Synthesis and properties of oligonucleotides having a phosphorus chiral center by incorporation of conformationally rigid 5'-cyclouridylic acid derivatives

Article abstract of DOI:10.1021/jo0002588

This paper describes the design and synthesis of a conformationally rigid dimer building block Umpc3Um as a chiral center at the phosphate group with the S/N junction where c3 refers to a propylene bridge linked between the uracil 5-position and 5'-phosphate group of pUre. The extensive H¹ NMR analysis of Umpc3Um suggests that the 5'-upstream Um has predominantly a C2'-endo conformation and the pc3Um moiety exists almost exclusively in a C3'-endo conformation. The absolute configuration of the diastereomers Umpc3Um(fast) (8a) and Umpc3Um(slow) (8b) was determined by CD spectroscopy as well as computer simulations. The oligonucleotides U₄[Umpc3Um-(fast)]U₄ (13a) and U₄[Umpc3Um(slow)]U₄ (13b) incorporating 8a and 8b were synthesized by use of the phosphoramidite building blocks 11a and 11b, respectively. The T(m) experiments of the duplexes formed between these modified oligomers and the complementary oligomers imply that the modified oligomer 13a having Umpc3Um(fast) has the Sp configuration at the chiral phosphoryl group.

Full text of DOI:10.1021/jo0002588

J . Org. Chem. 2000, 65, 6515-6524
6515
Syn th esis a n d P r op er ties of Oligon u cleotid es Ha vin g a
P h osp h or u s Ch ir a l Cen ter by In cor p or a tion of Con for m a tion a lly
Rigid 5′-Cyclou r id ylic Acid Der iva tives
Mitsuo Sekine,* Osamu Kurasawa, Koh-ichiroh Shohda, Kohji Seio, and Takeshi Wada
Department of Life Science, Tokyo Institute of Technology, Nagatsuta, Midoriku,
Yokohama 226-8501 J apan
msekine@bio.titech.ac.jp
Received February 24, 2000
This paper describes the design and synthesis of a conformationally rigid dimer building block
Umpc3Um as a chiral center at the phosphate group with the S/N junction where c3 refers to a
propylene bridge linked between the uracil 5-position and 5′-phosphate group of pUm. The extensive
H1 NMR analysis of Umpc3Um suggests that the 5′-upstream Um has predominantly a C2′-endo
conformation and the pc3Um moiety exists almost exclusively in a C3′-endo conformation. The
absolute configuration of the diastereomers Umpc3Um(fast) (8a ) and Umpc3Um(slow) (8b) was
determined by CD spectroscopy as well as computer simulations. The oligonucleotides U4[Umpc3Um-
(fast)]U4 (13a ) and U4[Umpc3Um(slow)]U4 (13b) incorporating 8a and 8b were synthesized by use
of the phosphoramidite building blocks 11a and 11b, respectively. The Tm experiments of the
duplexes formed between these modified oligomers and the complementary oligomers imply that
the modified oligomer 13a having Umpc3Um(fast) has the Sp configuration at the chiral phosphoryl
group.
In tr od u ction
We previously reported the synthesis of confomation-
ally rigid 5′-cyclouridylic acid derivatives 1 and 2 as
shown in Figure 1.1-3 These studies revealed that the
cyclouridylic acids 1 and 2 have the fixed torsion angles
of R, â, and γ, defined in the backbone structure in RNA
duplex, in a quite natural manner without strain.
Therefore, one can expect that, if compound 2 can be
F igu r e 1. Intramolecular hydrogen bonds found in tRNAs
and chemically synthesized covalent-bonding mimics 1 and 2.
incorporated into RNA, two stereoisomers with the R and
S configurations due to the phosphorus center will be
generated. One of them has a typical R torsion angle,
which corresponds to that of the A type-helical RNA
strand, while the other has a different torsion angle with
+120°. It is also expected that the ribose conformation
of a nucleoside 5′-upstream from the cyclouridylic acid
will be influenced in terms of the neutral phosphate
group that serves as an electron-withdrawing effect. Now,
we can expect that totally RNA oligonucleotides incor-
porating these R and S phosphorus centers would exhibit
unique structures conformationally fixed. From the crys-
tal structure of Escherichia coli tRNAPhe,4 we have also
noticed many characteristic bent motifs with disorder
structures of the base-orientation between proximal
positions 7-8, 9-10, and 19-20 or in the regions of
consecutive positions 16-18, 46-49, and 53-61. These
unique structures are apparently essential for compact
folding of the tRNA molecule. In other RNA loop regions,
a variety of backbone structures have been found by
several research groups.5 In general, RNA holding occurs
via change of torsion angles of the RNA backbone
structure and/or the ribose puckering, especially via
dramatic change of the torsion angle of the internucleo-
tidic phosphate diester linkage.4
Therefore, it is interesting if various RNA structural
motifs, formed between neighboring two ribonucleosides,
can be artificially fixed by chemical methods. Such
studies would provide a new insight in man-made func-
tionalization of RNA as well as understanding of the
structure-function relationship of RNA. To our best
knowledge, no comprehensive studies have appeared
(5) The U-turn structure was found in hammerhead ribozymes and
U2 snRNA: (a) Pley, H. W.; Flaherty, K. M.; McKay, D. B. Nature 1994,
372, 68-74. (b) Scott, W. G.; Finch, J . T.; Klug, A. Nucl. Acids Symp.
Ser. 1995, 34, 214-216. (c) Scott, W. G.; Finch, J . T.; Klug, A. Cell
1995, 81, 991-1002. (d) Scott, W. G.; Murray, J . B.; Arnold, J . R. P.;
Stoddard, B. L.; Klug, A. Science 1996, 274, 2065-2069.; (e) Stallings,
S. C.; Moore, P. B. Structure 1997, 5, 1173-1185. In the loop of
hepatitis delta virus antigenomic RNA a sharp reversed U-turn
structure was found: (f) Kolk, M. H.; Heus, H. A.; Hilbers, C. W. EMBO
J . 1997, 16, 3685-3692. The following papers reported the presence
of the U-turn structure in a RNA hairpin loop: (g) Huang, S.; Wang,
Y. X.; Draper, D. E. J . Mol. Biol. 1996, 258, 308-321. (h) J ucker, F.
M.; Pardi, A. RNA 1995, 1, 219-222.
(1) Seio, K.; Wada, T.; Sakamoto, K.; Yokoyama, S.; Sekine M. J .
Org. Chem. 1996, 61, 1500-1504.
(2) (a) Seio, K.; Wada, T.; Sakamoto, K.; Yokoyama, S.; Sekine M.
Tetrahedron Lett. 1995, 36, 9515-9518. (b) Seio, K.; Kurasawa, O.;
Wada, T.; Sakamoto, K.; Yokoyama, S.; Sekine M. Nucleosides Nucle-
otides 1997, 16, 1023-1032.
(3) Sekine, M.; Kurasawa, O.; Seio, K.; Wada, T. J . Org. Chem. 2000,
65, 3571-3578.
(4) (a) Quigley, G. J .; Rich, A. Science 1976, 194, 796-806. (b) Von
Ahsen, U.; Green, R.; Schroeder, R.; Noller, H. F. RNA 1997, 3, 49-
56.
10.1021/jo0002588 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/09/2000

6516 J . Org. Chem., Vol. 65, No. 20, 2000
Sekine et al.
F igu r e 2. Schematic images of oligonucleotides incorporating a conformationally rigid cyclouridylic acid derivative pc3Um.
about the synthesis of oligonucleotides stereo-controlled
simultaneously at the R, â, and γ positions to date.
In this paper, we report the synthesis of oligonucle-
otides incorporating a phosphorus chiral center by use
of a conformationally fixed dimer building unit that has
rigid sugar conformations, anti base-orientation, g+ or t
torsion angle around the P-O5′ bond, and the typical
A-type helical â and γ torsion angles.
which was estimated by the well-known equation of
P(C3′-endo) ) J 3′,4′/(J 1′,2′ + J 3′,4′).9 Generally, the average
value of the sum of J 1′,2′ and J 3′,4′ in many uridine
derivatives is 9.4 Hz ( 0.2.10 Deviation from this average
value is a good indicator for estimation of the structural
disorder. Particularly, the cyclouridylic acid 1 has proven
to have a natural, favorable structure without strain,
since this compound has actually a normal value (9.4 Hz)
of the sum of J 1′,2′ and J 3′,4′
.
On the basis of these promising conformational proper-
ties of 1, we have recently reported that a decathymid-
ylate pc3Um(pT)9 incorporating the 2′-O-methylcyclou-
ridylic acid derivative 2 (pc3Um) at the 5′-terminal end
can form a more stable duplex with A(pA)9 than the
unmodified decathymidylate T(pT)9.3 This study also
revealed that introduction of a 2′-O-methyl group into 1
can enhance the fractional population of the C3′-endo
conformation up to more than 89%.
Syn th esis of Um p c3Um (8) In volvin g th e 5′-Cyc-
lou r id ylic Acid Der iva tive p c3Um (2). To incorporate
a dimer motif having a phosphorus chiral center into
RNA oligomers, we synthesized Umpc3Um (8), which
contains a cyclic structure at the 3′-downstream site. The
2′-hydroxy group of the 3′-downstream uridine was in
advance methylated to circumvent the difficult selective
2′-O-protection. It is known that if the 2′-hydroxy group
is free and a dialkoxyphosphoryl [-P(O)(OR)2] group is
attached to the neighboring 3′-oxygen, 3′ f 2′ phosphoryl
migration inevitably occurs with cyclization of the phos-
phoryl group and elimination of an alcohol.11 Accordingly,
the 2′-hydroxy group of the 5′-upstream uridine was also
masked with a methyl group to avoid such side reactions
associated with the pc3Um residue.
Resu lts a n d Discu ssion
RNA Ben d in g Str u ctu r es. Originally, 5′-uridylic acid
derivatives 1-2 were designed as covalent-bonding mim-
ics of conformationally rigid structural motifs, such as
intramolecularly hydrogen-bonding 5-[(methylamino)-
methyl]uridine 5′-phosphate (pmnm5U)6 and water-
bridged pseudouridine 5′-phosphate (pψ)7 which are
known to exhibit predominantly the C3′-endo conforma-
tion (A-type or N-type RNA) in the ribose moiety. We also
disclosed that particularly the propylene-bridged 5′-
cyclouridylic acids derivatives 12 and 23,8 have typical C3′-
endo, anti, and g+ conformations for the ribose moiety,
glycoside bond, and C4′-C5′ torsion angle, respectively,
which are actually the same as that of the A-type RNA
strand.
This orientation, which produces a chiral phosphoryl
center at the junction with the Rp configuration, is the
same as that of the usual A-type RNA strand. The other
is depicted in the right side of Figure 2. This Sp config-
uration can apparently create a bending structure if all
the conformations of the cyclouridylic acid component can
be maintained as in the original structure.
Our previous studies revealed that the fractional
population of the C3′-endo conformation in 1 was 88%,2a
We adopted a straightforward method for the simul-
taneous phosphotriester bond formation between a nu-
cleoside and a nucleotide component (Scheme 1): The 2′-
O-methyluridine 3′-phosphorodiamidite derivative 4 was
prepared by reaction of 5′-O-DMTr-2′-O-methyluridine (3)
with bis(diisopropylamino)chlorophosphine according to
the method reported by van Boom et al.12 Condensation
of 4 with the 2′-O-methyluridine diol derivative 53 gave
(6) (a) Sakamoto, K.; Kawai, G.; Niimi, T.; Watanabe, S.; Niimi, T.;
Satoh, T.; Sekine, M.; Yamauti, Z.; Nisimura, S.; Miyazawa, T.;
Yokoyama, S. Eur. J . Biochem. 1993, 216, 369-375. (b) Sakamoto, K.;
Kawai, G.; Watanabe, S.; Niimi, T.; Hayasi, N.; Muto, Y.; Watanabe,
K.; Satoh, T.; Sekine, M.; Yokoyama, S. Biochemistry 1996, 35, 6533-
6538.
(7) Davis, D. R.; Poulter, C. D. Biochemistry 1991, 30, 4223-4231.
(8) In the case of 2 having the 2′-O-methyl group, the resonance
signal of 1′-H was observed as a singlet, suggesting the J 1′, 2′ coupling
constant is less than 1.0 Hz. J udging from the J 3′, value (8.1 Hz) of
(9) Altona, C.; Sundaralingham, M. J . Am. Chem. Soc. 1973, 95,
2333-2344.
4′
2, the fractional population of its C3′-endo conformation was expected
to be at least more than 89%.3 In this simulation, the sum of J 1′, 2′ and
(10) Davies, D. B. Prog. Nucl. Magn. Reson. Spectrosc. 1978, 12.
135-225.
J 3′, of 2 is estimated to be less than 9.1 Hz, which is relatively close
4′
to the standard value of 9.4 ( 0.2 Hz.10 Therefore, the sugar puckering
of 2 is faintly disordered because of the presence of both the cyclic
structure and the 2′-O-methyl group but is essentially similar to that
of 1.
(11) Sekine, M.; Tsuruoka, H.; Iimura, S.; Kusuoku, H.; Wada, T.
J . Org. Chem. 1996, 61, 4087-4100.
(12) Marugg, J . E.; Burik, A.; Tromp, M.; van der Marel, G. A.; van
Boom, J . H. Tetrahedron Lett. 1986, 27, 2271-2274.

Oligonucleotides Having a Phosphorus Chiral Center
J . Org. Chem., Vol. 65, No. 20, 2000 6517
Sch em e 1
the coupling product 6 as a diastereomeric mixture in
80% yield. It should be noted that this internucleotidic
phosphotriester bond formation proceeded effectively
without using any techniques such as the high dilution
method often used in the synthesis of macrocyclic com-
pounds.13 In other words, this result implies that the 12-
membered ring cyclic structure fits reasonably the natu-
rally occurring cyclic hydrogen bonding systems.
3′-oxygen of the neutral phosphoryl group increases its
electronegativity due to the electron-withdrawing effect
of the [-P(O)(OR)2] moiety so that the ribose conforma-
tion of the 5′-upstream Um can be expected to be strongly
affected. Therefore, we examined if this difference leads
to significant conformational change of both the 3′-
downstream cyclonucleoside and the 5′-upstream nucleo-
side in the modified dimer 8a ,b.
The 1′-H proton of the 3′-downstream pc3Um in 8a was
observed as a doublet with J 1′,2′ value of 1.7 Hz, which
shows 82% of C3′-endo conformation on the assumption
that the J 1′,2′ + J 3′,4′ value is equal to that of pU (9.2 Hz),
as shown in Table 1. Furthermore, it is noteworthy that
the 1′-H proton of the 3′-downstream pc3Um of 8b
appeared as a singlet within the resolution of the NMR
apparatus. This means that the sugar pucker of the 3′-
downstream pc3Um has an extremely predominant C3′-
endo conformation. From these results, it turned out that
the phosphotriester moiety did not affect the whole
conformation of the 3′-downstream pc3Um of 8a and 8b.
Expectedly, the sugar puckers of the 5′-upstream Um
in 8a and 8b were maintained predominantly in the C2′-
endo (B-type) conformation to the extent of 70% and 73%,
respectively, regardless of the presence of the 2′-O-
methoxy group, which is known to induce the C3′-endo
conformation.18
These results suggest that the dimers 8a and 8b have
conformationally locked C3′-endo and C2′-endo sugar
puckers at the 3′-downstream and 5′-upstream compo-
nents, respectively. In yeast tRNAPhe there exists a
sharply bent dimer motif with a sequence of U(7)pU(8)
having the C2′-endo and C3′-endo conformations at
positions 7 and 8, respectively.19 This fact is interesting
since our dimer model of either 8a or 8b could be used
as a conformationally restrained mimic of such a sharply
bent motif. Therefore, we needed to determine which
Deacetylation of 6 gave a pair of diastereomeric
products DMTrUmpc3Um (7a ,b) without damage of the
phosphotriester linkage. Fortunately, the stereoisomers
7a (fast-eluted product) and 7b (slow-eluted product)
could be readily separated by silica gel column chroma-
tography and successfully isolated in 46 and 43% yields,
respectively. Treatment of each diastereomer 7a or 7b
with 80% acetic acid gave the corresponding product
Umpc3Um(fast) (8a ) or Umpc3Um(slow) (8b).
Con for m a tion a l An a lysis of Um p c3Um (8a ,b) In -
volvin g a Cyclic P h osp h otr iester Str u ctu r e. First,
it should be taken into account that the neutral phos-
phoryl moiety in 8a ,b affects conformation of the original
ribose moieties of the Um and pc3Um components.
Generally, the dialkoxyphosphoryloxy [-OP(O)(OR)2]
group serves as a strong electron-withdrawing group,
while the monoalkoxyphosphoryloxy [-OP(O)(OR)(O-)]
group behaves as
a
weakly electron-withdrawing
group.14-16 Basically, the electronegativity of a substitu-
ent attached to the 2′ or 3′-carbon is a key factor for
control of the ribose pucker, as reported by Ikehara.17 The
(13) For a recent paper see Sinha, S. C.; Keinan, E. J . Org. Chem.
1997, 62, 377-386 and cited papers therein.
(14) Hansch, C.; Leo, A.; Taft, W. Chem. Rev. 1991, 91, 165-195.
For example, the diethoxyphosphoryl group exhibits a field/inductive
effect having the F value of 0.52,15 while the F values of the acetyl
and acetoxy groups are 0.33 and 0.42, respectively.17 Since the F value
of the dissociated phosphoryl group PO3H- is 0.19,17 monoalkoxyphos-
phoryl P(O)(OR)O- group should have a little lower F value than 0.19,
which means a weakly electron-withdrawing group.
(15) Freedman, L. D.; J affe, H. H. J . Am. Chem. Soc. 1955, 77, 920-
921.
(18) Kawai, G.; Yamamoto, Y.; Kamimura, T.; Masegi, T.; Sekine,
M.; Hata, T.; Iimori, T.; Watanabe, T.; Miyazawa, T.; Yokoyama, S.
Biochemistry 1992, 31, 1040-1046.
(19) Protein Databank, Protein Institute of Osaka University, J apan.
(20) Davies, D. B.; Danyluk, S. S. Biochemistry 1974, 13, 4417-4434.
(16) McDaniel, D. H.; Brown, H. C. J . Org. Chem. 1958, 23, 420-
427.
(17) Ikehara, M. Heterocycles 1984, 21, 75-90.

6518 J . Org. Chem., Vol. 65, No. 20, 2000
Sekine et al.
Ta ble 1. F r a ction a l P op u la tion s of C3′-En d o Con for m a tion of th e 5′-Up str ea m a n d 3′-Dow n str ea m Nu cleosid es of
Dim er s Con ta in in g a Cyclou r id ylic Acid Der iva tive a n d Rela ted Com p ou n d s
a
b
Reference 20. Reference 18. c The fractional populations of ribose residues were calculated according to ref 9.
stereoisomer 8a or 8b has a bent structure. This deter-
mination was done as described in a later section.
MacroModel ver 5.0.21 Each conformer was energy-
minimized with the implicit solvent model, GB/SA.22 This
calculation was carried out by constraining the 3′-
downstream c3Um in the C3′-endo conformation, which
has proven to be apparently predominant from the NMR
analysis. The 5′-upstream Um was also constrained as
the C2′-endo (major conformation) or C3′-endo form
(minor conformation). The results of these calculations
are shown in Figure 3A-D. At the upper left side of
Figure 3, UpU extracted from the typical A-type RNA
duplex is also shown.
In the case of the Rp isomer having only the 3′-endo
conformation in both the 5′-upstream Um and 3′-
downstream pc3Um, the most stable conformation (Fig-
ure 3A) was similar to that of UpU extracted from the
A-type RNA duplex. Therefore, it can be expected that
the dimer with this Rp configuration and the C3′-endo/
C3′-endo conformations can hybridize with ApA in a
natural way. In contrast, the most stable structure
(Figure 3B) of the Sp stereoisomer has a conformation
entirely different from that of UpU extracted from the
A-type RNA duplex. This conformer forms a normal
duplex in a straight direction only with difficulty.
On the other hand, the calculated structure (Figure 3C)
of the Rp stereoisomer having the C2′-endo and C3′-endo
conformations for the 5′-upstream Um and the 3′-
downstream pc3Um, respectively, showed the stacked
structure of the two uracil rings with different orienta-
tions with each other, suggesting that formation of the
duplex with ApA is impossible. The Sp isomer (Figure
3D) having the C2′-endo and C3′-endo conformations for
the 5′-upstream Um and the 3′-downstream pc3Um,
respectively, has a conformation considerably different
Electr on ic Effect of a 5′-Up str ea m 3′-Th iop h os-
p h or yl Su bstitu en t on Con for m a tion of th e Um
Ribose. To check the electronic effect of the phosphoryl
group on the conformation of the 5′-upstream Um, the
cyclic phosphorothioate derivatives Umspc3Um (10a ,b)
were synthesized via the intermediate 9, which could be
obtained by sulfurization of the common intermediate
obtained by condensation of 4 with 5. The cyclic phos-
phorothioates 10a (fast-eluted product) and 10b (slow-
eluted product) were separated and isolated. The 1H
NMR analysis of these stereoisomers showed essentially
similar results but a singlet signal for the 1′-H proton of
the 3′-downstream pc3Um in the fast-eluted isomer 10a
was observed, and the slow-eluted isomer 10b has the
J 1′,2′ value of 1.6 Hz (see Table 1).
These results suggest that the thiophosphoryl group
did not afford the conformational change of the 5′-
upstream Um appreciably, although the sulfur atom is
less electronegative than the oxygen.
Com p u t a t ion a l An a lysis of 3D-St r u ct u r es of
Um p c3Um 8 Ca p a ble of Du p lex F or m a tion w ith
RNA a n d DNA. The structures of the Rp and Sp
stereoisomers of Umpc3Um were examined by computa-
tional methods. As mentioned above, the 5′-upstream Um
ribose moieties of 8a and 8b predominantly exist in the
C3′-endo conformation while the 3′-downstream pc3Um
has the C3′-endo conformation. Accordingly, an initial 3D
model was constructed by using the B-type uridine 3′-
phosphate and the A-type uridine structures extracted
from the canonical B-type RNA and A-type RNA du-
plexes, respectively. The initial Rp and Sp model struc-
tures of Umpc3Um 8 were built by using these fragments
and adding the methyl groups, and many conformers
were generated according to the Monte Carlo method of
(21) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Caufield, C.; Chang, T.; Hendrickson, T.; Still, W. C. J . Comput. Chem.
1990, 11, 440-467.
(22) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J .
Am. Chem. Soc. 1990, 112, 6127-6129.

Oligonucleotides Having a Phosphorus Chiral Center
J . Org. Chem., Vol. 65, No. 20, 2000 6519
F igu r e 3. Energy-minimized structures of (Rp)-Umpc3Um and (Sp)-Umpc3Um in which the 5′-upstream ribose residue was
fixed in either C3′-endo or C2′-endo conformation and the 3′-downstream ribose residue was fixed in the C3′-endo conformation.
(A) (Rp)-Umpc3Um with C3′-endo/C3′-endo; (B) (Sp)-Umpc3Um with C3′-endo/C3′-endo; (C) (Rp)-Umpc3Um with C2′-endo/C3′-
endo; (D) (Sp)-Umpc3Um with C2′-endo/C3′-endo.
F igu r e 4. CD spectra of Umpc3Um(fast) (8a ) (A), Umpc3Um(slow) (8b) (B), and UpU (C) in phosphate buffer (1.0 A260/1 mL,
pH 7.0, 25, 50, 80 °C)
from that of UpU so that this isomer cannot form a stable
duplex with ApA.
7.0) at 25, 50, and 80 °C. These results are shown in
Figure 4A,B. It is clearly shown that the intensity of the
positive Cotton effect of 8b around 260 nm is ca. twice
that of 8a . The negative Cotton effects around 240 and
215 nm can be seen at 25 °C in 8b, but these features
typical for UpU cannot be clearly observed in 8a . These
results strongly suggested that the slow-eluted isomer
8b has the Rp configuration, which allows us to give a
structure closer to that of UpU, as suggested by the above
computer simulation.
In cor p or a tion of Cyclou r id ylic Acid Der iva tives
in to Deca u r id yla te a n d Deca th ym id yla te. To incor-
porate a conformationally rigid C2′-endo/C3′-endo (i.e.,
S/N) junction unit Umpc3Um into RNA and DNA oligo-
mers, the phosphoramidite units 11a and 11b were
synthesized in high yields from 7a and 7b, respectively.
The decauridylate derivatives U4[Umpc3Um(fast)]U4 (13a)
and U4[Umpc3Um(slow)]U4 (13b) incorporating the S/N
These calculations suggest that the Rp isomer of
Umpc3Um can form a normal duplex only when the 5′-
upstream Um takes the minor C2′-endo conformation,
but the Sp isomer cannot give such a normal linear
duplex in any cases. Therefore, it was expected that
duplexes formed between oligonucleotides containing Rp-
Umpc3Um and their complementary oligonucleotides
should have higher Tm values than those formed between
oligonucleotides containing Sp-Umpc3Um and their
complementary oligonucleotides. This prediction can be
used for determination of the absolute configuration of
8a and 8b.
CD Mea su r em en ts of 8a a n d 8b. To determine the
absolute configuration of the fast-eluted Umpc3Um 8a
and the slowly eluted Umpc3Um 8b, the CD spectra of
these compounds were measured in phosphate buffer (pH

6520 J . Org. Chem., Vol. 65, No. 20, 2000
Sekine et al.
Ta ble 2. Tm Va lu es of th e Du p lexes betw een Mod ified
Oligon u cleotid es a n d Deca a d en yla te or
Deca d eoxya d en yla te
melting temp of duplex (°C)
oligonucleotides
U(pU)9
U4[Umpc3Um(fast)]U4 (13a )
U4[Umpc3Um(slow)]U4 (13b)
T(pT)9
A(pA)9
∆Tm
d[A(pA)9]
∆Tm
16.7
<0
6.8
20.3
8.2
16.3
13.3
>-16.7 not obsd
-9.9 6.3
-7.0
27.0
T4[Umpc3Um(fast)]T4 (15a )
T4[Umpc3Um(slow)]T4 (15b)
-12.1 <5
-4.0 17.0
>-22.0
-10.0
Ta ble 3. Syn th esis of Din u cleotid es Con ta in in g a
Cyclou r id ylic Acid Un it
protected
dimer
HPLC ret
isolated yield
yield, % dimer time,b min yield, % by HPLC, %
6a ,b
7a
7b
80
46
43
F igu r e 5. Enzymatic digestion of the decauridylates 13a and
13b and decathymidylates 15a and 15b incorporating a
cyclouridylic acid unit of Umpc3Um(fast) or Umpc3Um(slow)
with venom phosphodiesterase and alkaline phosphatase. The
differences of Umpc3Um in retention time between A, B and
C, D arose from different HPLC apparatus used.
8a
8b
10a
10b
20.8
23.2
17.3
19.3
82
79
21
22
9a ,b
43
91
a
TLC was carried out by using the following solvent system:
b
CHCl3-Et2O-MeOH (20:10:3, v/v/v, three times). HPLC was
performed under the conditions described in the experimental
section.
Sch em e 2
junction units 8a and 8b were obtained in 22% and 18%
yields, respectively, by using the standard solid support
synthesis starting from an U-loaded CPG gel 12. Simi-
larly, decathymidylate derivatives T4[Umpc3Um(fast)]T4
(15a ) and T4[Umpc3Um(slow)]T4 (15b) incorporating the
S/N junction units 8a and 8b were obtained in 58% and
25% yields, respectively, from a T-loaded CPG gel 14. See
Scheme 2.
En zym a tic An a lysis of Oligon u cleotid es In cor -
p or a tin g a Cyclou r id ylic Acid Der iva tive. The cyc-
louridylic acid derivatives 1 and 2 were found to be
completely resistant to venom phosphodiesterase, spleen
phosphodiesterase, and nuclease P1 under the standard
conditions. Therefore, the cyclic bridge structure inhibits
the enzyme activity. This inhibition property can be used
for characterization of the structure of oligonucleotides
incorporating a cyclic structure. When U4[Umpc3Um-
(fast)]U4 (13a ) and U4[Umpc3Um(slow)]U4 (13b) were
incubated with venom phosphodiesterase followed by
alkaline phosphatase, the cyclic dimer fragments of
Umpc3Um(fast) (8a ) and Umpc3Um(slow) (8b) were
observed with the concomitant formation of uridine in
the correct ratios, as shown in Figure 5A,B, respectively.
Similarly, the enzyme digestion of T4[Umpc3Um(fast)]-
T4 (15a ) and T4[Umpc3Um(slow)]T4 (15b) gave the same
dimer fragments of Umpc3Um(fast) (8a ) and Umpc3Um-
(slow) (8b), as shown in Figure 5C,D, respectively.
F igu r e 6. Melting curves of the duplexes formed between
A(pA)9 or d[A(pA)9] and oligonucleotides 13 or 15. Panel A:
(O) A(pA)9-U4[Umpc3Um(fast)]U4(13a ), (×) A(pA)9-U4-
[Umpc3Um(slow)]U4(13b), (b) A(pA)9-U(pU)9. Panel B: (O)
d[A(pA)9]-U4[Umpc3Um(fast)]U4(13a ), (×) d[A(pA)9]-U4-
[Umpc3Um(slow)]U4(13b), d[A(pA)9]-U(pU)9. Panel C: (O)
A(pA)9-T4[Umpc3Um(fast)]T4(15a ), (×) A(pA)9-T4[Umpc3Um-
(slow)]T4U4(15b), (b) A(pA)9-T(pT)9. Panel D: (O) d[A(pA)9]-
T4[Umpc3Um(fast)]T4(15a), (×) d[A(pA)9]-T4[Umpc3Um(slow)]-
T4(15b), (b) d[A(pA)9]-T(pT)9.
Th er m a l Sta bility of Du p lexes of U4[Um p c3Um -
(fa st )]U4 (13a ) a n d U4[Um p c3Um (slow )]U4 (13b )
w ith A(p A)9 a n d d [A(p A)9]. We studied the thermal
stabilities of all possible four duplexes of U4[Umpc3Um-
(fast)]U4 (13a ) and U4[Umpc3Um(slow)]U4 (13b) with the
complementary oligonucleotides A(pA)9 and d[A(pA)9]
compared with those of the reference duplexes A(pA)9-
U(pU)9 and d[A(pA)9]-U(pU)9. It is also of great interest
to examine how the duplexes are affected by the presence
of the S/N junction point. The results of the Tm experi-
ments of all the possible duplexes are summarized in
Table 2 and Figure 6A,B. The duplex of A(pA)9-U4-
[Umpc3Um(fast)]U4(13a ) shows a little change in the
melting curve, as shown in Figure 6A. On the other hand,
the melting curve of A(pA)9-U4[Umpc3Um(slow)]U4(13b)
shows a more typical sigmoid curve, giving a higher Tm

Oligonucleotides Having a Phosphorus Chiral Center
J . Org. Chem., Vol. 65, No. 20, 2000 6521
of 6.8 °C, which is considerably low compared with that
of A(pA)9-U(pU)9.
The duplex of d[A(pA)9]-U4[Umpc3Um(fast)]U4(13a )
shows no clear Tm curve, as shown in Figure 6B. In
contrast, the duplex of d[A(pA)9]-U4[Umpc3Um(slow)]-
U4(13b) shows a sigmoid curve with a lower Tm value
(7.0 °C) than that of d[A(pA)9]-U(pU)9.
In these Tm experiments, the decauridylate 13b con-
taining Umpc3Um(slow) (8b) shows unexceptionally
higher Tm values than 13a containing Umpc3Um(fast)
(8a ). Therefore, these results indicate that Umpc3Um-
(slow) 8b has the Rp configuration and Umpc3Um(fast)
8a has the Sp configuration, which does not allow
favorable duplex formation. This conclusion is in agree-
ment with that from the CD analysis.
Th er m a l Sta bility of Du p lexes of T4[Um p c3Um -
(fa st)]T4 15a a n d T4[Um p c3Um (slow )]T4 15b w ith
A(p A)9 a n d d [A(p A)9]. Figure 6C shows the melting
curves of three kinds of duplexes, i.e., A(pA)9-
T4[Umpc3Um(fast)]T4(15a ), A(pA)9-T4[Umpc3Um(slow)]-
T4(15b), and A(pA)9-T(pT)9. As seen in Figure 6C, there
is a big difference in Tm between A(pA)9-T4[Umpc3Um-
(fast)]T4(15a ) and A(pA)9-T4[Umpc3Um(slow)]T4(15b).
The latter has the Tm value of 16.3 °C, which is lower by
4.0 °C than that of A(pA)9-T(pT)9, while the former
showed a very low Tm value of 8.2 °C. Figure 6D shows
the melting curves of the duplexes of d[A(pA)9]-
T4[Umpc3Um(fast)]T4(15a ), d[A(pA)9]-T4[Umpc3Um-
(slow)]T4(15b), and d[A(pA)9]-T(pT)9. In the case of
d[A(pA)9]-T4[Umpc3Um(slow)]T4(15b), the Tm value was
considerably low (17.0 °C) compared with that (27 °C) of
d[A(pA)9]-T(pT)9. Furthermore, the duplex of d[A(pA)9]-
T4[Umpc3Um(fast)]T4(15a ) has a much lower Tm value
(5.8 °C). This marked decrease (21.2 °C) is probably one
of the extreme cases reported so far. These results also
support our previous conclusion that Umpc3Um(fast) (8a )
should have the Sp configuration.
orientation of the phosphate moiety is unfavorable for
duplex formation. In other words, there is a possibility
that the backbone structure can be artificially bent by
use of this unfavorable stereoisomer as the bending point.
Finally, it should be noted that such an S/N junction
actually exists in naturally occurring RNA molecules.
After extensive search, we noticed that in the crystal
structure of E. coli tRNAPhe this S/N junction of U(7)pU-
(8) sequence is present and plays a role in changing the
orientation of the base vectors at the position between
the acceptor- and D-stems.4a Therefore, our structural
motif having a S/N junction can also be used to fix the
bending in certain RNAs.
It is also interesting to introduce a rigid C3′-endo
nucleoside into the 5′-upstream nucleoside of the UpU
dimer motif. As such a rigid nucleoside, bicyclonucleoside
derivatives recently reported by Imanishi23k,l and
Wengel23m,n are facinating, since the sugar conformation
of these locked nucleosides must be unaffected by the
presence of a 3′-phosphotriester moiety. Further studies
are now under investigation.
Exp er im en ta l Section
Gen er a l Meth od s. 1H, 13C, and 31P NMR spectra were
recorded at 270, 68, and 109 MHz, respectively. The chemical
shifts were measured from tetramethylsilane or DSS for 1H
NMR spectra, CDCl3 (77 ppm) or DSS (0 ppm) for 13C NMR
spectra, and 85% phosphoric acid (0 ppm) for 31P NMR spectra.
The 1H-1H coupling constants were measured at 400 MHz
using Varian Unity 400 spectrometer at 20 °C. UV spectra
were recorded on a U-2000 spectrometer. CD spectra were
recorded on J -500 C spectrometer using a 0.5 cm cell. The
synthesis of oligonucleotides was performed on a DNA syn-
thesizer model 381A or DNA/RNA synthesizer model 392. TLC
was performed by the use of Kieselgel 60-F-254 (0.25 mm).
Column chromatography was performed with silica gel C-200
purchased from Wako Co. Ltd., and a minipump for a goldfish
bowl was conveniently used to attain sufficient pressure for
rapid chromatographic separation. Reversed-phase column
chromatography was performed by the use of µBondapak C-18
silica gel (prep S-500). Reversed-phase HPLC was performed
using the following systems. System 1: An LC module 1 was
used with an M-741 data module and a µBondasphere 5 µ C18
Con clu sion s
In this study, it was found that the 5′-upstream 2′-O-
methyluridine having a cyclic structure has the C2′-endo
conformation predominantly over the C3′-endo conforma-
tion. On the other hand, the 3′-downstream 2′-O-cyclou-
ridine of the dimer exists in favor of the C3′-endo
conformation because of the rigid structure. All duplexes
between decathymidylates or decauridylates incorporat-
ing Umpc3Um(fast or slow) and the complementary
decadeoxyadenylates or decaadenylates showed decrease
of the Tm values. Particularly, decanucleotides having
Umpc3Um(fast) exhibited significantly lower Tm values
than those having Umpc3Um(slow).
Molecular mechanics calculations of two stereoisomers
of Umpc3Um indicate that the Rp isomer can form
relatively normal duplexes when the 5′-upstream uridine
turns back to the C3′-endo conformation but the Sp
isomer is unfavorable for stable duplex formation even
in any combination of the C2′-endo and C3′-endo confor-
mations in the 5′-upstream and 3′-downstream uridine
moieties. From these results, it could be concluded that
Umpc3Um(fast) is the Sp isomer and Umpc3Um(slow)
is the Rp isomer. Although the 5′-upstream nucleoside
residue tends to have the C2′-endo conformation, deca-
nucleotides having (Rp)-Umpc3Um are capable of duplex
formation with a disordered partial structure. In addi-
tion, it is also suggested that in Umpc3Um (Sp) the
(23) In connection with the present study, there have been reported
conformationally locked nucleosides: (a) Tarko¨y, M.; Bolli, M.; Sch-
weizer, B.: Leumann, C. Helv. Chim. Acta 1993, 76, 481-510. (b) Egli,
M.; Lubini, P.; Bolli, M.; Dobler, M.; Leumann, C. J . Am. Chem. Soc.
1993, 115, 5855-5856. (c) Tarko¨y, M.; Leumann, C. Angew. Chem.,
Int. Ed. Engl. 1993, 32, 1432-1434. (d) Tarko¨y, M.; Bolli, M.; Leumann,
C. Helv. Chim. Acta 1994, 77, 716-744. (e) Bolli, M.; Lubini, P.;
Leumann, C. Helv. Chim. Acta 1995, 78, 2077-2096. (f) Litten, C. J .;
Epple, C.; Leumann, C. J .; Bioorg. Med. Chem. Lett. 1995, 5, 1231-
1234. (g) Litten, J . C.; Leumann, C. Helv. Chim. Acta 1996, 79, 1129-
1146. (h) J ones, R. J .; Swaminathan, S.; Milligan, J . F.; Wadwani, S.;
Froehler, B. C.; Matteucci, M. D. J . Am. Chem. Soc. 1993, 115, 9816-
9817. (i) Zou, R.; Matteucci, M. D. Tetrahedron Lett. 1996, 37, 941-
944. (j) Altmann, K.-H.; Kesselring, R.; Francotte, E.; Rihs, G.
Tetrahedron Lett. 1994, 35, 2331-2334. (k) Altmann, K.-H.; Imwinkel-
ried, R.; Kesselring, R.; Rihs, G. Tetrahedron Lett. 1994, 35, 7625-
7628. (l) Herdewijn, P. Liebigs Ann. 1996, 1337-1348. (m) Egli, M.
Angew. Chem., Int. Ed. Engl. 1996, 35, 1894-1909. (n) Ezzitouni, A.;
Russ, P.; Marquez, V. E. J . Org. Chem. 1997, 62, 4870-4873. (o) Obika,
S.; Nanbu, D.; Hari, Y.; Morio, K.; In, Y.; Ishida, T.; Imanishi, T.
Tetrahedron Lett. 1997, 38, 8735-8738. (p) Obika, S.; Morio, K.;
Nanbu, D.; Imanishi, T. Chem. Commun. 1997, 1643-1644. (q) Obika,
S.; Nanbu, D.; Hari, Y.; Andoh, J .; Morio, K.; Doi, T.; Imanishi, T.
Tetrahedron Lett. 1998, 39, 5401-5404. (r) Singh, S. K.; Nielsen, P.;
Koshkin, A. A.; Wengel, J . Chem. Commun. 1998, 455-456. (s)
Koshkin, A. A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.; Kumar,
R.; Meldgaard, M.; Olsen, C. E.; Wengel, J . Tetrahedron 1998, 54,
3607-3630. (t) Christensen, N. K.; Petersen, M.; Nielsen, P.; J acobsen,
J . P.; Olsen, C. E.; Wengel, J . J . Am. Chem. Soc. 1998, 120, 5458-
5463. (u) Wang, G.; Gunic, E.; Girardet, J .-L.; Stoisavljevic, V. Bioorg.
Med. Chem. Lett. 1999, 9, 1147-1150.

6522 J . Org. Chem., Vol. 65, No. 20, 2000
Sekine et al.
100 Å column (3.9 × 150 mm) at 50 °C with a linear gradient
(0-30%) of CH3CN in 0.1 M NH4OAc, pH 7.0 at a flow rate of
1.0 mL/min for 30 min. System 1b (for 10a ,b): The same
conditions as described in system 1 were used except for the
conditions of a linear gradient (0-50%) of CH3CN in 0.1 M
NH4OAc, pH 7.0 for 25 min. System 2: A 2690 separation
module was used with a 996 photodiode array detector and a
Millennium 2010 chromatography manager. The other condi-
tions are the same as those of system 1. System 3: A Shimazu
SCL-6A was used with LC-6A, SPD-6A, and C-R3A. The other
conditions are the same as those of system 1 except for a flow
rate of 3.0 mL/min. Anion-exchange HPLC was done on an
LC module 1 with an M-741 data module and a column heater
with a 10-67% linear gradient of 25 mM phosphate, 1 M
sodium chloride buffer (pH 6.9) in 25 mM phosphate buffer
(pH 6.0) at a flow rate of 1.0 mL/min for 40 min. Pyridine was
distilled two times from p-toluenesulfonyl chloride and from
calcium hydride and then stored over molecular sieves 4 Å.
Elemental analyses were performed by the Microanalytical
Laboratory, Tokyo Institute of Technology at Nagatsuta.
Syn th esis of P a r tia lly P r otected 2′-O-Meth ylu r id ylyl-
(3′-5′)2′-O-m eth yld eoxyu r id in e Der iva tive 7a ,b Ha vin g
a Cyclic Str u ctu r e w ith a P r op ylen e Br id ge. The fully
protected dimer 6a ,b (721 mg, 0.748 mmol) was dissolved in
pyridine (10.0 mL) and 25% aqueous ammonia (10.0 mL). After
being stirred at room temperature for 3.5 h, the mixture was
evaporated under reduced pressure. The residue was chro-
matographed on a column of silica gel (70 g) with [CHCl3-
ether (4:1, v/v)]-MeOH (100:3, v/v) containing 1% pyridine to
give DMTrUmpc3Um (fast isomer) 7a (319 mg, 46%) and
DMTrUmpc3Um (slow isomer) 7b (298 mg, 43%) as foam.
DMT-Umpc3Um (fast isomer) 7a : TLC Rf 0.33 (CHCl3:Et2O:
1
CH3OH, 20:10:3, v/v/v, developed three times); H NMR (270
MHz, CDCl3) δ 1.70-1.95 (2H, m), 2.40-2.60 (2H, m), 3.45
(1H, m), 3.62-3.80 (15H, m), 3.88 (1H, d, J 2′,3′ ) 4.90 Hz),
4.15-4.27 (4H, m), 4.47 (1H, m), 4.68 (1H, m), 5.08 (1H, m),
5.23 (1H, d, J 5,6 ) 7.90 Hz), 5.91 (1H, s), 6.10 (1H, d, J 1′,2′
)
3.60 Hz), 6.84-6.87 (4H, m), 7.20-7.31 (9H, m), 7.80 (1H, s),
7.84 (1H, d, J 5,6 ) 8.20 Hz), 9.70 (1H, br), 9.84 (1H, br); 13C
NMR (68 MHz, CDCl3) δ 20.52, 26.81, 55.24, 58.60, 58.87,
61.33, 64.24, 65.68, 67.32, 73.77, 81.60, 82.10, 82.22, 82.41,
83.36, 86.45, 87.26, 87.48, 102.77, 110.24, 113.30, 127.44,
128.05, 128.25, 130.23, 134.45, 134.61, 137.34, 139.52, 143.63,
150.04, 150.60, 158.83, 163.14, 163.58; 31P NMR (109 MHz,
CDCl3) δ -2.29. Anal. Calcd for C44H49N4O16P‚H2O: C, 56.28;
H, 5.48; N, 5.96. Found: C, 56.21; H, 5.45; N, 5.89. DMT-
Umpc3Um (slow isomer) 7b: TLC Rf 0.22 (CHCl3:Et2O:CH3-
OH, 20:10:3, developed three times); 1H NMR (270 MHz,
CDCl3) δ 1.88-2.09 (2H, m), 2.35-2.57 (2H, m), 3.50-4.32
(23H, m), 5.09 (1H, m), 5.25 (1H, d, J 5,6 ) 7.90 Hz), 5.93 (1H,
s), 6.02 (1H, d, J 1′,2′ ) 2.30 Hz), 6.84-6.87 (4H, m), 7.24-7.34
(9H, m), 7.82 (1H, s), 7.94 (1H, d, J 5,6 ) 8.30 Hz), 9.80 (2H,
br); 13C NMR (68 MHz, CDCl3) δ 20.90, 27.04, 55.27, 58.65,
58.78, 60.72, 63.45, 66.05, 67.35, 73.71, 81.04, 81.18, 81.92,
82.03, 82.28, 83.50, 87.22, 87.37, 102.55, 110.37, 113.37,
127.42, 128.10, 128.30, 130.26, 130.33, 134.61, 134.68, 137.64,
139.62, 143.79, 150.13, 150.37, 158.79, 158.83, 163.43, 163.66;
5′-O-(4,4′-Dim et h oxyt r it yl)-2′-O-m et h ylu r id in e
3′-
(N,N,N′,N′-tetr a isop r op yl)p h osp h or od ia m id ite (4). 5′-O-
(4,4′-Dimethoxytrityl)-2′-O-methyluridine (1.12 g, 2.0 mmol)
was coevaporated three times with dry toluene and finally
dissolved in dry dioxane (10 mL). To this solution were added
triethylamine (834 ul, 6.0 mmol) and bis(diisopropylamino)-
chlorophosphine26 (1.05 g, 4.0 mmol). The resulting mixture
was stirred at room temperature for 2 h and then quenched
by addition of ethanol (1 mL). The mixture was diluted with
CHCl3 (50 mL). The CHCl3 solution was washed three times
with 5% NaHCO3 (50 mL), dried over Na2SO4, filtered, and
evaporated under reduced pressure. The residue was dissolved
in CHCl3 (2 mL), and this solution was added portionwise to
hexane (100 mL) with vigorous stirring. The resulting white
precipitate was collected by removal of the solvent by decanta-
tion and dried in vacuo to give 4 (930 mg, 59%): 31P NMR
(109 MHz, CDCl3) δ 119.12. This compound was used for the
next experiment without further purification because of its
extreme instability.
31P NMR (109 MHz, CDCl3) δ -1.67. Anal. Calcd for C44H49
-
N4O16P‚H2O: C, 56.28; H, 5.48; N, 5.96. Found: C, 56.22; H,
5.48; N, 5.96.
Syn t h esis of F u lly P r ot ect ed 2′-O-Met h yl-2′-d eox-
yu r id ylyl(3′-5′)2′-O-m eth yld eoxyu r id in e Der iva tive 6a ,b
Ha vin g a Cyclic Str u ctu r e w ith a P r op ylen e Br id ge. A
mixture of 5 (358 mg, 1.0 mmol) and 1H-tetrazole (420 mg,
6.0 mmol) was rendered anhydrous by coevaporation with dry
toluene and the residue was dissolved in dry acetonitrile (50
mL) and dioxane (50 mL). A solution of 4 (1.0 mmol) in
acetonitrile (2.0 mL) was added dropwise to the mixture over
a period of 5 min. The resulting mixture was stirred at room
temperature for 1 h and a solution of 4 (0.50 mmol) in
acetonitrile (1.0 mL) was added. After stirring was continued
for an additional 4 h, t-BuOOH (1.0 mL, 10.0 mmol) was added
to the mixture. The solution was evaporated under reduced
pressure after stirring for 30 min, and the residue was
extracted with CHCl3 (100 mL) and 5% NaHCO3 (100 mL).
The organic layer was collected, washed two times with 5%
NaHCO3 (10 mL), dried over Na2SO4, filtered, and evaporated
under reduced pressure. The residue was chromatographed
on a column of silica gel (50 g) with CHCl3-MeOH (100:2, v/v)
containing 1% pyridine to give DMTrUmpc3Um(OAc) 6a ,b as
foam (771 mg, 80%, the ratio of the diasteromers 1:1): 1H NMR
(270 MHz, CDCl3) δ 0.64-2.04 (4H, m), 2.14 (6H, s), 2.35-
2.55 (4H, m), 3.37-4.43 (37H, m), 4.67-4.72 (1H, m), 4.85-
4.96 (2H, m), 5.00-5.18 (2H, m), 5.21-5.29 (2H, m), 5.99-
6.04 (4H, m), 6.84-6.87 (8H, m), 7.15-7.34 (18H, m), 7.73 (2H,
s), 7.91-7.95 (2H, m), 9.78-9.90 (4H, m); 13C NMR (68 MHz,
CDCl3) δ 20.56, 20.81, 21.46, 27.01, 55.27, 58.74, 58.94, 60.65,
63.22, 64.10, 66.04, 68.68, 69.06, 72.74, 73.69, 79.53, 81.01,
81.15, 81.76, 81.85, 82.30, 82.68, 87.17, 87.40, 88.03, 88.21,
102.60, 110.58, 110.87, 113.35, 127.47, 128.10, 128.23, 128.34,
129.04, 130.28, 134.50, 134.54, 134.70, 137.28, 139.51, 139.60,
143.68, 143.75, 149.67, 150.11, 150.17, 150.33, 150.38, 158.87,
163.36, 163.52, 163.57, 170.01, 170.22; 31P NMR (109 MHz,
CDCl3) δ -2.32, -1.88. Anal. Calcd for C46H51N4O17P: C, 57.38;
H, 5.34; N, 5.82. Found: C, 57.00; H, 5.58; N, 5.71.
Un pr otected 2′-O-Meth ylu r idylyl(3′-5′)2′-O-m eth yldeox-
yu r id in e Der iva tive 8a ,b Ha vin g a Cyclic Str u ctu r e w ith
a P r op ylen e Br id ge. Compound 7a or 7b (55 mg, 60 µmol)
was dissolved in 80% acetic acid (2 mL). After being kept at
room temperature for 30 min, the mixture was evaporated
under reduced pressure. The residue was partitioned between
ether (2.5 mL) and water (2.5 mL). The aqueous layer was
collected, washed two times with ether, and lyophilized to give
8a (30 mg, 82%) or 8b (29 mg, 79%). Umpc3Um (fast isomer)
8a : 1H NMR (400 MHz, DMSO-d6) δ 1.68-1.71 (2H, m), 2.32-
2.35 (2H, m), 3.40 (3H, s), 3.46 (3H, s), 3.63 (2H, m), 3.86 (1H,
dd, J 1′,2′ ) 1.7 Hz, J 2′,3′ ) 5.0 Hz), 3.88-3.94 (1H, m), 4.02-
4.06 (2H, m), 4.12-4.17 (2H, m), 4.21-4.26 (2H, m), 4.45-
4.49 (1H, m), 4.97 (1H, seven, J 2′,3′ ) 4.90 Hz, J 3′,4′ ) 2.80 Hz,
J 3′,P ) 7.50 Hz), 5.30-5.36 (2H, m), 5.72 (1H, dd, J 5,6 ) 8.1
Hz, J 5,NH ) 2.2 Hz), 5.83 (1H, d, J 1′,2′ ) 1.70 Hz,), 5.90 (1H, d,
J 1′,2′ ) 6.60 Hz), 7.55 (1H, s), 7.88 (1H, d, J 5,6 ) 8.10 Hz), 11.36
(1H, br), 11.44 (1H, d, J NH-5H ) 2.2 Hz); 13C NMR (68 MHz,
DMSO-d6) δ 20.27, 26.51, 26.65, 49.19, 57.86, 60.49, 64.85,
65.09, 67.59, 74.47, 74.54, 80.38, 81.33, 81.46, 82.86, 83.65,
85.50, 87.01, 102.46, 109.42, 137.41, 140.29, 150.19, 150.59,
162.89, 163.34; 31P NMR (109 MHz, DMSO-d6) δ -0.98. Anal.
Calcd for C23H31N4O14P‚H2O: C, 43.40; H, 5.23; N, 8.80.
Found: C, 43.51; H, 5.23; N, 8.44. Umpc3Um (slow isomer)
8b: 1H NMR (400 MHz, DMSO-d6) δ 1.62-1.80 (2H, m), 2.25-
2.30 (1H, m), 2.36-2.41 (1H, m), 3.37 (3H, s), 3.47 (3H, s), 3.63
(2H, m), 3.85 (1H, d, J 2′,3′ ) 4.9 Hz), 3.93-3.97 (1H, m), 4.03-
4.05 (2H, m), 4.10 (1H, t, J 1′,2′ ) J 2′,3′ ) 4.90 Hz), 4.18-4.25
(3H, m), 4.34-4.37 (1H, m, J 5′,5" ) 11.80 Hz), 4.96 (1H, seven,
J 2′,3′ ) 4.80 Hz, J 3′,4′ ) 2.50 Hz, J 3′,P ) 7.10 Hz), 5.32-5.37
(2H, m), 5.72 (1H, d, J 5,6 ) 8.10 Hz), 5.83 (1H, s), 5.90 (1H, d,
J 1′,2′ ) 6.60 Hz), 7.65 (1H, s), 7.89 (1H, d, J 5,6 ) 8.10 Hz), 11.33
(1H, br), 11.44 (1H, br); 13C NMR (68 MHz, DMSO-d6) δ 20.53,
26.60, 26.71, 57.92, 60.51, 63.54, 65.34, 67.46, 75.19, 75.28,

Oligonucleotides Having a Phosphorus Chiral Center
J . Org. Chem., Vol. 65, No. 20, 2000 6523
80.54, 81.24, 81.37, 83.02, 83.79, 83.87, 85.20, 87.30, 102.48,
109.44, 137.18, 140.13, 150.15, 150.59, 162.89, 163.40; 31P
HPLC: 19.4 min (system 1b); UV (H2O-CH3CN, 100:1, v/v)
λmax 261 nm, λmin 231 nm; 1H NMR (400 MHz, DMSO-d6) δ
1.60-1.82 (2H, m), 2.28-2.43 (2H, m), 3.38 (3H, s), 3.41 (3H,
s), 3.58-3.65 (2H, m), 3.82 (1H, m, J 1′,2′ ) 1.22 Hz, J 2′,3′ ) 5.03
Hz), 4.02-4.08 (2H, m), 4.12-4.17 (3H, m), 4.29 (1H, m, J 2′,3′
) 5.18 Hz, J 3′,P ) 7.41 Hz), 4.53 (1H, ddd, J 4′,5′ ) 2.29 Hz, J 5′,5"
) 9.51 Hz, J 5′,P ) 6.71 Hz, 5.34-5.39 (2H, m), 5.72 (1H, d, J 5,6
NMR (109 MHz, DMSO-d6) δ -1.61. Anal. Calcd for C23H31
-
N4O14P‚1.5H2O: C, 42.79; H, 5.31; N, 8.68. Found: C, 42.89;
H, 5.10; N, 8.63.
Syn th esis of F u lly P r otected 2′-O-Meth ylu r id ylyl(3′-
5′)2′-O-m eth yld eoxyu r id in e P h osp h or oth ioa te Der iva -
tive 9a ,b Ha vin g a Cyclic Str u ctu r e w ith a P r op ylen e
Br id ge. A mixture of 5 (179 mg, 0.5 mmol) and 1H-tetrazole
(210 mg, 3.0 mmol) was rendered anhydrous by coevaporation
with dry toluene, and the residue was dissolved in dry
acetonitrile (25 mL) and dioxane (25 mL). A solution of 4 (0.5
mmol) in acetonitrile (2.0 mL) was added dropwise to the
mixture over a period of 5 min. The resulting mixture was
stirred at room temperature for 4 h and a solution of 4 (0.25
mmol) in CS2 (1.0 mL) was added. After stirring was continued
for an additional 5 h, the mixture was evaporated under
reduced pressure. The residue was dissolved in acetonitrile (5
mL) and a solution of elemental sulfur (641 mg, 2.5 mmol) in
acetonitrile (5.0 mL) was added to the mixture. The solution
was stirred at room temperature for 3 h, filtered, and evapo-
rated under reduced pressure. The residue was extracted with
CHCl3 (50 mL) and 5% NaHCO3 (50 mL). The organic layer
was collected, washed two times with 5% NaHCO3 (25 mL),
dried over Na2SO4, filtered, and evaporated under reduced
pressure. The residue was dissolved in pyridine (5 mL) and
25% aqueous ammonia (5 mL) was added. The mixture was
stirred at room temperature for 3 h and evaporated under
reduced pressure. The residue was chromatographed on a
column of silica gel (25 g) with CHCl3-MeOH (100:2, v/v)
containing 1% pyridine to give DMT-Umspc3Um 9a ,b as foam
(200 mg, 43%, the ratio of the diasteromers 1.1:1): 1H NMR
(270 MHz, CDCl3) δ 1.74-2.00 (4H, m), 2.34-2.86 (4H, m),
3.43-4.35 (45H, m), 4.66-4.69 (1H, m), 5.20-5.32 (2H, m),
5.88 (1H, s, 1′-H of pU), 6.00 (1H, s), 6.04 (1H, d, J 1′,2′ ) 2.64
Hz), 6.12 (1H, d, J 1′,2′ ) 4.62 Hz), 5.99-6.04 (4H, m), 6.84-
6.87 (8H, m), 7.25-7.41 (18H, m), 7.71 (1H, s), 7.82 (1H, d,
J 5,6 ) 8.25 Hz), 7.93 (1H, s), 8.01 (1H, d, J 5,6 ) 7.91 Hz), 10.02-
10.13 (4H, m); 13C NMR (68 MHz, CDCl3) δ 20.58, 20.88, 21.31,
26.11, 26.70, 54.65, 55.13, 58.51, 58.64, 58.80, 60.40, 61.69,
62.12, 63.00, 64.91, 65.84, 67.10, 67.78, 73.71, 74.68, 80.94,
81.74, 81.89, 82.19, 83.18, 83.49, 86.27, 87.21, 87.40, 102.50,
102.86, 109.90, 110.60, 113.23, 125.12, 127.24, 127.94, 128.10,
128.19, 128.86, 129.00, 130.06, 130.21, 134.43, 134.47, 134.56,
134.65, 137.47, 137.68, 139.46, 143.65, 143.68, 150.15, 150.39,
150.71, 158.63, 158.69, 163.38, 163.61, 163.76, 163.85; 31P
NMR (109 MHz, CDCl3) δ 67.17, 68.80. Anal. Calcd for
) 8.09 Hz), 5.89 (1H, d, J 1′,2′ ) 1.56 Hz), 5.90 (1H, d, J 1′,2′
)
6.87 Hz), 7.44 (1H, s), 7.87 (1H, d, J 5,6 ) 8.14 Hz), 11.35 (1H,
br), 11.43 (1H, br); 13C NMR (68 MHz, DMSO-d6) δ 20.54,
26.35, 26.47, 57.97, 60.70, 65.88, 68.22, 75.64, 79.18, 80.27,
81.03, 81.17, 83.06, 83.74, 85.48, 86.92, 102.68, 109.74, 137.74,
140.33, 150.26, 150.71, 162.98, 163.33; 31P NMR (109 MHz,
DMSO-d6) δ 67.90.
Syn th esis of 5′-O-(4,4′-Dim eth oxytr ityl)-2′-O-m eth ylu -
r id ylyl(3′-5′)2′-O-m et h ylu r id in e P h osp h or a m id it e De-
r iva tive (11a ,b). DMT-Umpc3Um (fast isomer) (7a ) (230 mg,
0.25 mmol) was rendered anhydrous by repeated coevapora-
tions with dry toluene and finally dissolved in CH2Cl2 (2.5 mL).
To the mixture were added triethylamine (105 µL, 0.75 mmol)
and 2-cyanoethoxydiisopropylaminochlorophosphine (89 mg,
0.375 mmol). The mixture was stirred at room temperature
for 2 h and then quenched by addition of ethanol (0.5 mL).
The mixture was diluted with CHCl3 (20 mL). The CHCl3
solution was washed three times with 5% NaHCO3 (20 mL),
dried over Na2SO4, filtered, and evaporated under reduced
pressure. The residue was dissolved in CHCl3 (2 mL) and this
solution was added portionwise to hexane (100 mL)-ether (30
mL)-pyridine (1.5 mL) with vigorous stirring. The resulting
white precipitate was collected by removal of the solvent by a
pipet and dried in vacuo to give DMTrUmpc3Ump(a,ce) (fast
isomer) (11a ) (241 mg, 86%). In a similar manner, from 7b
(230 mg, 0.25 mmol) DMTrUmpc3Ump(a,ce) (slow isomer)
(11b) (271 mg, 97%) was obtained. DMTrUmpc3Ump(a,ce)
(fast isomer) (11a ): 1H NMR (270 MHz, CDCl3) δ 1.14-1.28
(24H, m), 1.60-1.81 (4H, m), 2.36-2.64 (8H, m), 3.40-4.40
(48H, m), 4.69-4.75 (2H, m), 4.96-5.07 (2H, m), 5.17-5.24
(1H, m), 5.88-5.95 (2H, 2s), 6.02-6.06 (2H, m), 6.81-6.88 (8H,
m), 7.17-7.37 (18H, m), 7.70-7.73 (2H, 2s), 7.86-7.94 (2H,
2d, J 5,6 ) 8.40 Hz), 8.19 (4H, br); 13C NMR (68 MHz, CDCl3)
δ 20.25, 24.35, 24.46, 27.05, 42.98, 43.16, 55.08, 57.58, 57.94,
58.22, 58.42, 58.64, 60.47, 61.13, 63.70, 65.48, 68.29, 68.50,
68.81, 72.11, 80.52, 82.30, 82.55, 83.04, 86.85, 87.01, 87.17,
87.39, 88.23, 102.28, 110.08, 110.24, 113.10, 117.52, 117.77,
127.22, 127.87, 128.03, 129.99, 134.43, 134.54, 139.26, 143.65,
150.28, 158.65, 163.49, 163.90; 31P NMR (109 MHz, CDCl3) δ
-2.00, -1.09, 149.89, 151.11. MS (FAB+) calcd for C53H67N6O17P
(M + H) 1121.4038, found 1121.4049. DMTrUmpc3Ump(a,ce)
(slow isomer) (11b): 1H NMR (270 MHz, CDCl3, TMS) δ 1.17-
1.26 (24H, m), 1.70-2.01 (4H, m), 2.38-2.83 (8H, m), 3.44-
4.33 (54H, m), 5.09-5.19 (2H, m), 5.22-5.26 (2H, m), 5.95-
6.02 (4H, m), 6.84-6.88 (8H, m), 7.21-7.32 (18H, m), 7.68-
7.74 (2H, 2s), 7.90-7.96 (2H, m), 8.53 (4H, br); 13C NMR (68
MHz, CDCl3) δ 20.02, 20.11, 20.38, 24.35, 24.46, 26.83, 42.97,
43.15, 54.93, 57.81, 58.11, 58.29, 58.44, 60.27, 62.63, 65.66,
68.47, 68.64, 73.26, 80.54, 80.68, 82.10, 83.15, 86.97, 88.05,
102.28, 110.37, 110.53, 113.03, 117.59, 118.04, 127.08, 127.78,
128.03, 130.03, 134.29, 134.45, 136.96, 143.55, 150.31, 158.54,
163.49, 163.81; 31P NMR (109 MHz, CDCl3) δ -3.27, -3.06,
150.45, 151.26. MS (FAB+) calcd for C53H67N6O17P (M + H)
1121.4038, found 1121.4010.
C
44H49N4O15PS‚2H2O: C, 54.31; H, 5.49; N, 5.76. Found: C,
54.00; H, 5.38; N, 5.82.
Un pr otected 2′-O-Meth ylu r idylyl(3′-5′)2′-O-m eth yldeox-
yu r id in e P h osp h or oth ioa te Der iva tive 10a ,b Ha vin g a
Cyclic Str u ctu r e w ith a P r op ylen e Br id ge. Compound
9a ,b (83.7 mg, 89.3 µmol) was dissolved in 80% acetic acid (2
mL). After being kept at room temperature for 30 min, the
mixture was evaporated under reduced pressure. The residue
was partitioned between ether (10 mL) and water (10 mL).
The aqueous layer was collected, washed two times with ether,
and lyophilized to give 10a ,b (52 mg, 91%): Anal. Calcd for
C
23H31N4O13PS‚2H2O: C, 41.19; H, 5.26; N, 8.36. Found: C,
41.32; H, 5.02; N, 8.25. Reversed-phase HPLC of this mixture
gave 10a (365 A260, 21%) and 10b (396 A260, 22%). Umspc3Um
(fast isomer) 10a : reversed-phase HPLC 17.4 min (system 1b);
UV (H2O-CH3CN, 100:1, v/v) λmax 263 nm, λmin 232 nm; 1H
NMR (400 MHz, DMSO-d6) δ 1.58-1.88 (2H, m), 2.19-2.45
(2H, m), 3.44 (3H, s), 3.47 (3H, s), 3.57-3.65 (2H, m), 3.81 (1H,
d, J 2′,3′ ) 4.50 Hz), 3.98-4.04 (2H, m), 4.05-4.15 (4H, m),
4.24-4.36 (2H, m), 5.06 (1H, ddd, J 2′,3′ ) 4.85 Hz, J 3′,4′ ) 2.25
Typ ica l P r oced u r e for th e Syn th esis of Oligon u cle-
otid es In cor p or a tin g Um c3Um in th e Solid -P h a se Ap -
p r oa ch . The standard protocol described in the ABI manual
for ABI 381 synthesizer was used. For the incorporation of
Umpc3Um into the middle position of dodecaoligonucleotides,
the Umpc3Um amidite was used in a glass vessel with a filter.
A T-loaded CPG resin (1 µmol, 44 mmol/g, Glen Research, 500
Å) or U-loaded CPG resin (1 µmol, 19.7 µmol/g, Glen Research,
500 Å) was used. When the modified dimer is inserted, the
following manual manipulation procedure was used: (1)
detritylation with 3% TCA/CH2Cl2, (2) washing with pyridine,
(3) drying, CH3CN for 10 min, (4) condensation with the
amidite unit (0.1 M, 20 equiv) in the presence of 1H-tetrazole
(0.5 M,100 equiv)/CH3CN for 10 min, (5) washing with CH3-
Hz, J 3′,P ) 10.86 Hz), 5.21-5.39 (2H, m), 5.72 (1H, d, J 5,6
)
8.09 Hz), 5.76 (1H, s), 5.89 (1H, d, J 1′,2′ ) 6.85 Hz), 7.70 (1H,
s), 7.88 (1H, d, J 5,6 ) 8.14 Hz), 11.32 (1H, br), 11.44 (1H, br);
13C NMR (68 MHz, DMSO-d6) δ 20.74, 26.51, 26.63, 57.99,
60.61, 63.81, 65.71, 67.32, 76.25, 79.21, 80.52, 81.15, 81.30,
83.08, 83.97, 85.34, 87.59, 102.63, 109.17, 137.02, 140.20,
150.14, 150.68, 162.97, 163.56; 31P NMR (109 MHz, DMSO-
d6) δ 67.04. Umspc3Um (slow isomer) 10b: Reversed-phase

6524 J . Org. Chem., Vol. 65, No. 20, 2000
Sekine et al.
CN, (6) capping with Ac2O-Py (1:9, v/v) in the presence of 0.1
M DMAP for 1 min, (7) washing with pyridine, (8) oxidation
with 0.05 M I2/THF-Py-H2O for 1 min (7:2:1, v/v/v), (9)
washing with pyridine, (10) washing with CH2Cl2. After the
final condensation was finished, the CPG gel was removed
from the synthesizer and treated with 25% NH3 aq-Py (9:1,
v/v, 2 mL) for 20 h. The excess ammonia was removed under
reduced pressure, and the resin was suspended in water (1
mL). The mixture was evaporated under reduced pressure and
lyophilized. (When the synthesis of RNA oligomers was carried
out, the resin was treated with 1 M TBAF‚H2O in THF (1 mL)
for 16 h and thereafter desalting was performed by gel
filtration using Sephadex G-15. The eluate was lyophilized.)
The lyophilized material was separated by anion-exchange
HPLC using a FAX column followed by reversed-phase
HPLC: U4[Umpc3Um(fast)]U4 (13a ) (average yield 98%), 19.9
A260 (22%) (ꢀ 9.27 × 103); U4[Umpc3Um(slow)]U4 (13b) (aver-
age yield 98%), 16.5 A260 (18%) (ꢀ 9.27 × 103); T4[Umpc3Um-
(fast)]T4 (15a ) (average yield 99%), 44.2 A260 (58%) (ꢀ 7.65 ×
103); T4[Umpc3Um(slow)]T4 (15b) (average yield 99%), 18.9
A260 (25%) (ꢀ 7.65 × 103).
phodiesterase (1 unit/µL, 8 µL). The mixture was incubated
at 37 °C for 12 h and then calf intestinal alkaline phosphatase
(4 unit/mL, 5 µL) was added. The mixture was incubated at
37 °C for 6 h, and an aliquot was analyzed by reversed-phase
HPLC.
Str u ctu r al An alysis of Rp-Um pc3Um an d Sp-Um pc3Um
by Molecu la r Mech a n ics. MacroModel ver 6.0 was used in
a Silicon Graphics O2 workstation by using the AMBER* force
field. The PRCG program and GB/SA model were used for
energy minimization. Many conformers were generated using
5000 steps by Monte Carlo method and energy-minimized.
Ack n ow led gm en t. This work was partially sup-
ported by Toray Scientific Foundation and also sup-
ported by a Grant from “Research for the Future”
Program of the J apan Society for the Promotion of
Science (J SPS-RFTF97I00301) and a Grant-in-Aid for
Scientific Research from the Ministry of Education,
Science, Sports and Culture, J apan. We thank Dr. Sei-
ichiro Shigashiya for his measurement of FAB mass.
En zym a tic Assa y of Oligon u cleotid es In cor p or a tin g
Um p c3Um . A modified oligonucleotide (1.0 A260) was dissolved
in Tris-HCl buffer (80 µL, pH 8.0) and snake venom phos-
J O0002588