Evaluation of novel third-strand bases for the recognition of a C·G base pair in the parallel DNA triple-helical binding motif

Article abstract of DOI:10.1002/1522-2675(200202)85:2<502::AID-HLCA502>3.0.CO;2-L

We describe the synthesis and the incorporation into oligonucleotides of the novel nucleoside building blocks 9, 10, and 16, carrying purine-like double H-bond-acceptor bases. These base-modified nucleosides were conceived to recognize selectively a cytosine · guanine (C·G) inversion site within a homopurine · homopyrimidine duplex, when constituent of a DNA third strand designed to bind in the parallel binding motif. While building block 16 turned out to be incompatible with standard oligonucleotide-synthesis conditions, UV/triplex melting experiments with third-strand 15-mers containing β-D-nucleoside 6 (from 9) showed that recognition of the four natural Watson-Crick base pairs follows the order G·C ? C·G > A· T> T·A. The recognition is sequence-context sensitive, and G·C or C·G recognition does not involve protonated species of β-D-nucleoside 6. The data obtained fit (but do not prove) a structural model for C · G recognition via one conventional and one C - H··· O H-bond. The unexpected G · C recognition is best explained by third-strand base intercalation. A comparison of the triplex binding properties of these new bases with those of 4-deoxothymine (5-methylpyrimidine-2(1H)-one, ^4HT), previously shown to be C·G selective but energetically weak, is also described.

Full text of DOI:10.1002/1522-2675(200202)85:2<502::AID-HLCA502>3.0.CO;2-L

502
Helvetica Chimica Acta Vol. 85 (2002)
Evaluation of Novel Third-Strand Bases for the Recognition of a C ¥ G
Base Pair in the Parallel DNA Triple-Helical Binding Motif
by Isabelle Pre¬vot and Christian J. Leumann*
Department of Chemistry and Biochemistry, University of Bern, Freiestrasse 3, CH-3012 Bern
We describe the synthesis and the incorporation into oligonucleotides of the novel nucleoside building
blocks 9, 10, and 16, carryingpurine-like double H-bond-acceptor bases. These base-modified nucleosides were
conceived to recognize selectively a cytosine ¥ guanine (C ¥ G) inversion site within a homopurine ¥ homopyri-
midine DNA duplex, when constituent of a DNA third strand designed to bind in the parallel binding motif.
While buildingblock 16 turned out to be incompatible with standard oligonucleotide-synthesis conditions, UV/
triplex meltingexperiments with third-strand 15-mers containing b-d-nucleoside 6 (from 9) showed that
recognition of the four natural Watson-Crick base pairs follows the order G ¥ C ꢀ C ¥ G > A ¥ T> T¥ A. The
recognition is sequence-context sensitive, and G ¥ C or C ¥ G recognition does not involve protonated species of
b-d-nucleoside 6. The data obtained fit (but do not prove) a structural model for C ¥ G recognition via one
conventional and one CÀH ¥¥¥ O H-bond. The unexpected G ¥ C recognition is best explained by third-strand
base intercalation. A comparison of the triplex bindingproperties of these new bases with those of 4-
deoxothymine (5-methylpyrimidine-2(1H)-one, 4 HT), previously shown to be C ¥ G selective but energetically
weak, is also described.
Introduction. Strongand sequence-specific triple-helix formation of oligonucleo-
tides with genomic DNA can selectively interfere with gene expression on the level of
transcription and, therefore, is of interest in medicinal chemistry and biotechnology
[1][2]. Molecular recognition of a DNA duplex in the major groove via a third-strand
oligonucleotide in either the parallel [3][4] or the antiparallel [5][6] binding motif,
however, is restricted to homopurine À homopyrimidine DNA-sequence tracts. De-
spite considerable efforts over the past decade, a general recognition motif by third
strands for any given DNA sequence is still elusive [7].
Approaches to overcome this sequence limitation in the past included the
preparation and evaluation of new, unnatural bases that are designed to either target
a pyrimidine base via one H-bond [8][9], or complete pyrimidine ¥ purine base pairs
[10][11]. The use of abasic sites or aromatic heterocycles that unselectively contribute
to stackingwithout any base-readingcapability via H-bondingwas also evaluated,
however, with limited success [12].
We recently reported on the selective recognition of G ¥ C base pairs in DNA
duplexes by parallel complementary oligodeoxynucleotides containing the unnatural
nucleoside 7-(2'-deoxy-b-d-ribofuranosyl)hypoxanthine (7H) [13]. Although only one
7
7
H-bond between the bases H and G can be formed, the H ¥ G ¥ C base triple (Fig. 1)
‡
equals the stability of the canonical C ¥ G ¥ C base-triple (C ˆ 2'-deoxy-5-methylcyti-
7
dine) at pH 7.0. Empirically, we attributed the remarkable stability of the H ¥ G base
pair to the assistance of two additional CÀH ¥¥¥ O H-bonds, flankingthe conventional
one (Fig. 1). As a consequence of this result, we explored whether the pyrimidine base

Helvetica Chimica Acta Vol. 85 (2002)
503
uracil can be recognized by 7H in a similar way [14]. We were able to show that up to
three uracil units within a 15 base-pair triplex target sequence (amounting to a 20%
pyrimidine content) can be recognized, although with lower efficiency compared to a
canonical triplex structure. We have since expanded on that theme and reported that
selective recognition of a C ¥ G base pair within the parallel DNA triple-helical binding
motif can be achieved by a third strand containingthe base 5-methylpyrimidin-2(1 H)-
one (4 HT) [15]. The third strand affinities (KD) for a representative 15-mer duplex
sequence containingall four Watson-Crick base pairs opposite to this base were C ¥ G
(26 nm) ꢁ A ¥ T (270 nm) ꢀ T¥ A (350 nm) > G ¥ C (ca. 700 nm).
Fig. 1. A selection of stable, known mono- and bidentate base triples in the parallel triple-helical binding motif,
and the newly designed third-strand nucleosides 4 HT, Q, and R for the recognition of a C ¥ G base pair (box).
Potential dipolar CÀH ¥¥¥ O interactions between spatially close arrangements of carbonyl O-atoms and sp2
CÀH bonds are highlighted in bold.
Alongthese lines, we report here on the synthesis and triple-helix-forming
properties of the deoxynucleoside analogues Q (and R) with novel, purine-like
heterocyclic bases (Fig. 1). The rationale behind the design was to increase third-strand
affinity by increasingthe stackingsurface of the third-strand base while maintainingits
double H-bond acceptor character that has been shown to confer C ¥ G selectivity in the
case of 5-methylpyrimidin-2(1H)-one (4 HT).
Results and Discussion. Synthesis of Nucleosides Q (see 6). Startingfrom the
known 1H-imidazole-2-carboxamide (1) [16] (Scheme 1), base 3 was obtained in two

504
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 1. Synthesis of Nucleosides Q
BSA ˆ N,O-bis(trimethylsilyl)acetamide,
CEP-Cl ˆ 2-cyanoethyl
diisopropylphosphoramidochloridite,
DMTrCl ˆ (4,4'-dimethoxytriphenyl)methyl chloride
steps by treatment of 1 with BrCH2CH(OEt)2, followed by acid hydrolysis of the acetal
derivative 2 and condensation, in analogy to [17]. The base 3 in its silylated form was
subsequently used for nucleoside formation accordingto the one-pot procedure of
Vorbr¸ggen [18], in the presence of SnCl4 as the Lewis acid promotor and 4 as the sugar
component. The anomeric nucleoside-mixture 5 was thus obtained in 62% yield. As
expected, no other nucleoside isomers could be isolated. Subsequently, the anomer
mixture was saponified to give the unprotected nucleosides 6 (nucleosides Q) that were
tritylated to yield the mixture 7/8, which could be resolved by reversed-phase HPLC.
Both anomers 7 and 8 were then converted separately to the correspondingphosphor-
amidite buildingblocks 9 and 10 for oligonucleotide synthesis. The configurational
assignment (a-d,b-d) for 7 and 8 was made by comparison of the characteristic sugar-
proton resonances in the 1H-NMR spectra with those in the series of nucleosides R, for
which independent proof of configuration was obtained by X-ray analysis (vide infra).

Helvetica Chimica Acta Vol. 85 (2002)
505
Synthesis of Nucleosides R (see 14). The base triazolo[1,5-c]pyrimidin-5(3H)-one
12 (Scheme 2) that was needed for nucleoside synthesis1) was prepared in quantitative
yield in one step from the known 3-aminocytosine hydrochloride (11) [20] by treatment
with triethyl orthoformate and HCl. Nucleosidation was then effected again according
to the Vorbr¸ggen procedure [18] with the anomers methyl 2-deoxy-3,5-di-O-acetyl-d-
ribofuranoside (4) as the nucleosyl donor, the in situ persilylated base 12 as the
acceptor, and SnCl4 as the Lewis acid promotor, to give the anomeric-nucleoside
mixture 13 in 82% yield. From the three potentially nucleophilic ringN-atoms of 12,
only that of the pyrimidine-ringmoiety was found to be reactive under these conditions,
which is in accord with the literature [19]. No other isomeric nucleosides could be
isolated. Unfortunately, the anomer mixture could not easily be resolved at this stage,
which prompted us to continue the synthesis with the mixture. Ammonolysis of the
ester functions in 13 yielded the anomeric unprotected nucleosides 14 (nucleosides R)
which, again, since they could not be efficiently separated, were directly converted to
the trityl derivatives 15. At this stage, the two anomeric forms could be separated by
column chromatography. The b-d-anomer 15 was then further converted to phosphor-
amidite buildingblock 16.
Scheme 2. Synthesis of Nucleosides R
BSA ˆ N,O-bis(trimethylsilyl)acetamide, DMTrCl ˆ (4,4'-dimethoxytriphenyl)methyl chloride, CEP-Cl ˆ 2-
cyanoethyl diisopropylphosphoramidochloridite
To establish the point of attachment of the base and the configuration at the
anomeric center, crystals of a chromatographically enriched a-d-fraction of 13 were
obtained and subjected to X-ray analysis (Fig. 2)2). The crystal structure clearly
1
)
The synthesis of the b-d-ribonucleoside derivative of 12 and 2',3'-dideoxy variants thereof has already been
described [19].
Crystallographic data (excluding structure factors) for the structure reported in this paper have been
deposited with the Cambridge Crystallographic Data Centre as deposition No. CCDC-169141. Copies of the
data can be obtained, free of charge, on application to the CCDC, 12 Union Road, Cambridge CB2 1EZ
UK (fax: ‡ 44(1223)336033; e-mail: deposit@ccdc.cam.ac.uk).
2
)

506
Helvetica Chimica Acta Vol. 85 (2002)
confirmed the structural assignments and showed, in addition, the glycosidic bond to be
in the anti orientation. The sugar pucker is C(1')-exo, arranging the base into a
pseudoaxial position, a constellation that conforms well with the anomeric effect. The
bond lengths and angles in the molecule are normal within experimental error.
Fig. 2. Perspective view of the a-d-anomer of 13. Thermal ellipsoids are drawn at the 50% probability level.
Synthesis of Oligonucleotides. Oligonucleotides were prepared by the well-
established phosphoramidite protocol for automated DNA synthesis. The phosphor-
amidite buildingblock 16 of the b-d-nucleoside R seemed to be incorporated into
oligonucleotides with coupling yields of > 98% accordingto the trityl assay. It turned
out, however, that after completed synthesis and deprotection, no full-length
oligonucleotides could be detected by HPLC and ESI-MS. Instead, strand breakage
occurred at the position of incorporation of 16. Qualitative control experiments with 14
revealed that its nucleosidic bond is unstable under the detritylation conditions (2%
CCl3COOH in 1,2-dichloroethane) used duringoliognucleotide assembly. Thus, it
seems plausible that, after incorporation of 16, acid-catalyzed depurination occurs,
leadingto an abasic site, at which strand cleavage via b-elimination occurs during
ammonia treatment in the deprotection step.
Incorporation of buildingblocks 9 and 10 into oligonucleotides, on the other hand,
proceeded smoothly. To analyze the base-pairingproperties of 4 HT (the phosphor-
amidite buildingblock of which was prepared as described [21]) and nucleosides Q
towards double-stranded DNA, the oligonucleotide third strands and the DNA duplex
targets depicted in the Tables 1 4 were prepared. All oligonucleotides were purified by
DEAE ion-exchange HPLC, their purity controlled by reversed-phase HPLC, and
their structural integrity analyzed by ESI-mass spectrometry (see Exper. Part). Fig. 3

Helvetica Chimica Acta Vol. 85 (2002)
507
Fig. 3. Formulae of nonconventional nucleosides and their corresponding symbols used in this study (cf.
Tables 1 4).
gives an overview of the chemical formulae and symbols used for the evaluation of
triplex-bindingproperties, described in the followingsection.
Triplex-Forming Properties. First, insight into triplex stability and selectivity of 4 HT-
containingthird strands was obtained from UV/meltingcurves within the sequence
context depicted in Table 1. Possible structures of base triples for all four natural target
arrangements, as imposed by the parallel triple-helical binding motif, are depicted in
Fig. 4. As noted earlier [15], there is a strongpreference of 4 HT¥ C ¥ G base-triple
Fig. 4. Visualization of base-triple arrangements resulting from sequence design (cf. Table 1)

508
Helvetica Chimica Acta Vol. 85 (2002)
formation. The selectivity in target base-pair recognition follows the order C ¥ G > G ¥
C > A ¥ Tꢀ T¥ A over the whole pH range investigated. Differences in Tm between the
4 HT¥ G ¥ C base-triple and the next best arrangement amount to 4.7 7.5 K, at all pH
values. To probe the influence of the electrostatic environment at the interface of the
target base cytosine and 4 HT, meltingcurves were measured with a duplex, in which the
target deoxycytidine was replaced by 2'-deoxy-5-fluoro-cytidine (FC, Figs. 3 and 4). As
can bee seen from the corresponding Tm data (Table 1), this replacement strongly
destabilizes the triple helix (DTm ˆ 8.3 K, pH 6.0), underliningthe importance of the
local CÀH bond for stability. So far, the results obtained are in accord with a structural
model for C ¥ G recognition (Fig. 4, top, left) relyingon one conventional H-bond and,
perhaps, an additional dipolar interaction between HÀC(5) of cytosine and the
carbonyl O-atom of 4 HT (e.g., a CÀH ¥¥¥ O H-bond) weakly contributingto the stability
of the base triple. Thus, a similar situation as found previously for a 7H ¥ U ¥ A base triple
within a parallel triple helix is encountered [14].
Table 1. Tm Data [8] of Third-Strand Melting for the Indicated Triplexes. For buffer conditions, see Exper. Part.
Duplex target 5'-d(GCTAAAAAGAXAGAGAGATCG )-3'
3'-d(CGAT T T T TC TYTC TC TC TAGC)-5'
Third strand
5'-d(T T T T TCT ZTC TC TC T)-3'
pH
X ¥ Y (Z ˆ 4 HT)
A ¥ T
T¥ A
C ¥ G
G ¥ C
FC ¥ G
6.0
7.0
8.0
25.0
26.3
10.8
)
36.1
18.3
7.0
29.9
8.4
)
27.8
14.2
)
a
)
a
a
a
a
)
a) No Tm detectable.
Experiments towards the determination of the quality of the novel 4 HT analogues
b
aQ and Q (Fig. 3) in triple-helix formation were performed at first within the same
a
sequence context as before (Table 2). The corresponding Tm data for meltingof the Q-
containingtriplex, not unexpectedly, showed drastically reduced thermal stabilities
relative to the 4 HT-containingtriplexes, irrespective of the nature of the target base-
pair. With the exception of the G ¥ C target, Tm data are reminiscent of a mismatch
situation.
The picture becomes different when changing the anomeric configuration in Q. The
bQ-containing third strand, as expected, binds strongly to the C ¥ G target with affinities
that are similar to slightly inferior compared to 4 HT. Stability at pH 6.0 towards the
duplex target base pairs decreases in the order C ¥ G ꢀ G ¥ C > A ¥ T> T¥ A. A surprising
b
result is the bindingof Q to the G ¥ C target with similar thermal stability as to the C ¥ G
target. While C ¥ G recognition may be related to an analogous structure as postulated
for the case of 4 HT (Fig. 5,a), there is no obvious way for any H-bond formation
between bQ and G, due to the fact that both bases have only free H-bond acceptor sites
(Fig. 5,b). A closer understandingof this fact clearly needed further experimental
investigations.

Helvetica Chimica Acta Vol. 85 (2002)
509
Table 2. Tm Data [8] of Third-Strand Melting for the Indicated Triplexes. For buffer conditions, see Exper. Part.
Duplex target 5'-d(GCTAAAAAGAXAGAGAGATCG )-3'
3'-d(CGAT T T T TC TYTC TC TC TAGC)-5'
Third strand
5'-d(T T T T TCT ZTC TC TC T)-3'
pH
X ¥ Y (Z ˆ aQ)
X ¥ Y (Z ˆ bQ)
A ¥ T
T¥ A
C ¥ G
22.8
G ¥ C
A ¥ T
T¥ A
C ¥ G
G ¥ C
6.0
6.5
7.0
8.0
18.0
10.2
21.0
12.7
28.1
20.7
19.4
26.3
13.6
4.7
19.9
12.0
33.0
20.8
13.0
34.0
24.0
a
)
)
a
a
a
)
)
)
)
a)16.1
a
a
a
a
a
a
a
a
)
)
)
)
)
)
a) No Tm detectable.
To examine whether protonated species of bQ might be involved in guanine
recognition (e.g., Fig. 5,c), we changed to the triple-helix sequence context shown in
‡
Table 3, in which all charged C ¥ G ¥ C base-triple arrangements were replaced by non-
charged T¥ A ¥ T base triples. Thus, duplex recognition by third strands is intrisically pH
insensitive in this case. As can be seen from Table 3, third-strand meltingtemperatures
are virtually pH-independent in the pH range 6.0 8.0 for all duplex target base pairs,
rulingout protonated species of bQ to be involved in DNA bindingnear neutral pH.
There may be a weak dependence due to partial protonation of bQ, in the case of A ¥ T
base-pair recognition, which shows a tendency to increased triplex stability at
b
decreased pH. The relative order of stability of Q follows the line G ¥ C > A ¥ T> C ¥
G ꢀ T¥ A. Thus, the preferences are quite different compared to that of the previous
system (Table 2), indicatinga strongsequence dependence in the recongition
b
preferences of Q.
At this point, the question arose whether noncanonical duplex base pairs might be
b
involved in base-triple formation with Q. One such noncanonical arrangement could
arise from a conformational shift of the base from the anti- to the syn-form of the target
deoxyguanosine residue upon complexation with bQ (Fig. 5,d). Admittingprotonation
of the Watson-Crick partner cytosine, the thermodynamic drivingforce could be the
formation of a total of four instead of only three H-bonds in this base triple. It was
found earlier that guanosine mononucleotides prefer the syn-conformation of the base
[22]. Furthermore, deoxyguanosine residues with the base in the syn-conformation are
Table 3. Tm Data [8] of Third-Strand Melting for the Indicated Triplexes. For buffer conditions, see Exper. Part.
Duplex target 5'-d(GCTAAAAAAAXAAAAAAATCG)-3'
3'-d(CGAT T T T T T TYT T T T T T TAGC)-5'
Third strand
5'-d(T T T T T T TZ T T T T T T T)-3'
pH
X ¥ Y (Z ˆ bQ)
A ¥ T
T¥ A
C ¥ G
G ¥ C
6.0
6.5
7.0
8.0
27.0
26.3
25.6
25.4
21.0
22.1
22.6
22.8
23.1
22.0
22.2
22.9
32.7
32.7
33.3
33.3

510
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 5. Visualization of selected, putative base-triple arrangements in the triplexes described in Tables 2
4
encountered, e.g., in left-handed Z-DNA duplexes [22], or in duplexes containingG ¥ A
[23] or G ¥ G [24] mismatches. Introduction of bulky groups as, e.g., a Br-atom at C(8) of
guanine are known to enforce the syn-conformation of the base in the respective
nucleoside unit [22].
A test whether such noncanonical base triples (Fig. 5,d and e) might be created by a
bQ residue in the third strand was performed in the sequence context described in
Table 4. Forcingthe target base G into a syn-conformation, either by means of its 8-
bromo derivative 8BrG against C or G, or simply by creating a G ¥ G mismatch situation
in the target duplex, did lead to increases in thermal stability of only 2.9 3.2 K, relative
to the canonical G ¥ C base pair, at the best. It thus seems unlikely that such non-
canonical base triples were formed.
Conclusions. The results obtained in this study clearly show that, inspite of
increasingthe stackingsurface in bQ relative to 4 HT, no gain in affinity to a C ¥ G target
is observed within the parallel triple-helical bindingmotif. To the contrary, a loss of
selectivity was encountered. Various factors could be responsible for this. For example,
b
the geometry of the H-bond acceptor sites in Q is not identical to that of 4 HT, which

Helvetica Chimica Acta Vol. 85 (2002)
511
Table 4. Tm Data [8] of Third-Strand Melting for the Indicated Triplexes (duplex Tm×s in parenthesis). For buffer
conditions, see Exper. Part.
Duplex target 5'-d(GCTAAAAAGAXAGAGAGATCG )-3'
3'-d(CGAT T T T TC TYTC TC TC TAGC)-5'
Third strand
5'-d(T T T T TCT ZTC TC TC T)-3'
X ¥ Y (Z ˆ bQ)
pH 6.0
pH 7.0
pH 8.0
8BrG ¥ G
G ¥ G
37.2 (53.1)
36.9 (50.0)
31.0 (55.0)
34.0 (59.0)
18.8 (53.3)
18.0 (52.2)
15.0 (55.2)
13.0 (59.0)
a) (51.9)
a) (49.9)
a) (55.1)
a) (59.0)
8BrG ¥ C
G ¥ C
a) No Tm detectable.
might result in backbone distorsion upon base-pair formation with a C ¥ G target.
Alternatively it can not be excluded that other modes of recognition as, e.g., selective
intercalation of bQ in between target duplex base pairs does occur. In fact, the affinity
b
of Q for the G ¥ C base-pair may be the result of such an interaction.
It also remains open at this point whether the alignment of the carbonyl group of
4 HT with the HÀC(5) of its target base cytosine structurally occurs; and if so, whether
this arrangement contributes to the stability of the base triple, or has just to be
considered as a neutral Van der Waals contact. To determine this, more structural and
biophysical work as, e.g., an IR-spectroscopic investigation, is necessary.
Investigation of the pairing properties of 4 HT in the DNA triple-helix context
teached us that base ¥ base recognition relying on one H-bond only can be highly
selective. The major problem to be overcome in such cases, however, is the low target
affinity. Again in the context of DNA triplexes, it might, therefore, be of interest to
combine bases that recognize pyrimidines selectively by one H-bond, like 4 HT, with
sugar-backbone modifications that are known to enhance DNA affinity in a non-base-
dependent manner, e.g., via conformational restriction, as in locked nucleic acid (LNA)
[25][26] or via phosphate-charge screening, as in the case of 2'-aminoethyl-modified
RNA [27][28]. First attempts to use LNA in combination with such bases for targeting
pyrimidines in triple-helix formation are encouraging [29].
Financial support from the Swiss National Science Foundation (C.J.L) and the Roche Research Foundation
(I.P.) is gratefully acknowledged. We thank the BENEFRI Small Molecule Crystallography Service directed by
Prof. Helen Stoeckli-Evans for measuringthe X-ray structure.
Experimental Part
General. Solvents for chromatography and extractions: technical grade, distilled; Reagents: highest grade
available from Fluka or Aldrich. Standard conventions and equipment were used for analysis and character-
ization of new compounds (1H- and 13C-NMR: d in ppm, J in Hz. 31P-NMR (202 MHz); d in ppm relative to PPh3
as external standard).
1-(2,2-Diethoxyethyl)-1H-imidazole-2-carboxamide (2). Amide 1 (2.0 g, 18 mmol; prepared as described
[17]) was dissolved in dry DMF (50 ml). After addition of K2CO3 (6 g, 41 mmol) and BrCH2CH(OEt)2 (3 ml,
18 mmol), the soln. was heated to 1008 overnight. After filtration, the residual solid was washed with DMF
(50 ml) and the filtrate evaporated. CC (SiO2, AcOEt ! AcOEt/MeOH 9 :1) yielded 2 (3.39 g, 83%). White
solid. 1H-NMR (300 MHz, CDCl3): 7.22 (br. s, 1 H); 7.10 (s, 1 H); 6.99 (s, 1 H); 5.40 (br. s, 1 H); 4.68 (t, J ˆ 5.49,
1 H); 4.51 (d, J ˆ 5.16, 2 H); 3.69 (m, 2 H); 3.46 (m, 2 H); 1.14 (t, J ˆ 6.99, 6 H). 13C-NMR (300 MHz, CDCl3):

512
Helvetica Chimica Acta Vol. 85 (2002)
‡
‡
137.76, 127.42, 126.56, 101.74, 63.91, 50.98, 15.12. HR-LSI-MS: 228.134817 ([M ‡ H] , C10H18N3O3 ; calc.
228.134540).
Imidazo[1,2-a]pyrazin-8(7H)-one (3). To a suspension of 2 (3.40 g, 25 mmol) in H2O (68 ml), 5% HCl
soln. (17 ml) was added. After refluxingfor 3 h, the solvents were evaporated. The resultingproduct was
recrystallized from MeOH: 3 (1.90 g, 94%). White needles. 1H-NMR (300 MHz, (D6)DMSO): 12.21 (br. s,
1 H); 8.16 (d, J ˆ 1.47, 1 H); 7.99 (d, J ˆ 1.47, 1 H); 7.72 (d, J ˆ 5.52, 1 H); 7.22 (dd, J ˆ 5.88, 5.52, 1 H). 13C-NMR
‡
‡
(300 MHz, (D6)DMSO): 126.08, 120.33, 118.75, 107.56. HR-LSI-MS: 135.043262 ([M ‡ H] , C6H5N3O ; calc.
135.043010).
7-(3',5'-Di-O-acetyl-2'-deoxy-d-ribofuranosyl)imidazo[1,2-a]pyrazin-8(7H)-one (5). To a suspension of 3
(1.8 g, 13.3 mmol) in MeCN (150 ml), 4 (3.0 g, 13 mmol) and BSA (10.8 ml, 44 mmol) were added, and the
mixture was stirred at r.t. for 30 min. After coolingto 0 8, SnCl4 (5.5 ml, 52 mmol) was added and the ice bath
removed. Standard aq. workup after 3.5 h, followed by CC (SiO2, AcOEt/MeOH 9 :1) gave 5 (a-d/b-d 1 :1;
2.70 g, 62%). Yellowish oil. 1H-NMR (300 MHz, CDCl3): 7.47 (d, J ˆ 1.11, 2 H); 7.36 (d, J ˆ 1.1, 2 H); 7.18
(dd, J ˆ 6.24, 4.05, 2 H); 6.94 (d, J ˆ 6.24, 2 H); 6.61 (dd, J ˆ 8.46, 5.52, 1 H); 6.49 (dd, J ˆ 6.99, 2.22, 1 H); 5.18
(m, 2 H); 4.59 (m, 1 H); 4.31 (dd, J ˆ 4.41, 2.94, 2 H); 4.24 (m, 2 H); 4.16 (dd, J ˆ 4.41, 3.69, 1 H); 2.84 (m, 1 H);
2.54 (m, 1 H); 2.25 (m, 1 H); 2.10 (m, 1 H); 2.06, 2.05, 1.97, 1.92 (4s, 12 H). 13C-NMR (300 MHz, CDCl3):
170.20, 170.09, 169.80, 152.85, 152.78, 137.01, 136.77, 133.19, 133.11, 116.59, 116.36, 114.37, 113.80, 107.43, 106.45,
86.44, 84.26, 84.04, 81.99, 74.14, 74.09, 63.66, 63.56, 38.45, 37.59, 20.65, 20.63, 20.55, 20.53. HR-LSI-MS:
‡
‡
336.119420 ([M ‡ H] , C15H18N3O6 ; calc. 336.119561).
7-(2'-Deoxy-d-ribofuranosyl)imidazo[1,2-a]pyrazin-8(7H)-one (6). To a soln. of 5 (650 mg, 1.9 mmol) in
EtOH (100 ml), conc. NH3 soln. (100 ml) was added at À308. The soln. was allowed to reach r.t. overnight.
Evaporation yielded 6 (486 mg, 100%). Beige solid. 1H-NMR (300 MHz, (D6)DMSO): 8.43 (br. s, 0.5 H); 7.80
(s, 2 H); 7.60 (t, J ˆ 2.94, 2 H); 7.48 (d, J ˆ 2.94, 2 H); 7.37 (m, 2 H); 6.50 (dd, J ˆ 6.98, 6.62, 1 H); 6.43 (dd, J ˆ
7.72, 2.94, 1 H); 4.28 (br. m, 2 H); 4.26 (br. t, 2 H); 3.83 (m, 2 H); 3.61 (m, 2 H); 3.43 (d, J ˆ 4.78, 2 H); 2.66
(m, 1 H); 2.11 (m, 3 H). 13C-NMR (300 MHz, (D6)DMSO): 171.99, 133.04, 124.70, 117.88, 116.18, 115.26,
‡
107.95, 107.34, 91.34, 89.53, 87.79, 85.17, 83.52, 70.86, 62.04, 61.65, 22.71. HR-LSI-MS: 252.098770 ([M ‡ H] ,
‡
C11H14N3O4 ; calc. 252.098431).
7-{2'-Deoxy-5'-O-[(4,4'-dimethoxytriphenyl)methyl]-d-ribofuranosyl}imidazo[1,2-a]pyrazin-8(7H)-one (7/
8). To a soln. of 6 (400 mg, 1.6 mmol) in pyridine (5 ml) at 08, DMTrCl (600 mg, 1.7 mmol) was added in 6
portions. After stirringfor 2 h at r.t., the reaction was quenched with MeOH (1 ml), the mixture evaporated, and
the residue purified by CC (SiO2, CH2Cl2/MeOH/Et3N 97:2 :1): 7/8 (0.715 g, 80%). The mixture 7/8 was
separated by reversed-phase HPLC (gradient MeCN (1% Et3N/H2O).
b-d-Anomer 7: 320 mg(35%). White solid. 1H-NMR (300 MHz, CDCl3): 7.49 (d, J ˆ 1.11, 1 H); 7.39
(m, 3 H); 7.28 (m, 9 H); 6.82 (d, J ˆ 9.18, 4 H); 6.76 (d, J ˆ 5.88, 1 H); 6.70 (t, J ˆ 6.70, 1 H); 4.64 (m, 1 H); 4.11
(dd, J ˆ 7.35, 3.30, 1 H); 3.77 (s, 6 H); 3.48 (2dd, J ˆ 10.65, 10.29, 3.30, 2.94, 2 H); 2.60 (m, 1 H); 2.28 (m, 1 H).
13C-NMR (300 MHz, CDCl3): 138.12, 130.28, 130.08, 130.00, 128.22, 128.07, 127.93, 127.74, 126.90, 126.61, 116.53,
‡
115.81, 113.23, 113.02, 108.18, 105.96, 88.73, 88.04, 86.02, 82.53, 77.21, 55.24, 41.52. LSI-MS: 554 (14, M ). HR-
‡
‡
LSI-MS: 554.229250 ([M ‡ H] , C32H32N3O6 ; calc. 554.229111).
a-d-Anomer 8: 360 mg(40%). 1H-NMR (300 MHz, CDCl3): 7.43 (d, J ˆ 6.99, 3 H); 7.27 (m, 10 H); 7.04
(d, J ˆ 5.88, 1 H); 6.85 (d, J ˆ 8.82, 4 H); 6.48 (dd, J ˆ 7.35, 1.83, 1 H); 4.53 (m, 1 H); 3.80 (s, 6 H); 3.22 (dd, J ˆ
5.13, 4.80, 2 H); 2.87 (m, 1 H); 2.45 (d, J ˆ 14.7, 1 H). 13C-NMR (300 MHz, CDCl3): 158.5, 144.59, 135.798,
135.70, 132.91, 130.14, 130.12, 130.06, 130.0, 129.11, 128.22, 128.09, 127.92, 127.89, 127.81, 127.75, 127.04, 126.87,
116.59, 115.96, 113.16, 113.01, 106.03, 88.87, 87.94, 86.46, 84.29, 72.68, 71.57, 64.20, 55.22, 41.49. FAB-MS (pos.):
‡
‡
‡
554 (13, M ). HR-LSI-MS: 554.227480 ([M ‡ H] , C32H32N3O6 ; calc. 554.229111).
7-{2'-Deoxy-5'-O-[(4,4'-dimethoxytriphenyl)methyl]-b-d-ribofuranosyl}imidazo[1,2-a]pyrazin-8(7H)-one
3'-(2-Cyanoethyl Diisopropylphosphoramidite) (9). At r.t., 2-cyanoethyl diisopropylphosphoramidochloridite
(74 ml, 0.34 mmol) was added dropwise to a soln. of 7 (140 mg, 0.25 mmol) and iPr2NEt (110 ml, 0.66 mmol) in
THF (5 ml) at r.t. After 2 h at r.t., the reaction was quenched with sat. NaHCO3 soln. and the mixture extracted
with CH2Cl2 (2Â). After drying(MgSO 4) and evaporation of the org. phase, the residue was purified by CC
(SiO2, 10% Et3N in AcOEt/hexane 1:1): 9 (80%; 1:1 diastereoisomer mixture). White foam. 1H-NMR
(300 MHz, CDCl3): 7.50 7.20 (m, 43 H); 7.15 (d, J ˆ 3.63, 2 H); 7.10 (d, J ˆ 3.57, 1 H); 6.85 (d ‡ m, J ˆ 5.34,
13 H); 6.70 (dd, J ˆ 7.44, 1.14, 2 H); 6.56 (dd, J ˆ 3.18, 1.38, 1 H); 4.85 (m, 2 H); 4.73 (dd, J ˆ 2.37, 2.52, 1 H);
4.55 (dd, J ˆ 2.28, 2.52, 1 H); 4.51 (d, J ˆ 3.54, 1 H); 4.05 (m, 1 H); 3.87 (t, J ˆ 3.78, 3 H); 3.81 (s ‡ m, 20 H); 3.26
(m, 8 H); 3.18 (m, 2 H); 2.90 (m, 3 H); 2.60 (t, J ˆ 3.78, 3 H); 2.36 (m, 3 H); 1.10 1.35 (m, 26 H). 13C-NMR
(300 MHz, CDCl3): 156.34, 151.02, 142.42, 142.35, 135.26, 133.60, 133.52, 133.44, 131.23, 131.03, 130.98, 127.84,
127.82, 125.92, 125.91, 125.87, 125.83, 125.79, 125.69, 124.83, 124.78, 124.67, 114.40, 113.91, 113.86, 111.13, 111.11,

Helvetica Chimica Acta Vol. 85 (2002)
513
111.02, 104.25, 104.03, 86.51, 85.64, 85.57, 85.34, 84.81, 84.49, 84.35, 84.26, 75.26, 75.04, 74.84, 74.42, 72.15, 72.10,
70.41, 62.07, 61.89, 56.39, 56.35, 55.47, 53.03, 44.38, 43.29, 39.36, 38.63, 38.59, 19.76, 19.27, 16.81, 16.77, 6.37. 31P-
‡
‡
NMR (202 MHz, CDCl3): 140.1, 139.7. HR-LSI-MS: 770.332580 ([M ‡ H] , C41H49N5O8P ; calc. 770.331878).
7-{2'-Deoxy-5'-O-[(4,4'-dimethoxytriphenyl)methyl]-a-d-ribofuranosyl}imidazo[1,2-a]pyrazin-8(7H)-one
3'-(2-Cyanoethyl Diisopropylphosphoramidite (10). As described for 9, from 8 (140 mg, 0.25 mmol): 10 (207 mg,
1
79%). Colorless foam. H-NMR (300 MHz, CDCl3): 7.20 7.60 (m, 15 H); 6.60 6.90 (m, 3 H); 4.70 (m, 1 H);
3.78 (s, 3 H); 3.70 (m, 1 H); 3.75 (s, 3 H); 3.54 (m, 2 H); 2.2 2.8 (m, 4 H); 1.0 1.5 (m, 16 H). 13C-NMR
(300 MHz, CDCl3): 156.6, 156.59, 156.50, 156.47, 150.88, 142.19, 133.28, 133.19, 131.41, 131.21, 128.04, 128.00,
127.93, 126.15, 126.09, 126.05, 126.00, 125.83, 125.76, 125.72, 124.89, 124.84, 114.22, 113.99, 113.17, 113.13, 112.74,
112.71, 111.15, 111.08, 111.03, 104.80, 104.51, 104.46, 84.64, 83.18, 83.12, 83.05, 82.97, 81.95, 81.89, 81.58, 81.54,
75.03, 70.93, 70.70, 70.15, 69.92, 60.45, 60.12, 56.18, 56.17, 55.93, 55.92, 53.09, 53.08, 53.07, 53.05, 44.97, 43.65,
41.21, 41.13, 41.04, 40.96, 22.46, 22.39, 22.37, 22.35, 22.29, 22.25, 18.24, 18.14, 18.03, 17.93, 17.12, 17.09. 31P-NMR
‡
‡
(202 MHz, CDCl3): 148.59, 148.10. HR-LSI-MS: 770.33093 ([M ‡ H] , C41H49N5O8P ; calc. 770.331878).
[1,2,4]Triazolo[1,5-c]pyrimidin-5(1H)-one (12). To a soln. of 11 (530 mg, 3.25 mmol; prepared as described
[19]) in dry DMF (80 ml), HC(OEt)3 (16.6 ml, 0.1 mmol) and conc. HCl (1.6 ml) were added. The mixture was
heated overnight to 1308. After evaporation of solvents and purification by CC (SiO2, AcOEt/MeOH 10 :1), 2
1
(442 mg, 100%) was obtained. White solid. M.p. >2008. H-NMR (300 MHz, (D6)DMSO): 12.09 (br. s, 1 H);
8.41 (s, 1 H); 7.57 (d, J ˆ 5.88, 1 H); 6.76 (d, J ˆ 5.88, 1 H). 13C-NMR (300 MHz, (D6)DMSO): 154.31, 145.50,
‡
134.87, 95.86. EI-MS: 136 (38, M ), 108 (32), 95 (16), 81 (20).
6-(3',5'-Di-O-acetyl-2'-deoxy-d-ribofuranosyl)[1,2,4]triazolo[1,5-c]pyrimidin-5(6H)-one (13). To a susp. of
12 (305.4 mg, 2.24 mmol) in MeCN (30 ml), 4 (517 mg, 2.23 mmol) and BSA (1.2 ml, 4.93 mmol, 2.2 equiv.)
were added. After 30 min at r.t., the soln. was cooled to 08, and SnCl4 (0.525 ml, 4.93 mmol, 2.2 equiv.) was
added slowly. The mixture was allowed to warm to r.t., and, after 3.5 h, sat. NaCl soln. (75 ml) was added. The
mixture was extracted with CH2Cl2 (2Â), the combined org. layer dried (MgSO4) and evaporated, and the
residue purified by CC (SiO2, AcOEt/MeOH 10 :1): 13 (612 mg, 82%; a-d/b-d 1 :1). Colorless oil. 1H-NMR
(300 MHz, (D6)DMSO): 8.27 (s, 1 H); 7.57, 7.63 (2d, J ˆ 8.07, 1 H); 6.73 (t, J ˆ 7.17, 1 H); 6.51 (t, J ˆ 5.52, 0.5 H);
6.46 (d, J ˆ 5.88, 0.5 H); 5.25 (d, J ˆ 5.88, 1 H); 4.39 4.68 (m, 1 H); 4.23 4.37 (m, 2 H); 2.19 2.96 (m, 2 H).
13C-NMR (300 MHz, (D6)DMSO): 169.58, 154.69, 152.89, 144.41, 130.99, 97.10, 95.79, 88.63, 86.48, 85.14, 82.70,
‡
‡
73.79, 63.35, 38.13, 20.50. HR-LSI-MS: 337.114750 ([M ‡ H] ; C14H17N4O6 ; calc. 337.114810).
X-Ray Analysis of a-d-13. A suitable crystal (0.50 Â 0.30 Â 0.15 mm) was obtained from the soln. of a
chromatographically enriched a-d-fraction in CH2Cl2/hexane in the form of a colorless rod; monoclinic space
group P2(1). Intensity data were collected at 223 K on a Stoe Image Plate diffraction system (MoKa, l
0.71073 ä). Of The 2867 independent reflections (q ˆ 2.03 25 858), 2618 with F > 2s(I) were used in the
refinement. The structure was solved by direct methods with SHELXS-97 and refined with SHELXL-97. The
non-H-atoms were refined anisotropically, by means of weighted full-matrix least-squares on F2. H-Atoms were
included in calculated positions and treated as ridingatoms with SHELXL-97 default parameters (AFIX 137 for
the methyl H-atoms). The refinement converged at R ˆ 0.0264, s ˆ 0.993.
6-(2'-Deoxy-d-ribofuranosyl)[1,2,4]triazolo[1,5-c]pyrimidin-5(6H)-one (14). A soln. of 13 (564 mg,
1.67 mmol) in MeOH/conc. NH3 soln. 1:1 (180 ml) was allowed to warm up from À208 to r.t. overnight.
After evaporation, the crude product was washed with CH2Cl2 to give 14 (205 mg, 48%). White solid. 1H-NMR
(300 MHz, (D6)DMSO): 8.41 (d, J ˆ 2.22, 1 H); 7.98 (dd, J ˆ 7.71, 8.1, 1 H); 6.86 (dd, J ˆ 5.52, 2.19, 1 H); 6.30
6.38 (m, 1 H); 5.26 5.30 (m, 1 H); 5.10 (m, 1 H); 4.88 (t, J ˆ 4.59, 1 H); 3.87 4.28 (m, 1 H); 3.41 3.61
(m, 2 H); 2.04 2.93 (m, 2 H). 13C-NMR (300 MHz, (D6)DMSO): 181.63, 180.19, 171.63, 160.27, 122.34, 117.31,
‡
‡
113.09, 97.07, 88.00, 72.62. HR-LSI-MS: 253.094940 ([M ‡ H] ; C10H13N4O4 ; calc. 253.093680).
6-(2'-Deoxy-5'-O-[(4,4'-dimethoxytriphenyl)methyl]-d-ribofuranosyl)[1,2,4]triazolo[1,5-c]pyrimidin-
5(6H)-one (15). As described for 7 and 8; from 14 (400 mg, 1.58 mmol). CC (SiO2, CH2Cl2/MeOH/Et3N
97:2 :1) gave both 15 and its a-d anomer separately.
a-d-Anomer of 15: 166 mg(19%). 1H-NMR (300 MHz, CDCl3): 8.15 (s, 1 H); 7.83 (d, J ˆ 7.71, 1 H); 7.41
6.80 (m, 13 H); 6.64 (d, J ˆ 8.1, 1 H); 6.47 (d, J ˆ 5.88, 1 H); 4.52 (t, J ˆ 4.05, 1 H); 4.47 (d, J ˆ 5.52, 1 H); 3.77
(s, 6 H); 3.31 3.15 (m, 2 H); 2.82 (q, J ˆ 6.96, 1 H); 2.49 (q, J ˆ 7.35, 1 H). 13C-NMR (300 MHz, CDCl3):
158.55, 154.54, 153.41, 144.84, 144.39, 135.48, 133.39, 129.94, 127.91, 126.93, 113.20, 95.75, 89.49, 89.28, 86.58,
‡
‡
72.33, 63.96, 55.19, 41.63. HR-LSI-MS: 555.224670 ([M ‡ 1] , C31H31N4O6 ; calc. 555.224360).
b-d-Anomer 15: 210 mg(24%). 1H-NMR (300 MHz, CDCl3): 8.16 (s, 1 H); 8.00 (d, J ˆ 7.71, 1 H); 7.37
6.77 (m, 13 H); 6.52 (t, J ˆ 6.24, 1 H); 6.27 (d, J ˆ 8.07, 1 H); 4.65 (t, J ˆ 4.95, 1 H); 4.08 (m, 1 H); 3.73 (s, 6 H);
3.43 (m, 2 H); 2.56 (m, 1 H); 2.29 (m, 1 H). 13C-NMR (300 MHz, CDCl3): 158.68, 154.77, 153.20, 144.77, 144.20,

514
Helvetica Chimica Acta Vol. 85 (2002)
135.08, 132.55, 129.90, 127.89, 127.18, 113.26, 96.66, 87.13, 86.46, 86.37, 71.02, 62.58, 55.23, 41.80. FAB-MS (pos.):
‡
‡
631 (85, [M ‡ 76] ), 555 (23, [M ‡ 1] ), 303 (100).
6-{2'-Deoxy-5'-O-[(4,4'-dimethoxytriphenyl)methyl]-b-d-ribofuranosyl}[1,2,4]triazolo[1,5-c]pyrimidin-
5(6H)-one 3'-(2-Cyanoethyl Diisopropylphosphoramidite) (16). As described for 9, from 15 (b-d-anomer;
72 mg, 0.13 mmol). CC (SiO2, AcOEt/hexane 1:1 (1% Et3N) gave 16 (71 mg, 72.4%). White foam. H-NMR
1
(300 MHz, CDCl3): 8.22 (s, 1 H); 7.99 (dd, J ˆ 18, 8.1, 1 H); 7.40 6.80 (m, 13 H); 6.25 (dd, J ˆ 7.71, 2.94, 1 H);
6.53 (dd, J ˆ 10.68, 5.88, 1 H); 4.72 (m, 1 H); 4.17 (dd, J ˆ 11.76, 2.94, 1 H); 3.77 (s, 6 H); 3.63 3.35 (m, 5 H);
2.71 2.59 (m, 2 H); 2.46 2.42 (m, 2 H); 1.25 1.06 (m, 12 H). 13C-NMR (300 MHz, CDCl3): 158.67, 154.75,
153.19, 144.71, 144.12, 134.97, 132.42, 130.09, 127.94, 127.11, 117.32, 112.97, 96.69, 87.00, 86.28, 85.62, 71.86, 62.06,
57.96, 55.19, 43.09, 40.80, 24.41, 20.12. 31P-NMR (300 MHz, CDCl3): 149.2, 148.7. HR-LSI-MS: 755.341310
‡
‡
([M ‡ H] , C40H48N6O7P ; calc. 755.332212).
Oligonucleotide Synthesis. Oligonucleotides (Table 5) were synthesized with a Pharmacia Gene-Assembler-
Special DNA synthesizer usingstandard 2-cyanoethyl phosphoramidite chemistry. Reagents and concentrations
were as for the synthesis of natural DNA oligomers. The nonstandard phosphoramidite building blocks 8BrG, FC,
and C (Fig. 3) were from Glen Research. Syntheses were performed on a 1.3-mmol scale, accordingto the
manufacturer×s protocol. The only change made in the usual synthesis cycle was the prolongation of the coupling
time for the nonstandard buildingblocks to 6 min. Couplingyields were typically higher than 98% also for non-
standard phosphoramidites. After chain assembly and final detritylation, the oligomers were removed from the
support and deprotected by treatment with ca. 2 ml of conc. NH3 soln. (558, 16 h). The crude oligonucleotides
were purified by ion-exchange HPLC (MonoQ HR 10/10 (Pharmacia) column). The isolated oligonucleotides
were desalted over SEP-PAK-C-18 cartridges (Waters) accordingto standard procedures. The oligonucleotides
were further characterized by MALDI-TOF- or ESI-MS. Table 5 shows MS data of the oligonucleotides
containingnonstandard nucleotide units.
‡
Table 5. Calculated and Experimental Masses (M ) of Oligonucleotides from ESI-MS
Deoxyoligonucleotide
m/z
found
calc.
5'-d(TTTTTCT4 HTTCTCTCT)-3'
4480
4504
4504
6539
4509
6640
4476a)
4505
4505
6538
4509
6639
5'-d(TTTTTCTaQTCTCTCT)-3'
5'-d(TTTTTCTbQTCTCTCT)-3'
5'-d(GCTAAAAAGAFCAGAGAGATCG)-3'
5'-d(TTTTTTTbQ TTTTTTT)-3'
5'-d(GCTAAAAAGA8BrGAGAGAGATCG)-3'
a) Data from MALDI-TOF-MS.
UV/Melting Experiments. Meltingexperiments were performed on a Cary-3E-UV/VIS spectrophotometer
(Varian). Meltingcurves ( l 260 nm) were recorded for a consecutive heating(0 ! 908)-cooling-heating protocol
with a linear gradient of 0.58/min. All measurements were performed in buffered solns. (10 mm Na-cacodylate,
100 mm NaCl, 0.25 mm spermine). Stoichiometric amounts of each strand were used for triplex formation, and
the triplex conc. was kept at 1.7 2.0 mm. The final pH was adjusted to the desired values with 0.1m HCl or 0.1m
NaOH. Tm Data were obtained from the first derivative of the heatingcycle. As in almost all cases, hysteresis
formation was observed for third-strand meltingand annealing; these data do not reflect equilibrium binding
conditions.
REFERENCES
[1] S. Neidle, Anti-Cancer Drug Design 1997, 12, 433.
¬ ¡
[2] D. Praseuth, A. L. Guieysse, C. Helene, Biochim. Biophys. Acta 1999, 1489, 181.
[3] H. E. Moser, P. B. Dervan, Science (Washington, D.C.) 1987, 238, 645.
¬ ¡
[4] J.-C. FranÁois, T. Saison-Behmoaras, C. Helene, Nucleic Acids Res. 1988, 16, 11431.
[5] P. A. Beal, P. B. Dervan, Science (Washington, D.C.) 1991, 251, 1360.

Helvetica Chimica Acta Vol. 85 (2002)
515
[6] R. H. Durland, D. J. Kessler, S. Gunnell, M. Duvic, B. M. Pettitt, M. E. Hogan, Biochemistry 1991, 30, 9246.
[7] K. R. Fox, Curr. Med. Chem. 2000, 7, 17.
[8] H.-U. Stilz, P. B. Dervan, Biochemistry 1993, 32, 2177.
[9] D. M. Gowers, J. Bijapur, T. Brown, K. R. Fox, Biochemistry 1999, 38, 13747.
[10] T. E. Lehmann, W. A. Greenberg, D. A. Liberles, C. K. Wada, P. B. Dervan, Helv. Chim. Acta 1997, 80,
2002.
[11] C.-Y. Huang, G. Bi, P. S. Miller, Nucleic Acids Res. 1996, 24, 2606.
[12] R. H. Durland, T. S. Rao, V. Bodepudi, D. M. Seth, K. Jayaraman, G. R. Revankar, Nucleic Acids Res. 1995,
23, 647.
[13] J. Marfurt, S. P. Parel, C. J. Leumann, Nucleic Acids Res. 1997, 25, 1875.
[14] J. Marfurt, C. Leumann, Angew. Chem. 1998, 110, 184; Angew. Chem., Int. Ed. 1998, 37, 175.
¬
[15] I. Prevot-Halter, C. Leumann, Bioorg. Med. Chem. Lett. 1999, 9, 2657.
[16] J. P. Dirlam, R. B. James, E. V. Shoop, J. Heterocycl. Chem. 1980, 17, 409.
[17] J. Bradac, Z. Furek, D. Janezic, S. Molan, I. Smerkolj, B. Stanovnik, M. Tisler, B. Vercek, J. Org. Chem.
1977, 42, 4197.
[18] H. Vorbr¸ggen, B. Bennua, Chem. Ber. 1981, 114, 1279.
[19] V. Nair, A. G. Lyons, D. F. Purdy, Tetrahedron 1991, 47, 8949.
[20] K. Kohda, I. Kobayashi, K. Itano, S. Asano, Y. Kawazoe, Tetrahedron 1993, 49, 3947.
[21] −Oligonucleotides and Analogues A Practical Approach×, Ed. F. Eckstein, Oxford University Press,
Oxford, 1991.
[22] W. Saenger, −Principles of Nucleic Acid Structure×, Springer-Verlag, New York, 1984.
[23] T. Brown, G. A. Leonard, E. D. Booth, J. Chambers, J. Mol. Biol. 1989, 207, 455.
[24] N. Peyret, P. A. Seneviratne, H. T. Allawi, J. SantaLucia Jr., Biochemistry 1999, 38, 3468.
[25] A. A. Koshkin, S. K. Singh, P. Nielsen, V. K. Rajwanshi, R. Kumar, M. Meldgaard, C. E. Olsen, J. Wengel,
Tetrahedron 1998, 54, 3607.
[26] S. Obika, D. Nanbu, Y. Hari, J.-I. Andoh, K.-I. Morio, T. Doi, T. Imanishi, Tetrahedron Lett. 1998, 39, 5401.
[27] M. J. J. Blommers, F. Natt, W. Jahnke, B. Cuenoud, Biochemistry 1998, 37, 17714.
[28] B. Cuenoud, F. Casset, D. H¸sken, F. Natt, R. M. Wolf, K.-H. Altmann, P. Martin, H. E. Moser, Angew.
Chem. 1998, 110, 1350; Angew. Chem., Int. Ed. 1998, 37, 1288.
[29] S. Obika, Y. Hari, M. Sekiguchi, T. Imanishi, Angew. Chem. 2001, 113, 2149; Angew. Chem., Int. Ed. 2001,
40, 2079.
Received August 23, 2001