2-Benzamidoethyl group - A novel type of phosphate protecting group for oligonucleotide synthesis

Article abstract of DOI:10.1021/ja0016396

A number of 5′-O-(4,4′-dimethoxytrityl)thymidine N,N-diisopropylamino phosphoramidites protected at P(III) with derivatives of 2-benzamidoethanol were synthesized and incorporated into synthetic oligonucleotides. Depending on substitution patterns at the

Full text of DOI:10.1021/ja0016396

VOLUME 123, NUMBER 5
FEBRUARY 7, 2001
© Copyright 2001 by the
American Chemical Society
2-Benzamidoethyl Group - A Novel Type of Phosphate Protecting
Group for Oligonucleotide Synthesis
Andrei P. Guzaev* and Muthiah Manoharan
Contribution from the Department of Medicinal Chemistry, Isis Pharmaceuticals, Inc., 2292 Faraday AVe.,
Carlsbad, California 92009
ReceiVed May 12, 2000
Abstract: A number of 5′-O-(4,4′-dimethoxytrityl)thymidine N,N-diisopropylamino phosphoramidites protected
at P(III) with derivatives of 2-benzamidoethanol were synthesized and incorporated into synthetic oligonucle-
otides. Depending on substitution patterns at the alkyl chain, amido group, and phenyl ring, the time required
for removal of these protecting groups using concentrated ammonium hydroxide varied from 48 h at 55 °C to
25 min at 25 °C. Of the 11 groups studied, 2-[N-isopropyl-N- (4-methoxybenzoyl)amino]ethyl- (H) and
ω-(thionobenzoylamino)alkyl protections (I and K) were most easily removed. Derivatives of the 2-[N-methyl-
N-benzoylamino]ethyl group (E-G) demonstrated moderate stability, but those of the 2-(N-benzoylamino)-
ethyl group (A-C) were the most stable. For the most reactive group, H, a phosphitylating reagent, bisamidite
60, was synthesized and used in the preparation of four deoxynucleoside phosphoramidites 28 and 65-67,
plus the 2′-O-(2-methoxyethyl)-5-methyluridine phosphoramidite 68. All of these novel building blocks were
successfully tested in the preparation of natural, 20-mer oligonucleotides and their phosphorothioate analogues.
With the model phosphotriester 37, the mechanism of deprotection was studied and revealed, in the case of
group H, a pH-independent formation of the 2-oxazolinium cation 47. Under aqueous conditions, 47 gave 54,
which in turn was converted in the presence of ammonia to a number of identified products. It is important to
note that none of the products formed was reactive toward the oligonucleotide backbone or nucleic bases.
Thus, a general strategy for protection of internucleosidic phosphodiester groups is described, which may also
find application in synthetic organic chemistry of phosphorus(III) and (V).
Introduction
tionary approach allows an efficient preparation of DNA and
RNA fragments,3 as well as many types of modified oligo-
nucleotides,4 on a routine basis. Phosphoramidite building blocks
are most often protected at the phosphite moiety by a 2-cyano-
ethyl group.5 Final deprotection of the oligonucleotide with
ammonia effects â-elimination in the 2-cyanoethyl group, thus
(1) (a) Atkinson, T.; Smith, M. In Oligonucleotide Synthesis: A Practical
Approach; Gait, M. J., Ed.; Oxford University Press: Oxford, New York,
Tokyo, 1984; pp 35-81. (b) Sproat, B. S.; Gait, M. J. In Oligonucleotide
Synthesis: A Practical Approach; Gait, M. J., Ed.; Oxford University
Press: Oxford, New York, Tokyo, 1984; pp 83-115.
(2) (a) Matteucci, M. D.; Caruthers, M. H. J. Am. Chem. Soc. 1981, 103,
3185-3191. (b) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981,
22, 1859-1862.
For the past three decades, oligonucleotide synthesis on solid
support has made tremendous progress from the preparation of
milligram quantities using a manual, labor-intensive procedure1
to the manufacturing of the first antisense drug, an anti-CMV
21-mer 2′-deoxyoligonucleotide phosphorothioate, on kilogram
scale.
Currently, oligonucleotide synthesis on solid support employs
a phosphoramidite method2 and is carried out automatically by
stepwise coupling of nucleoside building blocks. This revolu-
* Corresponding author: Phone: (760) 603-2445. Fax: (760) 929-0036.
E-mail: aguzaev@isisph.com.
10.1021/ja0016396 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/10/2001

784 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
GuzaeV and Manoharan
releasing an internucleosidic phosphodiester moiety and acrylo-
nitrile, which is toxic and a potential carcinogen, as a side
product. In addition, acrylonitrile has been demonstrated to
alkylate nucleic bases under conditions of large-scale oligo-
nucleotide deprotection.6
Scheme 1
Recent interest in manufacturing oligonucleotides as drugs
led to a search for protecting groups without the drawbacks
associated with the 2-cyanoethyl and related groups. Of the base
labile protections whose removal is governed by mechanisms
different from â-elimination, several are worthy of note. Among
them, the 2-trialkylsilylethyl protecting group is removed by
â-fragmentation, yielding only the harmless products ethylene
and trialkylsilanol.7,8 Recently, ω-(trifluoroacetylamino)alkyl
protection has been introduced. It is cleaved by aqueous or
gaseous ammonia in two steps. First, the base labile trifluoro-
acetyl group is removed to yield the ω-aminoalkyl phosphate
triester, which in the second step undergoes cyclodeesterifica-
tion. This releases the final products, the internucleosidic
phosphodiester and an azacycloalkane.6,9,10 Allyl phosphate
protection is removed either by aqueous ammonia or, under
orthogonal conditions, by treatment with a Pd0 complex in the
presence of a nucleophile.11 However, there is no known
protecting strategy that allows one to control the kinetics of
deprotection of internucleosidic phosphates over a wide range.
Having such a methodology would allow protecting groups to
be chosen in accordance with anticipated deprotection conditions
and vice versa.
2-[(1-naphthyl)carbamoyloxy]ethyl protecting group in DNA
synthesis.12 Now we turned our attention to a reported intramo-
lecular nucleophilic substitution in 1 where 2-oxazoline 3 was
formed by releasing phosphodiester 2 as a leaving group.13
Similarly, solvolysis of thioureido analogues of 1 gave 2-thi-
azolines and phosphate esters or inorganic phosphate.14 Similar
reactions with more traditional leaving groups, such as halogens,
and methyl- or toluenesulfonate, have been studied in detail.15
To the best of our knowledge, use of N-acylaminoethyl-
protected nucleoside phosphoramidites 4 (Scheme 1) has not
been described in the literature. Nevertheless, oligonucleotides
similar to 5 have recently been synthesized by different routes
via cyclic phosphoramidites.16,17 When exposed to aqueous
ammonium hydroxide, all of them gave deprotected oligonucle-
otides 6, although this required a prolonged treatment at elevated
temperature. However, structural optimization of the protecting
group with respect to deprotection kinetics has not been
attempted; nor have the products derived from the protecting
groups been identified.
In our search for a more general methodology for protection
of internucleosidic phosphates, we reported recently the use of
(3) (a) Caruthers, M. H. Acc. Chem. Res. 1991, 24, 278-284. (b)
Beaucage, S. L.; Iyer, R. P. Tetrahedron 1992, 48, 2223-2311.
(4) (a) Manoharan, M. In Antisense Research and Applications; Crooke,
S. T., Lebleu, B. Eds.; CRC Press: Boca Raton, FL, 1993; pp 303-349.
(b) Beaucage, S. L.; Iyer, R. P. Tetrahedron 1993, 49, 1925-1963.
(5) (a) Sinha, N. D.; Biernat, J.; Ko¨ster, H. Tetrahedron Lett. 1983, 24,
5843-5846. (b) Sinha, N. D.; Biernat, J.; McManus, J.; Ko¨ster, H. Nucleic
Acids Res. 1984, 12, 4539-4557.
On the basis of these data, we aimed our study at the
N-acylaminoethyl moieties as protecting groups for internucleo-
sidic phosphates. To this end, we synthesized a number of novel
(12) Guzaev, A. P.; Manoharan, M. Tetrahedron Lett. 2000, 41, 5623-
5626.
(13) Ziodrou, C.; Schmir, G. L. J. Am. Chem. Soc. 1963, 85, 3258-
3264.
(6) Wilk, A.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L. J. Org.
Chem. 1999, 64, 7515-7522.
(7) (a) Krotz, A. H.; Wheeler, P.; Ravikumar, V. T. Angew. Chem., Int.
Ed. Engl. 1995, 34, 2406-2409. (b) Ravikumar, V. T. Synth. Commun.
1995, 25, 2161-2164. (c) Ravikumar, V. T.; Cole, D. L. Gene 1994, 149,
157-161. (d) Ravikumar, V. T.; Wyrzykiewicz, T. K.; Cole, D. L.
Tetrahedron 1994, 50, 9255-9266. (e) Ravikumar, V. T.; Sasmor, H.; Cole,
D. L. Bioorg. Med. Chem. Lett. 1993, 3, 2637-2640.
(14) (a) Cherbuliez, E.; Espejo, O.; Willhalm, B.; Rabinowitz, J. HelV.
Chim. Acta 1968, 51, 241-248. (b) Cherbuliez, E.; Willhalm, B.; Espejo,
O.; Jaccard, S.; Rabinowitz, J. HelV. Chim. Acta 1967, 50, 1440-1452. (c)
Cherbuliez, E.; Moll, H.; Baehler, B.; Rabinowitz, J. HelV. Chim. Acta 1967,
50, 1154-1158. (d) Cherbuliez, E.; Baehler, B.; Espejo, O.; Jaccard, S.;
Jindra, H.; Rabinowitz, J. HelV. Chim. Acta 1966, 49, 2408-2415.
(15) For a comprehensive review, see: (a) Lambert, J. B.; Mark, H. W.;
Holcomb, A. G.; Magyar, E. S. Acc. Chem. Res. 1979, 12, 317-324. (b)
Capon, B.; McManus, S. P. Neighboring Group Participation; Plenum
Press: New York, 1976; Vol. 1; p 280. (c) Capon, B. In Proton-Transfer
Reactions; Caldin, E., Gold, V., Eds.; Chapman & Hall: London, 1975; pp
339-384. (d) Page, M. I. Chem. Soc. ReV. 1973, 2, 295-323. (e) Goodman,
L. AdV. Carbohydr. Biochem. 1967, 22, 109-175. (f) Capon, B. Quart.
ReV. 1964, 18, 45-111.
(8) Wada, T.; Sekine, M. Tetrahedron Lett 1994, 35, 757-760.
(9) Cheruvallath, Z. S.; Capaldi, D. C.; Ravikumar, V. T.; Cole, D. L.
U.S. Patent 5,760,209 1998; Chem. Abstr. 129: 41382.
(10) Wilk, A.; Srinivasachar, K.; Beaucage, S. L. J. Org. Chem. 1997,
62, 6712-6713.
(11) (a) Bergmann, F.; Kueng, E.; Iaiza, P.; Bannwarth, W. Tetrahedron
1995, 25, 6971-6976. (b) Hayakawa, Y.; Hirose, M.; Hayakawa, M.;
Noyori, R. J. Org. Chem. 1995, 60, 925-930. (c) Hayakawa, Y.;
Wakabayashi, S.; Kato, H.; Noyori, R. J. Am. Chem. Soc. 1990, 112, 1691-
1696. (d) Hayakawa, Y.; Uchiyama, M.; Kato, H.; Noyori, R. Tetrahedron
Lett. 1985, 26, 6505-6508. (e) Bogdan, F.; Chow, C. S. Tetrahedron Lett.
1998, 39, 1897-1900. (f) Manoharan, M.; Lu, Y.; Casper, M. D.; Just, G.
Org. Lett. 2000, 2, 243-246.
(16) (a) Iyer, R. P.; Yu, D.; Devlin, T.; Ho, N.-H.; Agrawal, S. J. Org.
Chem. 1995, 60, 5388-5389. (b) Guo, M. J.; Yu, D.; Iyer, R. P.; Agrawal,
S. Bioorg. Med. Chem. Lett. 1998, 8, 2539-2544.
(17) Wilk, A.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L. J. Am.
Chem. Soc. 2000, 122, 2149-2156.

Phosphate Protecting Group
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 785
Scheme 2
phosphoramidites 4. These were incorporated into the synthetic
oligonucleotides 5 in which Y, Z, and R may vary to form
amido, thioamido, or thioureido groups that are capable of
efficient intramolecular nucleophilic assistance (Scheme 1). We
hypothesized that, once introduced into a synthetic oligonucle-
otide 5, these groups might subsequently attack an electrophilic
center, either phosphorus or carbon in the P-O-C fragment.
If, as with phosphotriester 1,13 the nucleophilic attack occurred
at the carbon atom, a cyclic intermediate, 7, would be formed,
and the internucleosidic phosphate would serve as a leaving
group. In this event, the deprotection of an oligonucleotide
backbone would occur and yield 6. To test this hypothesis,
deprotection of 5 with aqueous ammonia was studied. The most
promising phosphoramidites, 28 and 65-68 (Scheme 7) were
successfully employed in standard oligonucleotide synthesis for
the preparation of natural oligonucleotides and their antisense
phosphorothioate analogues. To illustrate the mechanism of
deprotection, the intermediate 47 and stable products formed
in the course of the reaction were characterized (Scheme 6).
Here we report the results of this study.
phosphoramidites, along with the isomeric mixture. Because
individual (R)p- and (S)p-isomers are of limited use in oligo-
nucleotide chemistry,19 assignment of the absolute configuration
of 21-31 at the P(III) was not performed. Instead, the individual
species were assigned as fast- and slow-eluting isomers in
accordance with their chromatographic mobility on silica gel.
Nevertheless, the availability of pure (R)p- and (S)p-isomers
was of paramount importance for spectral characterization of
21-31 by NMR. In general, phosphoramidites exhibit rather
Results and Discussion
Synthesis and Evaluation of Thymidine Phosphoramidites
Protected by 2-Benzamidoethyl, ω-(Thionobenzoylamino)-
alkyl, and 2-(N-Phenylthiocarbamoylamino) Ethyl Groups
at P(III). Of the many structural motifs capable of an intramo-
lecular nucleophilic cyclization,15 we chose to study the utility
of 2-benzamidoethyl (A-H), ω- (thionobenzoylamino)alkyl
(I and K), and the 2-(N-phenylthiocarbamoylamino)ethyl (J)
groups (Scheme 2). Several features justify their potential
advantage for phosphate protection. First, under basic conditions,
the carbonyl function in these systems is known to provide
remarkable anchimeric assistance.15f On the other hand, in
neutral and acidic media, A-K are chemically inert groups,
with no apparent reactivity toward phosphitylating and acylating
reagents that are employed in standard oligonucleotide synthesis.
Finally, to introduce A-K at the phosphite moiety of a
nucleoside phosphoramidite, the alcohols 10-20 required as
starting materials are stable compounds which are commercially
available or which may be readily synthesized with great
structural diversity. For the phosphoramidite synthesis, 8 was
first converted to a bisamidite, 9 (Scheme 2).18 Without isolation,
9 was reacted with 10-20 in the presence of 1H-tetrazole to
give thymidine phosphoramidites 21-31. These were chro-
matographed on a silica gel column to allow purification and,
except for 24, isolation of pure (R)p- and (S)p-isomers of the
complex H NMR and 13C NMR spectra, due to H and 13C
long-distance spin-spin couplings to phosphorus and to the
overall structural complexity of these molecules. In addition,
the amide groups of 21-31 apparently formed stable rotamers
at the C-N bond, which resulted in a multiplicity of NMR
signals. Similar observations have been recently reported for
4-[N-methyl-N-(trifluoroacetyl)amino]butyl6 and cyclic N-acyl17
nucleoside phosphoramidites. Furthermore, with the tertiary
amides 25-28, hindered rotation around the amide bond resulted
in very low signal intensity for the methylene group adjacent
to the amide nitrogen. Considering all of these factors, recording
the NMR spectra for individual isomers dramatically facilitated
assignment of the structures of the phosphoramidites 21-31.
The phosphoramidites 21-31 demonstrated excellent shelf
life and hydrolytic stability. As revealed by 31P NMR, all of
these compounds were > 95% pure after being stored as dry
samples for nine months at room temperature under argon.
Hydrolytic stability of a representative selection of phosphor-
amidites was also determined. As measured by 31P NMR, the
half-lives of disappearance of compounds 22, 26, and 28 (0.1
M in 95% aqueous CD3CN; 25 °C) were 200, 126, and 73 h,
1
1
(18) Marugg, J. E.; Burik, A.; Tromp, M.; van der Marel, G. A.; van
Boom, J. H. Tetrahedron Lett. 1986, 27, 2271-2274.
(19) Stec, W. J.; Zon, G. Tetrahedron Lett. 1984, 25, 5279-5282.

786 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
GuzaeV and Manoharan
Table 1. Deprotection of 33a-k with Concentrated Aqueous
Scheme 3a
Ammonium Hydroxide
deprotection
conditions
protecting
protected
group
phosphoramidite oligonucleotide time, h temp, °C
A
B
C
D
E
F
G
H
I
21
22
23
24
25
26
27
28
29
30
31
33a
33b
33c
33d
33e
33f
33g
33h
33i
48
48
48
8
6
5
55
55
55
55
55
55
25
25
25
25
25
1
<0.5
<1
1
J
K
33j
33k
1.5
expected products, either 35 or, for 21-24, mixtures of 34a-d
and 35 (for efficiency of deprotection, see below). For 29-31,
use of the standard iodine oxidizer led to formation of
unidentified side products. In contrast, oxidation of phosphite
triester intermediates derived from 29-31 and 32 with tert-
butyl hydroperoxide21 led only to 35. When the oxidation step
was replaced with sulfurization using 3H-1,2-benzodithiol-3-
one 1,1-dioxide,22 the phosphorothioate analogues of 35 were
easily obtained.
The above data suggested that each of the phosphoramidites
21-31 could be used for preparation of synthetic oligonucle-
otides. They also indicated that, depending on the nature of the
protecting group, the deprotection time that is required for
complete conversion of 33a-k to 35 differed substantially. To
study the removal of protecting groups in more detail, the solid-
support-bound compounds 33a-k were treated with concen-
trated aqueous ammonia at either 25 °C (33f-k) or 55 °C (33a-
g), and the progress of deprotection was followed by reverse-
phase HPLC. In all cases, the final product 35 could be easily
separated from the more hydrophobic oligonucleotides 34a-k,
and the time required for complete deprotection of 33a-k
(>99.5%) was determined (Table 1). The data in Table 1 reveal
dramatic differences in reactivity of the protecting groups studied
toward concentrated aqueous ammonia, which are discussed
below.
Of the protecting groups derived from 2-benzamidoethanol,
the tertiary amides E-H were removed dramatically faster than
the secondary ones A-D (cf. 33e-h and 33a-d). Deprotection
was additionally facilitated when bulky substituents were
introduced on the aminoethyl skeleton of a protecting group,
as predicted by the Thorpe-Ingold gem-dialkyl effect.23 Indeed,
the dimethyl-substituted compound 33d reacted markedly faster
than 33a. Similarly, the N-iPr derivative 33h was more reactive
than the N-Me analogue 33f.
Of the electron-donating or -accepting substituents introduced
on the phenyl ring, only the 4-dimethylamino group altered the
deprotection kinetics to any substantial extent (cf. 33e and 33g).
In contrast, the influence of NO2 or MeO groups was insig-
nificant (for secondary amides, cf. 33a, b, and c; for tertiary
ones, cf. 33e and f).
a (a) 1H-Tetrazole/MeCN; (b) for 21-28: I2/Py/THF/H2O; for 29-
31: t-BuOOH/MeCN.
respectively.20 These data suggest excellent lifetimes for solu-
tions of phosphoramidites while on the DNA synthesizers.
At the beginning of this work, no information was available
on the ability of the phosphoramidites 21-31 to support
oligonucleotide synthesis or the compatibility of the novel
protecting groups with the known conditions of oligonucleotide
deprotection. Therefore, a proper model system had to be chosen
that would allow one to assess the potential of the proposed
protecting strategy. In our opinion, this was best achieved by
using solid-support-bound oligonucleotides 33a-k as the model
compounds (Scheme 3). This model offered several advantages.
First, it permitted testing of a variety of phosphoramidites that
were prepared on a small scale. Regardless of the structure of
the protecting group or the success in removing it, the reaction
mixtures that were obtained were likely to be analyzed under
standardized conditions by HPLC. These two features allowed
a relatively high-throughput screening of a number of protecting
groups. Second, this model offered the possibility of testing the
phosphoramidites 21-31 specifically under conditions of oligo-
nucleotide synthesis on solid support. When the synthesis is
carried out on solid phase, the reagents are used in much larger
excess, and unwanted side reactions are more pronounced than
in solution. In addition, more drastic conditions are often
required to remove protecting groups from a negatively charged
oligonucleotide backbone than from dinucleoside monophos-
phates. Thus, with model oligonucleotides 33a-k, the choice
for the most suitable protecting groups could be founded on
more representative data. The ability of the phosphoramidites
21-31 to support oligonucleotide synthesis was tested first
(Scheme 3). Using a standard synthetic cycle and ancillary
reagents, we attached the phosphoramidites at the 5′-terminus
of solid-support-bound undecathymidylate, 32, to give 33a-k.
These were conventionally deprotected with ammonia (6 h/55
°C), and the products were analyzed by HPLC and electron-
spray MS. The following relevant observations were noted. As
revealed by the dimethoxytrityl assay, all phosphoramidites
demonstrated high coupling efficiency (>95%). Phosphoramid-
ites 21-28 were fully compatible with ancillary reagents that
were used in oligonucleotide synthesis, yielding only the
Comparison of the thionocarbonyl derivative 33i with the
corresponding oxygen analogue 33a indicates dramatically faster
deprotection of the former compound, which may reflect an
(21) Alul, R. H.; Singman, C. N.; Zhang, G.; Letsinger, R. L. Nucleic
Acids Res. 1991, 19, 1527-1532.
(22) Iyer, R. P.; Phillips, L. R.; Egan, W.; Regan, J. B.; Beaucage, S. L.
J. Org. Chem. 1990, 55, 4693-4699.
(20) Under these conditions, a commercial sample of 77 had a half-life
of 200 h.
(23) (a) Beesley, R. M.; Ingold, C. K.; Thorpe, J. F. J. Chem. Soc. 1915,
107, 1080-1106. (b) Ingold, C. K. J. Chem. Soc. 1921, 119, 305-329.

Phosphate Protecting Group
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 787
Scheme 4a
Scheme 5a
a (a) 1. EtOH/1H-Tetrazole/CH2Cl2; 2. (3-ClC6H4)CO3H. (b) MeCN,
80% aqueous MeCN, MeOH, or THF:2,6-lutidine:H2O (96:2:2).
Table 2. Deprotection of 40-43, 49, and 52 under Neutral
Conditions
half-life: τ1/2 × 10-2, min
conditions (25 °C)
MeCN
40 41 42 43
49
52
1.0 0.5 0.8 0.8a 23.0 171.0
80% aqueous MeCN
MeOH
THF:2,6-lutidine:H2O (96:2:2) 1.0 0.7
1.7 0.5 0.9
2.0 0.7 0.8
3.1
2.6
15.0
14.7
15.5
a Determined in MeCN:pyridine (95:5).
phosphorothioate24 to yield 39, which was then converted to
43. Additionally, two simpler model compounds were prepared
in solution. Phosphoramidite 28 was reacted with EtOH in the
presence of 1H-tetrazole, and the product was oxidized to give
49 (Scheme 5). Phosphotriester 52 was synthesized by phos-
phorylation of 17 with O,O-diethyl phosphorochloridate (Scheme
6). Both 49 and 52 were isolated by column chromatography
on silica gel.
The model compounds 40-43, 49, and 52 were treated under
the mild conditions shown in Table 2. For 40-43, the progress
of the deprotection was followed by 31P NMR in gel phase,
where their peaks were observed at 68.5-71 ppm. Regardless
of the conditions employed, deprotection resulted in the forma-
tion of the corresponding thionophosphodiesters. Thus, as seen
from Scheme 4, compounds 40-42 were converted to 44 (55-
61 ppm), and 43 yielded 45 (56.0-56.5 ppm). Simultaneously,
the protecting groups G and H were converted to the corre-
sponding 2-oxazolinium cations 46 and 47, and group I was
converted to the 2-thiazolinium cation 48 (for the fate of the
protecting groups, see below).
The deprotection of 49 and 52 was studied by conventional
31P NMR in solution. Similarly to the phosphorothioates above,
both 49 and 52 were smoothly converted to the phosphodiesters
50 and 53 (Schemes 5 and 6, respectively). In MeCN and THF,
the peaks of the starting material were observed at 1.5 to -1.0
ppm and were downfield of the peaks of the respective products
50 and 53 at 1.1 to -1.5 ppm. In 80% aqueous MeCN and
MeOH, the peaks of 49 and 52 (-1.2 to -1.4 ppm) were found
upfield with respect to the peaks of 50 and 53 (0.5 to -0.2
ppm). Additionally, to reveal the fate of group H, the depro-
tection of 52 was studied by reverse-phase HPLC (see below).
In terms of the half-lives of disappearance of 40-43, 49, and
52, several conclusions might be drawn from the results
summarized in Table 2. Comparing the half-lives of the
dinucleoside phosphorothioates 40-42, the reactivities of the
protecting groups G-I appear to be very similar. Under both
aqueous and nonaqueous conditions, the stability of the protect-
ing groups increased in the following order, H < I < G, that
a (a) R-OH/1H-Tetrazole/MeCN; (b) 3H-1,2-benzodithiol-3-one 1,1-
dioxide/MeCN; (c) MeCN, 80% aqueous MeCN, MeOH, or THF:2,6-
lutidine:H2O (96:2:2).
enhanced intramolecular participation of the more nucleophilic
thiocarbonyl group. In a very similar fashion, groups J and K
were removed under conditions as mild, or almost as mild, as
those for I.
Removal of the Protecting Groups G-I under Neutral
Conditions. It can be seen from Table 1 that the removal of
the protecting groups G-K required e1 h at 25 °C. Due to
experimental limitations, it is difficult to determine with
confidence their stability in concentrated ammonium hydroxide.
To provide a more accurate comparison of the deprotection
characteristics, removal of the representative groups G-I was
studied under less aggressive conditions. In preliminary experi-
ments, dithymidine monophosphates protected with G-I proved
to be very labile compounds, which prohibited their isolation.
Instead, model compounds 40-43 were synthesized on a high-
loaded Tentagel polymer support as depicted in Scheme 4.
Phosphoramidite building blocks 27-29 were first converted
to solid-support-bound 36-38. The stability of 37 and 38 in
MeCN was studied by 31P NMR spectroscopy in gel phase to
reveal that both compounds remained unchanged for at least
16 h at room temperature. Compounds 36-38 were then
sulfurized with 3H-benzodithiol-2-one 1,1-dioxide in MeCN22
to give 40-42. In a similar fashion, 28 was coupled to a solid-
support-bound, phosphate-deprotected hexathymidylate penta-
(24) Guzaev, A. P.; Manoharan, M. J. Org. Chem., in press.

788 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
GuzaeV and Manoharan
is, their relative reactivity remained unchanged in comparison
to that in concentrated ammonium hydroxide.
or if oxidation with iodine solutions can be avoided (I-K).
Finally, A, B, and C are not recommended for routine
oligonucleotide synthesis, although they may be suitable for
performing reactions in solution. Moreover, the dramatic
difference in deprotection time allows keeping A, B, or C intact
while the most labile group H is completely removed.25 Thus,
a combination of these means of protection may find use in
synthesizing selectively modified oligonucleotides.
The Fate of the 2-[N-Isopropyl-4-methoxybenzoylamino]-
ethyl and 2-(N-Thionobenzoylamino)ethyl Groups (H and
I) Upon Deprotection. In a recent communication, acrylonitrile,
which is released on cleavage of the traditional 2-cyanoethyl
phosphate protecting group, has been demonstrated to alkylate
nucleic bases.6 Although this side reaction may occur to a
measurable extent only on a manufacturing scale, the study
illustrates the importance of analyzing potential consequences
before novel protecting groups are introduced in laboratory
practice. Although the protecting groups H and I are labile under
the conditions of oligonucleotide synthesis, the extent of the
deprotection may depend on the length of the synthesis. For
preparations on a small scale, in which faster coupling protocols
are used, an oligonucleotide may remain extensively protected
at the end of the chain assembly. These protecting groups will
have to be removed at the same time as the nucleic base
protecting groups. Therefore, to evaluate possible drawbacks
of the proposed protecting strategy, the fate of the protecting
groups H and I and the ability of the deprotection products to
participate in side reactions were studied.
The half-lives of 40-42 in MeCN demonstrated that the
protecting groups G-I were labile to the extent that they might
be removed under the conditions of oligonucleotide synthesis.
Furthermore, in comparison with 41, removal of group H from
43 containing the negatively charged backbone was retarded
by a factor of < 2. This suggested that a partial deprotection of
an oligonucleotide might not inhibit the progress of the further
deprotection. For group H, this was verified by synthesizing
pentadecathymidylate phosphorothioate 51 on a 20-µmol scale
using the phosphoramidite 28 as a building block. On comple-
tion of the chain assembly, the extent of the protection of the
solid-support-bound 51 was immediately determined by 31P
NMR. The two major peaks of thionophosphates observed at
68 and 58 ppm were attributed to the triester groups protected
by H and the deprotected diester groups, respectively. The latter
accounted for ca. 90% of the total integrated area, which
revealed that the majority of the thionophosphate linkages were
deprotected at the end of the synthesis. The deprotection of 51
was completed by a brief treatment of the solid support with
10% pyridine in MeOH (1 h/RT). The 31P NMR spectrum of
the product displayed the signal of thionophosphate diester
groups and a minor signal of desulfurized phosphodiester groups
(-2 ppm, 3.4%). Importantly, no peaks of other side products
were detected, which demonstrated that the deprotection was
not accompanied by any unwanted reactions of 47 with the
internucleosidic thionophosphate residues.
To study the fate of group H, a model compound, 52, was
treated with either water or solutions of ammonium hydroxide
(0.5 M, 2.0 M, or 28% in water), and the progress of the
reactions was followed by HPLC. The products formed were
either isolated and characterized (47 and 54) or identified by
co-injection with authentic samples (17 and 56-58). The results
are summarized in Scheme 6. Regardless of the basicity of the
media, 52 rapidly reacted to form diethyl phosphate 53, which
Comparison of the half-lives for compounds 41, 49, and 52
that were protected with group H demonstrated that nucleosidic
residues dramatically facilitated the deprotection. One may
tentatively speculate that the bulky nucleosidic residues favor
the departure of the protecting group, which reduces steric
hindrance around the phosphate moiety.
1
was characterized by H, 13C NMR, and 31P NMR and other
products. The half-life of disappearance of 52 (380-400 min
at 25 °C) was independent of the concentration of ammonium
hydroxide. In contrast, the distribution of products other than
53 markedly depended on the concentration of ammonium
hydroxide.
Our attempts to clarify the influence of protic solvents on
the deprotection rates led to conclusive results only for the
reactions in homogeneous phase. For compounds 49 and 52,
whose stability was measured in solution, the protic solvents
accelerated deprotection by a factor of >10. However, for the
solid-support-bound compounds 40-42, either a very minor
negative effect (40) or no influence of the polar environment
on the half-lives of the deprotection was observed. This
discrepancy might be tentatively explained by the lack of
efficient solvation inside the matrix of the hydrophobic poly-
styrene solid support rather than by the reactivity of the
protecting group H in dinucleoside monophosphates 40-42.
The fact that the protecting groups G-I were removed under
the conditions of oligonucleotide synthesis was not considered
a disadvantage. We demonstrated recently that, upon conversion
of the phosphodiester residues to the 4-dimethylaminopyri-
dinium salt form, the coupling of 2-cyanoethyl phosphoramidite
building blocks to phosphate-unprotected oligonucleotides may
be carried out as efficiently as the standard coupling.24
Therefore, the protecting group H that provided the best
combination of synthetic availability and ease of deprotection
was selected for further investigation. Other groups can also be
employed as phosphate protecting groups in oligonucleotide
synthesis if slower deprotection kinetics is acceptable (D-G)
Under neutral conditions, 54 was the final product of the
reaction, although in early stages, the intermediate 47 ac-
cumulated in the reaction mixture. When 70-98% of starting
material 52 was consumed, 45-55% of the total integrated area
in the HPLC traces accounted for 47. At this stage, 47 was
isolated by HPLC in 90% purity, and its structure was proven
1
by LC/MS, H, and 13C NMR spectra. However, the stability
of 47 was insufficient to afford its complete characterization.
In water, 47 exhibited a half-life of 40 h and underwent
hydrolytic ring opening to form 54 in quantitative yield. The
latter product was indefinitely stable under neutral and acidic
conditions, which permitted its isolation by HPLC as a trifluo-
roacetate salt and its unambiguous characterization.
In the presence of dilute ammonium hydroxide (0.5 to 2.0
M), the product distribution dramatically changed. Compound
47 was rapidly hydrolyzed so that the extent of its accumulation
in the reaction mixture was < 1%. It has been reported that in
aqueous K2CO3, alkaline hydrolysis of similar N-methyl-2-
oxazolinium salts gave exclusively O-acyl N-methylethanol-
amines as kinetic products.26 According to very recent findings,
this regioselectivity resulted from bifunctional nucleophilic
(25) Guzaev, A.; Manoharan, M. Unpublished results.

Phosphate Protecting Group
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 789
Scheme 6
the deprotection was complete (Scheme 4). The solid support
was then washed with 5% aqueous NaHCO3 and ethyl acetate,
which converted 48 to the 2-phenyl-2-thiazoline-free base (59)
and eluted it from the solid support. On evaporation of the
organic phase, 59 was obtained in 85% yield. The structure of
59 was consistent with its 1H NMR and IR spectra, which were
in good agreement with those reported in the literature,28 and
with its 13C NMR spectrum. The treatment of 59 with
concentrated aqueous ammonium hydroxide for 5 days revealed
no appreciable chemical transformations, which demonstrated
its high stability under the conditions of oligonucleotide
deprotection.
Formation of 47, 54, and 59 demonstrated above confirms
the proposed intramolecular mechanism of deprotection where
the phosphodiester anion is substituted with the oxygen atom
of the amido group or the sulfur of the thionoamido group
(Scheme 1). Moreover, it is important to note that no products
that indicate reactivity of 52 by any other route were detected
in the reaction mixtures.
The occurrence of minor side reactions of the products derived
from group H with nucleic bases was studied. Each of the 5′-
O-(4,4′-dimethoxytrityl)-2′-deoxynucleosides (dA, dC, dG, and
T) was mixed with 52 (10-fold excess) and treated with
concentrated ammonium hydroxide for 24 h at 55 °C. When
the reaction mixtures were analyzed by reverse-phase HPLC,
no new nucleosidic products were observed. Of the compounds
presented in Scheme 7, only compound 47 may in principle be
considered as an alkylating agent. As seen from the reactivity
of 47 toward ammonia, the products of alkylation, that is, 55
and 58, were formed only in concentrated ammonium hydroxide
in low yield (<10%). Thus, 47 is a very weak alkylating agent
that is, however, highly labile toward hydrolysis. Rapid
consumption of 47 by hydrolysis under basic conditions further
reduces its chance to react with nucleosides. In view of these
considerations, the observed lack of modifications of nucleic
bases is not surprising.
Utility of the 2-[N-Isopropyl-4-methoxybenzoylamino]-
ethyl-Protected Nucleoside Phosphoramidites in Oligonucle-
otide Synthesis. The usefulness of the proposed protection
strategy for routine preparation of synthetic oligonucleotides
and their phosphorothioate analogues was rigorously tested as
described below. First, a novel phosphitylating reagent, bis-
amidite 60, was synthesized from 17, as depicted in Scheme 7.
After column purification, 60 was obtained in 84.5% yield as a
stable crystalline material. Under dry conditions, it could be
stored at -18 °C for >2 months without any detectable
decomposition. When the protected 2′-deoxynucleosides 8 and
61-63 were reacted with a slight excess of 60, the correspond-
ing 2′-deoxynucleoside phosphoramidites 28 and 65-67 were
synthesized in 91-98% yield. Similarly, the 2′-O-(2-methoxy-
ethyl)-5-methyluridine phosphoramidite 68 was prepared from
64 in 96% yield. The phosphoramidites 28 and 65-68 were
employed in the synthesis of the natural 20-mer oligonucleotides
69-71, the oligonucleotide phosphorothioates with proven
antisense activity, 72-75, and the 2′-O-(2-methoxyethyl)-
modified oligonucleotide phosphorothioate 76 (Table 3). For
comparison, the same sequences were likewise prepared using
commercial phosphoramidites 77-80 protected with a 2-cyano-
ethyl group at the phosphite moiety.
catalysis provided by the buffer that was used. More generally,
predominant formation of N-acylated N-methylethanolamines
as kinetic products along with O-acylated isomers has been
observed under basic conditions.27 In contrast to these findings,
47 was hydrolyzed to give mainly 54 as the kinetic product.
In turn, 54 was consumed by three concurrent reactions. First,
it underwent reversible OfN acyl migration to give 17. This
was proven by subjecting 17 to the same conditions. On
equilibration, a mixture of 17 and 54 (3:1) was obtained. Second,
basic hydrolysis of the O-anisoyl group led to 56. Finally, at a
concentration for ammonium hydroxide of 2.0 M and higher,
ammonolysis of 54 gave 57. It is reasonable to expect that
N-isopropylaminoethanol was also formed in the latter two
reactions, although no attempt was made to identify it.
In the presence of 28% aqueous ammonium hydroxide, the
product distribution was somewhat more complex. Along with
the products mentioned above, a new compound tentatively
assigned as 55 was formed in approximately 10% yield. On
longer exposure to the reaction conditions, 55 was converted
to 58, which was confirmed by co-injection with an authentic
sample synthesized by an independent route.
To study the fate of group I, the solid support 42 (40 µmol)
was treated with 80% aqueous MeCN as described above until
(26) Deslongchamps, P.; Dube, S.; Lebreux, C.; Patterson, D. R.;
Taillefer, R. J. Can. J. Chem. 1975, 53, 2791-2807.
(27) Perrin, C. L.; Engler, R. I.; Young, D. B. J. Am. Chem. Soc. 2000,
122, 4877-4881.
(28) (a) Ino, A.; Murabayashi, A. Tetrahedron 1999, 55, 10271-10282.
(b) Vorbru¨ggen, H.; Krolikiewicz, K. Tetrahedron 1993, 49, 9353-9372.

790 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
GuzaeV and Manoharan
Scheme 7
Solid-support-bound oligonucleotides were deprotected with
concentrated aqueous ammonium hydroxide for 2 h at room
temperature (69 and 76) or for 6 h at 55 °C (70-74).
Considering the rates of deprotection of compounds 41 and 43
(Table 2) and the fact that the release of diethyl phosphate from
52 occurred in a pH-independent manner, the use of a base might
be unnecessary for the removal of the remaining phosphate
protecting groups. However, these conditions were required for
the removal of the protecting groups from nucleic bases and
for the release of the oligonucleotides from the solid support.
After deprotection, all of the oligonucleotides were routinely
isolated by DMTr-On reverse-phase HPLC. When the 5′-
terminal DMTr group was removed, compounds 69-76 were
desalted and characterized by electrospray MS. Their purity was
confirmed by HPLC and capillary gel electrophoresis.
The utility of protecting group H for synthesis of oligonucle-
otides on a large scale was next tested. On a preparative scale,
an extensive optimization of the coupling protocol is required
for the successful preparation of 20-mer oligonucleotides.
Therefore, a lower yield of the product might be initially
expected for a set of novel phosphoramidite building blocks
that are used under unoptimized conditions. The oligonucleotide
75 was synthesized on a 150 µmol scale using 2.5-3 equiv. of
28 and 65-67 and a 20-min coupling time. Indeed, the yield
of 75 using 28 and 65-67 was lower than that obtained with
the standard building blocks 77-80 (43.9 and 55.0%, respec-
tively). When it was deprotected with concentrated ammonium
hydroxide, the crude product was analyzed by reverse-phase
HPLC and LC/MS. It is important to note that no products
indicating modification of 75 at the nucleic bases were detected
in the reaction mixture. This agreed with the observation
presented above that 52 was inert toward nucleosides in
concentrated ammonium hydroxide.
On a small scale, the coupling cycle was first optimized using
the oligonucleotide 69 as an example. The coupling of phos-
phoramidite 28 was slower than that of 77. To obtain the high
stepwise yields, the extended 5-min coupling time for 28 was
used. However, the selectivity of the reaction of 28 with the
5′-hydroxy functions in the presence of unprotected phosphate
moieties was higher than that of 77. We demonstrated recently
Conclusion
The data presented suggest that 2-benzamidoethyl, ω-(thiono-
benzoylamino)alkyl, and 2-(N-phenylthiocarbamoylamino)ethyl
groups are a novel, versatile class of protecting groups for
phosphodiester functions. The utility of these groups was
demonstrated by preparing building blocks for oligonucleotide
synthesis, the nucleoside phosphoramidites 21-31 and 65-68.
All of these compounds were synthesized in high yields using
well-established methods and demonstrated excellent solubility
in MeCN, stability in both solid state and in solution, and
coupling efficiency.
Studying the structure-activity relationship revealed that,
depending on substitution, removal of the 2-benzamidoethyl,
ω-(thionobenzoylamino)alkyl and 2-(N-phenylthiocarbamoyl-
amino)ethyl protection occurs under a wide range of conditions.
We demonstrated here the successful use of the most labile
protecting group for oligonucleotide synthesis on solid support.
Other more stable groups allow one to synthesize chimeric
oligonucleotides by deprotecting certain internucleosidic phos-
phate linkages while other selected groups are kept protected.
In addition, this may make the protecting strategy attractive for
synthetic applications whose requirements for deblocking condi-
tions differ from those of oligonucleotide chemistry.
that under standard conditions, the coupling of 77 to phosphate-
unprotected oligonucleotides resulted in a stepwise yield of
93.8-96.3%.24 As demonstrated above for 51, the protecting
group H was almost completely removed in the course of the
oligonucleotide synthesis. Nevertheless, using 28 as a building
block, the oligonucleotide 69 was synthesized in 97.4% stepwise
yield. When optimized conditions for the phosphoramidite
coupling to phosphate-unprotected oligonucleotides were used,24
the coupling yields could be further improved. In a modified
synthetic cycle, the detritylation of solid-support-bound oligo-
nucleotides was followed by a brief wash with a solution of
0.1 M 4-dimethylaminopyridinium 1H-tetrazolide (DMAP-Tet)
in MeCN. This neutralized the acidic phosphodiester groups.
Furthermore, the protonated DMAP that was retained on the
solid support as a counterion of phosphodiester residues served
as a nucleophilic catalyst in the following coupling reaction.
Under these conditions, 69 was synthesized in 98+% stepwise
yield. In a similar fashion, the mixed-base oligonucleotides 70-
74 were synthesized in yields consistent with a high coupling
efficiency. When the phosphoramidite 68 was used, a coupling
for 10 min was required to ensure 97+% coupling efficiency.
The fate of the most convenient protecting groups, H and I,
upon deprotection was studied. The mechanism of deprotection
was confirmed by isolation of the intermediate 47 or the stable
product 59. The products formed from 47 were identified, and
the dependence of their distribution on the concentration of
ammonium hydroxide in solution was determined. The lack of
reactivity of these compounds toward nucleic bases or the
oligonucleotide backbone was unambiguously demonstrated.

Phosphate Protecting Group
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 791
Table 3. Oligonucleotides Synthesized with the Aid of 28 and 65-68
compound
sequence (5′ f 3′)
backbone
scale, µmol
yield, %
target
69
70
71
72
73
74
75
76
T20
PdO
PdO
PdO
PdS
PdS
PdS
PdS
PdS
1
1
20
1
69.2
63.1
61.5
52.3
50.6
56.5
43.9
49.2
GC3A2GCTG2CATC2GTCA
TC3GC2(TG)2ACATGCAT2
GC3A2GCTG2CATC2GTCA
TC3GC2(TG)2ACATGCAT2
ATGCAT2CTGC5A2G2A
AGCT2CT3G(CA)2TGTA3
human ICAM-1
human C-raf
mouse C-raf
1
20
150
1
human MDM-2
a
T*20
a T* stands for 2′-O-(2-methoxyethyl)-5-methyluridine residue.
temperature, and the solvent was evaporated in vacuo. The residue was
dissolved in ice-cold water (200 mL). The solution was brought to pH
6 with concentrated hydrochloric acid and extracted with ethyl acetate
(5 × 100 mL). Extracts were washed with brine (100 mL), dried over
Na2SO4, and evaporated. Crude 2-(N-benzoylamino)ethanols (85-95%)
were used without further purification. For characterization, 10-13 and
17 were recrystallized from water or a mixture of toluene and hexane.
14 and 15 were purified by column chromatography on silica gel using
a gradient from 5 to 20% MeOH in CH2Cl2 to give colorless oils.
Compounds 10, 11, 13,30 14, and 15,26 were identical to those reported
previously.
One may finally conclude that the proposed protecting
strategy is a valuable addition to the portfolio of methods for
oligonucleotide synthesis and that it may eventually attain
general applicability in the organic chemistry of phosphorus.
Recently, synthesis of oligonucleotides without nucleoside-base
protection has been reported.29 Combining this methodology
with the use of the protecting group H that is removed under
neutral conditions may result in a novel synthetic strategy that
involves no deprotection procedures except for the release from
a solid support.
N-(2-hydroxyethyl)-N-isopropyl-4-methoxybenzamide (15). 88%
yield: white solid, mp 71.5-72 °C (toluene-hexane). 1H NMR
(CDCl3): δ 7.32 (2H, d, J ) 8.6 Hz), 6.89 (2H, d, J ) 8.6 Hz), 4.40
(1H, br s), 4.09 (1H, m), 3.80 (3H, s), 3.90-3.70 (1H, m), 3.53 (2H,
t, J ) 4.6 Hz), 1.13 (6H, d, J ) 6.8 Hz). 13C NMR (CDCl3): δ 173.56,
160.60, 128.71, 128.16, 113.89, 63.69, 55.39, 50.60, 44.45, 21.19.
FAB-HRMS: calcd for C13H19NO3, 238.1443 (MH+); found, 238.1444.
N,N,N′,N′-Tetraisopropyl-O-[2-[N-isopropyl-N-(4-methoxybenzoyl)-
amino] ethyl] phosphordiamidite (60). Chloro bis[(N,N,-diisopropyl)-
amino]phosphite (1.73 g, 6.5 mmol) in CH2Cl2 (20 mL) was added to
a stirred solution of 17 (1.19 g (5.0 mmol) and N,N-diisopropylethyl-
amine (1.03 g, 8.0 mmol) in CH2Cl2 (5 mL) dropwise under argon
atmosphere at -78 °C. The mixture was stirred at -78 °C for 10 min
and was allowed to warm to room temperature. The solution was treated
with triethylamine (2.5 mL) and hexane (50 mL). The mixture was
evaporated to dryness and coevaporated twice with triethylamine-
hexane (5:95; 25 mL). The residue was dissolved in triethylamine-
hexane (5:95; 25 mL), filtered, and applied on a short silica gel column.
The column was eluted with triethylamine-ethyl acetate-hexane (5:
5:90). Fractions were evaporated to give 60 (1.98 g, 84.5%) as a white
Experimental Section
Oligonucleotide synthesis. The oligonucleotide synthesis was
performed on an ABI 380B DNA Synthesizer on a 1-µmol scale
according to the manufacturer’s recommendations. The phosphoramid-
ites 21-31 were used as 0.1 M solutions in dry MeCN. When 29-31
were employed, the oxidation step was carried out with tert-butyl
hydroperoxide21 (10% in MeCN). For preparation of the phosphoro-
thioate oligonucleotides, 3H-1,2-benzodithiol-3-one 1,1-dioxide22 (0.05
M in MeCN) was used as a sulfur-transfer reagent. Solid-support-bound
33a-k were deprotected as depicted in Scheme 3 under conditions
specified in Table 1. The 5′-DMT-protected oligonucleotides were
analyzed by reverse-phase HPLC and characterized by ES/MS.
For preparation of 69-74 and 76 (Table 3) by the modified cycle,
a solution of 0.1 M DMAP and 0.1 M 1H-tetrazole in MeCN was placed
in positions 15 or 17 of the synthesizer (PS and PO cycles, respectively).
The standard detritylation subroutine was followed by a brief washing
with MeCN and flushing with argon. Next, the solution of DMAP-
TET was delivered to the columns for 15 or 45 s (1- and 20-µmol
cycles, respectively), followed by washing with MeCN and flushing
with argon. For the coupling step, phosphoramidites 28 and 65-68
(0.2 M in MeCN) were delivered in 15- and 5-fold excess over the
solid support (1- and 20-µmol cycles, respectively), and the phosphora-
midite condensation was carried out for 5 (28 and 65-67) or 10 min
(68). The solid-support-bound material was deprotected with concen-
trated aqueous ammonium hydroxide under standard conditions (8 h
at 55 °C for 70-75 or 2 h at RT for 69 and 76). The 5′-DMT-protected
oligonucleotides were isolated by reverse-phase HPLC. The 5′-DMT
group was cleaved using 10% aq AcOH (30 min), and the oligonucle-
otides were desalted by reverse-phase HPLC and characterized by ES/
MS. The purity of the isolated oligonucleotides was confirmed by HPLC
and capillary gel electrophoresis.
1
solid: mp 58.5-59.5 °C. H NMR (CDCl3): δ 7.35-7.29 (2H, m),
6.95-6.85 (2H, m), 3.82 (3H, s), 3.80-3.34 (9H, m), 1.40-1.05 (30H,
m). 13C NMR (CDCl3): δ 171.6, 160.4, 129.8, 128.2, 127.7, 113.8,
62.8, 62.4, 55.3, 45.2, 44.5, 44.2, 24.8, 24.6, 23.8, 23.7, 21.0. 31P NMR
(CDCl3): δ 124.6; (CD3CN): δ 130.9. FAB-HRMS: calcd for
C25H46N3O3P (M + Na+), 490.3174; found, 490.3166.
General Procedure for the Preparation of 28 and 65-68. N2-
Isobutyryl-5′-O-(4,4′-dimethoxytrityl)-3′-O-(N,N-diisopropylamino)-
[2-[N-isopropyl-N-(4-methoxybenzoyl)amino]ethoxy]phosphinyl-2′-
deoxyguanosine (67). 1H-Tetrazole (0.45 M in MeCN, 1.29 mL, 0.58
mmol) was added to a mixture of 60 (766 mg, 1.64 mmol), 63 (928
mg, 1.45 mmol), and CH2Cl2 (15 mL), and the resulting solution was
stirred for 2 h at room temperature. Aqueous NaHCO3 (5%, 10 mL)
was added, the emulsion was diluted with brine (50 mL), and the
product was extracted with ethyl acetate (3 × 75 mL). Extracts were
washed with brine (3 × 50 mL), dried over Na2SO4, and evaporated to
dryness. The residue was dissolved in toluene (25 mL), applied on a
silica gel column; and separated, eluting with a gradient from 40:55:5
ethyl acetate:hexane:triethylamine to 5:90:5 ethanol:ethyl acetate:
triethylamine. Collected fractions were evaporated, coevaporated with
dry MeCN (2 × 50 mL), and dried on an oil pump to give 67, fast
diastereomer (170 mg); 67, slow diastereomer (161 mg); and their
mixture (1006 mg) totaled in 1330 mg (91.2%) of 67. 67, fast
diastereomer, 1H NMR (CDCl3): δ 7.78 (1H, s), 7.44-7.15 (13H, m),
6.94-6.74 (6H, m), 6.28 (1H, m), 4.56 (1H, m), 4.23 (1H, m), 3.78
(3H, s), 3.76 (6H, s), 4.15-3.20 (10H, m), 2.80-2.65 (1H, m), 2.43-
2.25 (1H, m), 1.3-0.8 (24H, m). 13C NMR (CDCl3): δ 179.93, 172.53,
The synthesis of 75 was carried out on a MilliGene/Biosearch 8800
Reagent Delivery Module on a 150-µmol scale using 2.5-3 equivs of
28 and 65-67 (0.3 M in MeCN) and a 20-min coupling time. The
solid support was treated with concentrated aqueous ammonium
hydroxide overnight at room temperature. The solution obtained was
heated at 55 °C for 6 h and evaporated. The crude product was analyzed
by reverse-phase HPLC and LC/MS. Compound 75 was isolated and
characterized as described above for 70-75.
General Procedure for the Preparation of 10-15 and 17. Acyl
chloride (0.1 mol) in THF (100 mL) was added dropwise to a solution
of an aminoethanol (0.4 mol) in THF (200 mL) under magnetic stirring
at 4 to 10 °C. The reaction mixture was stirred for 2 h at room
(29) Hayakawa, Y.; Kataoka, M. J. Am. Chem. Soc. 1998, 120, 12395-
12401.

792 J. Am. Chem. Soc., Vol. 123, No. 5, 2001
GuzaeV and Manoharan
160.44, 158.62, 155.82, 148.19, 144.53, 135.58, 129.98, 129.03, 128.02,
127.94, 127.63, 126.96, 121.31, 113.89, 113.24, 86.52, 85.83, 84.53,
74.15, 73.75, 63.67, 62.57, 62.40, 55.34, 55.22, 43.26, 43.02, 40.85,
35.45, 24.75, 24.60, 24.45, 21.25, 18.90. 31P NMR (CDCl3): δ 146.72.
FAB-HRMS: calcd for C54H68N7O10P (M + Na+), 1028.4663; found,
1028.4630. 67, slow diastereomer, 1H NMR (CDCl3): δ 7.77 (1H, s),
7.48-7.12 (13H, m), 6.90-6.70 (6H, m), 6.30 (1H, dd, J ) 7.7, 5.5
Hz), 4.67 (1H, m), 4.38 (1H, m), 3.76 (3H, s), 3.74 (6H, s), 4.15-3.18
(10H, m), 2.84-2.60 (1H, m), 2.53-2.35 (1H, m), 1.3-0.8 (24H, m).
13C NMR (CDCl3): δ 179.44, 172.32, 160.46, 158.63, 155.82, 148.52,
148.15, 144.66, 135.71, 131.62, 130.04, 129.18, 128.12, 128.02, 127.95,
127.00, 121.54, 113.85, 113.21, 86.44, 85.65, 85.53, 84.12, 73.78, 73.51,
63.87, 61-62 (br m), 55.33, 55.24, 43.22, 43.98, 39.67, 35.74, 24.75,
24.61, 21.16, 18.86. 31P NMR (CDCl3): δ 144.07. FAB-HRMS: calcd
for C54H68N7O10P (M + Na+), 1028.4663; found, 1028.4628.
5′-O-(4,4′-dimethoxytrityl)-3′-O-(N,N-diisopropylamino)[2-[N-iso-
propyl-N-(4-methoxybenzoyl)amino]ethoxy]phosphinyl-2′-deoxythy-
midine (28) was synthesized analogously from 60 (468 mg, 1.1 mmol)
and 8 (545 mg, 1.0 mmol) and isolated by column chromatography to
afford 28, fast diastereomer (96 mg); 28, slow diastereomer (125 mg);
and their mixture (668 mg) totaled in 889 mg (97.6%) of 28 as a white
amorphous solid. 28, fast diastereomer, 1H NMR (CDCl3): δ 7.61 (1H,
s), 7.40-7.16 (12H, m), 6.90-6.74 (6H, m), 6.39 (1H, dd, J ) 8.0,
5.7 Hz), 4.63 (1H, m), 4.19 (1H, m), 3.77 (3H, s), 3.74 (6H, s), 3.95-
3.26 (9H, m), 2.45 (1H, ddd, J ) 13.0, 5.4, 1.5 Hz), 2.28 (1H, ddd, J
) 13.0, 7.2, 6.18 Hz), 1.35 (3H, s), 1.20-1.10 (18H, m). 13C NMR
(CDCl3): δ 171.88, 164.00, 150.43, 135.77, 111.16, 86.96, 85.98, 84.88,
73.71, 73.42, 63.45, 61.34, 61.04, 55.27, 43.15, 42.91, 40.19, 24.74,
24.62, 21.15, 11.74. 31P NMR: δ 147.64. FAB-HRMS: calcd for
C50H63N4O10P (M + Na+), 933.4180; found 933.4166. 28, slow
diastereomer, 1H NMR (CDCl3): δ 7.60 (1H, s), 7.41-7.18 (12H, m),
6.90-6.74 (6H, m), 6.43 (1H, br t), 4.66 (1H, m), 4.14 (1H, m), 3.79
(3H, s), 3.76 (6H, s), 3.96-3.25 (9H, m), 2.64-2.48 (1H, m), 2.42-
2.20 (1H, m), 1.40 (3H, s), 1.28-1.0 (18H, m). 13C NMR (CDCl3): δ
171.98, 163.98, 150.41, 135.79, 111.21, 86.93, 85.61, 85.52, 84.76,
73.96, 73.59, 63.42, 61.30, 60.99, 55.27, 43.09, 42.85, 40.19, 24.66,
24.54, 21.20, 11.73. 31P NMR: δ 147.96. FAB-HRMS: calcd for
C50H63N4O10P (M + Na+), 933.4180; found, 933.4166.
using a gradient from 25:70:5 ethyl acetate:hexane:triethylamine to 70:
25:5 ethyl acetate:hexane:triethylamine gave 66, fast diastereomer (1.46
g); 66, slow diastereomer (1.63 g); and their mixture (6.51 g) totaled
1
in 9.60 g (96.0%) of 66. 66, fast diastereomer, H NMR (CDCl3): δ
8.56 (1H, br s), 8.33 (1H, d, J ) 7.5 Hz), 7.88 (2H, m), 7.66-7.15
(15H, m), 6.90-6.74 (6H, m), 6.28 (1H, dd, J ) 5.9, 5.7 Hz), 4.67
(1H, m), 4.27 (1H, m), 3.79 (9H, s), 4.25-3.35 (9H, m), 2.79 (1H, m),
2.33 (1H, m), 1.30-1.10 (18H, m). 13C NMR (CDCl3): δ 171.87,
166.50, 162.08, 160.41, 158.71, 154.77, 144.82, 144.21, 135.58, 135.29,
133.29, 133.10, 130.20, 130.07, 129.46, 128.23, 128.05, 127.57, 127.14,
113.78, 113.34, 96.37, 87.28, 86.97, 86.20, 71.85, 71.51, 62.38, 61.41,
61.15, 55.51, 43.15, 42.92, 41.21, 24.76, 24.68, 21.30. 31P NMR
(CDCl3): δ 148.00. FAB-HRMS: calcd for C56H66N5O10P (M + Na+),
1022.4445; found, 1022.4487. 66, slow diastereomer, 1H NMR
(CDCl3): δ 8.64 (1H, br s), 8.29 (1H, d, J ) 7.3 Hz), 7.88 (2H, m),
7.66-7.10 (15H, m), 6.92-6.80 (6H, m), 6.32 (1H, t, J ) 5.8 Hz),
4.62 (1H, m), 4.23 (1H, m), 3.80 (9H, s), 4.18-3.35 (9H, m), 2.81
(1H, m), 2.31 (1H, m), 1.30-1.0 (18H, m). 13C NMR (CDCl3): δ
171.96, 166.65, 162.05, 160.4, 158.73, 154.70, 144.78, 144.14, 135.47,
135.26, 133.29, 133.11, 130.19, 129.47, 129.03, 128.05, 127.57, 127.17,
113.80, 113.33, 96.42, 87.21, 86.98, 86.01, 85.80, 72.52, 72.16, 62.51,
61.18, 60.91, 55.26, 43.14, 42.89, 41.43, 25.01, 24.87, 24.77, 24.51,
21.20. 31P NMR (CDCl3): δ 148.15. FAB-HRMS: calcd for
C56H66N5O10P (M + Na+),1022.4445; found, 1022.4488.
5-Methyl-5′-O-(4,4′-dimethoxytrityl) -2′-O-(2-methoxyethyl)- 3′-
O-(N,N-diisopropylamino)[2-[N-isopropyl-N-(4-methoxybenzoyl)-
amino]ethoxy]phosphinyluridine (68) was synthesized analogously
from 60 (0.98 g, 2.1 mmol) and 64 (1.21 g, 2.0 mmol). Column
separation using a gradient from 15:80:5 to 80:15:5 ethyl acetate:hexane:
triethylamine gave 68, fast diastereomer (0.15 g); 68, slow diastereomer
(0.41 g); and their mixture (1.33 g) totaled in 1.89 g (96.2%) of 68.
68, fast diastereomer, 1H NMR (CDCl3): δ 8.53 (1H, br s), 7.68 (1H,
s), 7.50-7.20 (11H, m), 6.93-6.78 (6H, m), 6.05 (1H, d, J ) 4.8 Hz),
4.48 (1H, ddd, J ) 10.0, 4.5, 4.5 Hz), 4.31 (1H, m), 4.25 (2H, t, J )
4.5 Hz), 3.81 (3H, s), 3.78 (6H, s), 3.33 (3H, s), 3.93-3.27 (12H, m),
1.33 (3H, s), 1.22-1.13 (12H, m), 1.08 (3H, d, J ) 7.3 Hz), 1.05 (3H,
d, J ) 7.3 Hz).13C NMR (CDCl3): δ 171.81, 163.99, 160.42, 158.74,
150.45, 144.39, 135.87, 135.51, 135.36, 129.41, 128.02, 127.70, 127.15,
113.78, 113.29, 110.78, 87.33, 86.96, 83.28, 81.61, 81.54, 72.32, 71.05,
70.79, 70.01, 62.31, 60.98, 60.68, 59.11, 55.27, 43.23, 42.99, 24.79,
24.68, 21.12, 11.70. 31P NMR (CDCl3): δ 149.35. FAB-HRMS: calcd
for C53H69N4O12P (M + Na+), 1007.4547; found, 1007.4529. 68, slow
diastereomer, 1H NMR (CDCl3): δ 8.15 (1H, br s), 7.66 (1H, s), 7.45-
7.20 (11H, m), 6.93-6.78 (6H, m), 6.08 (1H, d, J ) 5.0 Hz), 4.48
(1H, ddd, J ) 10.0, 4.5, 4.5 Hz), 4.27 (2H, t, J ) 4.9 Hz), 4.22 (1H,
m), 3.82 (3H, s), 3.78 (6H, s), 3.32 (3H, s), 3.92-3.26 (12H, m), 1.31
(3H, s), 1.20-1.15 (12H, m), 1.01 (6H, d, J ) 6.7 Hz).13C NMR
(CDCl3): δ 171.94, 164.02, 160.40, 156.76, 150.45, 144.27, 135.95,
135.42, 130.25, 129.56, 128.27, 127.18, 113.78, 113.49, 110.80, 87.44,
87.01, 83.16, 83.08, 82.37, 72.53, 70.60, 70.43, 70.26, 62.68, 61.80,
61.50, 59.11, 55.26, 43.02, 42.77, 24.76, 24.66, 21.19, 11.62. 31P NMR
(CDCl3): δ 149.56. FAB-HRMS: calcd for C53H69N4O12P (M + Na+),
1007.4547,; found, 1007.4529.
N6-Benzoyl-5′-O-(4,4′-dimethoxytrityl)-3′-O-(N,N-diisopropyl-
amino)[2-[N-isopropyl-N-(4-methoxybenzoyl)amino]ethoxy]phos-
phinyl-2′-deoxyadenosine (65) was synthesized analogously from 60
(5.14 g, 11.0 mmol) and 61 (6.58 g, 10.0 mmol). Column separation
using a gradient from 25:70:5 ethyl acetate:hexane:triethylamine to 96:5
ethyl acetate:triethylamine gave 65, fast diastereomer (1.88 g); 65, slow
diastereomer (1.01 g); and their mixture (6.74 g) totaled in 9.63 g
1
(94.0%) of 65. 65, fast diastereomer, H NMR (CDCl3): δ 8.99 (1H,
br s), 8.73 (1H, s), 8.20 (1H, s), 8.06-7.96 (2H, m), 7.65-7.45 (3H,
m), 7.45-7.15 (11H, m), 6.9-6.7 (6H, m), 6.53 (1H, dd, J ) 6.2, 6.4
Hz), 4.77 (1H, m), 4.41 (1H, m), 3.80 (3H, s), 3.76 (6H, s), 4.2-3.3
(9H, m), 3.04-2.86 (1H, m), 2.72-2.57 (1H, m), 1.3-1.1 (18H, m).
13C NMR (CDCl3): δ 171.91, 164.67, 160.44, 158.55, 152.56, 151.53,
149.50, 144.58, 142.25, 135.73, 133.81, 132.75, 130.08, 129.46, 128.87,
128.23, 127.90, 126.92, 123.59, 113.79, 113.17, 86.52, 86.38, 85.09,
74.65, 73.35, 63.56, 61.45, 61.12, 55.33, 55.24, 43.36, 42.93, 39.64,
24.78, 24.68, 21.20. 31P NMR (CDCl3): δ 147.30. FAB-HRMS: calcd
for C57H66N7O9P (M + Na+), 1046.4557,; found, 1046.4515. 65, slow
diastereomer, 1H NMR (CDCl3): δ 9.01 (1H, br s), 8.72 (1H, s), 8.18
(1H, s), 8.06-7.94 (2H, m), 7.65-7.45 (3H, m), 7.45-7.15 (11H, m),
6.9-6.7 (6H, m), 6.54 (1H, dd, J ) 7.3, 6.4 Hz), 4.76 (1H, m), 4.32
(1H, m), 3.80 (3H, s), 3.76 (6H, s), 4.2-3.3 (9H, m), 3.04-2.84 (1H,
m), 2.80-2.62 (1H, m), 1.3-1.05 (18H, m). 13C NMR (CDCl3): δ
171.99, 164.68, 160.44, 158.55, 152.51, 151.53, 149.49, 144.55, 141.77,
135.70, 133.81, 132.73, 130.07, 129.39, 128.21, 127.88, 127.74, 126.92,
123.59, 128.85, 113.80, 113.16, 86.50, 86.11, 86.01, 85.02, 74.14, 73.76,
63.60, 61.34, 61.04, 55.33, 55.24, 43.19, 42.95, 39.43, 24.73, 24.61,
21.23. 31P NMR (CDCl3): δ 147.94. FAB-HRMS: calcd for
C57H66N7O9P (M + Na+), 1046.4557; found, 1046.4515.
O,O-Diethyl-O-[2-[N-isopropyl-N-(4-methoxybenzoyl)amino] eth-
yl] phosphate (52). O,O-Diethyl phosphorochloridate (777 mg, 4.5
mmol) in CH2Cl2 (3 mL) was added to a stirred solution of 17 (712
mg, 3.0 mmol) and 1H-tetrazole (210 mg, 3.0 mmol) in pyridine (2
mL) and CH2Cl2 (3 mL) under argon atmosphere at ambient temper-
ature. The mixture was stirred for 30 min, evaporated, and treated with
Na2CO3 (5% aq, 25 mL). The product was extracted with ethyl acetate
(3 × 50 mL). Extracts were washed with brine (2 × 50 mL), dried
over Na2SO4, and evaporated. Purification on a silica gel column eluting
with a gradient of ethanol in CH2Cl2 (0-3%) afforded 52 (894 mg,
1
79.8%) as a colorless oil. H NMR (CDCl3): δ 7.35-7.20 (2H, m),
6.95-6.85 (2H, m), 4.3-4.05 (7H, m), 3.80 (3H, s), 3.57 (2H, t, J )
6.3 Hz), 1.31 (6H, t, J ) 7.1 Hz), 1.15 (6H, d, J ) 6.6 Hz). 13C NMR
(CDCl3): δ 172.1, 160.6, 129.1, 128.3, 113.8, 64.7, 64.6, 63.9, 63.8,
55.3, 50.0, 41.3, 41.2, 21.1, 16.2, 16.1. 31P NMR (CDCl3): δ -1.32.
Anal. Calcd for C17H28NO6P: C, 54.68; H, 7.56; N, 3.75. Found: C,
54.52; H, 7.51; N, 3.68.
N4-Benzoyl-5′-O-(4,4′-dimethoxytrityl)-3′-O-(N,N-diisopropyl-
amino)[2-[N-isopropyl-N-(4-methoxybenzoyl)amino]ethoxy]phos-
phinyl-2′-deoxycytidine (66) was synthesized analogously from 60
(5.14 g, 11.0 mmol) and 62 (6.18 g, 10.0 mmol). Column separation
Hydrolysis of 52 under Neutral Conditions and Preparation of

Phosphate Protecting Group
J. Am. Chem. Soc., Vol. 123, No. 5, 2001 793
2-[(4-Methoxybenzoyl)oxy]-N-isopropylethanaminium trifluoroace-
tate (40‚CF3COOH). A solution of 52 (224 mg, 0.6 mmol) in water
(50 mL) was kept at 25 °C, with progress of the hydrolysis followed
by HPLC. After 50 h, 52 disappeared (>98% conversion), and the
reaction mixture contained an intermediate, 47, and product, 54, in a
46:54 ratio. At this moment, part of the reaction mixture (10 mL) was
withdrawn, acidified with TFA to pH 2, and separated by HPLC. After
evaporation of the collected fractions, 47 trifluoroacetate was obtained
as a 90:10 mixture with 54 trifluoroacetate. LC/MS: calcd for (C13H18-
168.4, 164.3 (q), 132.5, 121.7, 114.7, 61.3, 56.2, 51.8, 44.1, 18.7. Anal.
Calcd for C15H20F3NO5: C, 51.28; H, 5.74; N, 3.99. Found: C, 51.09;
H, 5.71; N, 3.88.
2-Phenyl-2-thiazoline (59). Compound 29 (171 mg, 0.2 mmol) and
1H-tetrazole (63 mg, 0.9 mmol) in MeCN (2 mL) were reacted with a
detritylated commercial DMT-T derivatized Tentagel N (45 µmol; 0.208
mmol g-1) for 10 min. The solid support was washed with MeCN (3
mL), treated with 3H-1,2-benzodithiol-3-one 1,1-dioxide (60.0 mg, 0.3
mmol)22 in MeCN (3 mL) for 1 min, and detritylated with a solution
of dichloroacetic acid (3% in CH2Cl2). After washing with MeCN (3
mL), the solid support was suspended in 80% aq MeCN (5 mL) and
stirred overnight. The liquid phase was withdrawn, and the solid support
was alternately washed with ethyl acetate and 5% aq NaHCO3 (3 ×
10 mL each). The solution in 80% aq MeCN and aqueous washings
were combined and extracted with ethyl acetate (3 × 10 mL). The
ethyl acetate washings and extracts were dried over Na2SO4 and
evaporated to give 59 (ca. 90% pure, 6.3 mg, 85.7%). 1H NMR
(CDCl3): δ 7.89-7.80 (2H, m), 7.52-7.34 (3H, m), 4.46 (2H, t, J )
8.4 Hz), 3.42 (2H, t, J ) 8.4 Hz). 13C NMR (CDCl3): δ 163.7, 131. 4,
1
NO2)+, 220.29; found M+, 220.30. H NMR (1% TFA in D2O): δ
7.82 (2H, m), 7.20 (2H, m), 5.03 (2H, t, J ) 9.6 Hz), 4.69 (1H, sept,
J ) 6.6 Hz), 4.35 (2H, t, J ) 9.8 Hz), 3.93 (3H, s), 1.43 (6H, d, J )
6.6 Hz). 13C NMR (1% TFA in D2O): δ 171.3, 163.3 (q), 132.8, 119.8,
115.6, 113.0, 70.9, 56.4, 51.8, 46.0, 19.7.
After the disappearance of 47 (7 days) the mixture was evaporated,
coevaporated with MeCN, and dried to give >98% HPLC-pure,
1
crystalline 54 diethyl phosphate (186 mg, 99%), mp 102-103 °C. H
NMR (DMSO-d6): δ 9.35 (2H, br s); 8.03-7.99 (2H, m), 7.06-7.02
(2H, m), 4.45 (2H, t, J ) 5.2 Hz), 3.82 (3H, s), 3.65(4H, pent, J ) 7.2
Hz), 3.30-3.20 (3H, m), 1.23 (6H, d, J ) 6.4 Hz), 1.07 (6H, t, J )
128.7, 128.6, 65.5, 33.9. IR (Nujol): ν 1610 cm-1
.
1
7.2 Hz). H NMR (D2O): δ 7.93-7.89 (2H, m), 6.98-6.93 (2H, m),
Acknowledgment. We thank Professor M. E. Jung (Uni-
versity of California, Los Angeles) and Professor G. Just (McGill
University, Montreal, Canada) for helpful discussions.
4.48-4.44 (2H, m), 3.79 (4H, dt, J ) 7.2 Hz), 3.77 (3H, s), 3.40 (1H,
sept, J ) 6.4 Hz), 3.37 (2H, t, J ) 5.2 Hz), 1.22 (6H, d, J ) 6.4 Hz),
1.12 (6H, J ) 7.2 Hz). 13C NMR (D2O): δ 168.4, 164.3, 132.6, 121.8,
114.7, 62.9, 62.8, 61.3, 56.3, 51.8, 44.1, 18.7, 16.3. 31P NMR (DMSO-
d6): δ 3.21. 31P NMR (D2O): δ 1.82.
Supporting Information Available: Materials and methods,
HPLC techniques. Synthetic procedures and spectral data for
12, 16, 21-31, 49, and 58. 31P NMR spectra for solid-support-
Crude 54 was dissolved in water (10 mL) and acidified with TFA
(1% aq; 1 mL). The product was isolated by HPLC to afford crystalline
54 trifluoroacetate (147 mg, 88%), mp 123-125 °C. 1H NMR (DMSO-
d6): δ 9.18 (2H, br s), 8.02-7.98 (2H, m), 7.08-7.03 (2H, m), 4.44
(2H, t, J ) 5.2 Hz), 3.83 (3H, s), 3.44-3.31 (3H, m), 1.24 (6H, d, J
1
bound compounds 37, 39, 41, 43, 44, 45, and 51. H and 13C
NMR spectra for 47 and 59. ESMS Data and reverse-phase
HPLC profiles for oligonucleotides 69-76. This material is
available free of charge via the Internet at http://pubs.acs.org.
1
) 6.4 Hz). H NMR (D2O): δ 7.93-7.89 (2H, m), 6.98-6.93 (2H,
m), 4.48-4.44 (2H, m), 3.77 (3H, s), 3.40 (1H, sept, J ) 6.4 Hz), 3.38
(2H, t, J ) 5.0 Hz), 1.24 (6H, d, J ) 6.4 Hz). 13C NMR (D2O): δ
JA0016396