Design, Synthesis, and Self-Assembly of "Artificial Dinucleotide Duplexes"

Article abstract of DOI:10.1021/jo980194p

The synthesis of the A-U and G-C functionalized systems 1, 2, 3, and 4 has been accomplished using palladium-mediated cross-coupling reactions. These systems undergo self-association in nonpolar solvents such as CDCl₃ as judged from FABMS and NMR spectroscopic analyses.

Full text of DOI:10.1021/jo980194p

J . Org. Chem. 1998, 63, 4079-4091
4079
Design , Syn th esis, a n d Self-Assem bly of “Ar tificia l Din u cleotid e
Du p lexes”
J onathan L. Sessler* and Ruizheng Wang
Department of Chemisty and Biochemistry, The University of Texas at Austin, Austin, Texas 78712
Received February 3, 1998
The synthesis of the A-U and G-C functionalized systems 1, 2, 3, and 4 has been accomplished
using palladium-mediated cross-coupling reactions. These systems undergo self-association in
nonpolar solvents such as CDCl3 as judged from FABMS and NMR spectroscopic analyses.
Ch a r t 1
In tr od u ction
The specific hydrogen-bonding interactions that occur
between complementary DNA bases provide a time-
honored affirmation of how appropriately designed func-
tionality may be used to induce spontaneous assembly
in complex supramolecules. Indeed, the efficiency and
apparent simplicity whereby information is stored and
transferred in natural nucleic acids has inspired us1 and
others2-4 to reproduce the essential features of duplex
DNA using much simpler synthetic systems. Although
many self-complementary structures capable of undergo-
ing assembly have been synthesized of late,2,3 only a few
synthetic systems are known that employ DNA-like
complementary purine-pyrimidine base pairing (Chart
1) as the critical recognition motif.1,4 This seeming
unpopularity could reflect the fact that (1) the binding
affinities associated with adenosine(A)-thymidine(T) [or
uridine(U)] “dimerization” are too low (Ka ≈ 102 M-1 in
have high affinities for complementary association (Ka
≈ 104 M-1 in CHCl34e,f,6), are highly insoluble in organic
media and thus constitute species with which it is
difficult to work. Despite these potential difficulties, we
remain of the opinion that the continued study of nucleic
acid base derived systems is worthwhile; it could help
us to understand better the details of base pairing under
abiotic conditions and, perhaps, provide a means of
constructing complex, but well-defined synthetic arrays.
In this paper we discuss systems that are capable of
undergoing self-dimerization in accord with the general-
ized equilibrium shown in Scheme 1. We recently
reported, in Communication form,7 the synthesis and self-
assembling properties of I. In this paper we provide full
synthetic details for the mononuclear constituents of I
(from compound 1) and also report the synthesis and
solution-phase self-assembly properties of analogous
ensembles, namely, II (from 2), III (from 3), and IV (from
4) (Chart 2).
5
CDCl3 ) and (2) cytidine(C) and guanosine(G), which do
(1) (a) Sessler, J . L.; Magda, D. J .; Hugdahl, J . J . Incl. Phenomen.
1989, 7, 19. (b) Sessler, J . L.; Magda, D. In Inclusion Phenomena and
Molecular Recognition; Atwood, J . L., Ed., Plenum Press: New York,
1989; p 17. (c) Sessler, J . L.; Magda, D.; Furuta, H. J . Org. Chem. 1992,
57, 818.
(2) For recent reviews on self-assembly, see: (a) Lawrence, D. S.;
J iang, T.; Levett, M. Chem. Rev. 1995, 95, 2229. (b) Philp, D. S.;
Stoddart, J . F. Angew. Chem., Int. Ed. Engl. 1996, 35, 1154. (c)
Whitesides, G. M.; Simanek, E. E.; Mathias, J . P.; Seto, C. T.; Chin,
D. N.; Mammen, M.; Gordon, D. M. Acc. Chem. Res. 1995, 28, 37. (d)
Whitesides, G. M.; Mathias, J . P.; Seto, C. T. Science 1991, 254, 1312.
(e) Comprehensive Supramolecular Chemistry, Vol. 9; Atwood, J . L.,
Davies, J . E. D., Macnicol, D. D., Vogtle, F., Eds.; Elsevier: Exeter,
199(36). For recent examples on self-assembly, see: (a) Zimmerman, S.
C.; Zeng, F.; Reichert, D. E. C.; Kolotuchin, S. V. Science 1996, 271,
1095. (b) Kang, J .; Rebek, J ., J r. Nature 1996, 382, 239. (c) Grotzfeld,
R. M.; Branda, N.; Rebek, J ., J r. Science 1996, 271, 487. (d) Sessler, J .
L.; Andrievsky, A.; Gale, P. A.; Lynch, V. Angew. Chem., Int. Ed. Engl.
1996, 35, 2782. (e) Hamann, B. C.; Shimizu, K. D.; Rebek, J ., J r. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1326. (f) Tirumala, S.; Davis, J . T. J .
Am. Chem. Soc. 1997, 119, 2769. (g) Davis, J . T.; Tirumala, S. K.;
Marlow, A. L. J . Am. Chem. Soc. 1997, 119, 5271. (h) Nakamura, K.;
Sheu, C.; Keating, A. E.; Houk, K. N. J . Am. Chem. Soc. 1997, 119,
4321. (i) Meissner, R.; Garcias, X.; Mecozzi, S.; Rebek, J ., J r. J . Am.
Chem. Soc. 1997, 119, 77.
(4) (a) Schall, O. F.; Gokel, G. W. J . Am. Chem. Soc. 1994, 116, 6089.
(b) Schall, O. F.; Gokel, G. W. J . Chem. Soc., Chem. Commun. 1992,
748. (c) Schall, O. F.; Robinson, K.; Atwood, J . L.; Gokel, G. W. J . Am.
Chem. Soc. 1991, 113, 7434. (d) Kim, M.; Gokel, G. W. J . Chem. Soc.,
Chem. Commun. 1987, 1686. (e) Sessler, J . L.; Wang, B.; Harrimann,
A. J . Am. Chem. Soc. 1995, 117, 704. (f) Sessler, J . L.; Wang, B.;
Harrimann, A. J . Am. Chem. Soc. 1993, 115, 10418. (g) Harrimann,
A.; Kubo, Y.; Sessler, J . L. J . Am. Chem. Soc. 1992, 114, 388. (h) Kral,
V.; Sessler, J . L. Tetrahedron 1995, 51, 539.
Resu lts a n d Discu ssion
Molecu la r Design . The basic principles underlying
the design of the self-complementary systems 1-4 are
rigidity and solubility. The first of these design charac-
teristic is deemed essential since conformational flex-
ibility carries with it a large entropic “penalty” that must
be paid during the course of the self-assembly process.
Such a “payment” weakens the stability of the resulting
complexes and can also preclude entirely the formation
of the sought-after duplexlike dimer in the absence of
some further stabilizing interaction.4a-d,i The need for
solubility is even more straightforward. For a self-
dimerizing system to work, it must be rendered soluble
in a solvent where the putative hydrogen-bonding inter-
(6) (a) Murry, T. J .; Zimmerman, S. C. J . Am. Chem. Soc. 1992, 114,
4010. (b) Pranata, J .; Wierschke, S. O.; J orgensen, W. L. J . Am. Chem.
Soc. 1991, 113, 2810.
(5) (a) Kyogoku, Y.; Lord, R. C.; Rich, A. Proc. Biochim. Biophys.
Acta 1969, 179 10. (b) Kyogoku, Y.; Lord, R. C.; Rich, A. Natl. Acad.
Sci. U.S.A. 1967, 57, 250.
(7) Sessler, J . L.; Wang, R. J . Am. Chem. Soc. 1996, 118, 9808.
S0022-3263(98)00194-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 05/21/1998

4080 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
Sch em e 1
palladium-catalyzed cross-coupling reaction involves a
coupling between 1,8-diethynylanthracene 5 and 5-io-
douridine 6a . This reaction, carried out at room tem-
perature under argon, gives a mixture of both mono- and
bis-uridine- substituted anthracenes 7a and 8a . The
ratio of these two products depends largely on the ratio
of the two starting materials used. For instance, a 1:1
molar ratio of 5/6a gave a significant amount of the bis-
substituted compound 8a (>50%). While this material
could be useful as a control compound, it was not the
desired product in this particular synthesis. Fortunately,
it was found, after considerable experimentation, that
using a 3:1 molar ratio of 5/6a would give the mono-
uridine- substituted anthracene 7a as the major product
(89% yield based on 6a ). The alkyne functional group
in 7a could then be coupled with 8-bromoadenosine 9a
under conditions similar to those used in the first
coupling reaction (with the exception that a higher
temperature (60 °C) was necessary to complete the
transformation). Using this approach, the target com-
pound 1 was obtained in 71% yield after chromatographic
purification.
actions are appreciable. In practice, this means making
the target systems soluble in CHCl3 or similar solvents
that do not compete effectively for H-bond donor or
acceptor sites.
On the basis of an appreciation that rigidity and
solubility are critical, four initial target molecules 1-4
were chosen. These targets, designed to form self-
complementary duplexes, are similar in the general sense
that all of them combine the following features: (1)
complementary molecular recognition motifs involving
either adenosine/uridine or guanosine/cytidine base-
pairing subunits; (2) rigid aromatic spacer and an easy-
to-incorporate acetylene linker; (3) protecting groups on
the ribose subunits that render the target molecules
soluble in organic media.
Notwithstanding the similarities mentioned above,
these systems are structurally different. For instance,
1 differs from 4 in both the size of the protecting groups
and the number of hydrogen bonds; 2 differs from 1 and
4 in the size of the protecting groups and the number of
hydrogen bonds, respectively; and the only difference
between 1 and 3 is the degree of preorganization. Thus,
the study of these systems, it was expected, would not
only allow the effects of molecular rigidity to be evaluated
in terms of the effect such factors could have on aggrega-
tion behavior but also allow those associated with seem-
ingly minor variations in structure (e.g., number of
hydrogen bonds8) to be considered in a similar light.
Syn th esis of th e P r ecu r sor s to En sem bles I, II,
III, a n d IV. As mentioned earlier, all the monomeric
targets (i.e., 1-4) needed to prepare ensembles 1-IV
contain an acetylene linker. This structural motif serves
two functions: it provides both structural rigidity and
abets synthetic accessibility. Fortunately, this type of
linkage can now be introduced into many systems,
including monomers 1-4, as the result of palladium-
mediated cross-coupling chemistry.9-11 As it relates here,
this chemistry is summarized in Schemes 2-6. It is also
discussed explicitly below.
The same sequence of reactions was also used to
synthesize target molecule 2 (Scheme 2) from 6b and 9b15
in 45% overall yield.
Surprisingly and in contrast to what was true for the
syntheses of the analogous compounds 7a and 7b, the
reaction between 5-iodouridine 6a and 4,6-diethynyldiben-
zofuran 11 (Scheme 3), which was prepared in 88%
overall yield from 4,6-diiododibenzofuran 10,16 gave rise
to only a very low yield (5%) of mono-uridine-substituted
dibenzofuran 12 (i.e., the key intermediate in the syn-
thesis of target monomer 3). Instead, the bis-uridine-
substituted compound 13 was found to be the major
product (>90%), regardless of the bis-alkyne 11 to
iodouridine 6a ratio employed. An alternative route was
thus explored. Here, the basic idea was to switch the
functional groups on the two reactants in an attempt to
reduce their reactivity9 to the point where the reaction
would stop at the monosubstitution stage. In fact, as
summarized in Scheme 4, this approach worked well
when 5-ethynyluridine 14 was used as one reactant and
diiododibenzofuran 10 was used as the other.
As illustrated in Scheme 4, four successive palladium-
mediated couplings were needed to assemble the basic
skeleton of monomer 3. The first of these couplings
produced 5-ethynyluridine 15; it was obtained in 78%
overall yield from the reaction of 5-iodouridine 6a with
TMS-acetylene followed by a removal of the TMS group
with TBAF (to give 15). The second coupling reaction
involved the use of this precursor (15) and 4,6-diiodod-
ibenzofuran 10. This gave rise to a mixture of mono- and
bis-uridine-substituted products (compounds 16 and 13,
respectively). Here, in analogy to what was observed
earlier, the relative ratio of the resulting products was
found to depend on the relative ratio of the two starting
materials employed. This meant that, after optimization
of the reaction conditions, the mono iodo-substituted
dibenzofuran 16 could be obtained in 87% yield. The
same palladiun-mediated coupling sequence used to
prepare 15 was applied to this latter intermediate (i.e.,
The synthesis of 1,8-diethynylanthracene (5) from 1,8-
dichloroanthracene12 was described previously.13 5-Io-
douridines (6a ) and 8-bromoadenosines (9a ) were syn-
thesized according to the reported procedures.14 With
these intermediates in hand, the key C-C bond forma-
tion steps of target molecule 1 (i.e., the palladium-
catalyzed cross-coupling reactions) could be attempted.
Scheme 2 summarizes the relevant chemistry. The first
(8) J orgensen, W. L.; Pranata, J . J . Am. Chem. Soc. 1990, 112, 2008.
(9) Still, J . K. Angew. Chem., Int. Ed. Engl. 1986, 25, 1228.
(10) Schlosser, M., Ed. Organometallics in Synthesis; J ohn Wiley &
Sons: New York, 1994.
(11) (a) Heck, R. F. Palladium Reagents in Organic Syntheses;
Academic Press: London, 1985. (b) Heck, R. F. Org. React. 1982, 27,
345. (c) Sonogashira, K.; Hugihara, N. Tetrahedron Lett. 1975, 4467.
(12) (a) House, H. O.; Koepsell, D.; J aeger, W. J . Org. Chem. 1973,
38, 1167. (b) House, H. O.; Hrabie, J . A.; vanDerree, D. J . Org. Chem.
1986, 51, 921.
(13) Katz, H. E. J . Org. Chem. 1989, 54, 2179.
(14) (a) For sugar protection, see: Matsuda, A. Synthesis 1986, 385.
(b) For iodonation of uridines, see: Sy, W.-W. Synth. Commun. 1990,
20, 3391. (c) For bromination of adenosine/guanosine, see: Holmes,
R. E.; Robins, R. K. J . Am. Chem. Soc. 1964, 86, 1242.
(15) Wang, B. Ph.D. Dissertation, The University of Texas at Austin,
1994.
(16) Tsang, K. Y.; Diaz, H.; Graciani, N.; Kelly, J . W. J . Am. Chem.
Soc. 1994, 116, 3988.

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4081
Ch a r t 2
Sch em e 2
Sch em e 3
in 12 could then be coupled with the 8-bromoadenosine
9a at 60 °C to give the desired target monomer 3 in 67%
yield after chromatographic purification.
The synthesis of the G/C analogue 4 proved to be a
much more demanding task. Besides containing protect-
ing groups that are different than those found in other
targets, such as 3, this system also contains a guanosine
functionality. These two aspects complicated the syn-
16) to give the mono-uridine-substituted compound 12
in 80% overall yield. The alkyne functional group present

4082 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
Sch em e 4
Sch em e 5
thesis. After considerable experimentation, it was found
that the alkyne functional groups on the anthracene had
to be activated by forming the corresponding organostan-
nyl derivatives.17 Additionally, it was found that the
amino group of the guanosine derivative 18 had to be
protected. Fortunately the N,N-dimethylformamidine
functionality was found suitable for this purpose. The
synthetic sequence that resulted from these changes is
shown in Scheme 5. In terms of specifics, the amino
group of guanosine 18 was protected with N,N-dimeth-
ylformamidine in 84% yield by reacting with excess N,N-
dimethylformamide dimethyl acetal in anhydrous THF.
The coupling reaction between 5 and 2018 produced the
desired mono-cytidine-substituted compound 21 in good
yield (68%). The alkyne functional group of 21 was
converted to its organostannyl derivative 22 by being
heated in toluene at reflux in the presence of excess
tributyltin methoxide, a species that, in turn, was
synthesized from bis(tributyltin)oxide and dimethyl car-
bonate.19 The resulting organostannyl product 22 proved
to be unstable to column chromatographic purification.
Therefore, it was prepared and used in situ. This was
practical since toluene was the solvent used in both
reactions. Thus, a mixture of 19, 21, and tributyltin
methoxide was heated in toluene at reflux for 10 h in
the presence of Pd(PPh3)4 to give rise to 23 in 36% yield
after chromatographic purification. Finally, the amino
protecting group of compound 23 was removed by treat-
ment with saturated methanolic ammonia at room tem-
perature overnight. This gave target 4 in 81% yield after
silica gel column chromatographic purification.
An alternative approach to target molecule 4 was also
investigated. It is summarized in Scheme 6. In this case,
the starting materials are the bis-alkyne 5, tributyltin
methoxide, and 1 equiv of the 8-bromoguanosine deriva-
tive 19. After these materials were mixed together in
toluene and PdCl2(PPh3)2 added, the reaction was made
to proceed by being heated at reflux for 10 h. This gave
the mono-guanosine-substituted compound 25 in 52%
yield after silica gel column chromatographic purification.
Subsequent coupling of the resulting intermediate prod-
uct, 25, with 5-iodocytidine (20) gave compound 23 in
71% yield after column chromatographic purification.
Target 4 was then obtained in its final deprotected form
by being treated with ammonia in methanol.
Self-Assem bly of I, II, III, a n d IV. As stated before,
the primary goal of the work is to understand various
factors which influence the self-assembly of duplexes via
base pairing. Thus, once the synthesis of the target
monomers was complete, the next task was to find out
whether these presumed-to-be self-complementary com-
pounds would in fact self-assemble to form the expected
corresponding duplexes, namely, structures I,7 II, III, and
IV. Here, the goals were to determine how structural
variations such as the size of protecting groups, degree
of preorganization, and number of hydrogen bond donors
and acceptors in the monomers would affect the stability
and architecture of the resulting complexes. These issues
(17) Logue, M. W.; Teng, K. J . Org. Chem. 1982, 47, 2549.
(18) Sessler, J . L.; Brown, C. T.; Wang, R.; Hirose, T. Inorg. Chim.
Acta 1996, 251, 135.
(19) Davies, A. G.; Kleinschmidt, D. C.; Palan, P. R.; Vasishtha, S.
C. J . Chem. Soc. (B) 1971, 3972.

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4083
Sch em e 6
Ta ble 1. F ABMS Ch a r a cter iza tion of En sem bles I, II,
III, a n d IV
1 self-associates to form a tightly bonded duplex I and
does so in a way such that two solvent molecules are
associated with the duplex. This interpretation was
further supported by elemental analyses. For instance,
when compound 1 was recrystallized from 1,2-dichloro-
ethane, the following elemental analysis was obtained:
C, 56.51; H, 4.43; N, 8.94; Cl, 6.59. The calculated values
for bis-1,2-dichloroethane adduct of I (C49H43 N7O16)2‚(C2H4-
Cl2)2 are as follows: C, 56.5; H, 4.37; N, 9.05; Cl, 6.46.
Recrystallization of compound 1 from a mixture of
benzene and dichloromethane, on the other hand, pro-
duced these microanalysis values: C, 62.07; H, 4.65; N,
9.17. This corresponds to dimeric species I that either
incorporates two benzene molecules within its boxlike
walls or has two benzene molecules associated with it in
some fashion (C49H43N7O16)2‚(C6H6)2: C, 62.09; H, 4.64;
N, 9.21. In the absence of X-ray crystallographic data,
it is currently unknown which of these two limiting
scenarios pertains in the solid state. In solution the
trapped guests (i.e., 1,2-dichloroethane or benzene) are
in fast exchange with solvent molecules as evidenced by
the fact that unchanged chemical shifts were observed
observed peak
found
composition
calculation
I
1970.5520
2836.4712
1951.5210 C94H83N14O34
2866.4900 146H233N16O20Si12
C
C
98H86N14O32
146H231N14O20Si12
1970.5533
2836.4721
1951.5196
2866.4939
II + 1
III + 1
IV + 1
C
were explored using mass spectrometric, vapor pressure
osmometric (VPO), and NMR spectroscopic analyses.
Ma ss Sp ectr om etr y. Initial evidence that stable self-
assembled complexes I, II, III, and IV could be formed
from the monomeric precursors 1, 2, 3, and 4 came from
fast atom bombardment mass spectrometric (FAB-MS)
experiments. Both peaks corresponding to the individual
monomeric components and peaks attributable to their
dimeric complexes (i.e., I, II, III, and IV, respectively)
were observed. On the other hand, control experiments
conducted under the same conditions using mixtures of
7a /9a , 9a /13, and 1/4 did not give the corresponding
noncovalent dimeric peaks. This indicates that the
observed noncovalent dimeric species were formed as the
result of complementary base-pair interactions.
Interestingly, when a 1/1 mixture of compounds 1 and
3 was measured using the same conditions, a new peak
at m/ z 1962 in addition to those expected for I and III
was observed. This peak most likely belongs to the
mismatched heterodimeric species between 1 and 3. This
result was not surprising since the two compounds are
functionally complementary to each other. This type of
mismatched phenomenon was also observed by Rebek et
al.20 High-resolution mass spectrometric analysis of
these dimeric signals proved to be consistent with their
proposed chemical formulas. The results are summa-
rized in Table 1.
VP O Exp er im en ts. The average molecular weights
of I and III were also determined in solution using vapor
pressure osmometry (VPO) using 1, 2-dichloroethane as
the solvent and compound 13 as the molecular weight
standard. From these measurements, a Mw value of 2210
( 45 (from three independent measurements) was ob-
tained for I. This value is in agreement with the
molecular weight of dimer I containing two trapped or
bound dichloroethane molecules (Mw ) 2168). In con-
trast, the VPO measurements for system 3 gave a value
of 1190 ( 30, which is close to the molecular weight of
the monomeric species (Mw ) 950). Taken together, these
results provide an indication that I is a more tightly
bound complex than III.
1
by H NMR spectroscopy (see Supporting Information).
Thus, ensemble I cannot be considered as being a
permanent chemical “box” or self-assembled receptor for
small molecules in solution.20
Resu lts fr om 1H NMR. Further evidence that com-
pounds 1, 2, 3, and 4 self-associate through base pairing
came from 1H NMR spectroscopic studies. In a first
study, chemical shifts of the imino N-H protons of the
uridine moieties of compound 1, 2, and 3, and of gua-
nosine moiety of compound 4, were compared with those
of their control molecules 7a /8a , 12/13, and 19/23,
respectively. Two different solvents were used for these
comparisons: CDCl3, an aprotic noncompetitive solvent
which was expected to favor base pairing, and DMSO-
d6, an aprotic but highly competitive solvent which was
expected to disrupt putative hydrogen-bonding interac-
tions. The results of this study are summarized in Table
2.
The significant downfield shifts in CDCl3, ∆δ ) 5.4
ppm for 1, 4.1 ppm for 2, 3.6 ppm for 3, and 2.9 ppm for
4, as compared to their respective controls, are most
readily interpreted in terms of the presence of strong self-
associating interactions mediated through hydrogen-
bonds. Supporting this conclusion was the fact that such
downfield shifts were not seen in DMSO-d6 as evidenced
by the fact that the chemical shifts of all the comple-
mentary compounds are virtually the same as those of
their controls. Under these conditions the monomeric
species predominate. These are presumably stabilized
by solvating interactions involving the DMSO-d6.
Resu lt fr om Elem en ta l An a lysis. The VPO experi-
ments are consistent with a scenario wherein compound
(20) Valdes, C.; Spitz, U. P.; Toledo, L.; Kubik, S.; Rebek, J ., J r. J .
Am. Chem. Soc. 1995, 117, 12733.

4084 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
Ta ble 2. N-H Reson a n ce (p p m ) in CDCl3 a n d DMSO-d 6
compound
12
7a
8a
1/I
2/II
13
3/III
19
23
4/IV
CDCl3
DMSO-d6
9.22
11.74
9.30
11.98
14.68
11.60
13.58
11.84
9.60
11.92
9.61
11.86
13.22
11.91
10.11
11.15
10.06
11.10
13.52
11.12
F igu r e 2. Partial 1H NMR spectrum of 2/II (300 MHz, CDCl3).
F igu r e 3. 1H NMR spectrum of 3/III (500 MHz, CDCl3).
protecting groups on the ribose subunits. Apparently,
the larger protecting groups present in monomer 2 lead
to the formation of a less-stabilized duplex (II).
F igu r e 1. 15N-1H HMQC spectrum of I at room temperature
in CDCl3. A single imino NH proton at 14.7 ppm and two
unequivalent amino NH2 separated by 2.5 ppm are attached
to the same nitrogen.
The duplex formed from 3, i.e., complex III, was also
found to be less stable than that formed from 1. Specif-
ically, the upfield shift of the imino proton, observed in
the 1H NMR spectrum of III in CDCl3 (Figure 3) as
compared with that of I, and the equivalence of the two
amino protons of III are interpreted in terms of the
dibenzofuran-based analogue III being less stable than
the anthracene-based analogue I. This conclusion was
further supported by 2D-NOESY experiments (Figure
4): For I, strong cross-couplings are observed between
the imino N-H signal of the uridine moiety and the two
inequivalent amino NH2 signals, as well as between the
imino proton N-H and H2 of adenosine. By contrast,
such cross-couplings are not seen in the case of II and
III under identical conditions. These results are thus
interpreted in terms of the duplex I being strongly
associated and complexes II and III being loosely bound.
As pointed out earlier, the only difference between 1
and 3 is the degree of preorganization of the two recogni-
tion units. The two nearly parallel recognition units in
the anthracene-based system I (according to CPK model)
seem to be oriented in the ideal geometry for duplex
formation. In the dibenzofuran-based system 3, these
same recognition units are constrained to a “V” config-
uration, where the hydrogen bonds might have to distort
to some degree in order to form the duplex III. Since
hydrogen bonds are directional interactions, the reduced
preorganization in the case of 3/III is believed to be at
the root of the lowered duplex stability.
1
The H-15N HMQC spectrum of I in CDCl3 (Figure 1)
provides evidence for the protons resonating at 9.5 and
7.0 ppm being located on the same nitrogen; they are thus
assigned to the amino group, NH2. Both the inequiva-
lence and large separation (∼2.5 ppm) of these two amino
protons are consistent with rotation about the amino-
C4 bond being slow on the NMR time scale in ensemble
I. Such a slow bond rotation is the result, presumably,
of strong hydrogen bond interactions involving the amino
group. Support for this conclusion comes from the
finding that when DMSO-d6 is used as the solvent, the
two amino protons become equivalent. This result
provides prima facie evidence that the inequivalence of
the amino group of I in CDCl3 is caused by the presence
of strong hydrogen bonds.
1
The H NMR spectrum of 2 in CDCl3 is quite different
from that of 1 (Figure 2). For instance, the very sharp
peak at 14.7 ppm, ascribed to the imino N-H of the
uridine moiety of 1, is replaced by a much broader and
upfield shifted peak at 13.7 ppm in the case of 2.
Further, in 2 the two amino protons are found to be
equivalent. Considered jointly, these findings are taken
as an indication that ensemble II is a loosely bound,
hydrogen-bonded complex. As pointed out earlier, the
only difference between 1 and 2 is the size of the

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4085
not the C2 carbonyl group, participating in the hydrogen
bond process. In other words, it is a normal Watson-
Crick base-pairing interaction and not a reversed one
that dominates the recognition processes.
1H NMR Titr a tion . Attempts to estimate quantita-
tively the thermodynamic parameters, including binding
constants for formation of I, II, III, and IV, failed due to
the simple fact that we were unable to find solution phase
conditions for any of the four ensembles in question under
which they were dissociated to such a degree that
accurate measurements of the binding constants could
be carried out. For instance, the chemical shifts of the
imino N-H protons for these systems are virtually
concentration independent (the chemical shifts of 2 and
3 show small concentration dependences in CDCl3, i.e.,
∼0.8 ppm from 10-2 to 10-4 M) in nonpolar solvents such
as CDCl3.22 Therefore, the dilution method in nonpolar
solvent could not be used accurately in this instance.23
Although the four duplex systems under consideration
do show concentration dependence when mixtures of
DMSO-d6/CDCl3 are used as solvents, they do not display
concentration dependence in the same solvent mixtures.
For instance, 3/III displayed concentration dependence
in 5% DMSO-d6/CDCl3, under which conditions a self-
aggregation binding constant of 10.6 M-1 was obtained
(see Supporting Information for details). By contrast, in
this same solvent mixture, no concentration dependence
was observed for the NH signals of 1/I. In fact, to
calculate a binding constant for I accurately, a solvent
mixture containing at least 35% DMSO-d6 in chloro-
form-d had to be used. Under these conditions, however,
complexes II, III, and IV were found to be totally
dissociated into their respective monomers. Despite
these limitations, titration experiments involving the
addition of DMSO-d6 to a solution of I, II, III, and IV in
CDCl3 were carried out; these studies, it was thought,
would still provide valuable information about their self-
association, at least in a qualitative sense. Specifically,
it was expected that the stronger the complexes, the more
DMSO-d6 would be needed to disrupt the hydrogen-bond
interactions. Further details of these experiments are
given below.
F igu r e 4. 2D-NOESY spectrum of I (500 MHz, CDCl3)
showing the strong cross-coupling of the imino NH with the
exocyclic amino NH2 and the H2 proton of adenosine moiety.
Ch a r t 3
Ta ble 3. 13C Reson a n ces of C4 Ca r bon yl a n d C2
Ca r bon yl of th e Ur id in e Moieties Con ta in ed in
En sem bles I a n d III
compound
1/I
7a
8a
3/III
12
13
C4
C2
162.41
149.40
160.48
149.35
160.85
149.40
162.79
149.00
160.85
149.20
160.76
149.27
13C NMR Stu d ies. Although 13C NMR spectroscopy
is not frequently used to study hydrogen-bonded com-
plexes21 because of the smaller changes in 13C chemical
shift observed on binding, it sometimes provides unique
The relevant region of the 1H NMR spectra of ensemble
I obtained from the DMSO-d6 titration experiments are
reproduced in Figure 5. In accord with the results
discussed earlier, the signal at 14.6 ppm is ascribed to
the hydrogen-bonded ensemble I while the signal at 11.6
ppm is assigned to the monomeric species 1. To the
extent that such assignments are correct, these spectra
clearly indicate that ensemble I remains intact with up
to 30% (v/v) DMSO-d6 as cosolvent. However, this same
ensemble becomes completely dissociated when the rela-
tive DMSO-d6 concentration reaches 60% or higher. In
the regime where the DMSO-d6 concentration in CDCl3
falls between 35% and 50% v/v, two signals for the imino
N-H resonances, corresponding to monomer and duplex-
like dimer, were observed.
1
information that can complement that obtained from H
NMR spectroscopic analyses. In the present instance, it
was considered that 13C NMR spectroscopy might provide
a means of differentiating between Watson-Crick and
reversed Watson-Crick binding modes (Chart 3). In the
case of the Watson-Crick base pairing model, the imino
NH and the C4 carbonyl group of the uridine participate
in hydrogen bonding with adenosine. On the other hand,
in the reversed Watson-Crick model, it is the C2
carbonyl group plus the imino NH of uridine that interact
with adenosine. Therefore, it might be expected that the
chemical shifts of C4 and C2 of the uridine moieties in I
and III might differ from those observed in the respective
control compounds, namely, 7a /8a and 12/13. Actual
experimental findings are summarized in Table 3. These
results serve to show that in both I and III it is the signal
of C4 that experiences a significant downfield shift (∼2
ppm) relative to control. By contrast, the chemical shifts
of C2 in I and III remain largely unchanged. These
findings are consistent with the C4 carbonyl group, but
1
The fact that two separate signals are seen in the H
NMR spectrum of ensemble I is consistent with exchange
between the solvated monomer 1 and dimer I being slow
(22) Compounds 2 and 3 show a small concentration dependence;
the chemical shifts changed by 0.8 and 0.3 ppm, respectively, upon
moving from 10-2 to 10-4 M at rt in CDCl3.
(23) (a) Connors, K. A. Binding Constants; J ohn Wiley & Sons: New
York, 1987. (b) Feeney, J .; Batchelor, J . G.; Albrand, J . P.; Roberts, G.
C. K. J . Magn. Reson. 1979, 33, 519.
(21) Iwahashi, H.; Kyogoku, Y. J . Am. Chem. Soc. 1977, 99, 7761.

4086 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
F igu r e 5. Stacked plots derived from 1H NMR spectroscopic
titration studies of 1/I. The DMSO-d6 composition in CDCl3
from bottom to top: 10%; 20%; 30%; 35%; 38%; 40%; 42%; 45%;
48%; 50%; 60%.
F igu r e 6. Stacked plots derived from 1H NMR spectroscopic
titration studies of 2/II. The DMSO-d6 composition in CDCl3
from bottom to top: 0%; 5%; 10%; 15%; 20%; 25%; 30%; 35%;
40%; 50%.
on the NMR time scale. Such slow exchange is usually
observed in cases of very tightly bound complexes.23 In
45% (v/v) DMSO-d6/CDCl3, the two uridine NH peaks
corresponding to monomer 1 and dimer I are sufficiently
similar in size that their areas may be determined
accurately by integration. This gives the relative ratio
of species 1 and I, after accounting for stoichiometry.
With a knowledge of the total amount of starting ligand
1 (3.94 × 10-2 molar), a self-association constant of 35 (
5 M-1, corresponding to the formation of I, could thus be
calculated directly. It was found from this calculation
(see Experimental Section for details) that 4 equiv of
DMSO are involved in the equilibrium; in other words,
each monomeric 1 molecule is solvated by four DMSO
molecules. Therefore, the calculated self-association
constant of ∼35 M-1 is an effective one relevant only
under these particular mixed solvent conditions. None-
theless, these studies lead us to conclude that self-
association of I is favorable not only in pure CDCl3 but
also under these mixed solvent conditions.
As demonstrated above, ensemble I shows remarkable
resistance toward dissociation in the presence of DMSO.
By contrast, solutions of ensembles II, III, and IV in
CDCl3 can easily be dissociated by the addition of DMSO-
d6. For instance, Figure 6 shows that the very broad
signal from the imino NH observed at 13.0 ppm in the
1H NMR spectrum of II became sharp when DMSO-d6
was added. Further, the signal quickly became insensi-
tive to DMSO-d6 concentration (δ ) 11.98 ppm) when
30% DMSO v/v was added to what was initially a pure
CDCl3 solution. Such findings are consistent with en-
semble II being able to survive only a relatively low
(<30%) concentration of DMSO before becoming totally
dissociated to its mononeric constituent 2. Such a
conclusion, in turn, supports the notion that protecting
group steric size plays a critical role in regulating the
monomer-dimer equilibrium.
F igu r e 7. Stacked plots derived from 1H NMR spectroscopic
titration studies of 3/III. The DMSO-d6 composition in CDCl3
from bottom to top: 0%; 10%; 15%; 20%; 25%; 30%; 35%.
pure CDCl3 was shifted to 11.9 ppm in 25% DMSO-d6/
CDCl3 (Figure 7) and that the same chemical shift value
was obtained when pure DMSO was used as the solvent.
These experimental results can be rationalized in terms
of the preorganization arguments discussed earlier.
Figure 8 shows the DMSO-based 1H NMR titration
profile for the GC analogue IV. Unlike ensemble II,
where the NH signal becomes sharper upon DMSO-d6
addition, the signal of the imino proton of IV, at 13.5 ppm
in pure CDCl3, broadens when DMSO-d6 is added. In
fact, this signal disappears (at a δ value of ca. 12.4 ppm)
once a DMSO-d6 concentration g40% is attained.
It is not surprising that both ensembles II and III are
more easily dissociated upon DMSO addition than en-
semble I. However, it is somewhat unexpected that it
takes less DMSO to dissociate ensemble IV than it does
to break up duplex I, since the former contains 50% more
hydrogen bonds than the latter. However, as pointed out
earlier, I differs from IV not only in the number of
hydrogen bonds but also in the size of protecting groups.
These latter can play a significant role as demonstrated
earlier in the case of I vs II. Apparently, the destabiliz-
ing steric effects from the larger protecting groups
present in IV are sufficient to offset the enthalpy gained
On the basis of a similar analysis, it was determined
that ensemble III is even less stable; it can only survive
25% DMSO before becoming fully dissociated to its
monomer 3. This conclusion is inferred from the fact that
the signal of the imino N-H proton at δ ) 13.2 ppm in

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4087
1
F igu r e 9. Partial H NMR spectra obtained in the course of
carrying out a temperature-dependence study of compound I-
(1) (500 MHz, CDCl3).
F igu r e 8. Stacked plots derived from 1H NMR spectroscopic
titration studies of 4/IV. The DMSO-d6 composition in CDCl3
from bottom to top: 0%; 5%; 10%; 15%; 20%; 25%; 30%; 35%;
40%; 45%; 50%.
from the 50% increase in the number of hydrogen-
bonding interactions. When the aggregation properties
of 2 and 4, systems that differ only in the number of
available H-bonding sites, are compared, it becomes
apparent that more DMSO (∼40% vs ∼30%) is needed
to dissociate the dimer corresponding to 4 (ensemble IV)
than that derived from 2 (ensemble II). This provides a
cogent “reminder” that the number of hydrogen bonds
still plays a critical role in mediating the aggregation
properties of these kinds of self-assembling systems.
Although the exact nature of the above steric hindrance
effect is not known, we speculate that it reflects a
conformational change involving the two nucleosides. In
order for compounds 1-4 to form duplexlike structures
cooperatively, the two recognition units have to adopt
“syn” conformations. The adoption of such conformations
may be possible for a system with small protecting groups
such as 1. For compounds 2 and 4, however, the two
bulky ribose groups are forced to adopt a more favorable
“anti” conformation to avoid the steric repulsions between
the two ribose subunits. As a result, an extra energy
barrier has to be overcome before the duplexlike dimer
can form.
F igu r e 10. Partial 1H NMR spectra obtained in the course
of carrying out a temperature-dependence study of compound
II(2) (500 MHz, CDCl3).
Tem p er a tu r e-Dep en d en ce Stu d ies. In an effort to
study more fully the equilibria leading to the formation
1
of ensembles I, II, III, and IV, variable temperature H
F igu r e 11. Partial 1H NMR spectra obtained in the course
of carrying out a temperature-dependence study of compound
III(3) (500 MHz, CDCl3).
NMR studies were undertaken. For ensemble I, tem-
perature has little effect on the monomer (1) to dimer (I)
equilibrium in the nonpolar solvent CDCl3,24 as judged
1
from the invariant H NMR spectra recorded between 328
into two separate peaks at 320 K; the same resonance
becomes even more complicated when the temperature
decreases. At 220 K, at least six imino NH resonance
signals are observed. Such complicated spectra, appar-
ently the results of the presence of a large protecting
group on the ribose subunit as discussed earlier, is best
interpreted in terms of a scenario wherein the duplex-
like ensemble II is not the only hydrogen-bonded complex
in solution. In other words, some other complex geom-
etries (i.e., trimer, oligomer, etc.) may also exist. These
aggregated forms become stable enough at low temper-
atures that multiple resonances from each corresponding
and 218 K (Figure 9).
By contrast, changes in temperature have a drastic
effect on the monomer/dimer equilibria of II, III, and IV
in nonpolar solvents such as CDCl3 and CD2Cl2. For
instance, in contrast to system 1 in CDCl3, where only
the duplexlike ensemble I is observed, system 2 shows
strong temperature dependence under the same condi-
tions (Figure 10). A very broad resonance at 280 K splits
(24) System 1/I shows a clear temperature dependence in 45%
DMSO-d6/CDCl3: low temperature favors the dimeric species I, while
high temperature favors the monomeric species 1. A van’t Hoff analysis
of the data reveals that ensemble I is 14.8 kcal mol-1 enthalpically
more favorable than the corresponding monomeric species 1; however,
ensemble I is also less favored entropically under these particular
conditions by 42.3 cal K-1 (see Supporting Information for details).
1
imino NH are observed in the H NMR spectrum.
Figure 11 shows the spectra obtained from the tem-
perature-dependence study of system III. In analogy to

4088 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
romethane when recrystallized from a mixture of both
solvents? Presently, we are working to prepare yet-larger
systems in an effort to address these issues.
Exp er im en ta l Section
Gen er a l In for m a tion . The nucleic acid base precursors
were purchased from Sigma Chemical Co. All other solvents
and reagents were of reagent grade quality and used as
received except as noted below. Toluene and tetrahydrofuran
(THF) when used as solvents for Pd-catalyzed cross-coupling
reactions were distilled from sodium and sodium/benzophe-
none, respectively. Anhydrous N,N-dimethylformamide (DMF)
was purchased from Aldrich Chemical Co. (Sure-Seal). Thin-
layer chromatography (TLC) was performed on commercially
prepared silica gel 60 F254 plates. Column chromatography
was performed using Merck silica gel 60 as the solid support
unless indicated otherwise.
F igu r e 12. Partial 1H NMR spectra obtained in the course
of carrying out a temperature-dependence study of compound
IV(4) (500 MHz, CD2Cl2).
Gen er a l P r oced u r e 1: P d -Ca ta lyzed Cr oss-Cou p lin g
R ea ct ion s bet w een
a Ter m in a l Alk yn e a n d a n Ar yl
system II, only a single broad resonance ascribable to
the imino proton of III was observed in the 1H NMR
spectrum recorded in CDCl3 at 328 K. As the tempera-
ture was lowered, increasingly complicated spectra were
observed; there are at least six resonances assignable to
the imino proton that are observed at 218 K. Such
complicated spectra are likely the result of a less ideal
preorganization than that which influences the formation
of duplexlike structure I, as discussed earlier. To the
extent this is true, it is considered likely that some other
types of complexes, possibly oligomers, are present and
that these are competing with the formation of the
duplexlike complex III. This is especially true at low
temperatures, where the exchange of the imino proton
among these different states slows down enough on the
1H NMR time scale to give rise to the observed multiple
resonances.
Ha lid e. To a solution of an appropriate alkyne and halide
(bromide or iodide) in an argon-degassed mixture of triethy-
lamine (Et3N) and acetonitrile (MeCN) or THF was added bis-
(triphenylphosphine)palladium(II) chloride (PdCl2(PPh3)2) or
tetrakis(triphenylphosphine)palladium(0) (Pd(PPh3)4) (3-6
mol % of the limiting reagent) and copper(I) iodide (6-10%
mol % equiv). The resulting mixture was then stirred at room
temperature (rt) or at an elevated temperature for 2-24 h.
After completion of the reaction (monitored by TLC), the
solvents were removed under reduced pressure using a rotary
evaporator to give a crude product which was purified chro-
matographically using silica gel as the solid support. Ap-
propriate mixtures of ethyl acetate/hexanes were used as
eluent so as to obtain the desired product.
Gen er a l P r oced u r e 2: Rem ova l of th e TMS gr ou p (s).
To a stirred solution of a trimethylsilyl (TMS)-protected alkyne
in methanol or THF was added (1) TBAF (1.2 equiv) at 0 °C
and then the solution was warmed to room temperature (2)
or KOH (1.5 equiv, 1 M aqueous) at room temperature. After
complete consumption of the starting material (determined by
TLC), the solution was concentrated by rotary evaporation to
give a crude product which was purified by column chroma-
tography on silica gel eluting with the appropriate solvent(s).
Gen er a l P r oced u r e 3: P d -Ca ta lyzed Cr oss-Cou p lin g
R ea ct ion s bet w een t h e Tr ib u t ylt in Der iva t ive of a
Ter m in a l Alk yn e a n d th e Gu a n osin e Br om id e 19. To a
round-bottomed flask containing an argon-degassed solution
of an alkyne, 8-bromoguanosine 19, and tributyltin methoxide
(excess) in toluene (50 mL) was added PdCl2(PPh3)2 or Pd-
(PPh3)4 (5-8 mol % of the limiting reagent). The resulting
mixture was heated at reflux for 10-18 h or until the reaction
was deemed complete using TLC analysis. The solvent was
then removed under reduced pressure using a rotary evapora-
tor to give a crude product which was purified by column
chromatography on silica gel, eluting with ethyl acetate/
hexanes and/or ethyl acetate/methanol, as appropriate.
4-H yd r oxy-5-[(8-e t h yn yl-1-a n t h r a ce n yl)e t h yn yl]-1-
(2′,3′,5′-tr i-O-acetyl-â-D-r ibofu r an osyl)pyr im idin -2-on e (7a)
a n d 1,8-Bis[4-h yd r oxyl-5-eth yn yl-1-(2′,3′,5′-tr i-O-a cetyl-
â-D-r ibofu r a n osyl)pyr im id in -2-on e]a n th r a cen e (8a ). This
compound was prepared by following General Procedure 1,
using 1,8-diethynyl anthracene 5 (3.2 g, 14.1 mmol), 5-iodo-
2′,3′,5′-tri-O-acetyluridine 6a (3.0 g, 6 mmol), 0.25 g (0.36
mmol) of PdCl2(PPh3)2, 0.11 g (0.6 mmol) of CuI, 50 mL of Et3N,
and 50 mL of MeCN (rt, 5 h). Silica gel chromatography first
gave (3:2 ethyl acetate/hexanes eluent) product 7a (3.2 g) in
89% yield and then 8a (0.3 g) (ethyl acetate eluent) in 10%
yield as a byproduct.
In analogy to system 2, variable temperature 1H NMR
spectroscopy of system 4 in CD2Cl2 (Figure 12) revealed
a single resonance at room temperature and multiple
resonances ascribable to the imino N-H’s at low tem-
peratures. This temperature-dependent behavior is eas-
ily rationalized in terms of the possible existence of
conformational isomers (syn and anti) as discussed above.
The syn conformation favors the formation of the du-
plexlike structure IV while the anti conformation, it is
predicted, leads to formation of linear oligomers.
Con clu sion
In summary, we have succeeded in achieving the self-
assembly of DNA-like artificial dinucleotide duplex struc-
tures by increasing the rigidity and solubility of their
components. The structural variations of preorganiza-
tion, protecting groups, and hydrogen bonds among these
systems allowed us to assess better how subtle changes
in component structures can dramatically affect the
structures and stability of the resulting complexes:
Compound 1 self-assembles to form ensemble I, a very
stable and discrete complex. On the other hand, the
complexes derived from the other three monomers (2-
4) (ensembles II-IV) are not only less stable but also
less discrete.
Despite these very concrete conclusions, several in-
triguing issues regarding ensemble I remain unanswered
at present: (1) How does I bind 2 equiv of 1,2-dichloro-
ethane and benzene? Is this via inclusion or in the form
of a clathrate type complex? (2) Why does ensemble I
selectively bind 2 equiv of benzene instead of dichlo-
For 7a . 1H NMR (CDCl3): δ 2.01 (s, 3H), 2.10 (s, 3H), 2.12
(s, 3H), 3.86 (s, 1H), 4.40 (m, 3H), 5.34 (t, 1H), 5.46 (t, 1H),
6.14 (d, 1H), 7.38 (m, 2H), 7.71 (d, 2H), 7.94 (d, 2H), 7.99 (s,
1H), 8.38 (s, 1H), 9.37 (bs, 1H), 9.40 (s, 1H) ppm. 13C NMR
(CDCl3): δ 20.4, 20.4, 20.8, 62.8, 70.0, 70.1, 80.2, 81.6, 83.7,
85.4, 87.5, 92.3, 101.8, 120.5, 123.8, 125.0, 127.5, 129.3, 129.4,

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4089
130.8, 131.2, 131.3, 131.6, 141.2, 149.4, 160.5, 169.6, 170.0
ppm. LRMSFAB: 594 (M+), HRCIMS calcd for C33H27N2O9
595.1717, found 595.1729 (M + 1)+.
(m, 2H), 8.07 (d, J ) 16.5 Hz, 1H), 8.21 (s, 1H), 8.45 (s, 1H),
9.71 (s, 1H), 13.82 (bs, 1H) ppm. 13C NMR (CDCl3): δ -5.8,
-5.5, -5.4, -5.0, -4.9, -4.8, -4.7, -4.5, -4.4, 17.5, 17.7, 17.9,
18.1, 18.3, 18.7, 25.4, 25.6, 25.73, 25.9, 26.0, 26.3, 62.9, 63.3,
72.2, 72.7, 73.1, 83.4, 86.0, 86.14, 87.3, 87.6, 89.3, 91.6, 93.9,
101.2, 119.9, 120.0, 122.0, 124.6, 125.0, 127.4, 128.7, 129.2,
131.2, 131.2, 131.3, 131.5, 132.0, 132.9, 136.2, 140.6, 149.5,
150.2, 152.1, 154.8, 164.1 ppm. LRMSFAB: 1418 (M + 1)+,
HRMSFAB calcd for C73H116N7O10Si6 1418.7399 [M + 1]+ of
2, found 1418.7449. LRMSFAB: 2836 ((M + 1)+ of II),
HRMSFAB calcd for C146H231N14O20Si12 2836.4721 ((M + 1)+
of II), found 2836.4712.
For 8a . 1H NMR (CDCl3): δ 2.07 (s, 6H), 2.10 (s, 6H), 2.12
(s, 6H), 4.39 (m, 6H), 5.39 (t, 2H), 5.49 (t, 2H), 6.06 (d, 2H),
7.39 (dd, 2H), 7.68 (d, 2H), 7.96 (d, 2H), 7.99 (s, 2H), 8.38 (s,
1H), 9.45 (s, 1H), 9.76 (bs, 2H) ppm. 13C NMR (CDCl3): δ 20.4,
20.7, 62.7, 70.1, 73.5, 80.3, 85.7, 88.6, 91.9, 101.3, 120.8, 124.0,
124.9, 127.5, 129.3, 130.8, 131.2, 142.0, 149.5, 160.9, 169.1,
170.1, 170.3 ppm. LRMSFAB: 963 (M + 1)+, HRMSFAB calcd
for C48H42N4O18 962.2489, found 962.2494. Anal. Calcd for
C48H42N4O18: C, 59.88; H, 4.4; N, 5.82. Found: C, 59.73; H,
4.48; N, 5.64.
4,6-Dieth yn yld iben zofu r a n (11). This compound was
prepared by following General Procedure 1, using 10 (10.7 g,
25.5 mmol), TMS-acetylene (5.0 g, 51 mmol), 1.5 g (1.3 mmol)
of Pd(PPh3)4, 0.5 g (2.6 mmol) of CuI, and 100 mL of Et3N (50
°C, 15 h). Silica gel chromatography (hexanes eluent) gave
product (4,6-dibenzofuranyldiethynylene)bis(trimethylsilane)
as an intermediate in quantitative yield. LRMSCI(+): 361,
HRMSCI(+) calcd for C22H24OSi2 360.1366, found 360.1365.
The TMS group of this intermediate was removed by
reaction with KF (5.35 g, 92 mmol) while the solution was
heated in ethanol (300 mL) at reflux for 3 h. After the reaction
was allowed to cool to room temperature, the solvent was then
removed and the residue taken up in H2O (200 mL) and
extracted with CDCl3 (3 × 100 mL). The organic layer was
dried over Na2SO4 and evaporated to dryness. The residue
was purified by silica gel column chromatography using CHCl3
as the eluent. This gave product 11 (3.8 g) in 80% yield. 1H
NMR (CDCl3): δ 3.47 (s, 2H), 7.30 (t, J ) 7.8 Hz, 2H), 7.61 (d,
J ) 7.8 Hz, 2H), 7.92 (d, J ) 8.1 Hz, 2H) ppm. 13C NMR
(CDCl3): δ 82.9, 83.1, 107.5, 121.7, 123.2, 124.5, 131.7, 131.9
ppm.
5-Tr im e t h ylsilyle t h yn yl-4-h yd r oxyl-1-(2′,3′,5′-t r i-O-
a cetyl-â-D-r ibofu r a n osyl)p yr im id in -2-on e (14). This com-
pound was prepared by following General Procedure 1, using
6a (10 g, 20 mmol), TMS-acetylene (3.64 g, 37.1 mmol), 0.84 g
(1.2 mmol) of PdCl2(PPh3)2, 0.38 g (2 mmol) of CuI, 50 mL of
Et3N, and 50 mL of MeCN (50 °C, 15 h). Silica gel chroma-
tography using 4:3 ethyl acetate/hexanes as the eluent gave
the TMS-protected product 14 (0.91 g) in 91% yield. 1H NMR
(CDCl3): δ 0.11 (s, 9H), 2.00 (s, 3H), 2.02 (s, 3H), 2.11 (s, 3H),
4.28 (m, 3H), 5.24 (m, 2H), 6.01 (d, 1H), 7.70 (s, 1H), 9.83 (bs,
1H) ppm. 13C NMR (CDCl3): δ -0.5, 20.1, 20.2, 20.6, 62.7,
69.9, 73.0, 80.0, 87.1, 95.0, 99.6, 101.1, 141.9, 149.2, 160.7,
169.3, 169.4, 169.8 ppm. LRCI(+)MS: 467 ([M + 1]+), HRCI-
(+)MS calcd for C20H27N2O9Si 467.1486, found 467.1484.
5-Eth yn yl-4-h yd r oxy-1-(2′,3′,5′-tr i-O-acetyl-â-D-ribofura-
nosyl)pyrimidin-2-one (15). This compound was prepared by
following General Procedure 2, using 14 (0.91 g, mmol), TBAF
(20 mL, 1 M in THF, 20 mmol), and THF (200 mL) (rt, 4 h).
Silica gel chromatography (ethyl acetate eluent) gave product
15 (6.4 g) in 88% yield. 1H NMR (CDCl3): δ 2.05 (s, 3H), 2.12
(s, 3H), 2.20 (s, 3H), 3.23 (s, 1H), 4.40 (m, 3H), 5.36 (m, 2H),
6.09 (d, 1H), 7.31 (s, 1H), 7.87 (s, 1H) ppm. 13C NMR
(CDCl3): δ 20.2, 20.3, 20.7, 62.8, 69.9, 73.1, 74.5, 80.1, 82.3,
87.5, 100.0, 142.7, 149.2, 161.0, 169.4, 169.5, 170.0 ppm.
LRMSCI: 395 (M+), HRCIMS(+) calcd for C17H19N2O9 395.1091,
found 395.1085.
1-[4-Hyd r oxyl-5-eth yn yl-1-(2′,3′,5′-tr i-O-a cetyl-â-D-r ibo-
fu r an osyl)pyr im idin -2-on e]-8-[6-am in o-8-eth yn yl-9-(2′,3′,5′-
tr i-O-acetyl-â-D-r ibofu r an osyl)pu r in ]an th r acen e (1). This
compound was prepared by following General Procedure 1,
using 7a (3.2 g, 5.39 mmol), 8-bromoadenosine 9a (3.8 g, 8.05
mmol), 0.23 g (0.32 mmol) of PdCl2(PPh3)2, 0.10 g (0.54 mmol)
of CuI, 50 mL of Et3N, and 50 mL of MeCN (60 °C, 20 h). Silica
gel chromatography using first 3:1 ethyl acetate/hexane (to
remove unreacted 9a ) and then ethyl acetate as the eluent
gave product 1 (3.76 g) in 71% yield. 1H NMR (CDCl3): δ 1.74
(s, 3H), 1.87 (s, 3H), 1.94 (s, 3H), 2.16 (s, 3H), 2.20 (s, 3H),
2.40 (s, 3H), 3.96 (m, 2H), 4.11 (m, 2H), 4.58 (m, 2H), 4.87 (m,
1H), 5.26 (m, 1H), 5.47 (m, 1H), 6.02 (d, J ) 6.8 Hz, 1H), 6.50
(d, J ) 5.3 Hz, 1H), 6.60 (d, J ) 5.2 Hz, 1H), 6.70 (m, 1H),
7.10 (m, 2H), 7.20 (t, J ) 7.7 Hz, 1H), 7.52 (d, J ) 6.8 Hz,
1H), 7.85 (t, J ) 8.5 Hz, 2H), 8.30 (s, 1H), 8.35(s, 1H), 9.22 (s,
1H), 9.49 (bs, 1H), 14.68 (s, 1H) ppm. 13C NMR (CDCl3): δ
20.1, 20.2, 20.6, 21.2, 63.5, 63.7, 70.3, 70.3, 73.0, 74.6, 79.0,
80.2, 83.0, 84.5, 84.7, 88.4, 90.5, 94.3, 99.9, 119.2, 120.0, 120.3,
124.3, 124.9, 126.9, 128.3, 128.4, 130.6, 130.8, 131.1, 131.3,
131.5, 132.2, 136.1, 139.8, 139.8, 148.1, 149.4, 153.2, 155.3,
162.4, 168.4, 169.0, 169.8, 169.9, 170.5 ppm. LRMSFAB: 985
(M+), HRMSFAB calcd for C49H43N7O16 985.2766 (M+ of 1),
found 985.2757. LRMSFAB: 1970 (M+ of I), HRMSFAB calcd
for C98H86N14O32 1970.5533 (M+ of I), found 1970.5520. Anal.
Calcd for (C49H43N7O16)2‚(C6H6)2: C, 62.09; H, 4.64; N, 9.21.
Found: C, 62.07; H, 4.65; N, 9.17. Anal. Calcd for
(C49H43N7O16)2‚(C2H4Cl2)2: C, 56.5; H, 4.37; N, 9.05; Cl, 6.46.
Found: C, 56.51; H, 4.43; N, 8.94; Cl, 6.59.
4-H yd r oxy-5-[(8-e t h yn yl-1-a n t h r a ce n yl)e t h yn yl]-1-
(2′,3′,5′-tr i-O-ter t-bu tyld im eth ylsilyl-â-D-r ibofu r a n osyl)-
p yr im id in -2-on e (7b). This compound was prepared by
following General Procedure 1, using 1,8-diethynylanthracene
5 (1.0 g, 4.4 mmol), 5-iodo-2′,3′,5′-tri-O-tert-butyldimethylsi-
lyluridine 7b (1.5 g, 2.1 mmol), 0.090 g (0.12 mmol) of PdCl2-
(PPh3)2, 0.04 g (0.21 mmol) of CuI, and 50 mL of Et3N (rt, 5
h). Silica gel chromatography (1:4 ethyl acetate/hexanes
eluent) gave product 7b (1.2 g) in 70% yield. 1H NMR
(CDCl3): δ 0.04-0.12 (multiple-singlet (ms), 18H), 0.87-0.94
(ms, 27H), 3.72 (s, 1H), 3.74 (m, 1H), 3.97 (d, J ) 11.1 Hz,
1H), 4.11 (s, 2H), 4.25 (m, 1H), 6.09 (d, J ) 6.3 Hz, 1H), 7.41
(m, 2H), 7.73 (t, J ) 7.8 Hz, J ) 7.2 Hz, 2H), 7.99 (d, J ) 8.4
Hz, 2H), 8.19 (s, 1H), 8.40 (s, 1H), 8.43 (s, 1H), 9.55 (s, 1H)
ppm. 13C NMR (CDCl3): δ -5.4, -5.2, -4.7, -4.5, -4.4, 18.0,
18.1, 18.4, 25.8, 25.8, 26.1, 63.0, 72.7, 76.0, 81.5, 83.6, 85.9,
86.5, 88.4, 92.2, 101.0, 120.8, 121.1, 124.4, 124.9, 125.0, 127.4,
128.0, 128.2, 129.3, 130.4, 131.1, 131.4, 131.5, 131.8, 142.1,
149.3, 160.6 ppm. LRMSFAB: 810 (M+), HRMSFAB calcd for
5-[(6-Iod o-4-d ib e n zofu r a n yl)e t h yn yl-4-h yd r oxyl-1-
(2′,3′,5′-tr i-O-acetyl-â-D-r ibofu r an osyl)pyr im idin -2-on e (16)
a n d 4,6-Bis[4-h yd r oxyl-5-eth yn yl-1-(2′,3′,5′-tr i-O-a cetyl-
â-D-r ibofu r a n osyl)p yr im id in -2-on e]a n th r a cen e (13). This
compound was prepared by following General Procedure 1,
using 10 (5 g, 12 mmol), 15 (2.3 g, 6 mmol), 0.1 g (0.36 mmol)
of PdCl2(PPh3)2, 0.12 g (0.6 mmol) of CuI, 80 mL of Et3N, and
80 mL of MeCN (rt, 24 h). Silica gel chromatography using
first 2:1 ethyl acetate/hexane and then ethyl acetate as the
eluent gave 3.5 g of product 16 (87% yield) and 0.5 g of
compound 13 (9% yield) as a byproduct.
C
45H62N2O6Si3 810.3915, found 810.3906.
1-[4-H yd r oxy-5-et h yn yl-1-(2′,3′,5′-t r i-O-ter t-b u t yld im -
eth ylsilyl-â-D-r ibofu r an osyl)pyr im idin -2-on e]-8-[6-am in o-
8-et h yn yl-9-(2′,3′,5′-t r i-O-ter t-b u t yld im et h ylsilyl-â-D-r i-
bofu r a n osyl)p u r in e]a n th r a cen e (2). This compound was
prepared by following General Procedure 1, using 7b (0.8 g,
1.0 mmol), 9b (0.8 g, 1.1 mmol), 0.04 g (0.06 mmol) of PdCl2-
(PPh3)2, 0.02 g (0.1 mmol) of CuI, and 30 mL Et3N (60 °C, 4
h). Silica gel chromatography (3:1 ethyl acetate/hexanes
eluent) gave product 2 (0.91 g) in 65% yield. 1H NMR
(CDCl3): δ -0.65 to 1.05 (ms, 90 H), 3.74-4.22 (m, 9H), 5.10
(m, 1H), 6.17 (d, J ) 6.0 Hz, 1H), 6.40 (d, J ) 7.5 Hz, 1H),
6.70 (bs, 2H), 7.48 (q, J ) 8.1 Hz, 2H), 7.67 (d, J ) 6.9 Hz,
1H), 7.89 (d, J ) 6.9 Hz), 1H), 7.99 (d, J ) 8.4 Hz, 1H), 8.05
For 16. LRMSFAB: 687 (M + 1)+, HRCIMS(+) calcd for
29H23N2O10I 686.0397, found 686.0391. This compound was
C
used without further purification in the ensuing step.
For 13. 1H NMR (CDCl3): δ 2.11 (s, 6H), 2.13 (s, 6H), 2.25
(s, 6H), 4.43 (m, 6H), 5.52 (m, 4H), 6.12 (d, 2H), 7.34 (t, 2H),

4090 J . Org. Chem., Vol. 63, No. 12, 1998
Sessler and Wang
7.62 (d, 2H), 7.94 (d, 2H), 8.10 (s, 2H), 9.25 (bs, 2H) ppm. 13C
NMR (CDCl3): δ 20.3, 20.39, 20.9, 62.7, 70.2, 73.2, 80.4, 85.4,
87.7, 88.3, 100.9, 107.3, 121.3, 123.1, 124.0, 130.7, 149.2, 155.7,
160.9, 169.6, 169.7, 170.5 ppm. LRMSFAB: 953 (M+), HRMS-
FAB calcd for C46H40N4O19 953.2365, found 953.2348. Anal.
Calcd for C46H40N4O19: C, 57.99; H, 4.23; N, 5.88. Found: C,
57.61; H, 4.28; N, 5.82.
Silica gel chromatography using first 3:1 hexanes/ethyl acetate
and then ethyl acetate as the eluent gave product 21 (3.5 g)
in 68% yield. LRMSFAB(+): 810, HRMSFAB(+) calcd for
C45H64N3O5Si3 810.4150, found 810.4124. This compound was
used directely in the ensuing step without further purification.
1-[4-Am in o-5-eth yn yl-1-(2′,3′,5′-tr i-O-ter t-bu tyld im eth -
ylsilyl-â-D-r ib ofu r a n osyl)p yr im id in -4-on e]-8-[N2-(N′,N′-
d im et h ylfor m a m id e)-8-et h yn yl-9-(2′,3′,5′-t r i-O-ter t-b u -
t y ld im e t h y ls ily l-â-D -r ib o fu r a n o s y l)p u r in -6-o n e ]a n -
th r a cen e (23). This compound was prepared by following
General Procedure 3, using 8-bromoguanosine 19 (3.6 g, 4.7
mmol), mono-cytidine-substituted derivative 21 (3.5 g, 4.3
mmol), tributyltin methoxide (3.2 mL, 10 mmol), 0.0.25 g (0.22
mmol) of Pd(PPh3)4, and 80 mL of toluene (110 °C, 10 h). Silica
gel chromatography using first 1:1 and 2:1 ethyl acetate/
hexanes (to remove the unreacted 19 and 21) and then 5:1
ethyl acetate/hexanes as the eluent gave product 23 (2.3 g) in
36% yield. MSFAB(+): 1488, HRFAB(+) calcd for C76H122N9O10-
Si6 1488.7930, found 1488.7957. The product 23 so obtained
was used directly without further purification in the following
step.
5-[(6-TMS-eth yn yl-4-diben zofu r an yl)eth yn yl]-1-(2′,3′,5′-
tr i-O-a cetyl-â-D-r ibofu r a n osyl]p yr im id in -2-on e (17). This
compound was prepared by following General Procedure 1,
using 16 (3.5 g, 5.1 mmol), TMS-acetylene (1.0 g, 10.6 mmol),
0.22 g (0.31 mmol) of PdCl2(PPh3)2, 0.9 g (0.5 mmol) of CuI,
20 mL of Et3N, and 20 mL of MeCN (rt, 12 h). Silica gel
chromatography (3:2 ethyl acetate/hexanes eluent) gave prod-
uct 17 (3.2 g) in 96% yield. LRMSFAB: 657 (M + 1)+,
HRMSFAB calcd for C34H32N2O10Si 656.1826, found 656.1823.
This compound was used without further purification.
5-[(6-Eth yn yl-4-d iben zofu r a n yl)eth yn yl]-1-(2′,3′,5′-tr i-
O-a cetyl-â-DD-r ibofu r a n osyl]p yr im id in -2-on e (12). This
compound was prepared by following General Procedure 2,
using 17 (3.2 g, mmol), TBAF (7.5 mL, 1 M in THF, 7.5 mmol),
and THF (50 mL) (rt, 5 h). Silica gel chromatography (3:2
ethyl acetate/hexane eluent) gave product 12 (2.4 g) in 84%
yield. 1H NMR (CDCl3): δ 2.05 (s, 3H), 2.14 (s, 3H), 2.28 (s,
3H), 3.62 (s, 1H), 4.39 (m, 3H), 5.44 (m, 2H), 6.20 (d, 1H), 7.33
(m, 2H), 7.58 (d, 1H), 7.64 (d, 1H), 7.92 (dd, 2H), 8.05 (s, 1H),
9.50 (bs, 1H) ppm. 13C NMR (CDCl3): δ 20.3, 20.4, 21.0, 60.3,
62.6, 69.9, 73.2, 77.8, 88.1, 101.3, 107.2, 107.0, 115.1, 121.4,
123.0, 123.1, 124.0, 131.5, 141.5, 141.5, 149.3, 155.5, 156.2,
160.8, 169.5, 169.6, 170.3 ppm. LRMSFAB: 585 (M+), HR-
CIMS(+) calcd for C31H25N2O10 585.1509, found 585.1500.
4-[(5-E t h yn yl-4-h yd r oxyl-1-(2′,3′,5′-t r i-O-a cet yl-â-D-r i-
b ofu r a n osyl)p yr im id in -2-on e]-6-[6-a m in o-8-et h yn yl-9-
(2′,3′,5′-tr i-O-a cetyl-â-D-r ibofu r a n osyl)p u r in e]d iben zofu -
r a n (3). This compound was prepared by following General
Procedure 1, using 12 (2.0 g, 3.4 mmol), 9a (2.5 g, 5.3 mmol),
0.15 g (0.21 mmol) of PdCl2(PPh3)2, 0.7 g (0.37 mmol) of CuI,
20 mL of Et3N, and 20 mL MeCN (60 °C, 18 h). Silica gel
chromatography using first 4:1 ethyl acetate/hexanes (to
remove the unreacted 9a ) and then ethyl acetate as the eluent
gave product 3 (2.24 g) in 67% yield. 1H NMR (CDCl3): δ
1.87-2.35 (ms, 18H), 4.07-4.46 (m, 6H), 5.21-5.28 (m, 2H),
5.80 (m, 1H), 6.25 (m, 1H), 6.41 (m, 1H), 6.47 (m, 1H), 7.06
(bs, 2H), 7.33 (t, 1H), 7.39 (t, 1H,), 7.45 (d, 1H, J ) 7.69 Hz),
7.74 (d, 1H), 7.81 (s, 1H) 7.91 (dd, 1H), 8.00 (d, 1H, J ) 7.69
Hz), 8.28 (s, 1H,), 13.02 (bs, 1H) ppm. 13C NMR (CDCl3): δ
20.3, 20.4, 20.4, 20.6, 20.7, 21.0, 62.7, 63.1, 69.6, 70.4, 73.4,
74.1, 79.6, 79.7, 82.3, 86.4, 86.6, 88.1, 88.6, 91.0, 100.5, 106.2,
108.2, 119.0, 121.1, 122.7, 123.1, 123.8, 124.4, 129.5, 131.2,
134.9, 141.0, 148.6, 149.0, 152.6, 154.5, 156.36, 156.7, 162.8,
169.3, 169.3, 169.5, 169.8, 170.1, 170.6 ppm. LRMSFAB: 976
(M+), HRMSFAB calcd for C47H42N7O17 976.2637, found
976.2634. Anal. Calcd for C47H42N7O17: C, 57.85; H, 4.23; N,
10.05. Found: C, 57.48; H, 4.27; N, 9.92.
1-[4-Am in o-5-eth yn yl-1-(2′,3′,5′-tr i-O-ter t-bu tyld im eth -
ylsilyl-â-D-r ibofu r a n osyl)p yr im id in -4-on e]-8-[2-a m in o-8-
et h yn yl-9-(2′,3′,5′-t r i-O-ter t-b u t yld im et h ylsilyl-â-D-r ib o-
fu r a n osyl)p u r in -6-on e]a n th r a cen e (4). Compound 23 (2.3
g) was treated with ammonia-saturated methanol (50 mL) at
room temperature overnight. The product was purified via
silica gel column chromatography using 2:1 hexanes/ethyl
acetate as the eluent. This gave 4 in the form of a yellow,
fluorescent solid (1.8 g; 81% yield). 1H NMR (CDCl3): δ 0.07-
0.2 (ms, 36H), 0.78-0.97 (ms, 54H), 3.77 (bs, 1H), 3.82 (d, J )
9.37 Hz, 2H), 3.99-4.11 (m, 6H), 4.17 (t, J ) 4.45 Hz, J )
4.57 Hz, 1H), 4.33 (bs, 1H), 5.37 (bs, 1H), 6.10 (d, J ) 4.45
Hz, 1H), 6.24 (d, J ) 10.00 Hz, 1H), 6.46 (bs, 1H), 7.46 (m,
2H), 7.81 (dd, J ) 6.34 Hz, J ) 16.46 Hz, 2H), 8.09 (m, 2H),
8.47 (s, 1H), 9.14 (bs, 1H), 9.27 (s, 1H), 13.37 (s, 1H) ppm. 13
C
NMR (CDCl3): δ -5.3, -5.7, -5.2, -4.9, -4.8, -4.6, -4.6,
-4.4, -4.3, -4.1, 17.9, 17.9, 18.1, 18.2, 18.4, 18.6, 25.8, 25.9,
26.0, 26.0, 26.3, 62.7, 63.3, 71.5, 71.8, 73.3, 84.4, 85.0, 86.2,
86.9, 89.0, 89.1, 92.2, 92.5, 93.4, 118.4, 119.6, 121.2, 123.0,
124.9, 125.0, 128.2, 129.2, 129.9, 130.6, 130.9, 131.3, 131.7,
131.7, 132.1, 132.8, 144.1, 152.4, 153.6, 155.0, 160.1, 163.8
ppm. MSFAB(+): 1434, HRFAB(+) calcd for C73H117N8O10
-
Si6 1433.7508, found 1433.7527. For the dimer IV: MSFAB-
(+) 2870, HRFAB(+) calcd for C146H233N16O20Si12 2866.4939,
found 2866.4900. Anal. Calcd for C73H117N8O10Si6: C, 61.13;
H, 8.15; N, 7.81. Found: C, 61.24; H, 8.16; N, 7.87.
N2-(N′,N′-Dim eth ylfor m a m id e)-8-[(8-eth yn yl-1-a n th r a -
cen yl)eth yn yl]-9-(2′,3′,5′-tr i-O-ter t-bu tyld im eth ylsilyl-â-
D-r ibofu r a n osyl)p u r in -6-on e (25). This compound was
prepared by following General Procedure 3, using 8-bromo-
guanosine 19 (5 g, 6.6 mmol), compound 5 (1.5 g, 6.6 mmol),
tributyltin methoxide (3.2 g, 10 mmol), 0.0.38 g (0.33 mmol)
of Pd(PPh3)4, and 80 mL of toluene (110 °C, 18 h). Silica gel
chramatography (1:1 ethyl acetate/hexanes eluent) gave prod-
uct 25 (3.1 g) in 52% yield. 1H NMR (CDCl3): δ -0.35, -0.32,
-0.28, -0.19, -0.08, -0.04 (ms, 18 H), 0.62, 0.64, 0.73 (ms,
27 H), 3.00 (s, 6H), 3.81 (s, 1H), 3.92 (m, 3H), 4.20 (d, 1H),
5.08 (m, 1H), 6.25 (d, 1H), 7.47 (m, 2H), 7.68 (dd, 2H), 7.86
(dd, 2H), 8.25 (s, 1H), 8.36 (s, 1H), 9.39 (s, 1H), 10.75 (bs, 1H)
ppm. 13C NMR (CDCl3): δ -5.9, -5.7, -5.2, -4.8, -4.6, -4.6,
17.7, 17.7, 17.9, 25.5, 25.6, 35.1, 41.1, 63.1, 72.2, 72.6, 80.8,
84.5, 84.7, 85.3, 87.7, 92.2, 119.5, 120.7, 121.5, 123.8, 124.6,
124.9, 127.4, 128.1, 128.3, 128.9, 131.3, 131.8, 150.6, 156.9,
157.3, 157.9 ppm. MSFAB(+): 905, HRMSCI(+) calcd for
C49H69N6O5Si3 905.4637, found 905.4651.
N2-(N′,N′-Dim eth ylfor m a m id e)-8-br om o-9-(2′,3′,5′-tr i-O-
ter t-bu tyldim eth ylsilyl-â-D-r ibofu r an osyl)pu r in -6-on e (19).
To a 250 mL round-bottomed flask containing a solution of
8-bromoguanosine 18 (7.1 g, 20 mmol) in THF (100 mL) was
added N,N-dimethylformamide dimethyl acetal (4 equiv). The
resulting mixture was stirred at room temperature overnight.
The residue was purified by chromatographicy (silica gel; 4:1
ethyl acetate/hexane eluent) to give product 19 (3.76 g) in 84%
yield. 1H NMR: δ -0.35 to -0.01 (18H), 0.71-0.90 (27H), 3.10
(s, 6H), 3.74 (2H, m), 3.94 (1H, m), 4.26 (1H, m), 5.17 (1H, m),
5.96 (1H, d), 8.38 (1H, s), 10.86 (1H, s). 13C NMR: δ -5.61,
-5.45, -5.29, -4.71, -4.58, 17.68, 17.84, 18.10, 25.51, 25.69,
35.20, 41.23, 62.90, 71.25, 72.25, 85.45, 88.55, 121.00, 122.64,
151.66, 156.48, 157.23, 157.43. MSCI(+): 759, HRCI(+) found
759.3160, calcd 759.3116 for C31H60N6O5Si3Br.
Alter n a tive Syn th esis of 23. Following General Proce-
dure 1, compounds 25 (1.8 g, 2 mmol) and 20 (1.8 g, 2.5 mmol)
were treated with 70 mg of PdCl2(PPh3)2, 40 mg of CuI, 30
mL of Et3N, and 20 mL of MeCN (rt, overnight). Silica gel
chromatography (2/1 hexanes/ethyl acetate eluent) then gave
23 (2.1 g) in 71% yield.
4-Am in o-5-[(8-eth yln yl-1-an th r acen yl)eth yn yl)-1-(2′,3′,5′-
tr i-O-ter t-bu tyldim eth ylsilyl-â-D-r ibofu r an osyl)pyr im idin -
4-on e (21). This compound was prepared by following General
Procedure 1, using 5 (2.3 g, 10 mmol), 5-iodocytidine 20 (4.5
g, 6.3 mmol), 0.26 g (0.37 mmol) of PdCl2(PPh3)2, 0.12 g (0.6
mmol) of CuI, 50 mL of Et3N, and 50 mL of MeCN (rt, 2 h).

Synthesis and Self-Assembly of “Artificial Dinucleotide Duplexes”
J . Org. Chem., Vol. 63, No. 12, 1998 4091
tions; quantification of DMSO solvation of the monomeric
species 1; binding constant determination of 3/III by 1H NMR
spectroscopic means (12 pages). This material is contained
in libraries on microfiche, immediately follows this article in
the microfilm version of the journal, and can be ordered from
the ACS; see any current masthead page for ordering
information.
Ack n ow led gm en t. We thank the Robert A Welch
Foundation (Grant F-1018 to J .L.S.) for financial sup-
port.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra for 2; MSFAB spectrum of 1/1 mixtures of 1 and 3; 1H
spectrum of 1 and its 1,2-dichloroethane adduct; binding
constant data for a mixture of 1/I in 45% DMSO-d6/CDCl3 and
a summary of resultant thermodynamic parameter determina-
J O980194P