Deoxyadenosine and thymidine bases held proximal and distal by means of a covalently-linked dimensional analogue of dA·dT: Intramolecular vs intermolecular hydrogen bonding

Article abstract of DOI:10.1021/ja9618255

Deoxyadenylic acid and deoxythymidilic acid have been attached to a covalently-linked cross section, dAχdT, on the same side (proximal) and on opposite sides (distal) by synthetic sequences involving various combined protection/deprotection steps. The str

Full text of DOI:10.1021/ja9618255

10744
J. Am. Chem. Soc. 1996, 118, 10744-10751
Deoxyadenosine and Thymidine Bases Held Proximal and
Distal by Means of a Covalently-Linked Dimensional Analogue
of dA‚dT: Intramolecular vs Intermolecular Hydrogen Bonding1
Balkrishen Bhat,† Nelson J. Leonard,*,‡ Howard Robinson,§ and
Andrew H.-J. Wang*,§
Contribution from the Departments of Chemistry and Cell & Structural Biology,
UniVersity of Illinois at Urbana-Champaign, Urbana, Illinois 61801
ReceiVed May 30, 1996X
Abstract: Deoxyadenylic acid and deoxythymidilic acid have been attached to a covalently-linked cross section,
dAYdT, on the same side (proximal) and on opposite sides (distal) by synthetic sequences involving various
∧
combined protection/deprotection steps. The structures in aqueous solution buffered at pH 7.0 have been examined
by 2D-NOESY NMR spectroscopy. The covalently-linked cross section provides a template that stabilizes dA‚dT
base pairing in the proximal isomer at 2 °C. In the distal isomer, it contributes to favorable conformations for
intermolecular association.
Introduction
deoxyadenosine and deoxythymidine units held in place by the
template dAYdT that is a dimensional mimic of a fixed dA‚dT
∧
The forces that stabilize associations between nucleic acid
bases, namely, hydrogen bonding and base stacking, have been
well studied,2 and various cross-linking methods have been
utilized to examine structural interactions between bases or
strands of bases.3-14 Hydrogen bonding between imides and
adenine or adenosine has been examined and identified in
structures consisting of aromatic stacking surfaces rendered
water soluble.15-17 We have now provided water soluble
constructs that allow us to determine interactions between
base pair.18 The proximal isomer (1) locates the dA and
† Department of Chemistry.
‡ Present address: Division of Chemistry and Chemical Engineering,
California Institute of Technology, Pasadena, CA 91125.
§ Department of Cell & Structural Biology.
dT units so that they can bond intramolecularly, where-
as the distal isomer (1′) fixes these units where intramolec-
ular bonding is unfavorable but where they are free to asso-
ciate intermolecularly. In effect, the distal 1′ serves as a con-
trol for the intramolecularity of any observed dA-dT interac-
tion in 1.
X Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) Abbreviations used: A*,N6-(allyloxycarbonyl)-2′-deoxyadenosine;
dAYdT, 3,9-bis(2′-deoxy-â-D-ribofuranosyl)-11-methyl-3H-pyrimido[1′′,
∧
6′′:1′,2′]imidazo[4′,5′:4,5]imidazo[2,1-i]purin-8(9H)-one; AOC, allyloxy-
carbonyl; DMAP, 4-(dimethylamino)pyridine; DMT, 4,4′-dimethoxytrityl;
Pd2(dba)3‚CHCl3, tris(dibenzylideneacetone)dipalladium(0) chloroform
adduct; TBAF, tetrabutylammonium fluoride; TBDMS, tert-butyldimeth-
ylsilyl; 2D-NOESY, two-dimensional nuclear Overhauser effect spec-
troscopy.
(2) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag:
New York, 1983; pp 116-140.
(3) Manoharan, M.; Guinosso, C. J.; Cook, P. D. Tetrahedron Lett. 1991,
32, 7171-7174.
(4) Wang, H.; Osborne, S. E.; Zuiderweg, E. R. P.; Glick, G. D. J. Am.
Chem. Soc. 1994, 116, 5021-5022.
(5) Cowart, M.; Gibson, K. J.; Allen, D. J.; Benkovic, S. J. Biochemistry
1989, 28, 1975-1983.
(6) Cowart, M.; Benkovic, S. J. Biochemistry 1991, 30, 788-796.
(7) Ferentz, A. E.; Keating, T. A.; Verdine, G. L. J. Am. Chem. Soc.
1993, 115, 9006-9014.
(8) Glick, G. D.; Osborne, S. E.; Knitt, D. S.; Marino, J. P., Jr. J. Am.
Chem. Soc. 1992, 114, 5447-5448.
(9) Richardson, P. L.; Schepartz, A. J. Am. Chem. Soc. 1991, 113, 5109-
5111.
(10) Cload, S. T.; Schepartz, A. J. Am. Chem. Soc. 1991, 113, 6324-
6326.
(11) Horne, D. A.; Dervan, P. B. J. Am. Chem. Soc. 1990, 112, 2435-
2437.
(12) Ono, A.; Chen, C.-N.; Kan, L.-S. Biochemistry 1991, 20, 9914-
9921.
(13) Froehler, B. C.; Terhorst, T.; Shaw, J.-P.; McCurdy, S. N.
Biochemistry 1992, 31, 1603-1609.
(14) Prakash, G.; Kool, E. T. J. Am. Chem. Soc. 1992, 114, 3523-3257.
(15) Kato, Y.; Conn, M. M.; Rebek, J., Jr. J. Am. Chem. Soc. 1994, 116,
3279-3284.
(16) Rotello, V. M.; Viani, E. A.; Deslongchamps, G.; Murray, B. A.;
Rebek, J., Jr. J. Am. Chem. Soc. 1993, 115, 797-798.
(17) Kato, Y.; Conn, M. M.; Rebek, J., Jr. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 1208-1212.
(18) Bhat, B.; Leonard, N. J. J. Am. Chem. Soc. 1992, 114, 7407-7410.
S0002-7863(96)01825-2 CCC: $12.00 © 1996 American Chemical Society

CoValently-Linked Dimensional Analogue of dA‚dT
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10745
Results and Discussion
Synthesis. The starting material for the synthesis of the
proximal isomer (1) was 3-(3′,5′-di-O-acetyl-2′-deoxy-â-D-
ribofuranosyl)-11-methyl-9-[3′,5′-O-(1,1,3,3-tetraisopropyldisi-
loxanediyl)-2′-deoxy-â-D-ribofuranosyl]-3H-pyrimido[1′′,6′′:
1′,2′]imidazo[4′,5′:4,5]imidazo[2,1-i]purin-8(9H)-one (2),18
selectively protected on each side of dAYdT. The disiloxane
∧
protection was removed by stirring in tetrahyrofuran with
tetrabutylammonium fluoride at 25 °C for 1.5 h to give 3,
protected on the deoxyadenosine side by O-acetyl groups.
Compound 3, as the acetonitrile solvate monohydrate, was
determined by X-ray analysis to be very close to coplanar for
the atoms constituting the central pentacyclic N-aromatic system
(root-mean-square deviation 0.03 Å).19 The further steps in the
synthesis of the bis-triethylammonium salt (17) of the desired
proximal isomer 1 are outlined in Scheme 1 (see the Experi-
mental Section). We found it advantageous to use a combina-
tion of the conventional protection/deprotection reagents em-
ployed in the synthesis of oligodeoxyribonucleotides in solution,
along with allyl/allyloxycarbonyl protection on phosphite and
phosphate and on the 6-NH2 of deoxyadenosine, respectively.20
The mild Hayakawa-Noyori methodology, including hydro-
genolytic cleavage of these protecting groups,20 was conserving
of structures and quantities throughout. The average yield at
each step was 84%, and the overall yield from 3 to 17 was
10%. The structures of the intermediates and product (17), once
they were determined to be homogeneous, were established by
the sequential synthetic procedure, by their low- and/or high-
resolution mass spectra, and by 31P NMR, as deemed necessary.
The final purification of the proximal example (1) for NMR
study was achieved by reverse phase HPLC using aqueous
sodium phosphate and a methanol gradient. This was followed
by desalting on the same HPLC column. The appropriate
fractions were pooled, and the sample was dried under reduced
pressure and kept frozen.
Figure 1. Chemical structures of proximal 1 and distal 1′. The total
structures are given as viewed from the major groove.
The synthesis of the distal isomer (1′) proceeded from the
common intermediate in both syntheses (11 in Scheme 1 ) 2′
in Scheme 2). The yield from 2′ to product 10′ was 21%,
comprised of an average yield of 78% for each step. The
structures of the intermediates, once determined to be homo-
geneous, were established by the sequential synthetic procedure
and by their mass spectra. The final product (1′), first obtained
as the disodium salt (10′), was purified, desalted, and preserved
as for the proximal isomer. The high-resolution mass spectrum
confirmed the structure.
This extended exercise in synthesis has further purpose.
Thus, the proximal isomer 1 is a potential substrate for blunt-
end ligation. The distal isomer 1′, suitably derivatized, is a
possible candidate for self assembly.21,22 If the latter intention
is realized, a stiff, rodlike oligodeoxynucleotide analogue of
poly(dA)‚poly(dT) would be created, one that offers little
opportunity for bending and none for slippage. In the present
study, the surface of dAYdT provides an undissociable equiv-
∧
alent of dA‚dT. The proximal model 1, which prevents any
fraying beyond the terminal dA and dT, thus serves as a limiting
case. Fraying of a duplex, wherein the second base pair, dA‚-
dT, is dissociable, has been thoroughly examined in the case
of [d(TAGCGCTA)]2.23
NMR. The complete formulas of the proximal 1 and distal
1′ isomers are shown in Figure 1, depicted as viewed from the
(19) Bhat, B.; Wilson, S. R.; Leonard, N. J. Acta Crystallogr. 1993, C49,
921-924.
Figure 2. Temperature-dependent 1D-NMR spectra of the proximal-
type (1) and the distal-type (1′) oligonucleotides. The T1H3 imino
proton of 1 is clearly detected at 12.20 ppm at 2 °C, whereas that of 1′
is seen as a broad resonance centered at 11.2 ppm.
(20) Hayakawa, Y.; Wakabayashi, S.; Kato, H.; Noyori, R. J. Am. Chem.
Soc. 1990, 112, 1691-1696.
(21) Sievers, D.; von Kiedrowski, G. Nature 1994, 369, 221-224.
(22) Li, T.; Nicolaou, K. C. Nature 1994, 369, 218-221.

10746 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Bhat et al.
Figure 3. The observed nonexchangeable proton 2D-NOESY spectrum showing the expanded aromatic to H1′ and H2′/H2′′ region of the proximal-
type 1 at 2 °C. NOE’s a, b, and c are from A4H8 to A3H1′, A4H8 to A3H2′, and A4H8 to A3H2′′, respectively.
Figure 4. The NOE-restrained refined structure of the proximal-type 1. The upper stereoscopic view (A) is from the major-groove direction. The
lower stereoscopic view (B) is from the minor-groove direction.
major groove of a potential helical structure. The terminal
deoxythymidine and deoxyadenosine are designated as T1 and
A4, respectively. The pyrimidine and purine moieties in the
proximal 1 have been confirmed by using the nonexchangeable
proton 2D-NOESY and TOCSY spectra (data not shown)
recorded at 2 °C. The T1H3 imino proton is clearly detected
at 12.20 ppm, which indicates that T1 forms a Watson-Crick
fused base, dAYdT, are designated as T2 and A3, and the
∧
hydrogens therein are numbered as they would be for the
separate components. This system of lettering and numbering
makes the proton interactions more readily traceable in the
discussion of the nuclear magnetic resonance. As background,
the proton resonances in the deoxyadenosine portion of dA
base pair with A4 and is stacked on the fused base dAYdT,
∧
forming a miniduplex at 2 °C. Intramolecular association of
the imino proton persists at least up to 20 °C, which is
remarkable for such a small miniduplex that is analogous to a
pair of dinucleoside monophosphates. On the basis of the
sharpened signal at 2 °C for T1H3, it is possible to state
qualitatively that equilibration with open forms is not occurring
rapidly. The shape of the signals for the temperature range
2-20 °C is the combined result of the hydrogen-bonded a
nonbonded equilibrium and the rates of unbonding and ex-
change.
Y
dT have been previously assigned as 9.73 for “H2” (actually
∧
H5 in the full pentacyclic numbering system) and 8.62 for “H8”
(actually H2). The aromatic proton resonance in the thymidine
portion has been assigned as 7.74 for “H6” (actually H10),
also in DMSO as the solvent.18 These, in turn, were checked
by comparison with assignments in the dAYdU series that
∧
were established by long-range heteronuclear 1H/13C cor-
relations.24
Of the aromatic protons, A3H2 shows the greatest chemical
shift, 8.95 ppm, at 2 °C in the aqueous buffer. Base stacking
at the lowest temperature is indicated also by this proton
resonance. As the temperature is raised to 40 °C, the A3H2
signal progresses downfield to 9.28 ppm. The deshielding with
rising temperature is consistent with the loss of the ring current
effect as the T1-A4 base pair becomes unstacked, i.e., with
greater participation of open conformations. There appears to
be a crossover of T2H6 and T1H6 from coincidence at about
In Figure 2 are compared the temperature-dependent be-
haviors of the imino proton and aromatic proton resonances
from their 1D-NMR spectra. The resonance assignments for
(23) Nonin, S.; Leroy, J.-L.; Gue´ron, M. Biochemistry 1995, 34, 10652-
10659.
(24) Devadas, B.; Leonard, N. J. J. Am. Chem. Soc. 1990, 112, 3125-
3135.

CoValently-Linked Dimensional Analogue of dA‚dT
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10747
explosions during distillation.32 Accordingly, we feel that it is important
to describe the synthesis of allyl phosphorodichloridite in detail,
applying the precautions indicated. At eVery stage, the preparation
and purification were carried out in a well-Ventilated hood behind a
protectiVe shield. In a 100-mL three-necked, round-bottomed flask
equipped with an immersion thermometer, dropping funnel, and calcium
chloride drying tube was placed 50 g (0.36 mol) of 99.99+%
phosphorous trichloride under an argon atmosphere, and the mixture
was cooled to approximately -55 °C in dry ice/isopropyl alcohol. The
dropping funnel was equipped with a rubber septum through which
argon could be purged. The PCl3 was cooled for 15 min to attain a
-55 °C inside temperature. Allyl alcohol (21.11 g, 0.36 mol, 99+%)
was added slowly (2 h) through the dropping funnel, during which
time the temperature of the reaction mixture was allowed to rise to
-30 °C. It was then allowed to rise to 20 °C, and stirring was continued
overnight (20 h) under a slow stream of argon. The volume of the
reaction mixture decreased by about 10%. The thermometer and drying
tube were replaced with stoppers, and the dropping funnel was replaced
by a fractionating column. The reaction mixture was heated by an oil
bath to a final temperature of 80 °C during fractionation. The first
fraction of distillate (4.7 g) was collected at 20 °C and 50 mmHg
(discarded), the second (2.5 g) was collected at 35 °C (discarded), and
a third fraction (7.0 g) was collected at 40 °C and 15 mmHg, and a
final fraction (15 g) at 51-54 °C, giving a total yield of 39%.
Distillation was discontinued while the reaction mixture still remained
colorless. The residual oil was cooled to -80 °C, and to it was added
30 mL of CHCl3-1-butanol (1:1). The temperature in the flask was
allowed to rise to 20 °C, and the flask was left for about 15 h, during
which time a bright yellow precipitate settled out. Final (safe)
decomposition of the contents of the flask was effected by cooling to
-70 °C and addition of 10 mL of 2 M NaOH.
7.53 ppm at 2 °C to a difference in chemical shift of ca. 0.16
ppm at 40 °C.
For the distal isomer 1′, broad resonance was observed,
centered at 11.2 ppm, at 2 °C, which indicates that T1H3 may
be loosely involved in hydrogen-bonding interaction intermo-
lecularly with acceptor molecules (Figure 2). While the A3H2
is at higher field at 2 °C than at 40 °C, the signal at 2 °C is
broadened, suggesting a family of partially stacked conforma-
tions. Other nonexchangeable resonances also exhibit broad
signals that sharpen as the temperature is raised. The differences
may be the result of a large change in the spin-spin relaxation
times, and the broadening at low temperature is consistent with
the existence of different degrees of aggregation.
Conclusive evidence for the duplex structure of proximal 1
is found in the observed internucleotide NOE crosspeaks
between A4H8 and (a) A3H1′ (Figure 3), (b) A3H2′, and (c)
A3H2′′, and between A3H1′ and A4H5′/5′′ (data not shown).
These NOE crosspeaks are consistent with a right-handed twist
to the helix.
The 3D solution structure of proximal 1 was obtained by a
combined SPEDREF25 and NOE-constrained molecular dynam-
ics refinement.26 In the refined duplex, the T1 base is stacked
significantly with the fused base, whereas only the N6 amino
group of A4 is stacked above the fused base (Figure 4). The
sugar pucker in the refined structure is of the S-type for the T1
and A3 nucleotides, and of the N-type for the T2 and A4
nucleotides. The measured J1′,2′/J1′,2′′ coupling constants of the
four sugar rings from the PE-COSY spectrum are 7.4/6.5, 7.1/
9.5, 8.0/5.5, and 6.0/8.4 Hz, respectively, for T1, T2, A3, and
A4 units. Those values are consistent with the sugar puckers
seen in the refined structure.
(Allyloxy)bis(diisopropylamino)phosphine, CH2dCHCH2OP-
[N(i-Pr)2]2, was prepared from CH2dCHCH2OPCl2,20 bp 115-116 °C/2
mmHg, yield 71%, purity >99% (31P NMR).
Compounds 2 and 3 were synthesized as described previously.18
4. After step b (see legend for Scheme 1), sequential manipulation
A (see General, above), and the use of 5% CH3OH in CHCl3 containing
a trace of Et3N in chromatography, appropriate fractions were pooled
and concentrated. The residual oil was taken up in anhydrous CH2-
Cl2, and addition of hexane furnished a colorless precipitate. Solvent
was removed under reduced pressure, and the colorless powder was
dried under high vacuum to give 4, C47H46N8O11, in 92% yield: low-
resolution FAB MS m/e 899.2 (MH+).
5. After step c and method A, the use of 2-4% CH3OH in CHCl3
in chromatography, pooling of fractions, and concentration, 5,
C53H60N8O11Si, was obtained as a colorless foam: yield 93%; low-
resolution FAB MS m/e 1013.6 (MH+).
6. After step d, the solution was poured into saturated NaHCO3
solution and extracted with ethyl acetate. The ethyl acetate was dried
over anhydrous Na2SO4 and filtered, and the solvent was removed under
reduced pressure. The residual oil was purified by chromatography
on silica gel using 3-7% CH3OH in CHCl3. Concentration of
appropriate fractions was followed by precipitation of the product with
CH2Cl2-hexane (1:9). Removal of solvent under vacuum gave
compound 6, C32H42N8O9Si, as a colorless solid: yield 82%; low-
resolution FAB MS m/e 711.2 (MH+).
5′-O-(tert-Butyldimethylsilyl)thymidine 3′-O-(Allyl N,N-diisopro-
pylphosphoramidite), C25H46N3O6PSi, 7. To a solution of 5′-O-(tert-
butyldimethylsilyl)thymidine (330 mg, 0.92 mmol) in anhydrous
CH3CN (6 mL, dried over CaH2 prior to use) was added diisopropyl-
amine (0.142 mL, 1.02 mmol), 1H-tetrazole (71 mg, 1.02 mmol), and
Our structural study shows that the fused base pair dA
Y
dT
∧
provides an excellent template to stabilize substantially the
formation of adjacent dA‚dT base pairs. When dA and dT are
attached on opposite sides of the template, intermolecular
association is detected and the prospect of self-assembly is thus
recognized, possibly through appropriate 3′-O-phosphoramidite
substitution.
Experimental Section
General. The synthesis section can be abbreviated because the
reaction conditions for the synthesis of all intermediates are given in
the legends to Schemes 1 and 2. The intermediates, which are
interrelated sequentially, were characterized by their chemical conver-
sions, by comparative TLC, and, where we considered it necessary, by
mass spectrometry and NMR. Many of the intermediates in the
sequence were isolated by method A, as follows: The reaction mixture
was poured into cold water and extracted with ethyl acetate. The
combined organic layers were dried over anhydrous Na2SO4, filtered,
concentrated, and purified by chromatography over silica gel. Where
there were variations from this methodology for individual intermedi-
ates, the separate conditions are mentioned. The quantities used ranged
from hundreds of milligrams down to tens of milligrams. More specific
details of each step can be made available by communication with the
authors, but the experimental procedures will be evident to anyone
versed in blocking/deblocking sequences.20,27
Compound 1
(28) Martin, D. R.; Pizzolato, P. J. J. Am. Chem. Soc. 1950, 72, 4584-
4586.
Allyl Phosphorodichloridite, CH2dCHCH2OPCl2. This reagent
was prepared according to the method described for methoxyphospho-
rodichloridite.28 Although the reagent has been prepared and used by
the Noyori group,20,29-31 there have been reports of low yields and of
(29) Hayakawa, Y.; Uchiyama, M.; Kato, H.; Noyori, R. Tetrahedron
Lett. 1985, 26, 6505-6508.
(30) Hayakawa, Y.; Kato, H.; Uchiyama, M.; Kajino, H.; Noyori, R. J.
Org. Chem. 1986, 51, 2400-2402.
(31) Hayakawa, Y.; Kato, H.; Nobori, T.; Noyori, R.; Imai, J. J. Nucl.
Acids Symp. Series 1986, No. 17, 97-100.
(25) Robinson, H.; Wang, A. H.-J. Biochemistry 1992, 31, 3524-3533.
(26) Bru¨nger, A. X-PLOR (version 3.1) (1993), The Howard Hughes
Medical Institute and Yale University, New Haven, CT.
(27) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22, 1859-
1862.
(32) Private communication from Profesor Yoshihiro Hayakawa based
upon experience with methyl phosphorodichloridite (methoxydichlorophos-
phine). See also: Kamber, M.; Just, G. Can. J. Chem. 1985, 63, 823-827,
wherein the reagent was obtained in 17% yield.

10748 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Bhat et al.
Scheme 1a
a (a) TBAF on silica gel, THF, 1.5 h, 25 °C; (b) DMTCl, Py, 20 h, 25 °C; (c) TBDMSCl, imidazole, DMF, 4 h, 25 °C; (d) Cl2CHCOOH, CH2Cl2,
3 min, 25 °C; (e) 7, 1H-tetrazole, CH3CN, 2 h, 25 °C; (f) excess t-BuOOH, CH2Cl2, 3.5 h, 0 °C; (g) t-BuNH2, CH3OH, 15 min, 0 °C, then 4 h, 25
°C; (h) DMTCl, Py, DMAP, 2.5 h, 25 °C; (i) CH2dCHCH2OP[N(i-Pr)2]2, HN(i-Pr)2, 1H-tetrazole, CH3CN, 2.5 h, 25 °C; (j) 3′-O-TBDMS-A*,
1H-tetrazole, CH3CN, 1.5 h, 25 °C; (k) t-BuOOH, CH2Cl2, 1 h, 0-25 °C; (l) Cl2CHCOOH, CH2Cl2, 3 min, 25 °C; (m) TBAF on silica gel, THF,
14 h, 25 °C; (n) (C6H5)3P, THF, C4H9NH2, HCOOH, Pd2(dba)3·CHCl3, 5 h, 25 °C. Note: In terms of double-strand helicity, these formulas are
depicted as viewed from the major groove.

CoValently-Linked Dimensional Analogue of dA‚dT
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10749
(allyloxy)bis(diisopropylamino)phosphine (0.43 mL, 1.4 mmol). The
reaction mixture was stirred at 20 °C for 1.5 h, taken up in ethyl acetate
(30 mL), washed with brine (2 × 20 mL), dried over anhydrous Na2-
SO4, filtered, concentrated, and purified by chromatography on silica
gel (5 × 1.5 in. column) using hexane-ethyl acetate (1:1) as eluent.
Removal of solvent under reduced pressure and eventually under high
vacuum yielded 7 (95%) as a colorless gum, the purity of which was
checked by 31P NMR, giving a chemical shift value identical with that
described for the 5′-O-dimethoxytrityl derivative.
8. After step e and method A, with 5% CH3OH in CHCl3 as the
eluting chromatographic solvent, appropriate fractions were combined
and concentrated, and CH2Cl2-hexane (1:9) was added to effect
precipitation. Removal of solvent and drying for 3 h under high vacuum
yielded 8, C51H73N10O15PSi2, as a light yellow solid: yield 99%;
homogeneous by a single spot on TLC.
14 was changed, in which case the bisphosphate was obtained after
purification and desalting (see below).
17 (Preferred). After steps m, n, and l, ice water was added to the
reaction flask, and dichloromethane was removed under reduced
pressure. The aqueous layer was neutralized with 1 M NaOH to pH 7
and purified on a preequilibrated (0.01 M TEAB) column using 0.01-
0.7 M TEAB as linear gradient. Appropriate fractions were com-
bined and lyophilized. The residue was dissolved in deionized
water, transferred to a smaller flask, and relyophilized to give 17,
C56H79N17O19P2, as a colorless fluffy solid: yield 50% based on 14;
low-resolution FAB MS (negative ion) m/e 1228.2 (M - Et3N+H - 1
) 1228.6).
The deblocking sequence, i.e., removal of TBDMS, allyl, AOC, and
DMT groups, and subsequent purification of 1 are as described in
parallel and in greater detail for 1′ (see below).
9. After step f, CH3OH was added before concentration. Purification
was effected by chromatography on silica gel, with 5% CH3OH in
CHCl3, and the usual workup to furnish the desired intermediate 9,
C51H73N10O16PSi2: yield 79%; homogeneous single spot on TLC.
10. After step g, CH3OH was removed. Purification was effected
on a short silica gel column using 10-15% CH3OH in CHCl3 as eluent.
Appropriate fractions were combined and concentrated, followed by
precipitation of the residual oil with CH2Cl2-hexane (1:20). Removal
of the solvent furnished the deacetylated product 10, C47H69N10O14-
PSi2, as a colorless solid: yield 77%; homogeneous by a single spot on
TLC.
Compound 1′
Compound 2′, C68H87N10O16Si2P, was prepared as described above
(11).
After step a (see legend for Scheme 2), the reaction mixture was
quenched with water and extracted with ethyl acetate. The extract was
dried over anhydrous Na2SO4, filtered, and purified by chromatography
on silica gel, using 5% CH3OH in CHCl3 with a trace of Et3N as eluent.
Appropriate fractions were pooled and concentrated to give 3′,
C74H101N10O16PSi3, as a colorless foam: yield 92%; homogeneous by a
single spot on TLC.
11. After step h, addition of CH3OH, and concentration under
reduced pressure, the residual oil was co-evaporated with anhydrous
toluene. This residue was purified by flash chromatography on silica
gel using 5-7% CH3OH in CHCl3 followed by 15% CH3OH in CHCl3.
Appropriate fractions were combined and concentrated to give inter-
mediate 11, C68H87N10O16PSi2, as a colorless powder: yield 79%;
homogeneous by a single spot on TLC.
12. After step i and method A, including purification on silica gel
using hexane-ethyl acetate-CH3OH (45:45:10) for elution, appropriate
fractions were combined and concentrated to give intermediate 12,
C77H106N11O17P2Si2, as a colorless foam: yield 82%; homogeneous by
single spot on TLC.
After step b, the reaction mixture was poured into saturated sodium
bicarbonate solution and extracted with ethyl acetate. The organic layer
was dried over anhydrous Na2SO4, filtered, and concentrated. The
residue was purified by flash chromatography on silica gel, using 7%
CH3OH in CHCl3, and concentration of the appropriate fractions. Solid
4′, C53H83N10O14PSi3, was obtained by solution of the residue in a
minimum amount of CH2Cl2 and addition of hexane, followed by
solvent removal under reduced pressure: yield 84%; homogeneous by
a single spot on TLC.
Compound 5′, N6-(Allyloxycarbonyl)-5′-O-(4,4′-dimethoxytrityl)-
2′-deoxyadenosine 3′-(allyl N,N-diisopropylphosphoramidite) was
synthesized according to the literature procedure.20
13. After step j and method A, 5-7% CH3OH in CHCl3 eluent
fractions were concentrated to give 13, C91H121N15O22P2Si3, as a
colorless solid: yield 83%; low-resolution FAB MS m/e 1922.5 (MH+),
1938.1 (MH + O)+.
14. After step k, CH3OH was added and the combined solvent was
removed in vacuo. The residue was dissolved in 10% CH3OH in CHCl3
and chromatographed on a short silica gel column. Appropriate
fractions were combined and concentrated, and the residual oil was
precipitated with CH2Cl2-hexane (1:20). Removal of the solvent in
vacuo furnished the fully protected 14, C91H121N15O23P2Si3, as a
colorless solid: yield 94%; low-resolution FAB MS m/e 1938.9 (MH+).
15. The first of the protecting groups was removed by step l. The
reaction mixture was poured into saturated NaHCO3 solution and
extracted with ethyl acetate. The organic layer was dried over
anhydrous Na2SO4, filtered, and concentrated to give 15, C70H103N15O21P2-
Si3 (yield approximately 85%), which was used directly under condi-
tions in step m without further purification.
16. After step m, the solvent was removed under reduced pressure,
and the residue was triturated three times with methanol and filtered
through cotton wool. The solvent was removed from the filtrate, and
the residue was purified on a short silica gel column, with 20% CH3-
OH in CHCl3 as the eluent. Appropriate fractions were combined and
concentrated. Addition of CH3CN gave a colorless precipitate. Solvent
was removed under reduced pressure to give 16, C52H61N15O21P2: yield
59%; homogeneous by a single spot on TLC; low-resolution FAB MS
m/e 1294.4 (MH+).
Compound 6′, C91H121N15O22P2Si3, was obtained by reaction of 4′
with 5′ (10% molar excess) in 1H-tetrazole and anhydrous CH3CN after
stirring for 3 h at 25 °C. The reaction mixture was diluted with ethyl
acetate and washed with brine, and the organic layer was dried over
anhydrous Na2SO4. After filtration and flash chromatography on silica
gel, using 5% CH3OH in CHCl3 plus a trace of Et3N as eluent, the
appropriate fractions were concentrated to give 6′: yield 76%; homo-
geneous by a single spot on TLC. Approximately 12% of unreacted 4
was recovered.
After step c, the reaction mixture was quenched with CH3OH and
the solution was concentrated. The residue was purified on a silica
gel column, elution with 2-7% CH3OH in CHCl3. Concentration of
the collected fractions gave 7′, C91H121N15O23P2Si3, as a light yellow
foam: yield 86%; homogeneous by a single spot on TLC; low-resolution
FAB MS m/e 1938.7 (MH+).
After step d, the silica gel was removed by filtration and washed
repeatedly with THF and CH3OH. Purification of the concentrated
filtrate and washings was effected by chromatography on a silica gel
column, elution with CHCl3 followed by 10-20% CH3OH in CHCl3.
Appropriate fractions were combined and concentrated to give a light
yellow gum. Trituration with 1:1 ether-hexane caused the separation
of a solid. The supernatant was decanted, and the solid was washed
with hexane and dried under high vacuum to give 8′, C73H79N15O23P2,
as an off-white solid: yield 79%; homogeneous by a single spot on
TLC; low-resolution FAB MS m/e 1596.7 (MH+).
After step e, ethyl acetate was added and solvent was removed under
vacuum. The residue was azeotroped three times with anhydrous
toluene and was finally dried under vacuum for 3 h to give intermediate
9′, C63H67N15O21P2, which was used directly in the removal of the DMT
protecting group.
17. After step n, the residue was taken up in ethyl acetate which
was extracted with water several times. The aqueous layer was
rotoevaporated for 10 min to remove any dissolved ethyl acetate. It
was purified on a Sephadex A-25 column preequilibrated with 0.01 M
TEAB buffer using 0.01-0.7 M linear gradient of TEAB. This method
gave 17 in extremely low yield. Therefore, an alternative method was
used to prepare compound 17. The sequence of reaction steps after
After step f, acetic acid was added and the mixture was stirred for
2 h. Solvent was removed in vacuo. The residue was dissolved, and
the pH was quickly adjusted to 7 by addition of 1 M NaOH. The

10750 J. Am. Chem. Soc., Vol. 118, No. 44, 1996
Bhat et al.
Scheme 2a
a (a) TBDMSCl, imidazole, anhydrous DMF, 3.5 h, 25 °C; (b) Cl2CHCOOH, CH2Cl2, 3 min, 25 °C; (c) excess t-BuOOH, CH2Cl2, 25 h, 0 °C;
(d) TBAF on silica gel, THF, 2.5 h, 25 °C; (e) (C6H5)3P, THF, C4H9NH2, HCOOH, Pd2(dba)3‚CHCl3, 4 h, 25 °C; (f) Cl2CHCOOH, CH2Cl2, 5 min,
25 °C. Note: In terms of double-strand helicity, these formulas are depicted as viewed from the major groove.
aqueous layer was extracted with ethyl acetate. The combined organic
extracts were then extracted with water. The aqueous layers were
combined and concentrated to dryness. The solid that separated upon
treatment with 90% EtOH was collected, washed with 90% EtOH, and
quickly transferred to a vial and dried under high vacuum at 25 °C:
10′, C42H47N15O19P2Na2, yield 53% based on 8′.
Purification and desalting for the NMR study of 1′ was conducted
as follows. (This was also done for compound 1.) We used an ODS
18 Altex reverse-phase HPLC column and a gradient of 20-70% B in
20 min (A ) 50 mM sodium phosphate at pH 5.6; B ) 95% MeOH),
kept 70% B constant for 15 min, and re-equilibrated the column to
20% B in 10 min. The flow rate for preparative separations was 2
mL/min with the UV detector set at 256 nm. The major fractions,
17-20 min, were collected, pooled, lyophilized, and desalted on the
same reverse phase column. The column was equilibrated with
deionized water (20 column volumes). Samples were dissolved in water
and injected into the column. The column was washed with water for
10-15 min at 2 mL/min, followed by a gradient of 0-100% B in 5
min. The product eluted with 100% B after 5-6 min (A ) H2O, B )
95% MeOH). Appropriate fractions were pooled and lyophilized to a
solid which was transferred to a small sintered funnel using ether as a
solvent. It was then placed in an Eppendorf tube, dried overnight under
reduced pressure, and stored in the freezer until use: 1′, C42H49N15O19P2:
UV λmax (H2O) 328 nm, 260, 236; low-resolution FAB MS m/e 1130.6
(MH+); high-resolution FAB MS m/e 1130.2910 (C42H50N15O19P2
requires 1130.2882 amu).
NMR Spectroscopy. Solutions of DNA oligomer were prepared
as described earlier.25 Lyophilized powder (2.0 mg) was dissolved in
0.55 mL of H2O containing 20 mM phosphate buffer at pH 7.0, resulting
in a 3.0 mM duplex solution. NMR spectra were collected on a Varian
VXR500 500 MHz spectrometer, and the data were processed with
FELIX v1.1 (Hare Research, Woodinville, WA). The nonexchangeable
2D NOE spectra were collected at 2 °C at a mixing time of 0.05 s and
a total recycle delay of 6.41 seconds where the average T1 relaxation

CoValently-Linked Dimensional Analogue of dA‚dT
Table 1. Chemical Shifts (ppm) for Proximal 1 at 2 °C
J. Am. Chem. Soc., Vol. 118, No. 44, 1996 10751
HMe/2
H6/8
H1′
H2′
H2′′
H3′
H4′
H5′
H5′′
H1/3
H6(a)a
H6(b)a
T1
T2
1.64
2.12
7.53
7.51
5.84
6.58
2.27
2.58
2.59
2.54
4.71
4.74
4.13
4.17
3.78
4.17
3.78
4.25
12.2
-
A3
A4
8.95
7.70
8.44
8.26
6.18
6.34
2.64
2.79
2.87
2.55
4.93
4.79
4.33
4.31
3.81
4.21
3.81
4.21
-
na
-
na
a H6(a) are base-pair hydrogen-bonded amino protons, H6(b) are not.
was 1.8 s. The data were collected by the States/TPPI technique33 with
512 t1 increments and 2048 t2 complex points each the average of 24
transients. TOCSY (total correlated spectroscopy) spectra were used
together with the NOE spectra to derive the assignment using the
standard sequential assignment procedure. The chemical shifts of all
assigned resonances are listed in Table 1.
the R factor in the range of ∼21%. The refined structures based on
the two starting models have a root mean square deviation of 1.1 Å
between them, suggesting a reasonable convergence of the refinement
process.
Exchangeable proton 1D spectra at 2 °C in 90% H2O were collected
with the 1 not 1 pulse sequence as the read pulse.35 The excitation
offset was set to one-quarter of the spectral bandwidth, which was set
to 12 000 Hz. The PE-COSY spectrum was collected following the
procedure of Bax and Lerner.36
A starting model was built using the atomic coordinates of the crystal
structure of dA
Y
dT19 on which the remaining nucleotides were added
∧
with a B-DNA conformation using program MIDAS.34 The duplex
model was energy-minimized by conjugate gradient minimization using
X-PLOR.26 X-PLOR’s all atom force field for DNA was used with
explicit hydrogen bond potentials. Refinement of the starting model
was carried out by the SPEDREF procedures.25 Minimization of the
residual errors was performed by conjugate gradient minimization with
the NOE constraints within the program X-PLOR25 employing 40 cycles
of SPEDREF refinement. The isotropic correlation time (τc) for each
refinement was empirically determined to be 2 ns. Another starting
model was built using canonical A-DNA and was similarly refined. A
family of converged models was obtained from the refinement with
Acknowledgment. This work was supported by an NIH
grant to A.H.-J.W. (GM41612). The Varian VXR500 NMR
spectrometer at UIUC was supported in part from the NIH
shared instrumentation grant 1S10RR06243. We thank Dr.
Neelima Bhat for her contributions to the synthetic sequences.
JA9618255
(34) Ferrin, T. E.; Huang, C. C.; Jarvis, L. E.; Langridge, R. J. Mol.
Graphics 1988, 6, 13-27.
(35) Hore, P. J. J. Magn. Reson. 1983, 55, 283-300.
(36) Bax, A.; Lerner, L. J. Magn. Reson. 1988, 79, 429-438.
(33) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson. 1982,
48, 286-292.