
C)
∆Tm
(°
C)
5′-d(TAGGTCAATACT) (33)
3′-d(ATCCAGTTATGA) (28)
5′-d(TAGGTiCAATACT) (37)
3′-d(ATCCA1TTATGA) (38)
5′-d(TAGGTiCAATACT) (37)
3′-d(ATCCA2TTATGA) (39)
5′-d(TAGGTiCAATACT) (37)
3′-d(ATCCA3TTATGA) (40)
5′-d(TAGGTiCAATACT) (37)
3′-d(ATCCA4TTATGA) (41)
51
54
54
57
57
3′-d(TAGGTCAATACT) (33a)
5′-d(ATCCAGTTATGA) (28a)
3′-d(TA11TiCAATAiCT) (27)d
5′-d(ATiCiCA1TTAT1A) (32)
3′-d(TA33TiCAATAiCT) (30)
5′-d(ATiCiCA3TTAT3A) (35)
3′-d(TA44TiCAATAiCT) (31)
5′-d(ATiCiCA4TTAT4A) (36)
49
60
67
68
0
+1.8
+2.0
+3.0
+3.0
a Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60 mM Na-cacodylate buffer, pH 7.0, with 5 µM + 5 µM single-strand concentration. b d(iC) ) MeiCd )
5-methyl-2′-deoxyisocytidine. c Tm increase per modification. d See Seela et al.12a
Table 7. Tm Values and Thermodynamic Data of aps-Duplexes
with Mismatches Opposite to 1-4a
5
′
-d(TAGGTXAATACT)-3
-d(ATCCAYTTATGA)-5
′
′
5
′
-d(AGTATTGAXCTA)-3
′
3
′
3
′
-d(TCATAACTYGAT)-5′
b
b
X
‚
Y
duplex
Tm
(
°C)
∆Tm
(°
C)
duplex
Tm
(°
C)
∆Tm (°C)
iC‚1c,d
C‚1d
G‚1d
T‚1d
A‚1c,d
iC‚2c,d
C‚2d
G‚2d
T‚2d
A‚2d
iC‚3c
C‚3
37‚38
33‚38
46‚38
47‚38
48‚38
37‚39
33‚39
46‚39
47‚39
48‚39
37‚40
33‚40
46‚40
47‚40
48‚40
37‚41
33‚41
46‚41
47‚41
48‚41
54
44
44
42
32
54
43
42
42
29
57
47
44
40
28
57
47
45
40
27
45‚49
28‚49
42‚49
43‚49
44‚49
45‚50
28‚50
42‚50
43‚50
44‚50
45‚51
28‚51
42‚51
43‚51
44‚51
45‚52
28‚52
42‚52
43‚52
44‚52
54
44
44
45
35
52
43
44
44
31
56
48
45
43
29
56
47
45
43
33
-10
-10
-12
-22
0
-11
-12
-12
-25
-10
-10
-9
-19
-9
-8
-8
Figure 7. “iGd-iCd” base-pair motifs in duplexes with antiparallel chain
-21
orientation.
Table 6. Tm Values and Thermodynamic Data of ps-Duplexes
with Mismatches Opposite to 1-4a
-10.0
-13.0
-17.0
-29.0
-8.0
-11.0
-13.0
-27.0
G‚3
T‚3
b
b
5
′
-d(TiCATAAiCTXXAT)-3
′
5
′
-d(ATiCiCAXTTATXA)-3
′
A‚3
5
′
-d(AG TATT GAYCTA)-3
′
5
′
-d(TA G GTYAATACT)-3
′
iC‚4c
C‚4
c
c
X
‚Y
duplex
Tm
(°
C)
∆Tm
(°
C)
duplex
Tm
(
°C)
∆Tm (°C)
-10.0
-12.0
-17.0
-30.0
-9.0
-11.0
-13.0
-23.0
1‚Cd
1‚Gd
1‚Td
1‚Ad
1‚iCb,d
2‚Cd
2‚Gd
2‚Td
2‚Ad
2‚iCb,d
3‚C
27‚28
27‚42
27‚43
27‚44
27‚45
29‚28
29‚42
29‚43
29‚44
29‚45
30‚28
30‚42
30‚43
30‚44
30‚45c
31‚28
31‚42
31‚43
31‚44
31‚45
44
36
33
34
40
44
33
31
32
37
48
34
31
31
44
49
36
31
31
45
32‚33
32‚46
32‚47
32‚48
32‚37
34‚33
34‚46
34‚47
34‚48
34‚37
35‚33
35‚46
35‚47
35‚48
35‚37
36‚33
36‚46
36‚47
36‚48
36‚37c
39
-
31
34
30
39
-
29
29
-
43
-
28
30
36
44
-
29
28
36
G‚4
T‚4
-8
-11
-10
-4
-
A‚4
-8
-5
-9
a Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60 mM Na-cacodylate
buffer, pH 7.0, with 5 µM + 5 µM single-strand concentration. b ∆Tm
)
-11
-13
-12
-7
-
-10
-10
-
Tbmase mismatch - Tmbase match c d(iC) ) MeiCd ) 5-methyl-2′-deoxyisocytidine.
.
d See Seela et al.12a
dA increased in the order: 3 ≈ 4 < 1 ≈ 2. This can be clearly
seen from Figure 8A-D. According to Figure 8C,D, it is
apparent that compounds 3 and 4 show much stronger discrimi-
nation with dA or dT than compounds 1 and 2 (Figure 8A,B)
as indicated by the order of |∆Tm| (1 ≈ 2 < 3 ≈ 4).
3‚G
-14
-17
-17
-4
-
-15
-13
-7
3‚T
3‚A
3‚iC
4‚C
Regarding the various base-pair motifs formed by mispairing
in duplexes with parallel or antiparallel chain orientation, the
enol tautomer of 1 is responsible for two major mismatches
(Figure 9). A stable tridentate mispair of 1 with dT is formed
in the parallel duplexes according to motif V, while motif VI
causes the mispairing in antiparallel DNA.5a,11,12a These motifs
are likely to be formed also in the case of 2. Apparently, the
KTAUT ≈ 103 of compound 2 is not sufficient to suppress
mispairing. On the contrary, the halogenated compounds 3 and
4 (KTAUT ≈ 104) are not able to form such a mispair, at least
not to a significant extent due to the low content of the enol
form. Another mismatch formed in parallel DNA, which results
from the relatively high Tm values shown in Table 6, is that
between isoGd or 7-deaza-2′-deoxyisoG and iCd. The likely
base-pair motif in this case is VII. A corresponding base pair
between isoGd and dC as shown in the motif X seems to be of
minor importance in the aps-duplexes according to the ∆Tm
4‚G
-13
-18
-18
-4
-
-15
-16
-8
4‚T
4‚A
4‚iC
a Measured in 1.0 M NaCl, 0.1 M MgCl2, and 60 mM Na-cacodylate
buffer, pH 7.0, with 5 µM + 5 µM single-strand concentration. b d(iC) )
MeiCd ) 5-methyl-2′-deoxyisocytidine. c ∆Tm ) Tbmase match - Tmbase match
d See Seela et al.12a
.
Next, duplexes with antiparallel chain orientation were
investigated. As listed in Table 7, the 3-MeiCd and 4-MeiCd
are the most stable base pairs as indicated from the highest Tm
of the corresponding duplexes, while 3 or 4 located opposite to
dC, dG, dA, or dT reduce duplex stability with the least stable
duplex for the “iGd”-dA pair. It can be further seen that
compounds 3 and 4 are much more discriminatory against
mispair formation than the nonfunctionalized nucleosides 1 and
2. The stability of duplexes incorporating “iGd”-dT or “iGd”-
9
7746 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
Figure 8. Graphs showing the stability of oligonucleotide duplexes containing mismatches with compounds 1-4 opposite to isoCd, dC, dG, dT, and dA.
The insets show the |∆Tm| of Tbase mismatch - Tmbase match for the evaluation of the discriminatory effects.
m
increased in antiparallel DNA or with dC in parallel DNA.
Moreover, the base discrimination against the four canonical
DNA constituents, in particular against adenine and thymine,
is significantly improved. This is underlined by the favored keto
tautomer population of compounds 3 and 4 in aqueous solution
compared to the nonhalogenated nucleosides 1 or 2. Also, low
glycosylic bond stability of compound 1 compared to that of
the canonical nucleosides (acidic conditions) was overcome.
Earlier, it was reported that triphosphates of 7-halogenated
or 7-alkynylated 7-deaza-2′-deoxyguanosines show good ac-
ceptance by various DNA polymerases.28 Considering the
favorable properties of 7-deazapurine 2′-deoxyribonucleosides,
it is expected that the triphosphates of compounds 3 and 4 are
good substrates as well-forming DNA fragments with antipar-
allel chain orientation. Also, a higher accuracy of nucleotide
incorporation is expected for the triphosphates of 3 and 4
compared to those of 1 or 2. Similarly, the ribonucleosides
Figure 9. Odd base-pairs motifs related to “iGd-iCd” (ps), “iGd-dT”-
(ps), “iGd-dC” (aps), and “iGd-dT” (aps).
related to the structure of 3 or 4 or other 7-functionalized
7-deazaisoguanine nucleosides with 7-substituents of moderate
size (e.g., propynyl residues) should show similar favorable
properties compared to those containing a purine skeleton. Work
in this area is now in progress.
values shown in Table 7. These results reveal that, from the
viewpoint of thermodynamic stability, compounds 3 and 4
reduce the mispairing during hybridization, a phenomenon that
should also minimize the chance of mispair formation of dT
opposite to iGd in DNA replication.
Experimental Section
Materials. All chemicals were purchased from Sigma-Aldrich
(Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany). Solvents were
Conclusion
of laboratory grade. Snake venom phosphodiesterase (EC 3.1.15.1,
Crotallus adamanteus) and alkaline phosphatase (EC 3.1.3.1, Escheri-
chia coli) were generous gifts from Roche Diagnostics GmbH,
Germany. The synthesis of 2, 3, and 4 has been previously published.12a,18
Oligonucleotides containing 2 or 7-halogenated derivatives
(3 and 4) are synthesized using the phosphoramidites 5-7. No
particular oxidation conditions are required during solid-phase
synthesis as recommended for oligonucleotides containing
7-deaza-2′-deoxyguanosine. Compared to the purine nucleoside
1, the stability of the base pairs of the halogenated compound
3 or 4 with 2′-deoxy-5-methylisocytidine is significantly
(28) (a) Seela, F.; Feiling, E.; Gross, J.; Hillenkamp, F.; Ramzaeva, N.;
Rosemeyer, H.; Zulauf, M. J. Biotechnol. 2001, 86, 269-279. (b) Augustin,
M. A.; Ankenbauer, W.; Angerer, B. J. Biotechnol. 2001, 86, 289-301.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7747

A R T I C L E S
Seela et al.
Procedures (Organic Synthesis). General. Thin-layer chromatog-
raphy (TLC) was performed on TLC aluminum sheets covered with
silica gel 60 F254 (0.2 mm, VWR International, Darmstadt, Germany).
Column flash chromatography (FC): silica gel 60 (VWR International,
Darmstadt, Germany) at 0.4 bar. UV spectra were recorded on a U-3200
spectrophotometer (Hitachi, Japan). NMR spectra were measured on
an Avance-250 or AMX-500 spectrometers (Bruker, Rheinstetten,
Germany). Chemical shifts (δ) are in ppm relative to internal Me4Si or
external H3PO4 (31P). The J values are given in hertz. Elemental analyses
were performed by the Mikroanalytisches Laboratorium Beller, Go¨t-
tingen, Germany.
H2-C(5′)], 3.77-3.78 [1H, m, H-C(4′)], 4.27-4.28 [1H, m, H-C(3′)],
5.20 [2H, br. s, HO-(5′), HO-C(3′)], 6.06 [1H, d, J ) 5.6 Hz,
H-C(5)], 6.25-6.26 [1H, m, H-C(1′)], 7.02 [1H, d, J ) 5.6 Hz,
H-C(6)], 10.82 (1H, br. S, NH); Anal. Calcd for C15H21N5O4 (335.4):
C 53.72, H 6.31, N 20.88; found: C 53.86, H 6.28, N 20.72.
5-Chloro-7-(2-deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethyl-
aminoethylidene)amino]-7H-pyrrolo[2,3-d]pyrimidine-2-one (13b).
Compound 3 (301 mg, 1.0 mmol) was used, and 13b was obtained as
a colorless foam (333 mg, 90%); TLC (silica gel, 5:1 CH2Cl2/MeOH)
Rf 0.57; UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1): 238 (24 400); 268
1
(9400); 341 (10 700); H NMR (DMSO-d6, 500 MHz): δ 2.03-2.06
[4H, m, (CH3)2N(CH3)Cd, H-C(2′)], 2.28-2.37 [1H, 1m, H-C(2′)],
3.04, 3.10 [6H, 2s, N(CH3)2], 3.43-3.47 [2H, m, H2-C(5′)], 3.75-
3.78 [1H, m, H-C(4′)], 4.25-4.27 [1H, m, H-C(3′)], 5.08 [1H, br. s,
HO-C(5′)], 5.26 [1H, d, J ) 3.9 Hz, HO-C(3′)], 6.27 [1H, ‘t’, J )
6.2 Hz, H-C(1′)], 7.17 [1H, s, H-C(6)], 10.98 (1H, br. S, NH); Anal.
Calcd for C15H20ClN5O4 (369.8): C 48.72, H 5.45, N 18.94; found: C
48.80, H 5.55, N 18.79.
4-Amino-5-bromo-7-(2-deoxy-â-D-erythro-pentofuranosyl)-2-
[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine (11). To a
suspension of compound 4 (280 mg, 0.81 mmol) in dry pyridine (3.0
mL), diphenylcarbamoyl chloride (254 mg, 1.10 mmol) and N,N-
diisopropylethylamine (0.2 mL, 1.15 mmol) were added and stirred
for 2 h at room temperature. Then, the mixture was poured into 5%
aqueous NaHCO3 (10 mL) and extracted with CH2Cl2 (3 × 10 mL).
The CH2Cl2 layers were combined, dried over Na2SO4, and filtered.
After evaporation of the solvent, the residue was applied to FC (silica
gel, column 3 × 10 cm). Elution with CH2Cl2 followed by CH2Cl2/
MeOH (95:5 to 9:1) gave 11 as a colorless foam (340 mg, 78%). TLC
(silica gel, 9:1 CH2Cl2/MeOH) Rf 0.43; UV λmax (MeOH)/nm (ꢀ/dm3
5-Bromo-7-(2-deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethyl-
aminoethylidene)amino]-7H-pyrrolo[2,3-d]pyrimidine-2-one (13c).
Methods as described for 13a, with 4 (500 mg, 1.45 mmol) in MeOH
(20 mL) and N,N-dimethylacetamide dimethyl acetal (1.0 mL, 6.84
mmol). FC (silica gel, column 4 × 10 cm, elution with CH2Cl2/MeOH,
95:5 to 9:1) resulted in a colorless foam (13c: 534 mg, 89%). TLC
(silica gel, 9:1 CH2Cl2/MeOH) Rf 0.57; UV λmax (MeOH)/nm (ꢀ/dm3
mol-1 cm-1) 239 (22 500), 268 (8300), 341 (10 200); 1H NMR (DMSO-
d6, 500 MHz): δ 2.02-2.04 [4H, m, (CH3)2N(CH3)Cd, H-C(2′)],
2.30-2.35 [1H, m, H-C(2′)], 3.05, 3.10 [6H, 2s, N(CH3)2], 3.50-
3.52 [2H, m, H2-C(5′)], 3.75-3.78 [1H, m, H-C(4′)], 4.26-4.28 [1H,
m, H-C(3′)], 5.06 [1H, br. s, HO-(5′)], 5.24 [1H, d, J ) 3.9 Hz,
HO-C(3′)], 6.26 [1H, ‘t’, J ) 6.1 Hz, H-C(1′)], 7.22 [1H, s, H-C(6)],
10.96 (1H, br. s, NH); Anal. Calcd for C15H20BrN5O4 (414.3): C 43.49,
H 4.87, N 16.91; found: C 43.52, H 4.76, N 16.61.
1
mol-1 cm-1): 276 (12 100), 230 (30 400); H NMR (DMSO-d6, 500
MHz): δ 2.12-2.19, 2.37-2.46 [2H, m, H2-C(2′)], 3.44-3.58 [2H,
m, H2-C(5′)], 3.80-3.81 [1H, m, H-C(4′)], 4.31-4.33 [1H, ‘t’, J )
3.8 Hz, H-C(3′)], 4.98 [1H, ‘t’, J ) 3.8 Hz, HO-C(5′)], 5.30 [1H, d,
J ) 3.9 Hz, HO-C(3′)], 6.39 [1H, ‘t’, J ) 7.6 Hz, H-C(1′)], 7.32-
7.46 (12H, m, NH2, arom. H,), 7.62 [1H, s, H-C(6)]; Anal. Calcd for
C24H22BrN5O5 (540.4): C 53.34, H 4.10, N 12.96; found: C 53.40, H
4.20, N 12.82.
5-Bromo-7-(2-deoxy-â-D-erythro-pentofuranosyl)-4-[(2-methyl-
propanoyl)amino]-2-[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]-
pyrimidine (12). To a solution of 11 (260 mg, 0.48 mmol) in dry
pyridine (3.0 mL), chlorotrimethylsilane (0.2 mL, 1.58 mmol) was
added. After the mixture was stirred for 30 min, isobutyryl chloride
(0.16 mL, 1.53 mmol) was introduced and the stirring was continued
for 4 h. The mixture was cooled in an ice bath. Upon addition of H2O
(1.0 mL), three portions of 25% NH3/H2O (3 mL) were added within
5 h under stirring in the ice bath. The mixture was poured into 5%
aqueous NaHCO3 (10 mL) and extracted with CH2Cl2 (3 × 10 mL).
The CH2Cl2 layers were combined, dried over Na2SO4, and filtered.
After evaporation of the solvent, the residue was submitted to FC (silica
gel, column 3 × 10 cm). Elution with CH2Cl2 followed by CH2Cl2/
MeOH (98:2 to 95:5) gave 12 as a colorless foam (29 mg, 10%).
(Starting material 11 (200 mg) was recovered.) TLC (silica gel, 9:1
CH2Cl2/MeOH) Rf 0.51; 1H NMR (DMSO-d6, 500 MHz): δ 1.15, 1.17
[6H, 2s, CH(CH3)2], 2.23-2.26, 2.70-2.78 [2H, 2m, H2-C(2′)], 3.50-
3.56 [2H, m, H2-C(5′)], 3.80-3.84 [1H, m, H-C(4′)], 4.31-4.35 [1H,
m, H-C(3′)], 4.99 [1H, ‘t’, J ) 3.8 Hz, HO-C(5′)], 5.34 [1H, d, J )
3.8 Hz, HO-C(3′)], 6.54 [1H, ‘t’, J ) 6.1 Hz, H-C(1′)], 7.32-7.46
(10H, m, arom. H), 7.98 [1H, s, H-C(6)], 10.52 (1H, br. s, NH); Anal.
Calcd for C28H28BrN5O6 (610.5): C 55.09, H 4.62, N 11.47; found: C
54.92, H 4.62, N 11.51.
7-(2-Deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethylaminoeth-
ylidene)amino]-2-[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyri-
midine (14a). General Procedure for the Preparation of 14a-c. To
a suspension of compound 13a (280 mg, 0.84 mmol) in dry pyridine
(7.0 mL), diphenylcarbamoyl chloride (254 mg, 1.1 mmol) and N,N-
diisopropylethylamine (0.2 mL, 1.15 mmol) were added and stirred
for 1 h at room temperature. Then, the mixture was poured into 5%
aqueous NaHCO3 (10 mL) and extracted with CH2Cl2 (3 × 10 mL).
The CH2Cl2 layers were combined, dried over Na2SO4, and filtered.
After evaporation of the solvent, the residue was applied to FC (silica
gel, column 3 × 10 cm). Elution with CH2Cl2 followed by CH2Cl2/
MeOH (98:2 to 95:5) gave 14a as a colorless foam (401 mg, 91%).
TLC (silica gel, 9:1 CH2Cl2/MeOH) Rf 0.50; UV λmax (MeOH)/nm (ꢀ/
1
dm3 mol-1 cm-1) 233 (35 000), 307 (17 300); H NMR (DMSO-d6,
500 MHz): 2.13 [3H, s, (CH3)2N(CH3)Cd], 2.20-2.24, 2.44-2.50 [2H,
m, H2-C(2′)], 3.08, 3.12 [6H, 2s, N(CH3)2], 3.50-3.57 [2H, m, H2-
C(5′)], 3.82-3.85 [1H, m, H-C(4′)], 4.34-4.38 [1H, m, H-C(3′)],
4.95 [1H, ‘t’, J ) 5.4 Hz, HO-C(5′)], 5.32 [1H, d, J ) 4.0 Hz, HO-
C(3′)], 6.42 [1H, d, J ) 5.3 Hz, H-C(5)], 6.49 [1H, ‘t’, J ) 7.2 Hz,
H-C(1′)], 7.30-7.45 (10H, m, arom. H), 7.52 [1H, d, J ) 5.3 Hz,
H-C(6)]; Anal. Calcd for C28H30N6O5 (530.6): C 63.38, H 5.70, N
15.84; found: C 63.56, H 5.80, N 15.73.
7-(2-Deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethylaminoeth-
ylidene)amino]-7H-pyrrolo[2,3-d]pyrimidine-2-one (13a). General
Procedure for the Preparation of 13a-c. A suspension of 2 (266
mg, 1.0 mmol) in MeOH (10.0 mL) was stirred with N,N-dimethylac-
etamide dimethyl acetal (0.5 mL, 3.42 mmol) for 30 min at room
temperature. After evaporation, the residue was applied to FC (silica
gel, column 4 × 10 cm, elution with CH2Cl2/MeOH, 95:5 to 9:1),
yielding 13a as a colorless foam (305 mg, 91%); TLC (silica gel, 5:1
CH2Cl2/MeOH) Rf 0.54; UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1):
5-Chloro-7-(2-deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethyl-
aminoethylidene)amino]-2-[(diphenylcarbamoyl)oxy]-7H-pyrrolo-
[2,3-d]pyrimidine (14b). Methods as described for 14a, with 13b (320
mg, 0.87 mmol) in dry pyridine (7.0 mL), diphenylcarbamoyl chloride
(254 mg, 1.1 mmol), and N,N-diisopropylethylamine (0.2 mL, 1.15
mmol). The FC (silica gel, column 3 × 10 cm) gave 14b as a colorless
foam (434 mg, 89%). TLC (silica gel, 9:1 CH2Cl2/MeOH) Rf 0.51;
UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1) 233 (36 800), 312 (15 300);
1H NMR (DMSO-d6, 500 MHz): 2.16 [3H, s, (CH3)2N(CH3)Cd],
2.20-2.22, 2.44-2.46 [2H, m, H2-C(2′)], 3.13 [6H, s, N(CH3)2], 3.52-
3.59 [2H, m, H2-C(5′)], 3.82-3.85 [1H, m, H-C(4′)], 4.34-4.36 [1H,
1
231 (26 100); 264 (11 300); 341 (12 700); H NMR (DMSO-d6, 250
MHz): δ 2.04-2.06 [4H, m, (CH3)2N(CH3)Cd, H-C(2′)], 2.37-2.38
[1H, 1m, H-C(2′)], 3.06, 3.09 [6H, 2s, N(CH3)2], 3.50-3.51 [2H, m,
9
7748 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
m, H-C(3′)], 5.01 [1H, ‘t’, J ) 5.4 Hz, HO-C(5′)], 5.34 [1H, d, J )
4.0 Hz, HO- C(3′)], 6.45 [1H, ‘t’, J ) 7.2 Hz, H-C(1′)], 7.30-7.45
(10H, m, arom. H), 7.61 [1H, s, H-C(6)]; Anal. Calcd for C28H29-
ClN6O5 (565.0): C 59.52, H 5.17, N 14.87; found: C 59.49, H 5.05,
N 14.57.
3.14 [8H, m, H2-C(5′), N(CH3)2], 3.69 (6H, m, 2 CH3O), 3.91-4.00
[1H, m, H-C(4′)], 4.35-4.36 [1H, m, H-C(3′)], 5.38-5.39 [1H, m,
HO-C(3′)], 6.46 [‘t’, J ) 5.9 Hz, H-C(1′)], 6.81-7.42 [23H, m, arom.
H], 7.49 [s, H-C(6)]; Anal. Calcd for C49H47BrN6O7 (911.8): C 64.54,
H 5.20, N 9.22; found: C 64.57, H 5.25, N 9.15.
5-Bromo-7-(2-deoxy-â-D-erythro-pentofuranosyl)-4-[(dimethyl-
aminoethylidene)amino]-2-[(diphenylcarbamoyl)oxy]-7H-pyrrolo-
[2,3-d]pyrimidine (14c). Methods as described for 14a, with 13c (476
mg, 1.15 mmol) in dry pyridine (8.0 mL), diphenylcarbamoyl chloride
(319 mg, 1.38 mmol), and N,N-diisopropylethylamine (0.3 mL, 1.72
mmol). The FC (silica gel, column 3 × 10 cm) yielded 14c as a colorless
foam (621 mg, 89%). TLC (silica gel, 9:1 CH2Cl2/MeOH) Rf 0.51;
UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1) 234 (36 900), 312 (14 400);
1H NMR (DMSO-d6, 500 MHz): δ 2.15 [3H, s, (CH3)2N(CH3)Cd],
2.16-2.19, 2.40-2.44 [2H, 2m, H2-C(2′)], 3.14 [6H, s, N(CH3)2],
3.52-3.55 [2H, m, H2-C(5′)], 3.81-3.83 [1H, m, H-C(4′)], 4.33-
4.34 [1H, m, H-C(3′)], 5.00 [1H, ‘t’, J ) 5.3 Hz, HO-C(5′)], 5.35
[1H, d, J ) 3.9 Hz, HO-C(3′)], 6.44 [1H, ‘t’, J ) 7.4 Hz, H-C(1′)],
7.30-7.43 (10H, m, arom. H), 7.67 [1H, s, H-C(6)]; Anal. Calcd for
C28H29BrN6O5 (609.5): C 55.18, H 4.80, N 13.79; found: C 55.07, H
5.00, N 13.65.
7-[2-Deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-erythro-
pentofuranosyl]-4-[(dimethylaminoethylidene)amino]-2-[(diphenyl-
carbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine 3′-(2-cyanoethyl)-N,N-
diisopropylphosphoramidite (5). A Typical Procedure for the
Preparation of 5-7. Compound 15a (450 mg, 0.54 mmol) dissolved
in anhydrous CH2Cl2 (3.0 mL) under Ar atmosphere was reacted with
2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (150 µL, 0.67
mmol) in the presence of (iPr)2NEt (130 µL, 0.25 mmol) at room
temperature. After 30 min, the reaction mixture was diluted with CH2-
Cl2, and the solution was washed with a 5% aqueous NaHCO3 solution,
followed by brine. The organic solution was dried over Na2SO4 and
concentrated, and the residue was submitted to FC (column 3 × 9 cm,
CH2Cl2/acetone, 98:2), yielding a colorless foam (508 mg, 91%). TLC
(silica gel, 97:3 CH2Cl2/acetone) Rf 0.22, 0.25; 31P NMR (CDCl3, 500
MHz): δ 149.9, 149.7.
5-Chloro-7-[2-deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-
erythro-pentofuranosyl]- 4-[(dimethylaminoethylidene)amino]-2-
[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine 3′-(2-cya-
noethyl)-N,N-diisopropylphosphoramidite (6). Methods as described
for 5, with compound 15b (360 mg, 0.42 mmol) in anhydrous CH2Cl2
(3.0 mL), 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (130 µL,
0.58 mmol), and (iPr)2NEt (130 µL, 0.75 mmol). The FC (column 3 ×
9 cm, CH2Cl2/acetone, 98:2) yielded a colorless foam (412 mg, 93%).
TLC (silica gel, 97:3 CH2Cl2/acetone) Rf 0.23, 0.27; 31P NMR (CDCl3,
500 MHz): δ 149.9, 149.8.
7-[2-Deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-erythro-
pentofuranosyl]- 4-[(dimethylaminoethylidene)amino]-2-[(diphenyl-
carbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine (15a). A Typical Pro-
cedure for the Preparation of 15a-c. Compound 14a (380 mg, 0.72
mmol) was coevaporated with anhydrous pyridine (three times) and
then dissolved in pyridine (2.0 mL). To this solution, 4,4′-dimethoxy-
triphenylmethyl chloride (313 mg, 0.92 mmol) was added, and the
mixture was stirred at room temperature for 5 h. The reaction was
quenched by addition of MeOH, and the mixture was evaporated to
dryness. It was dissolved in CH2Cl2 (3.0 mL) and subjected to FC
(column 4 × 9 cm, elution with AcOEt/petroleum ether, 1:2 to 2:3) to
give 15a as a colorless foam (503 mg, 84%). TLC (silica gel, 95:5
CH2Cl2/MeOH) Rf 0.36; UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1):
5-Bromo-7-[2-deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-
erythro-pentofuranosyl]-4-[(dimethylaminoethylidene)amino]-2-
[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine 3′-(2-cya-
noethyl)-N,N-diisopropylphosphoramidite (7). Method as described
for 5, with compound 15c (430 mg, 0.47 mmol) in anhydrous CH2Cl2
(6.0 mL), 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite (150 µL,
0.67 mmol), and (iPr)2NEt (148 µL, 0.85 mmol) under Ar. Compound
7 was obtained as a colorless foam (470 mg, 90%). TLC (silica gel,
97:3 CH2Cl2/acetone) Rf 0.24, 0.28; 31P NMR (CDCl3, 500 MHz):
149.9, 149.8.
1
233 (55 800), 282 (12 400), 306 (17 100); H NMR (DMSO-d6, 500
MHz): δ 2.12 [3H, s, (CH3)2N(CH3)Cd], 2.21-2.27, 2.48-2.52 [2H,
2m, H2-C(2′)], 3.08-3.12 [8H, m, H2-C(5′), N(CH3)2], 3.71 (6H, s,
2 CH3O), 3.90-3.92 [1 H, m, H-C(4′)], 4.333-4.36 [1H, m, H-C(3′)],
5.36 [d, J ) 4.5 Hz, HO-C(3′)], 6.38 [1H, d, J ) 3.7 Hz, H-C(5)],
6.49 [1H, ‘t’, J ) 7.2 Hz, H-C(1′)], 6.81-7.37 (23H, m, arom. H),
7.42 [1H, d, J ) 3.7 Hz, H-C(6)]. Anal. Calcd for C49H48N6O7
(832.9): C 70.66, H 5.81, N 10.09; found: C 70.58, H 5.85, N 9.99.
5-Chloro-7-[2-deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-
erythro-pentofuranosyl]- 4-[(dimethylaminoethylidene)amino]-2-
[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine (15b). Meth-
ods as described for 15a, with 14b (400 mg, 0.71 mmol) in anhydrous
pyridine (2.0 mL) and 4,4′-dimethoxytriphenylmethyl chloride (313 mg,
0.92 mmol). FC (column 4 × 9 cm, elution with AcOEt/petroleum
ether, 1:2 to 2:3) resulted in 15b as a colorless foam (491 mg, 80%).
TLC (silica gel, 95:5 CH2Cl2/MeOH) Rf 0.36; UV λmax (MeOH)/nm
Oligonucleotides Synthesis. The oligonucleotides 16-52 were
prepared in a 1 µmol scale (trityl on mode) on an ABI 392-08
synthesizer employing phosphoramidite chemistry.24 The phosphora-
midites 5-7, 5′-O-DMT-N2-[(dimethylamino)methylidene]-2′-deoxy-
5-methylisocytidine 3′-O-(2-cyanoethyl)phosphormidite,4d as well as
standard phosphoramidites were employed in solid-phase synthesis. The
average coupling yield of the modified phosphoramidites 5-7 was
always higher than 95%. After cleavage from the solid support, the
oligonucleotides were deprotected in 25% aqueous NH3 for 20-24 h
at 60 °C. The DMT-containing oligonucleotides were purified by
reversed-phase HPLC (RP-18) with the following solvent gradient
system [A: 0.1 M (Et3NH)OAc (pH 7.0)/MeCN 95:5; B: MeCN;
gradient I: 0-3 min 10-15% B in A, 3-15 min 15-50% B in A,
15-20 min 50-10% B in A, flow rate 1.0 mL/min]. Then, the mixture
was evaporated to dryness, and the residue was treated with 2.5% Cl2-
CHCOOH/CH2Cl2 for 4 min at room temperature to remove the 4,4′-
dimethoxytrityl residues. The detritylated oligomers were purified by
reversed-phase HPLC with the gradient II: 0-25 min 0-20% B in A,
25-30 min 20% B in A, 30-35 min 20-0% B in A, flow rate 1.0
mL/min. The oligomers were desalted on a short column (RP-18, silica
gel) using H2O for elution of the salt, while the oligomers were eluted
with MeOH/H2O 3:2. The oligonucleotides were lyophilized on a Speed
Vac evaporator to yield colorless solids that were stored frozen at -18
°C. Although the monomers of 7-deazpurines are more easily oxidized
than purines, no degradation was found when oligonucleotides were
1
(ꢀ/dm3 mol-1 cm-1): 234 (55 700), 275 (12 100), 311 (15 000); H
NMR (DMSO-d6, 500 MHz): δ 2.15 [3H, s, (CH3)2N(CH3)Cd], 2.25-
2.29, 2.50-2.51 [2H, 2m, H-C(2′)], 3.12-3.13 [8H, m, H-C(5′),
N(CH3)2], 3.71 (6H, s, 2 CH3O), 3.90-3.91 [1H, m, H-C(4′)], 4.36-
4.38 [1H, m, H-C(3′)], 5.41 [d, J ) 4.0 Hz, HO-C(3′)], 6.48 [‘t’, J
) 5.9 Hz, H-C(1′)], 6.81-7.42 (23H, m, arom. H), 7.46 [s, H-C(6)].
Anal. Calcd for C49H47ClN6O7 (867.4): C 67.85, H 5.46, N 9.69;
found: C 68.15, H 5.64, N 9.80.
5-Bromo-7-[2-deoxy-5-O-(4,4′-dimethoxytriphenylmethyl)-â-D-
erythro-pentofuranosyl]-4-[(dimethylaminoethylidene)amino]-2-
[(diphenylcarbamoyl)oxy]-7H-pyrrolo[2,3-d]pyrimidine (15c). Com-
pound 14c (400 mg, 0.66 mmol) was used, and 15c was obtained as a
colorless foam (493 mg, 82%). TLC (silica gel, 95:5 CH2Cl2/MeOH)
Rf 0.36; UV λmax (MeOH)/nm (ꢀ/dm3 mol-1 cm-1): 234 (55 600), 275
(11 800), 312 (14 200); 1H NMR (DMSO-d6, 500 MHz): δ 2.13 [3H,
s, (CH3)2N(CH3)Cd], 2.14-2.16, 2.48-2.50 [2H, m, H-C(2′)], 3.12-
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7749

A R T I C L E S
Seela et al.
Figure 10. HPLC profiles of the nucleosides 3 and 4 (A) and HPLC profiles of the enzymatic analysis of the oligonucleotide 18 (B) containing nucleoside
3 and oligonucleotide 36 (C) containing nucleoside 4, obtained by digestion with snake venom phosphodiesterase followed by alkaline phosphatase in 0.1
M Tris‚HCl buffer (pH 8.3) at 37 °C. The nucleoside mixtures were analyzed by reversed-phase HPLC at 260 nm on a RP-18 column (200 × 10 cm).
Gradient: 0-20 min 100% A, 20-40 min 0-65% B in A, 40-60 min 65-0% B in A, flow rate 0.7 mL/min [A: 0.1 M (Et3NH)OAc (pH 7.0)/MeCN 95:5;
B: MeCN] (MeiCd ) 5-methyl-2′-deoxyisocytidine).
Table 8. Molecular Masses Determined from MALDI-TOF Mass Spectra and Extinction Coefficients of Oligonucleotides
1
1
MH+ (calcd)
MH+ (found)
ꢀ260b (M-1 cm-
)
ꢀ260c (M-1 cm-
)
5′-d(1T)6-3′ (20)
5′-d(2T)6-3′ (22)
5′-d(3C3C3C)-3′ (18)
5′-d(3T)6-3′ (23)
5′-d(AGTATT3ACCTA)-3′ (40)
5′-d(TAG3TCAATACT)-3′ (51)
5′-d(ATiCiCA3TTAT3A)-3′ (35)a
5′-d(TiCATAAiCT33AT)-3′ (30)a
5′-d(4C4C4C)-3′ (19)
3739.2
3733.2
1894.6
3940.2
3678.9
3678.9
3740.5
3740.5
2027.9
4206.9
3723.4
3723.4
3827.5
3827.5
3553.1
3739.2
3645.4
3645.4
3739.8
3734.5
1894.9
3940.6
3680.4
3680.4
3741.2
3740.2
2029.3
4207.2
3723.7
3724.2
3827.6
3828.2
3552.9
3740.1
3645.2
3645.2
78 600
97 200
42 000
62 900
87 500
37 800
91 200
86 600
130 200
130 200
122 200
122 200
40 200
104 200
104 200
110 000
110 000
36 200
5′-d(4T)6-3′ (24)
87 600
83 200
5′-d(AGTATT4ACCTA)-3′ (41)
5′-d(TAG4TCAATACT)-3′ (52)
5′-d(ATiCiCA4TTAT4A)-3′ (36)a
5′-d(TiCATAAiCT44AT)-3′ (31)a
5′-d(CA)6-3′ (21)
5′-d(GT)6-3′ (25)
5′-d(AGTATTGACCTA)-3′ (28)
5′-d(TAGGTCAATACT)-3′ (33)
129 500
129 500
121 000
121 000
138 000
123 000
135 400
135 400
103 600
103 600
108 900
108 900
110 400
98 400
108 300
108 300
a d(iC) ) MeiCd ) 5-methyl-2′-deoxyisocytidine. b Calculated extinction coefficients. c The extinction coefficients corrected by the hypochromicity.
stored for longer periods or when measurements were performed as
described in the manuscript.
spectrometer in the reflector mode (Bruker Saxonia, Leipzig, Germany).
They were in agreement with the calculated values (Table 8).
Thermodynamic Measurements. UV thermal denaturation curves
were measured on a Cary 1/3 UV/vis spectrophotometer (Varian,
Australia) equipped with a Cary thermoelectrical controller. All
measurements were performed in 1.0 M NaCl, 0.1 M MgCl2, and 60
mM Na-cacodylate buffer, at pH 7.0 ( 0.2, with 5 µM + 5 µM single-
strand concentration. The concentrations of the oligonucleotides were
determined by measuring the absorbances at 260 nm. The extinction
coefficients at 260 nm of the oligonucleotides were calculated from
the sum of the extinction coefficients of the monomeric 2′-deoxyribo-
nucleosides corrected by the hypochromicity (Table 8). The hypochro-
Oligonucleotides Characterization. The nucleoside composition of
oligomers was determined by enzymatic hydrolysis with snake venom
phosphodiesterase (EC 3.1.15.1, C. adamanteus) followed by alkaline
phosphatase (EC 3.1.3.1, E. coli from Roche Diagnostics GmbH,
Germany). The mixture was analyzed on reversed-phase HPLC (RP-
18) (Figure 10). The absorbances were quantified at 260 nm by
measuring the peak areas under consideration of the molar extinction
coefficients of each monomer (ꢀ260: c7iGd 7400, Cl7c7iGd 6400, Br7c7-
iGd 5800, iGd 4300, dA 15 400, dT 8800, dG 11 700, dC 7600, m5iCd
6300). Figure 10A indicates brominated nucleoside 4 is more hydro-
phobic than compound 3. Figure 10B,C shows the HPLC profiles of
the enzymatic hydrolysates of the oligonucleotides 18 and 36. Quan-
tification of the material was made on the basis of the peak areas, which
were divided by the extinction coefficients of the nucleoside constituents
(Figure 10). The peak areas of oligonucleotide 18 are 840 633 for dC
and 766 055 for compound 3, then dC/3 ) 1.1:1. In the case of
oligonucleotide 36: m5iCd (1 479 552), dT (4 138 803), dA (6 789 902),
nucleoside 4 (1 460 000), accordingly m5iCd/dT/dA/4 equals 1:2:1.9:
1.1.These data fit with the calculated data, which shows that the
oligonucleotides were completely hydrolyzed.
micity (h ) [(ꢀmonomer - ꢀpolymer) ‚ (ꢀmonomer)
-1]‚100%) was determined
from the absorbance before and after enzymatic digestion with SVPD.
The hypochromicity was 20% for unmodified oligonucleotides or for
those with one modification (20, 21, 25, 28, 33, 40, 41, 51, and 52),
10% for highly modified oligonucleotides (18, 19, 22, 30, 31, 35, and
36), and around 5% for compounds 23 and 24. Absorbance versus
temperature spectra were collected at 260 nm over a range of 10-95
°C with 0.1 °C increments and a heating rate of 1.0 °C/min. Samples
were annealed by heating rapidly to 95 °C for 10-15 min, followed
by cooling slowly to 10 °C. The thermodynamic data were calculated
using the program Meltwin 3.0.29
The molecular masses of all synthesized oligonucleotides were
determined by MALDI-TOF mass spectra measured on a Biflex-III
(29) McDowell, J. A.; Turner, D. H. Biochemistry 1996, 35, 14077-14089.
9
7750 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Oligonucleotide Duplexes with ps and aps Chain Orientation
A R T I C L E S
Acknowledgment. We thank Dr. Helmut Rosemeyer and Dr.
Xiaomei Zhang for the measurement of the NMR spectra, Dr.
Peter Leonard for the supply of nucleobase precursors, Dr. Yang
He for the synthesis of the oligonucleotides, and Khalil I. Shaikh
for the MALDI-TOF mass spectra. We also appreciate the
assistance of Mrs. Monika Dubiel and Mrs. Elisabeth Michalek
during the preparation of the manuscript. Financial support by
Roche Diagnostics GmbH, Germany, is gratefully acknowl-
edged.
Supporting Information Available: The half-life measure-
ment of compound 12. The synthesis of the fixed-keto com-
pound 4k and the fixed-enol isomer 4e. pKa measurements of
compounds 2-4. Determination of the tautomeric equilibrium
data of compounds 2-4. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA0425785
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7751