Synthesis of purine and pyrimidine 3′-amino-3′-deoxy- and 3′-amino-2′,3′-dideoxyxylonucleosides

Article abstract of DOI:10.1021/jo9606259

A general procedure to obtain the 3′-aminoxylonucleosides 13a,b and 17a,b is presented. The synthetic scheme is based on the 5′ directed intramolecular nucleophilic substitution at the 3′-activated position of the nucleoside. The approach of the incoming group to this position takes place regio- and stereoselectively from the most hindered face of the nucleoside. The methodology presented is applicable to ribonucleosides and 2′-deoxyribonucleosides, regardless of their nitrogenated base.

Full text of DOI:10.1021/jo9606259

6980
J . Org. Chem. 1996, 61, 6980-6986
Syn th esis of P u r in e a n d P yr im id in e 3′-Am in o-3′-d eoxy- a n d
3′-Am in o-2′,3′-d id eoxyxylon u cleosid es
Luis F. Garc´ıa-Alles, J ulia Magdalena, and Vicente Gotor*
Departamento de Quı´mica Orga´nica e Inorga´nica, Facultad de Qu´ımica, Universidad de Oviedo,
33071 Oviedo, Spain
Received April 3, 1996X
A general procedure to obtain the 3′-aminoxylonucleosides 13a ,b and 17a ,b is presented. The
synthetic scheme is based on the 5′ directed intramolecular nucleophilic substitution at the 3′-
activated position of the nucleoside. The approach of the incoming group to this position takes
place regio- and stereoselectively from the most hindered face of the nucleoside. The methodology
presented is applicable to ribonucleosides and 2′-deoxyribonucleosides, regardless of their nitro-
genated base.
In tr od u ction
2-carbonyl group of the pyrimidine base.5 For ribo-
nucleosides, an added complication is found since both
the 3′- and the 2′-positions are capable of undergoing
substitution. Despite the fact that the 3′- position is
generally the most reactive one, in most of the reported
cases small amounts of the 2′-substituted products were
also found.2e,g
Polyoxin and puromycin are important examples of
compounds which can be classified as aminosugar nu-
cleosides. These and some other derivatives are known
to possess strong antibacterial, anticancer, and biosyn-
thetic inhibitory properties.1
Considerable effort has been devoted to the preparation
of this kind of compounds. Generally, the most successful
procedures have been those describing their synthesis via
reduction of the sugar-substituted azido analogs.2 Some
other approaches have also been reported.3 Among them,
two successful strategies took advantage of the presence
of a neighboring hydroxyl group to deliver intramolecu-
larly the amine nucleophile to the new position.3c,d
However, all of these procedures are strongly depend-
ent on the nature of the nitrogenated base present in the
starting nucleoside. An illustrating example can be the
work reported by Matsuda et al., where treatment of an
analog of 3′-O-mesylthymidine with NaN3 gave rise to a
mixture of several compounds coming from intervention
of the base in different ways.4
The objective is even more complicated when one
wishes to obtain a 3′-aminoxylonucleoside. In this case
the nucleophile must approach the 3′-position from the
most hindered side of the nucleoside: the â-face. While
for purine nucleosides, attacks from this face are still
possible, pyrimidine nucleosides usually resist such
transformation, since the incoming nucleophile must
compete with the favorable intramolecular attack of the
It would be important, from our point of view, to
develop a general methodology which could be applied
to obtain 3′-amino-3′-deoxy- and 3′-amino-2′,3′-dideoxy-
xylonucleosides. There are two main features of the
procedure we describe in this paper. (1) It has been
extended to ribonucleosides and 2′-deoxyribonucleosides,
containing both purine and pyrimidine bases in their
structures. (2) The amino substitution is delivered to the
3′-position from the most hindered face of the nucleoside
with total regio- and stereoselectivity.
Resu lts a n d Discu ssion
Syn t h esis of 3′-Am in o-2′,3′-d id eoxyxylon u cleo-
sid es 13 a ,b. The synthetic strategy we have developed
is based on the introduction of the amino function at the
3′-position from the neighboring 5′-position of the sugar
skeleton of the nucleoside.6 To achieve this goal, we
functionalized the 5′-position with a group able to release
the amino function at the 3′-position: the carbamoyl
group had been previously used with a similar purpose.3c,7
On the other hand, it was also necessary to activate the
3′-position for nucleophilic substitution to take place.
Bearing all this in mind, the first step was to obtain
compounds 4-8 (see Scheme 1).
Candida antarctica lipase (CAL) was used as the
synthetic tool which permits one to obtain carbonates 2
and 3 from the starting 2′-deoxyribonucleosides 1 (B )
Th, Ad) with high regioselectivity. The potential power
of this and other enzymes in nucleoside chemistry had
been previously studied in our group.8 We have used
acetone O-[(phenyloxy)carbonyl]oxime for thymidine 1,
since the regioselectivity of the process is better than that
X Abstract published in Advance ACS Abstracts, August 15, 1996.
(1) (a) Suhadolnik, R. J . Nucleoside Antibiotics; Wiley-Inter-
science: New York, 1970; pp 1-95. (b) Suhadolnik, R. J . Nucleosides
as Biological Probes; Wiley-Interscience: New York, 1979; pp 93-104,
146, 195. (c) Armstrong, V. W.; Eckstein, F. Eur. J . Biochem. 1976,
70, 33. (d) Iwai, Y.; Nakagawa, A.; Nagai, A. J . Antibiot. 1979, 32,
1367.
(2) (a) Horwitz, J . P.; Tomson, A. J .; Urbanski, J . A.; Chua, J . J .
Org. Chem. 1962, 27, 3045. (b) Horwitz, J . P.; Chua, J .; Noel, M. J .
Org. Chem. 1964, 29, 2076. (c) Verheyden, J . P. H.; Wagner, D.; Moffatt,
J . G. J . Org. Chem. 1971, 36, 250. (d) Glinski, H. P.; Khan, M. S.;
Kalamas, R. L.; Sporn, M. B. J . Org. Chem. 1973, 38, 4299. (e) Robins,
M. J .; Fouron, Y.; Mengel, R. J . Org. Chem. 1974, 39, 1564. (f) Lin, T.
S.; Prusoff, W. H. J . Med. Chem. 1978, 21, 109. (g) J ack Chen, Y.-C.;
Hansske, F.; J anda, K. D.; Robins, M. J . J . Org. Chem. 1991, 56, 3410.
(h) Herdewijn, P.; Balzarini, J .; Pauwels, R.; J anssen, G.; Van Aerschot,
A.; De Clercq, E. Nucleosides Nucleotides 1989, 8, 1231-1257.
(3) (a) Miller, N.; Fox, J . J . J . Org. Chem. 1964, 29, 1772. (b)
Mitsunobu, O.; Takizawa, S.; Morimato, H. J . Am. Chem. Soc. 1976,
98, 7858. (c) Samano, M. C.; Robins, M. J . Tetrahedron Lett. 1989, 30,
2329. (d) McGee, D. P. C.; Sebesta, D. P.; O’Rourke, S. S.; Mart´ınez,
R. L.; J ung, M. E.; Piecken, W. A. Tetrahedron Lett. 1996, 37, 1995.
(4) Matsuda, A.; Watanabe, K. A.; Fox, J . J . J . Org. Chem. 1980,
45, 3274.
(5) Many syntheses of 3′- (or 2′-) amino ribonucleosides have been
reported to proceed via the 2,3′-(or the 2,2′-)anhydronucleosides.2c,d,3a,d
(6) This kind of methodology had been previously used to invert
configuration at the 3′-position of purine nucleosides: Herdewijn, P.
A. M. J . Org. Chem. 1988, 53, 5050.
(7) (a) Minami, N.; Ko, S. S.; Kishi, Y. J . Am. Chem. Soc. 1982, 104,
1109. (b) Roush, W. R.; Adam, M. A. J . Org. Chem. 1985, 50, 3752.
(8) (a) Mor´ıs, F.; Gotor, V. J . Org. Chem. 1992, 57, 2490. (b) Mor´ıs,
F.; Gotor, V. Tetrahedron 1992, 48, 9869. (c) Garc´ıa-Alles, L. F.; Gotor,
V. Tetrahedron 1995, 51, 307.
S0022-3263(96)00625-1 CCC: $12.00 © 1996 American Chemical Society

Synthesis of Purine and Pyrimidine Nucleosides
J . Org. Chem., Vol. 61, No. 20, 1996 6981
Sch em e 1
Sch em e 3
reaction with 6a (when R2 ) Ms). An intermediate
situation was found when 5a ,b (R1 ) allyl) were used.
In this case, the pathway followed depends dramatically
on the R2 group, resulting almost exclusively 9b from 5a
(path a in Scheme 2) and 10a from 5b (path b). The
reasons for this behavior have not been studied in detail
at this point.
Once we knew the optimum conditions to perform the
key step of the synthetic strategy, we completed the
synthesis of the 3′-amino-3′-deoxyxylothymidine 13a as
follows (see Scheme 3). Carbamates 10a ,b were decar-
bonylated with LiOH, yielding 3′-N-allyl- and 3′-N-
benzylamino-3′-deoxyxylothymidine 12a ,b. Finally, 13a
was quantitatively formed when 12b was submitted to
catalytic transfer hydrogenation, using formic acid in
presence of palladium black.10 This amino nucleoside
was also obtained from 12a using (PPh3)3RhCl,11 but
with poorer results (longer reaction time and approxi-
mately 60% yield by TLC). Data for product 13a ,
obtained following this synthetic scheme, was shown to
be in complete accord with data reported by Matsuda et
al. for the compound synthesized following an alternative
strategy.4
Sch em e 2
As far as the synthesis of the adenosine derivative 13b
is concerned, the intramolecular displacement step at the
3′-position does not present the aforementioned problem
of the pyrimidine nucleosides. When 7a ,b (R1 ) H) were
submitted to reaction (NaH/THF), the expected amino
sugar nucleoside was not obtained. However, the 5′-O-
(N-benzylcarbamoyl)-3′-O-methylsulfonyl-2′-deoxyade-
nosine derivative 8a yielded 11 with good yield (Scheme
3), together with a small amount of a second unknown
product (<10%). Formation of this undesired product
was avoided using the tosyl derivative 8b, showing once
again the suitability of this leaving group. The synthetic
scheme was then completed from 11 to give 13b.2h
The identification of the 3′-amino xylonucleoside de-
rivatives was accomplished by 13C-NMR spectroscopy.
The main difference was a ca. 15-20 ppm upfield shift
of the peak corresponding to the 3′-carbon atom with
regard to the starting 2′-deoxyribonucleosides. Complete
1H- and 13C-NMR spectral data are given in Tables 1 and
2, respectively.
On the other hand, the stereochemistry of the sugar
ring was confirmed by means of NOE difference experi-
ments performed on the rigid cyclic carbamates 10b and
11. As can be seen in Table 3 (see also Chart 1), 3.2-
5.6% and 5.0-9.2% enhancements of the H3′ signal were
observed when the H4′ and H2′R signals of 10b and 11
were irradiated. This is a clear indication of the relative
disposition of these hydrogen atoms toward the same side
of the sugar framework.
obtained using the vinyl carbonate, and in consequence
the final yield of 2 is increased. Treatment of 2 or 3 with
the corresponding amine yields the 5′-carbamate,8c and
finally, reaction with mesyl or tosyl chloride gave the 3′-
activated compounds 4a ,b, 5a ,b, and 6a ,b for thymidine
(Th) and 7a ,b and 8a ,b for adenosine (Ad).
As previously mentioned, it is well-known that intro-
duction of nucleophiles at the 3′-activated position of
pyrimidine nucleosides is difficult to perform, since the
main product usually results to be the 2, 3′-anhydro-
nucleoside. For this reason, we studied the conditions
which would favor formation of products 10 (through path
b in Scheme 2) over formation of the 2,3′-anhydro
derivatives 9. From the very beginning it seemed clear
that using basic aqueous media the 2,3′-anhydro deriva-
tive was always obtained as the main product of reaction,
regardless of the R1 and R2 groups present in the starting
material. This was in complete agreement with other
reported results.9 However, we found that product 10
could be obtained when NaH in anhydrous THF was used
as the proton abstraction system,3c,7b and both the R1 and
R2 groups were adequately selected. Thus, whereas 9a
was the only product obtained from 4a ,b (R1 ) H), 6a ,b
(R1 ) Bn) yielded almost quantitatively 10b. Small
amounts of product 9c were also isolated from the
(9) (a) Fox, J . J .; Miller, N. C. J . Org. Chem. 1963, 28, 936. (b)
Horwitz, J . P.; Chua, J .; Urbanski, J . A.; Noel, M. J . Org. Chem. 1963,
28, 942.
(10) ElAmin, B.; Anantharamaiah, G. M.; Royer, G. P.; Means, G.
E. J . Org. Chem. 1979, 44, 3442.
(11) Laguzza, B. C.; Ganem, B. Tetrahedron Lett. 1981, 22, 1483.

6982 J . Org. Chem., Vol. 61, No. 20, 1996
Ta ble 1. 1H-NMR Sp ectr a l Da ta (on ly n on in ter ch a n ga ble sign a ls)a in δ (p p m )
Garc´ıa-Alles et al.
base ring
sugar moiety
H2′â H3′
product H6(H2) H5(H8)
Me
H1′
H2′R
H4′
H5′(2 H)
other
10a
7.25 (s)
1.88 (s) 6.08 (t)
1.90 (s) 6.03 (t)
2.73 (m) 2.06 (m) 4.12 (m) 4.50 (m) 4.41-4.19 (m) 5.73 (m), 5.2-5.3 (2H,m),
4.14 (m), 3.6 (m)
2.54 (m) 1.95 (m) 3.99 (m) 4.39 (m) 4.45-4.21 (m) 7.31 (5H, m), 4.86 (d), 4.16 (d)
10b
11
12a
7.25 (s)
8.21 (s) 8.27 (s)
8.20 (s)
6.35 (dd) 2.85 (m) 3.02 (m) 4.28 (m) 4.67 (m) 4.45 (m)
2.07 (s) 6.30 (t)
7.35 (5H, m), 4.86 (d), 4.32 (d)
6.15 (m), 5.6-5.4 (2H, m),
4.02 (m), 3.55 (m)
2.83 (m) 2.35 (m) 3.65 (m) 4.35 (m) 4.15 (m)
12b
14
8.35 (s)
2.05 (s) 6.29 (t)
6.30 (t)
2.72 (m) 2.23 (m) 3.79 (m) 4.30 (m) 4.12 (m)
2.71 (m) 2.55 (m) 3.65 (m) 4.17 (m) 3.88 (m)
7.50 (5H, m), 3.95 (2H, m)
7.45 (5H, m), 3.85 (2H, m)
8.22 (s) 8.57 (s)
13b2h 7.80 (s) 7.89 (s)
6.02 (dd) 2.32 (m) 2.90 (m) 3.97 (m) 4.12 (m) 3.87 (m)
22
24
7.41 (d) 6.21 (d)
8.31 (s) 8.48 (s)
7.66 (d) 5.80 (d)
8.22 (s) 8.32 (s)
8.33 (d) 5.90 (d)
8.25 (s) 8.55 (s)
7.74 (d) 5.79 (d)
6.58 (d) 5.54 (d)
6.40 (s)
5.73 (d)
5.94 (d)
6.01 (d)
5.91 (d)
5.60 (d)
5.59 (d)
4.73 (m) 4.62 (m) 4.23 (m)
4.65 (d) 4.43 (d) 4.48 (m) 4.50-4.12 (m) 7.35 (5H, m), 4.25 (2H, m)
7.45 (5H, m), 4.40 (2H, m)
25a
25b
26a
26b
17a
4.39 (m) 3.97 (dd) 4.78 (m) 4.45 (m)
5.11 (m) 4.04 (dd) 4.83 (m) 4.45 (m)
4.57 (dd) 3.69 (dd) 4.51 (m) 4.08 (m)
4.76 (dd) 3.49 (dd) 4.35 (m) 3.85 (m)
4.65 (dd) 3.92 (dd) 4.35 (m) 3.95-3.75 (m)
4.52 (dd) 3.68 (dd) 4.27 (m) 3.77 (m)
7.44 (5H, m), 4.83 (d), 4.50 (d)
7.43 (5H, m), 4.91 (d), 4.40 (d)
7.46 (5H, m), 4.15 (2H, m)
7.45 (5H, m), 4.00 (2H, m)
17b2e 7.66 (s) 7.91 (s)
a
Solvents: DMSO-d6, except for 10a and 10b (CDCl3), 12a , 12b, 22, and 26a (CD3OD), and 13b, 17a , and 17b (D2O). 1H-NMR signals
of the sugar moiety were assigned using selective homodecoupling experiments.
Ta ble 2. 13C-NMR Ch em ica l Sh iftsa in δ (p p m )
base ring
C5
sugar moiety
C2′ C3′ C4′
product
C2
C4
C6
Me/C8
C1′
C5′
other
10a
10b
11
150.34 163.87 111.18 133.98 12.21
150.30 163.81 111.41 133.86 12.37
153.08 149.53 119.64 156.48 138.56 83.81 36.75 57.04 74.60 65.72 153.92, 137.35, 128.90, 128.09, 127.70, 49.97
83.79 38.19 56.03 74.60 65.96 153.37, 131.80, 118.99, 49.53
83.69 38.09 55.73 74.62 66.18 154.15, 135.49, 128.71, 128.18, 127.99, 50.36
12a
12b
14
13a
13b
22
152.67 166.51 111.98 139.99 12.76
152.94 167.00 111.52 139.66 13.08
152.47 148.90 119.32 156.28 140.69 82.93 37.46 57.94 81.83 61.16 140.17, 128.30, 128.09, 126.80, 51.73
150.79 164.13 110.15 137.08 12.40 83.08 35.67 50.38 78.24 59.29
152.27 147.40 119.39 155.57 142.05 84.36 36.97 51.92 80.19 60.04
86.84 35.84 59.25 80.43 61.49 132.24, 122.12, 50.91
85.77 39.28 59.53 83.23 62.81 141.56, 129.99, 129.84, 128.69, 53.75
162.85 176.24 110.85 139.66
93.00 91.87 77.30 89.09 65.75 158.64, 141.10, 130.33, 129.06, 129.00, 46.27
24
152.91 149.14 119.03 156.20 139.58 82.37 57.58 58.30 78.31 63.66 156.05, 139.94, 128.32, 127.07, 126.85, 43.88
25a
25b
26a
27b
17a
17b
150.56 163.26 102.14 139.74
152.71 149.37 119.25 156.11 138.72 88.65 77.45 63.19 72.69 65.75 153.40, 136.91, 128.46, 127.56, 127.23, 49.47
152.72 166.34 102.78 143.94 91.53 81.65 65.53 79.35 62.03 140.05, 129.77, 128.69, 53.05
152.41 148.97 119.34 156.28 140.83 88.74 80.75 64.08 78.09 61.03 140.47, 128.26, 128.00, 126.74, 51.38
152.83 166.22 103.88 144.29 92.43 77.99 58.81 77.24 61.27
152.25 147.73 119.03 155.28 141.60 89.32 79.57 57.47 77.92 60.99
89.59 77.83 63.43 72.94 65.99 153.48, 137.17, 128.62, 127.69, 127.45, 49.92
a
Solvents: DMSO-d6, except for 10a and 10b (CDCl3), 12a , 12b, 17a , 22, and 26a (CD3OD), and 13b and 17b (D2O). 13C-NMR signals
1
of the sugar moiety were identificated on the basis of their J C-H coupling constants (see: Seela, F.; Stecker, H. Helv. Chim. Acta 1985,
68, 563) and by means of INEPT experiments. Other signals were assigned in comparison with those reported in 13C-NMR spectral data;
Verlag-Chemie: Weinheim, 1981.
Ta ble 3. Resu lts of 1H-NMR NOE Exp er im en tsa
Ch a r t 1
signal
signal
measured
compound
irradiated
enhanced
enhancement (%)
10b
H2′R
H4′
H2′R
H4′
H4′
H4′
H2′
H3′
H3′
H3′
H3′
H3′
H3′
H3′
5.0
3.2
9.2
5.6
7.0
4.3
0.8
11
25a
25b
a
Solvents: DMSO-d6, except for 10b (CDCl3 degassed by the
freeze-pump-thaw procedure). For signal assignment, see:
Table 1. H2′R in 10b and 11 presented NOE effect when the H1′
signal was saturated (5 and 4.1%, respectively).
We tried to force the still 3′-activated 2,2′-anhydronucleo-
sides 18 (R ) Ms or Ts) to react, adding a second
equivalent of NaH, but we observed the appearance of
two other products, containing a 2′,3′-lyxo-epoxide struc-
ture and differing only in the base substitution (one of
them was shown to be compound 19). The 2′,3′-epoxy-
lyxosyl structure probably arises from cleavage of the
2,2′-anhydronucleoside bond, as it has been proposed by
other authors.12
Syn t h esis of 3′-Am in o-3′-d eoxyxylon u cleosid es
17a ,b. We tried to prepare the 3′-amino-3′-deoxyxylo-
nucleosides in a manner similar to that followed for the
synthesis of compounds 13a ,b. For this reason, we
synthesized the corresponding 5′-O-(N-benzylcarbamoyl)-
2′,3′-sulfonylated derivatives 15a ,b (see Scheme 4).
The uridine derivatives 15a (R ) Ms or Ts) always
gave rise, under basic conditions (NaH in THF), to the
2,2′-anhydronucleosides 18 instead of the cyclic carbam-
ates 16a . These 2,2′-anhydronucleosides are so easily
formed that no chance for fomation of products 16a exists.
(12) Yung, N. C.; Burchenal, J . H.; Fecher, R.; Duschinsky, R.; Fox,
J . J . J . Am. Chem. Soc. 1961, 83, 4060.

Synthesis of Purine and Pyrimidine Nucleosides
J . Org. Chem., Vol. 61, No. 20, 1996 6983
Sch em e 4
Sch em e 5
Sch em e 6
On the other hand, when the same methodology was
applied to the purine nucleosides, which did not present
the problem of the easy anhydronucleoside formation, we
could see that the reactions were even more complicated.
Thus, when 15b (R ) Ms or Ts) were treated with NaH
in THF, mixtures of several unidentified products ap-
peared. The 13C-NMR spectra of those mixtures revealed
the presence of several carbonyl signals around δ 206.13
In an attempt to determine whether the lack of
formation of product 16b was due to conformational
reasons, we decided to perform the reaction in other
anhydrous organic solvents which could drive the con-
formational equilibrium of the sugar ring toward a more
favorable situation for intramolecular nucleophilic sub-
stitution. The conformation of the sugar ring of 15a ,b
(with R ) Ms or Ts) was roughly studied by measuring
3
the J H-H coupling constants between several hydrogen
atoms of the carbohydrate in different deuterated sol-
vents (CDCl3, acetone-d6, CD3OD, and DMSO-d6). This
3
brief study showed that, for instance, the J H1′-H2′ cou-
pling constant increased with the polarity of the solvent
(from around 2 Hz in CDCl3 to 5-7 Hz in DMSO-d6),
regardless of the R group and the base present. This may
be interpreted to be due to a shift of the averaged sugar
conformation toward dispositions closer to the C2′ endo
conformation (where the dihedral angle between both the
H1′ and H2′ is closer to the 180° value). This conforma-
tion would favor the intramolecular nucleophilic attack
from the neighboring 5′-position. However, when the
reactions were performed in these solvents, the same
results as those in THF were obtained.
A new strategy for the synthesis of the 3′-amino-3′-
deoxyxylonucleosides had to be devised. Some reported
results served as a good guide. In 1958, Brown et al.
showed that the reaction of 2,2′-anhydrouridine with
sodium ethyl sulfide gave 3′-deoxy-3′-ethylthioxylouri-
dine, this result presumably arising from intervention
of the 2′,3′-riboepoxide.14 Later, it was proposed that the
2,2′-anhydronucleoside and the riboepoxide may be re-
garded as tautomers, with the equilibrium shifted by the
basicity of the medium (see Scheme 5).15
neighboring 5′-O-(N-benzylcarbamoyl) group would be to
drive that equilibrium toward the 2′,3′-riboepoxide, which
subsequently would suffer intramolecular cyclization
from the 5′-position to give the desired product. In fact,
when the 2,2′-anhydronucleoside 22, synthesized from
the carbamate 21,16 was heated with a slight excess of
sodium hydride in anhydrous DMF, the desired product
25a was formed in a fast and quantitative manner (see
Scheme 6). The structure of this cyclic carbamate was
confirmed by 1H- and 13C-NMR spectra (see Tables 1 and
2). In order to corroborate the correct assignment of the
13C-NMR spectra, a 1H-13C heteronuclear correlation
experiment was performed. This showed the crosspeak
1
between the 13C signal at 63 ppm and the H signal at
3.97 ppm corresponding to the H3′ hydrogen atom.
Furthermore, the relative configuration of the sugar was
confirmed by NOE experiments (see Table 3), as for the
2′-deoxynucleosides. The H3′ signal was enhanced by
7.0% when H4′ signal was irradiated. However, irradia-
tion of H2′ gave no enhancement of H3′ signal, indicating
their relative trans disposition.
Taking this into account, we believed that a good way
to deliver the amino function to the 3′-position from the
In the case of purine ribonucleosides, the possibility
of establishing the tautomeric equilibrium 2′,3′-riboep-
(13) 2′-Keto-3′-deoxy nucleosides have been described to be formed
from compounds similar to 15a ,b through base-catalyzed desulfony-
loxylation-desulfonylation reactions: (a) Sasaki, T.; Minamoto, K.;
Suzuki, H. J . Org. Chem. 1973, 38, 598. (b) Sasaki, T.; Minamoto, K.;
Tanizawa, S. J . Org. Chem. 1973, 38, 2896. (c) Kawana, M.; Kuzuhara,
H. Tetrahedron Lett. 1987, 28, 4075. (d) An alternative explanation
would be an initial desulfonylation followed by a [1,2]-hydride shift
rearrangement with Walden displacement of the remaining sul-
fonate: Hansske, F.; Robins, M. J . J . Am. Chem. Soc. 1983, 105, 6736.
(14) Brown, D. M.; Parihar, D. B.; Todd, A.; Varadarajan, S. J . Chem.
Soc. 1958, 3028.
(15) Chattopadhyaya, J . B.; Reese, C. B. J . Chem. Soc., Chem.
Commun. 1976, 860.
(16) (a) Hampton, A.; Nichol, A. W. Biochemistry 1966, 5, 2076. (b)
We chose the procedure described in ref 16a instead of the one followed
by Verheyden et. al.2c Despite the fact that HMPT solvent gives bet-
ter results than DMF in this kind of reactions (88 vs 59% yield of
2,2′-anhydrouridine from uridine), we preferred not to use this toxic
solvent.

6984 J . Org. Chem., Vol. 61, No. 20, 1996
Garc´ıa-Alles et al.
Ta ble 4. Isola ted Globa l Yield sa
and solvents were distilled. THF and DMF were dried by
reflux over and distillation from sodium and CaH2, respec-
tively. Pyridine was freshly distilled from KOH. Acetone
O-(phenyloxycarbonyl)oxime was prepared by treating acetone
oxime with phenylchloroformate and purified by conventional
procedures. 1H-NMR data listed in the following manner:
chemical shift (multiplicity, number of protons, assignment).
For other experimental details, see ref 8c.
starting
nucleoside
carbamate
(%)
alkylamino
nucleoside (%)
amino nucleoside
(%)
1 B ) Th
1 B ) Th
1 B ) Ad
20a
10a (62)
10b (64)
11 (46)
12a (53)
12b (57)
14 (37)
13a (56)
13b (34)
17a (44)
17b (45)
25a (54)
25b (66)
26a (46)
26b (49)
20b
En zym a tic Syn th esis of 5′-O-((P h en yloxy)ca r bon yl)-
th ym id in e (2) [P r oced u r e A]. One millimole of 1 (B ) Th),
3 mmol of acetone O-((phenyloxy)carbonyl)oxime, and 0.4 g of
lipase from C. antarctica (CAL) were suspended in 20 mL of
THF. The mixture was allowed to react at 60 °C and 250 rpm
for 36 h (monitored by TLC until almost complete disappear-
ance of 1). Then, the enzyme was filtered and washed with
MeOH, and the residue was evaporated under vacuum and
subjected to flash chromatography (AcOEt) to yield 2 (0.29 g,
a
Calculated with respect to the starting material.
Ch a r t 2
1
80%): mp 163-164 °C; IR (KBr, cm-1) 1767; H NMR (CD3-
OD) δ 7.81 (s, 1, H6), 7.62 (m, 2, Ph), 7.48 (m, 1, Ph), 7.39 (m,
2, Ph), 6.52 (t, 1, H1′), 4.69 (m, 2, H5′,5′′), 4.64 (m, 1, H3′),
4.33 (m, 1, H4′), 2.59-2.42 (m, 2, H2′,2′′), 2.02 (s, 3, Me); 13C
NMR (CD3OD) δ 166.83 (C4), 155.38 (CdO), 153.04 (Ph),
152.78 (C2), 138.14 (C6), 131.15 (Ph), 127.83 (Ph), 122.61 (Ph),
112.30 (C5), 86.89 (C1′), 86.02 (C4′), 72.48 (C3′), 69.41 (C5′),
41.14 (C2′), 13.05 (Me); MS (70 eV), m/z (rel intensity) 362 (M+,
1), 237 (6), 94 (100). Anal. Calcd for C17H18N2O7: C, 49.72;
H, 4.97; N, 7.73. Found: C, 49.65; H, 5.10; N, 7.62.
5′-O-((Vin yloxy)ca r bon yl)a d en osin e (3). Treatment of
1 mmol of 2′-deoxyadenosine 1 (B ) Ad) by procedure A (using
acetone O-((vinyloxy)carbonyl)oxime and performing the reac-
tion during 8 h at 30 °C) gave 3 (243 mg, 78%) after
chromatography (AcOEt-MeOH 92:8).8c
Gen er a l P r oced u r e for th e Syn th esis of th e 5′-O-(N-
Alk ylca r b a m oyl)-3′-O-m et h ylsu lfon yl(a n d 3′-O-p -t olu -
en esu lfon yl)-2′-d eoxyn u cleosid es (5b, 6a ,b, 8a ,b). A solu-
tion of carbonate 2 or 3 (2 mmol) and the corresponding amine
(8 mmol) in 15 mL of THF was stirred at reflux until no
starting material was detected by TLC (less than 1 day).
Solvent and most of the amine were then removed by evapora-
tion under vacuum. The residue was precipitated and washed
several times by addition of ethyl ether (for the 2′-deoxyad-
enosine derivatives, the residue was submitted to flash chro-
matography, AcOEt-MeOH 9:1). Some of these intermediate
carbamates have been previously described by us.8c They were
dissolved in 10 mL of dry pyridine and treated with 6 mmol
of the corresponding sulfonyl chloride (in the case of the
p-toluenesulfonyl derivatives, an additional 2 mmol of p-
toluenesulfonyl chloride were added 10 h later to complete the
reaction). The mixture was stirred overnight, and the reaction
stopped by addition of 1 mL of ice water when complete by
TLC (about 15-20 h). The pyridine was then carefully
evaporated. Water (30 mL) was added to the residue, and
the solution was extracted (CH2Cl2, 2 × 50 mL). The combined
organic phase was washed twice with 5% NaHCO3 and
with water, dried, filtered, and evaporated to give products
5b (860 mg, 90%), 6a (760 mg, 84%), 6b (845 mg, 80%), and
8a (740 mg, 80%), as colorless foams. For compound 8b, the
solid obtained was chromatographed (AcOEt-MeOH 91:9), due
to the appearance of derivatives resulting from tosylation of
the adenine base. Compound 8b was obtained in 55% yield
(592 mg).
oxide T anhydronucleoside is restricted to 8-oxyadenos-
ine derivatives.15 However, the high yield synthesis of
the riboepoxide 23 can be easily achieved by means of
the procedure described by Robins et al.17 Carbamate
24 was then obtained and heated with NaH in DMF to
yield the corresponding cyclic carbamate 25b almost
quantitatively (see Table 3 for data about NOE experi-
ments performed on this compound).18
Subsequent decarbonylation of 25a ,b with LiOH and
catalytic transfer hydrogenation of the resulting com-
pounds 26a ,b yielded the corresponding 3′-amino-3′-
deoxyxylouridine 17a and 3′-amino-3′-deoxyxyloadenos-
1
ine 17b (see Tables 1 and 2 for H- and 13C-NMR spectral
data of these and the preceding compounds).2e Once
again, the most remarkable difference is the upfield shift
of ca. 12 ppm of the C3′ carbon atom signal with regard
to the starting ribonucleosides.
Su m m a r y
Table 4 summarizes the results obtained with respect
to the starting natural ribonucleosides and 2′-deoxyri-
bonucleosides. As can be seen, final aminosugar nucleo-
sides are attained with yields ranging from 30 to 55%.
In conclusion, we have presented a novel approach to
the synthesis of 3′-aminoxylonucleosides that involves the
intramolecular delivery of the masked amino nucleophile
from the 5′-position regiospecifically to the 3′-position of
the sugar.
The described aminosugar nucleosides are unique
precursors for isonucleoside analogs of the general struc-
ture 27 and 28 (Chart 2). The synthesis and pharma-
cological evaluation of several of these compounds will
be reported in the near future.
For 5b: mp softened at 68 °C (the compound does not have
a clear melting point) ; IR (KBr, cm-1) 1709; 1H NMR (CDCl3)
δ 7.77 (d, 2, Ts), 7.31 (d, 2, Ts), 7.12 (s, 1, H6), 6.10 (t, 1, H1′),
5.75 (m, 1, CH), 5.58 (m, 1, H3′), 5.11 (dd, 1, CH2), 5.02 (dd, 1,
CH2), 4.24-4.02 (m, 2, H5′,5′′), 4.23 (m, 1, H4′), 3.70 (m, 1,
CH2), 2.45 (m, 1, H2′), 2.39 (s, 3, MePh), 2.18 (m, 1, H2′′), 1.86
(s, 3, Me); 13C NMR (CDCl3) δ 164.03 (C4), 155.35 (CdO),
150.23 (C2), 145.43 (Ts), 136.06 (C6), 135.19 (CH), 132.61 (Ts),
129.96 (Ts), 127.57 (Ts), 116.05 (CH2), 110.89 (C5), 85.28 (C1′),
81.99 (C4′), 79.87 (C3′), 63.36 (C5′), 43.27 (CH2), 37.43 (C2′),
21.46 (MePh), 12.33 (Me); MS (70 eV), m/z (rel intensity) 479
(M+, 10), 307 (10). Anal. Calcd for C21H25N3O8S: C, 52.60;
H, 5.26; N, 8.76. Found: C, 52.52; H, 5.18; N, 8.70. For 6a :
mp softened at 80 °C (the compound does not have a clear
melting point); IR (KBr, cm-1) 1706; 1H NMR (CD3OD) δ 7.61
Exp er im en ta l Section
Gen er a l. Lipase from C. antarctica SP 435L (8.200 PLU/
g), generously donated by Novo Nordisk Company, was kept
for 2 days under vacuum prior to use. Nucleosides and other
chemicals were purchased from Sigma and Aldrich Chemie,
(17) Robins, M. J .; Wilson, J . S.; Madej, D.; Low, N. H.; Hansske,
F.; Wnuk, S. F. J . Org. Chem. 1995, 60, 7902.
(18) A simpler synthetic scheme would be to treat compound 23 with
benzyl isocyanate and follow the same synthetic scheme. However, this
sequence has been rejected, since the last step of hydrogenation of the
analog of product 26b, containing a N-benzylcarbamoyl group attached
to the exocyclic amino group of the adenine, did not give the expected
results.

Synthesis of Purine and Pyrimidine Nucleosides
J . Org. Chem., Vol. 61, No. 20, 1996 6985
(s, 1, H6), 7.49 (m, 5, Ph), 6.46 (dd, 1, H1′), 5.52 (m, 1, H3′),
4.61 (m, 1, H4′), 4.58-4.48 (m, 2, H5′,5′′), 4.49 (d, 2, CH2), 3.38
(s, 3, Me), 2.79 (m, 1, H2′), 2.65 (m, 1, H2′′), 1.95 (s, 3, Me);
13C NMR (CD3OD) δ 166.63 (C4), 158.64 (CdO), 152.62 (C2),
140.73 (Ph), 137.78 (C6), 130.01 (Ph), 128.85 (Ph), 128.76 (Ph),
112.45 (C5), 86.84 (C1′), 84.33 (C4′), 82.32 (C3′), 65.55 (C5′),
46.06 (CH2), 38.99 (C2′), 38.78 (Me), 13.12 (Me); MS (70 eV),
m/z (rel intensity) 453 (M+, 1), 357 (10), 150 (50), 91 (100).
Anal. Calcd for C19H23N3O8S: C, 50.33; H, 5.11; N, 9.27.
Found: C, 50.21; H, 5.06; N, 9.35. For 6b: mp 151-154 °C;
then removed under vacuum. The resulting residue was
treated with 30 mL of 10% NH4Cl solution and extracted with
CH2Cl2 (3 × 30 mL). The combined extracts were washed
twice with 5% NaHCO3 solution, once with water, dried (Na2-
SO4), filtered, and concentrated to afford 10b (360 mg, 92%).
3′-N-Ben zyla m in o-2′,3′-d id eoxyxyloa den osin e 3′,5′-Ca r -
ba m a te (11). From 8a : A solution of 370 mg (0.8 mmol) of
8a in 10 mL of dry THF was treated with a suspension of 77
mg (3.2 mmol) of NaH in THF (10 mL) at 30 °C under N2
atmosphere. The reaction was continued for 4 h, and then 5
mL of MeOH were added to the solution. Solvent was
evaporated under vacuum, and the residue was chromato-
graphed (AcOEt-MeOH 8:2). Appropriate fractions were
combined and evaporated to give 11 (216 mg, 74%): mp
236-238 °C; IR (KBr, cm-1) 1690; MS (70 eV), m/z (rel
intensity) 366 (M+, 70), 136 (100), 91 (100). Anal. Calcd for
C18H18N6O3: C, 59.01; H, 4.95; N, 22.94. Found: C, 58.95; H,
4.88; N, 22.79.
From 8b: A solution of 377 mg (0.7 mmol) of 8b in 8 mL of
dry THF was cooled to 0 °C and treated with a suspension of
68 mg (2.8 mmol) of NaH in THF (8 mL). The reaction was
stirred under N2 atmosphere for 3 h, at which point 5 mL of
MeOH were added. The solvent was evaporated under vacuum,
and the residue was chromatographed (AcOEt-MeOH 8:2).
The product was concentrated to give 11 (235 mg, 92%).
3′-N-Allyla m in o- a n d 3′-N-Ben zyla m in o-2′,3′-d id eoxy-
xylon u cleosid es (12a ,b, 14) [P r oced u r e B]. One millimole
of 10a ,b or 11 was dissolved in 30% aqueous EtOH (20 mL).
LiOH (800 mg, 33.3 mmol) was added, and the mixture was
heated at reflux for 1-3 h (until TLC revealed complete
disappearance of the starting nucleoside). Solvents were
removed under vacuum, and the precipitate was eluted with
CH2Cl2-MeOH 9:1 through a pad of silica gel (CH2Cl2-MeOH
85:15 in the case of product 12a ). For 12a (238 mg, 85%): mp
softened at 85 °C (the compound does not have a clear melting
point); IR (KBr, cm-1) 1695; MS, m/z (FAB+, rel intensity) 282
(M + H, 70)+, 154 (100). Anal. Calcd for C13H19N3O4: C,
55.51; H, 6.81; N, 14.94. Found: C, 55.43; H, 6.77; N, 14.79.
For 12b (298 mg, 90%): mp 141-143 °C; IR (KBr, cm-1) 1674;
MS (70 eV), m/z (rel intensity) 331 (M+, 20), 224 (60), 91 (100).
Anal. Calcd for C17H21N3O4: C, 61.62; H, 6.39; N, 12.68.
Found: C, 61.58; H, 6.30; N, 12.61. For 14 (275 mg, 81%):
mp 151-153 °C; IR (KBr, cm-1) 1676; MS (70 eV), m/z (rel
intensity) 340 (M+, 10), 235 (60), 136 (60), 91 (100). Anal.
Calcd for C17H20N6O2: C, 59.99; H, 5.92; N, 24.69. Found: C,
59.85; H, 5.80; N, 24.48.
1
IR (KBr, cm-1) 1707; H NMR (CDCl3) δ 7.76 (d, 2, Ts), 7.35
(d, 2, Ts), 7.25 (m, 5, Ph), 7.15 (s, 1, H6), 6.11 (t, 1, H1′), 5.10
(m, 1, H3′), 4.30-4.10 (m, 2, H5′,5′′), 4.30 (m, 1, H4′), 4.28 (d,
2, CH2), 2.41 (s, 3, MePh), 2.40 (m, 1, H2′), 2.25 (m, 1, H2′′),
1.70 (s, 3, Me); 13C NMR (CDCl3) δ 163.92 (C4), 155.50 (CdO),
150.13 (C2), 145.31 (Ts), 137.88 (Ph), 135.31 (C6), 132.48 (Ts),
129.84 (Ts), 128.23 (Ph), 127.44 (Ph), 127.09 (Ts), 110.73 (C5),
85.16 (C1′), 81.81 (C4′), 79.84 (C3′), 63.32 (C5′), 44.64 (CH2),
37.15 (C2′), 21.31 (MePh), 12.07 (Me); MS (70 eV), m/z (rel
intensity) 529 (M+, 5), 396 (10), 357 (20). Anal. Calcd for
C25H27N3O8S: C, 56.70; H, 5.14; N, 7.93. Found: C, 56.43; H,
5.00; N, 7.90. For 8a : mp 185-187 °C; IR (KBr, cm-1) 1718;
1H NMR (CDCl3) δ 8.09 (s, 1, H8), 7.83 (s, 1, H2), 7.23 (m, 5,
Ph), 6.29 (t, 1, H1′), 5.50 (m, 1, H3′), 4.41 (m, 1, H4′), 4.40-
4.25 (m, 2, H5′,5′′), 4.38 (d, 2, CH2), 3.10 (m, 1, H2′), 3.08 (s, 3,
Me), 2.70 (m, 1, H2′′); 13C NMR (CDCl3) δ 155.87 (C6), 155.53
(CdO), 152.64 (C2), 148.87 (C4), 138.94 (C8), 138.02 (Ph),
128.35 (Ph), 127.13 (Ph), 119.47 (C5), 83.87 (C1′), 82.22 (C4′),
79.19 (C3′), 62.87 (C5′), 44.72 (CH2), 38.01 (Me), 38.96 (C2′);
MS (70 eV), m/z (rel intensity) 462 (M+, 20), 366 (20), 119 (30).
Anal. Calcd for C19H22N6O6S: C, 49.34; H, 4.79; N, 18.17.
Found: C, 49.22; H, 4.76; N, 18.05. For 8b: mp 166-168 °C;
1
IR (KBr, cm-1) 1720; H NMR (CDCl3) δ 8.07 (s, 1, H8), 7.82
(s, 1, H2), 7.75 (d, 2, Ts), 7.29 (d, 2, Ts), 7.20 (m, 5, Ph), 6.25
(t, 1, H1′), 5.43 (m, 1, H3′), 4.30 (m, 1, H4′), 4.25-4.15 (m, 2,
H5′,5′′), 4.22 (d, 2, CH2), 2.95 (m, 1, H2′), 2.51 (m, 1, H2′′),
2.30 (s, 3, MePh); 13C NMR (CDCl3) δ 155.58 (C6), 155.52
(CdO), 152.43 (C2), 148.77 (C4), 145.22 (Ts), 138.94 (C8),
137.94 (Ph), 132.39 (Ts), 129.76 (Ts), 128.18 (Ph), 127.74 (Ph),
127.07 (Ts), 119.44 (C5), 83.91 (C1′), 81.99 (C4′), 79.87 (C3′),
63.03 (C5′), 44.57 (CH2), 36.50 (C2′), 21.26 (MePh); MS (70 eV),
m/z (rel intensity) 538 (M+, 20), 369 (70). Anal. Calcd for
C25H26N6O6S: C, 55.75; H, 4.87; N, 15.60. Found: C, 55.59;
H, 4.80; N, 15.48.
3′-N-Allyla m in o-3′-d eoxyxyloth ym id in e 3′,5′-Ca r ba m -
a te (10a ). A solution of 0.86 g (1.8 mmol) of 5b in 15 mL of
dry THF was treated with a suspension of 0.13 g (5.4 mmol)
of NaH in THF (15 mL). The resulting mixture was stirred
under nitrogen atmosphere at 25 °C for 20 h (until 5b
completely disappeared by TLC), at which point solvent was
carefully removed under vacuum and the resulting residue was
submitted to flash chromatography (AcOEt-MeOH 82:18) to
3′-Am in o-3′-d eoxyxyloth ym id in e (13a ) [P r oced u r e C].10
The benzyl derivative 12b (100 mg) dissolved in 5 mL of 4.4%
formic acid-methanol was added to a 5 mL round-bottom flask
containing 100 mg of palladium black catalyst and 5 mL of
4.4% formic acid-methanol. The mixture was continuously
stirred under a nitrogen atmosphere. Reaction was complete
within 40-60 min as determined by thin layer chromatogra-
phy analysis of samples taken at different times. Products
were isolated by filtering off with a pad of Celite, and the pad
was washed well with MeOH. Solvent was removed in vacuo,
and the resulting residue was dissolved in 3% HCl/MeOH. The
solvent was then removed to give the HCl salt of 13a (82 mg,
98%).4
3′-Am in o-2′,3′-d id eoxyxyloa d en osin e (13b). Treatment
of product 14 (100 mg) by procedure C in the presence of a
large excess of palladium black catalyst (1 g), showed after 1
h complete absence of the starting material. Careful filtration
was performed, washing the catalyst several times with
MeOH. The residue obtained after evaporation of the solvent
was dissolved in 2 mL of 3% HCl/MeOH, and the solvent was
removed to give the HCl salt of 13b (77 mg, 92%):2h MS, m/z
(FAB+, rel intensity) 251 (M + H, 85)+, 136 (100).
afford 10a (480 mg, 87%): mp 156-158 °C; IR (KBr, cm-1
)
1693; MS (70 eV), m/z (rel intensity) 307 (M+, 40), 182 (100).
Anal. Calcd for C14H17N3O5: C, 54.72; H, 5.58; N, 13.67.
Found: C, 54.69; H, 5.56; N, 13.60.
3′-N-Ben zyla m in o-3′-d eoxyxylot h ym id in e 3′,5′-Ca r -
ba m a te (10b). From 6a : A solution of 136 mg (0.3 mmol) of
6a in 2.5 mL of dry THF was cooled to 0 °C and treated with
a suspension of 22 mg (0.9 mmol) of NaH in THF (2.5 mL).
The cooling bath was removed after 2 h, and the reaction was
continued at room temperature and under N2 atmosphere for
20 h (until 6a completely disappeared), at which point solvent
was evaporated under vacuum. The residue was treated with
10 mL of 10% NH4Cl solution and extracted with CH2Cl2 (3 ×
10 mL). The combined extracts were washed twice with 5%
NaHCO3 solution, once with water, dried (Na2SO4), filtered,
and concentrated to afford 10b (75 mg, 70%): mp 194-196
°C; IR (KBr, cm-1) 1691; MS (70 eV), m/z (rel intensity) 357
(M+, 20), 232 (40), 91 (100). Anal. Calcd for C18H19N3O5: C,
60.50; H, 5.36; N, 11.76. Found: C, 60.43; H, 5.30; N, 11.55.
From 6b: A solution of 582 mg (1.1 mmol) of 6b in 10 mL
of dry THF was cooled to 0 °C and treated with a suspension
of 84 mg (3.5 mmol) of NaH in THF (10 mL). The cooling bath
was removed after 2 h, and the reaction was continued at room
temperature and under N2 atmosphere for 5 h. Solvent was
5′-O-(N-Ben zylca r ba m oyl)u r id in e (21). Treatment of
uridine 20a (244 mg, 1 mmol) with acetone O-((vinyloxy)-
carbonyl)oxime by procedure A at 60 °C, gave after 24 h a
mixture of two products, as shown by thin layer chromatog-
raphy analysis. After filtration of the catalyst and evaporation
of the solvent, the residue was subjected to flash chromatog-
raphy (AcOEt-MeOH 9:1) and two different products were
collected. The less polar product was the 5′-O-((vinyloxy)-

6986 J . Org. Chem., Vol. 61, No. 20, 1996
Garc´ıa-Alles et al.
carbonyl) uridine derivative (195 mg, 62%), as shown by 1H
NMR analysis, while the other one was the 5′-acetonoxime
carbonate (125 mg, 36%). Both products were dissolved in dry
THF (8 mL) and treated with 4 mmol of benzylamine. The
solution was stirred at reflux temperature for 5 h, at which
point the solvent and most of the remaining amine were
evaporated in vacuo. Flash chromatography (AcOEt-MeOH
9:1) of the resulting residue yielded 21 (320 mg, 85%).8c
2,2′-An h yd r o-1-[5-O-(N-ben zylca r ba m oyl)-â-D-a r a bin o-
fu r a n osyl]u r a cil (22). Uridine carbamate 21 (4 mmol, 1.51
g) was dissolved in dry DMF (4 mL) and treated with diphenyl
carbonate (1.1 g, 5.15 mmol) and NaHCO3 (20 mg).16 The
mixture was heated at 150 °C for 5 min, the solvent was
evaporated under vacuum and the resulting gum was chro-
matographed (AcOEt-MeOH 85:15) to yield 22 (1.0 g, 70%):
mp 144-145 °C; IR (KBr, cm-1) 1720; MS (70 eV), m/z (rel
intensity) 359 (M+, 5), 248 (10), 91 (100). Anal. Calcd for
C17H17N3O6: C, 56.82; H, 4.77; N, 11.69. Found: C, 56.74; H,
4.75; N, 11.61.
3′-N-Ben zyla m in o-3′-d eoxyxylou r id in e 3′,5′-Ca r ba m -
a te (25a ). 2,2′-Anhydrouridine derivative 22 (198 mg, 0.55
mmol) was dissolved in dry DMF (5 mL) and treated with NaH
(40 mg, 1.6 mmol). The mixture was heated under nitrogen
atmosphere at 100-110 °C for 10 min, at which point the
solvent was evaporated under vacuum. Flash chromatography
of the resulting residue (AcOEt-MeOH 9:1) yielded a colorless
solid identified as 25a (180 mg, 91%): mp 234-237 °C; IR
(KBr, cm-1) 1695; MS (70 eV), m/z (rel intensity) 359 (M+, 10),
189 (20), 91 (70), 44 (100). Anal. Calcd for C17H17N3O6: C,
56.82; H, 4.77; N, 11.69. Found: C, 56.78; H, 4.79; N, 11.56.
9-[2,3-An h yd r o-5-O-(N-ben zylca r ba m oyl)-â-D-a r a bin o-
fu r a n osyl]a d en in e (24). Treatment of 23 (200 mg, 0.8
mmol)17 with acetone O-((vinyloxy)carbonyl)oxime (2.4 mmol)
in 20 mL of THF in presence of 0.4 g of CAL at 60 °C gave
after 24 h complete disappearance of the starting material. A
batch of seven reactions like this were performed at the same
time (1.4 g, 5.6 mmol of 23). After filtration of the catalyst
and evaporation of the solvent, the total residue was subjected
to flash chromatography (CH2Cl2-MeOH 95:5). A mixture of
two products was obtained (5′-vinyl and 5′-acetonoxime car-
bonates), concentrated (1.72 g), dissolved in dry THF (30 mL),
and treated with 16.8 mmol of benzylamine. The solution was
stirred at reflux temperature for 8 h. The solvent and most
of the remaining amine were then removed by evaporation in
vacuo. The resulting gum was precipitated and washed with
EtOH (4 × 4 mL) to give 1.6 g of a white solid. Evaporation
of the filtrate and precipitation with EtOH yielded 0.15 g more
of the same colorless solid 24 (1.75 g, 82%): mp 166-168 °C;
IR (KBr, cm-1) 1690; MS (70 eV), m/z (rel intensity) 382 (M+,
5), 248 (10), 91 (100). Anal. Calcd for C18H18N6O4: C, 56.54;
H, 4.74; N, 21.98. Found: C, 56.48; H, 4.76; N, 21.84.
3′-N-Ben zylam in o-3′-deoxyxyloaden osin e 3′,5′-Car bam -
a te (25b). A solution of 191 mg (0.5 mmol) of 24 was dissolved
in dry DMF (5 mL) and treated with NaH (36 mg, 1.5 mmol).
The mixture was heated under nitrogen atmosphere at 100-
110 °C for 10 min, at which point the solvent was evaporated
under vacuum and the resulting residue was submitted to
flash chromatography (AcOEt-MeOH 9:1) to afford 25b (168
mg, 88%): mp softened at 135 °C (the compound does not have
a clear melting point but slowly colorized above 160 °C); IR
(KBr, cm-1) 1691; MS (70 eV), m/z (rel intensity) 382 (M+, 10),
164 (100), 136 (90). Anal. Calcd for C18H18N6O4: C, 56.54;
H, 4.74; N, 21.98. Found: C, 56.41; H, 4.66; N, 21.87.
3′-N-Allyla m in o- a n d 3′-N-Ben zyla m in o-3′-d eoxyxylo-
n u cleosid es (26a ,b). Treatment of 1 mmol of 25a ,b by
procedure B gave after 1-2 h complete disappearance of the
starting material. The solvents were removed under vacuum,
and the residue was chromatographed through a pad of silica
gel first with CH2Cl2-MeOH 9:1 and finally with MeOH. For
26a (286 mg, 86%): mp 189-192 °C; IR (KBr, cm-1) 1690; MS
(70 eV), m/z (rel intensity) 333 (M+, 1), 204 (20), 91 (100). Anal.
Calcd for C16H19N3O5: C, 57.65; H, 5.75; N, 12.61. Found: C,
57.58; H, 5.73; N, 12.49. For 26b (267 mg, 75%): mp 162-
164 °C; IR (KBr, cm-1) 1649; MS, m/z (FAB+, rel intensity)
357 (M + H, 95)+, 136 (100). Anal. Calcd for C17H20N6O3: C,
57.29; H, 5.66; N, 23.58. Found: C, 57.17; H, 5.59; N, 23.40.
3′-Am in o-3′-d eoxyxylon u cleosid es (17a ,b). Treatment
of products 26a ,b (100 mg) by procedure C (for the debenzy-
lation of the adenosine derivative 26b, a great excess of
palladium black catalyst (1 g) had to be used) showed after 1
h complete absence of the starting material. Products were
isolated by filtering the catalyst with a pad of Celite, and the
pad was washed well with MeOH. Solvent was then removed
in vacuo. The resulting residue was dissolved in 3% HCl/
MeOH, and the solvent was evaporated to give the HCl salt
of 17a ,b. For 17a (79 mg, 95%): mp softened at 70 °C (the
compound does not have a clear melting point but slowly
colorized above 150 °C); IR (KBr, cm-1) 1693; MS, m/z (FAB+,
rel intensity) 244 (M + H, 60)+, 136 (70). Anal. Calcd for
C9H14N3O5Cl: C, 38.65; H, 5.05; N, 15.02. Found: C, 38.58;
H, 4.99; N, 14.91. For 17b (77 mg, 91%):2e MS, m/z (FAB+,
rel intensity) 267 (M + H, 55)+, 136 (100).
Ack n ow led gm en t. Financial support of this work
by CICYT (project BIO-95-0687) is gratefully acknowl-
edged. L.F.G.A. also thanks to the FICYT (Asturias) for
a predoctoral scholarship.
J O9606259