Synthesis and properties of diuridine phosphate analogues containing thio and amino modifications

Article abstract of DOI:10.1021/jo960795l

Several analogues of diuridine phosphate (UpU) were synthesized in order to investigate why replacing the 2'-hydroxyl with a 2'-amino group prevents hydrolysis. These analogues were designed to investigate what influence the 2'-substituent and 5'-leaving group have upon the rate of hydrolysis. All the analogues were considerably more labile than UpU toward acid-base-catalyzed hydrolysis. In the pH region from 6 to 9, the rate of hydrolysis of uridylyl (3'-5') 5'-thio-5'deoxyuridine (UpsU) hydrolysis rose, in a log linear fashion, from a value of 5 x 10^-6 s^-1 at pH 6 to 3200 x 10^-6 s^-1 at pH 9, indicating that attack on the phosphorus by the 2'-oxo anion is rate-limiting in the hydrolysis mechanism. In contrast, the rate of uridylyl (3'-5') 5'-amino-5'deoxyuridine (UpnU) hydrolysis fell from a value of 1802 x 10^-6 s^-1 at pH 5 to 140 x 10^-6 s^-1 at pH 7.5, where it remained constant up to pH 11.5, thus indicating an acid-catalyzed reaction. The analogue 2'-amino-2'-deoxyuridylyl (3'-5') 5'-thio-5'-deoxyuridine (amUpsU) was readily hydrolyzed above pH 7, in contrast to the hydrolytic stability of amUpT, with rates between 85 x 10^-6 s^-1 and 138 x 10^-6 s^-1. The hydrolysis of 2'-amino-2'-deoxyuridylyl (3'-5') 5'-amino-5'-deoxythymidine (amUpnT) rose from 17 x 10^-6 s^-1 at pH 11.5 to 11 685 x 10^-6 s^-1 at pH 7.0, indicating an acid-catalyzed reaction, where protonation of the 5'-amine is rate limiting. The cleavage rates of UpsU, UpnU, and amUpsU were accelerated in the presence of Mg²⁺,Zn²⁺, and Cd²⁺ ions, but a correlation with interaction between metal ion and leaving group could only be demonstrated for amUpsU. UpsU and UpnU are also substrates for RNase A with UpsU having similar Michaelis-Menten parameters to UpU. In contrast, UpnU is more rapidly degraded with an approximate 35-fold increase in catalytic efficiency, which is reflected purely in an increase in the value of κ(cat).

Full text of DOI:10.1021/jo960795l

J . Org. Chem. 1996, 61, 6273-6281
6273
Syn th esis a n d P r op er ties of Diu r id in e P h osp h a te An a logu es
Con ta in in g Th io a n d Am in o Mod ifica tion s
J ames B. Thomson,† Bhisma K. Patel,† Victor J ime´nez,‡ Klaus Eckart, and Fritz Eckstein*
Max-Planck-Institut fu¨r experimentelle Medizin, Hermann-Rein-Str.3, D-37075 Go¨ttingen, Germany
Received April 30, 1996X
Several analogues of diuridine phosphate (UpU) were synthesized in order to investigate why
replacing the 2′-hydroxyl with a 2′-amino group prevents hydrolysis. These analogues were designed
to investigate what influence the 2′-substituent and 5′-leaving group have upon the rate of
hydrolysis. All the analogues were considerably more labile than UpU toward acid-base-catalyzed
hydrolysis. In the pH region from 6 to 9, the rate of hydrolysis of uridylyl (3′-5′) 5′-thio-5′-
deoxyuridine (UpsU) hydrolysis rose, in a log linear fashion, from a value of 5 × 10-6 s-1 at pH 6
to 3200 × 10-6 s-1 at pH 9, indicating that attack on the phosphorus by the 2′-oxo anion is rate-
limiting in the hydrolysis mechanism. In contrast, the rate of uridylyl (3′-5′) 5′-amino-5′-
deoxyuridine (UpnU) hydrolysis fell from a value of 1802 × 10-6 s-1 at pH 5 to 140 × 10-6 s-1 at
pH 7.5, where it remained constant up to pH 11.5, thus indicating an acid-catalyzed reaction. The
analogue 2′-amino-2′-deoxyuridylyl (3′-5′) 5′-thio-5′-deoxyuridine (amUpsU) was readily hydrolyzed
above pH 7, in contrast to the hydrolytic stability of amUpT, with rates between 85 × 10-6 s-1 and
138 × 10-6 s-1. The hydrolysis of 2′-amino-2′-deoxyuridylyl (3′-5′) 5′-amino-5′-deoxythymidine
(amUpnT) rose from 17 × 10-6 s-1 at pH 11.5 to 11 685 × 10-6 s-1 at pH 7.0, indicating an acid-
catalyzed reaction, where protonation of the 5′-amine is rate limiting. The cleavage rates of UpsU,
UpnU, and amUpsU were accelerated in the presence of Mg2+, Zn2+, and Cd2+ ions, but a correlation
with interaction between metal ion and leaving group could only be demonstrated for amUpsU.
UpsU and UpnU are also substrates for RNase A with UpsU having similar Michaelis-Menten
parameters to UpU. In contrast, UpnU is more rapidly degraded with an approximate 35-fold
increase in catalytic efficiency, which is reflected purely in an increase in the value of kcat
.
In tr od u ction
protocols to obtain nuclease-resistant RNA aptamers for
therapeutic applications.13-15 In order to understand this
hydrolytic stability we have undertaken a study to
determine whether cleavage could be obtained with the
2′-amino group by variation of the leaving group. We
report here a comparative study of the hydrolysis of UpU
with that of UpsU, amUpsU, UpnU, and amUpnT
(Figure 1) as a function of pH, divalent metal ions, and
RNaseA.
The interest in RNA-catalyzed reactions has stimu-
lated structural1-6 and chemical modification7 studies of
RNA in order to better understand both the structure-
function relationships and the mechanisms of such
reactions. One such catalytic RNA is the hammerhead
ribozyme, which most efficiently cleaves its substrate
after a GUC sequence to yield a 2′,3′-cyclic phosphate and
a 5′-hydroxyl termini.8 We have recently shown that
replacement of the 2′-hydroxy at the cleavage site by an
amino group prevents this cleavage9 and such a modifi-
cation also stabilizes the RNA toward alkaline and RNase
A-mediated hydrolysis.9-12 Nucleotides bearing this
modification have been utilized for in vitro selection
Resu lts
The four modified dinucleotides were synthesized by
the reaction schemes outlined in Schemes 1 and 2.
Additionally, TpsT and TpnT were prepared by similar
routes for use as control compounds containing no 2′-
nucleophile.
Hyd r olysis of Din u cleotid es. In order to establish
whether the buffer could enhance the hydrolysis rate of
the dimers, the dependence of kobs on the buffer concen-
tration was established for UpsU and UpnU at pH 7.5
using HEPES buffer. Their rate of hydrolysis demon-
strated a slight positive linear dependence upon the
buffer concentration, as a result of the weak general base
properties of HEPES. Extrapolating both plots to zero
† Both authors have contributed equally to this publication.
‡ Genetic Engineering and Biotechnology Center, PO Box 6162
Havana, Cuba.
* Author to whom correspondance should be addressed. Tel: +49-
551-3899-274. Fax: +49-551-3899-388. E-mail: eckstein@exmed1.
dnet.gwdg.de.
X Abstract published in Advance ACS Abstracts, August 15, 1996.
(1) Pley, H. W.; Flaherty, K. M.; McKay, D. B. Nature 1994, 372,
68-74.
(2) Scott, W. G.; Finch, J . T.; Klug, A. Cell (Cambridge, Mass.) 1995,
81, 991-1002.
(3) Tuschl, T.; Gohlke, C.; J ovin, T. M.; Westhof, E.; Eckstein, F.
Science 1994, 266, 785-788.
(4) Tanner, N. K.; Schaff, S.; Thill, G.; Petit-Koskas, E.; Crain-
Denoyelle, A.-M.; Westhof, E. Curr. Biol. 1994, 4, 488-498.
(5) Westhof, E.; Altman, S. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
5133-5137.
(6) Cech, T. R.; Damberger, S. H.; Gutell, R. R. Nature Struct. Biol.
1994, 1, 273-280.
(7) Grasby, J . A.; Gait, M. J . Biochimie 1994, 76, 1223-1234.
(8) Uhlenbeck, O. C. Nature 1987, 328, 596-600.
(9) Pieken, W. A.; Olsen, D. B.; Benseler, F.; Aurup, H.; Eckstein,
F. Science 1991, 253, 314-317.
(10) Hobbs, J .; Sternbach, H.; Sprinzl, M.; Eckstein, F. Biochemistry
1973, 12, 5138-4145.
(11) Heidenreich, O.; Benseler, F.; Fahrenholz, A.; Eckstein, F. J .
Biol. Chem. 1994, 269, 2131-2138.
(12) Sigurdsson, S. T.; Tuschl, T.; Eckstein, F. RNA 1995, 1, 575-
583.
(13) J ellinek, D.; Green, S. L.; Bell, C.; Lynott, C. K.; Gill, N.;
Vargeese, C.; Kirschenheuter, G.; McGee, D. P. C.; Abesinghe, P.;
Pieken, W. A.; Shapiro, R.; Rifkin, D. B.; Moscatelli, D.; J anjic, N.
Biochemistry 1995, 34, 11363-11372.
(14) Eaton, B. E.; Pieken, W. A. Annu. Rev. Biochem. 1995, 64, 837-
863.
(15) Gold, L.; Polisky, B.; Uhlenbeck, O.; Yarus, M. Annu. Rev.
Biochem. 1995, 64, 763-797.
S0022-3263(96)00795-5 CCC: $12.00 © 1996 American Chemical Society

6274 J . Org. Chem., Vol. 61, No. 18, 1996
Thomson et al.
Ta ble 1. 31P NMR Ch em ica l Sh ifts a n d HP LC Reten tion
Tim es of Ur id in e Din u cleotid es a n d Th eir Hyd r olysis
P r od u cts
dimer and
degradation
products
a
δ (ppm)
tR (min)
UpU
0 (q, J 1 ) 2.9 Hz, J 2 ) 4.6 Hz)
4 (d, J ) 7.8 Hz)
11
2′(3′)-UMP
2′,3′-cUMP
uridine
UpsU
10
11
UpnU
5′-aminoU
amUpnT
amUpsU
8.5
6.6
7.2
13
13.6
21
11.4
1.6
10.7
12.6
6.1
7.2
5.3
20 (q, J 1 ) 6.2 Hz, J 2 ) 12.1 Hz)
20.4
9.3
9.3
20.2 (q, J 1 ) 11.2 Hz, J 2 ) 21.9 Hz)
28 (m)
7.5 (d, J ) 6.2 Hz)
4.2 (d, J ) 8.4 Hz)
12
13
9
F igu r e 1. Modified uridine dimers investigated in this study.
U ) uridine and T ) thymidine.
a
All HPLC analysis were performed using buffer system 1 and
Sch em e 1.a Syn th esis of Up sU a n d a m Up sU
gradient 1.
remains constant up to pH 11.5. UpnU is exclusivly
hydrolyzed to 2′,3′-cUMP and 5′-amino-5′-deoxyuridine
(Table 1).
The hydrolysis of the dinucleotide amUpsU is pH
independent between pH 7 and 11.5, maintaining a
steady rate of 85 -138 × 10-6 s-1 (Figure 2). However,
below pH 7 the log rate of hydrolysis decreases, with a
slope of 0.4, which indicates inactivation by acid and is
consistent with protonation of the 2′-amine (pKa ) 6.2)17,18
to give the non-nucleophilic ammonium species. Hy-
drolysis of amUpsU follows two distinct pathways, with
the final hydrolysis products being dependent upon the
pH of the reaction mixture (Scheme 3). Hydrolysis at
pH 6 yields uridine 3′-monophosphate (9), 5′-thio-5′-
deoxyuridine (10) and its dimer 11. The cyclic phosphor-
amidate intermediate 12 was not observed, presumably
a result of its rapid degradation to 9. Cleavage of the
cyclic P-N bond is clearly more facile than that observed
for an acyclic P-N bond, such as in TpnT, where it is
stable to hydrolysis down to pH 5. This is consistent with
the observation that five-membered cyclic phosphates are
hydrolyzed at a faster rate than their acyclic counter-
parts.19 On following the course of hydrolysis of amUpsU
at pH 10, compounds 12, 10, and 11 can be detected first.
After 2 h incubation 13 appears, which reaches the same
concentration as 12 after 10 h of incubation, after which
time amUpsU is completely hydrolyzed. After incubation
for 12 h, 2′-amino-2′-deoxyuridine (14) and inorganic
phosphate (15) can be detected. Thus, under alkaline
conditions, where the nitrogen atom of the cyclic phos-
phoramidate 12 is not protonated, it is more stable and
degrades exclusively with P-O bond cleavage. There is
a precedence for both P-O and P-N cleavage in five-
membered 1,3,2,-oxazaphospholidin-2-ones, the level of
which is very dependent upon the nature of both the
phosphorus and nitrogen substituents.20
a
(i) 3-Hydroxypropionitrile, tetrazole; (ii) 3H-1,2-benzodithiol-
3-one 1,1-dioxide (Beaucage reagent); (iii) ammonia; (iv) 80%
AcOH; and (v) 5′-iodo-5′-deoxyuridine; (vi) triethylammonium
trihydrogen fluoride.
buffer strength gave kobs values of 92 and 94 × 10-6 s-1
,
for UpsU and UpnU, respectively. This is very close to
the values of 97 and 140 × 10-6 s-1 obtained in 100 mM
HEPES. This is also supported by the agreement be-
tween our data and that obtained in a study16 where
TRIS was used as the buffering agent. For the purposes
of this study, the buffer effect was therefore ignored.
The rate of hydrolysis of the dinucleotides, kobs, was
carried out in the pH range from 5 to 11.5 (Figure 2).
The hydrolytic cleavage of UpU was extremely slow
under the assay conditions, and degradation to uridine
and 2′,3′-cUMP was only observed at pH 11.5 (kobs ) 14
× 10-6 s-1). In contrast, UpsU, amUpsU, UpnU, and
amUpnT were readily hydrolyzed but gave markedly
different pH profiles as discussed below.
The plot of log kobs versus pH, for UpsU (Figure 2)
increases linearly, with increasing pH in the region from
pH 6 to 9, with a gradient of 0.9 indicating that hydrolysis
is base-catalyzed. This is consistent with the observa-
tions made by Liu and Reese,16 also with UpsU. The
hydrolysis of UpsU proceeded exclusively to 2′,3′-cUMP,
5′-thio-5′-deoxyuridine (10), and its disulfide (11) (Table
1). In contrast, UpnU hydrolysis is log linear in the pH
range from 5 to 6, with a slope of -0.8, which indicates
acid catalysis in this region. By pH 7.5 the hydrolysis
rate has plateaued to a value of 140 × 10 -6 s-1, where it
The hydrolysis of dinucleotide amUpnT increases log
linearly with decreasing pH between pH 11.5 and 7, with
a gradient of -0.7, indicating that the reaction is acid-
catalyzed. Hydrolysis proceeded to yield 5′-amino-5′-
(17) Aurup, H.; Tuschl, T.; Benseler, F.; Ludwig, J .; Eckstein, F.
Nucleic Acids Res. 1994, 22, 20-24.
(18) Miller, P. S.; Bhan, P.; Kan, L.-S. Nucleosides Nucleotides 1993,
12, 785-792.
(19) Kumamoto, J .; Cox, J . R.; Westheimer, F. H. J . Am. Chem. Soc.
1956, 78, 4858-4860.
(16) Liu, X.; Reese, C. B. Tetrahedron Lett. 1995, 36, 3413-3416.
(20) Hall, C. R.; Inch, T. D. Tetrahedron 1980, 36, 2059-2095.

Diuridine Phosphate Analogues
J . Org. Chem., Vol. 61, No. 18, 1996 6275
Sch em e 2.a Syn th esis of Up n U a n d a m Up n T
a
(i) (N,N′-diisopropylamino)dimethoxyphosphane, tetrazole; (ii) 5′-azido-5′-deoxynucleoside, LiCl, pyridine; (iii) thiophenol, triethylamine;
(iv) 80% AcOH; (v) tetrabutylammonium fluoride, and (vi) Na2CO3/MeOH/H2O.
Sch em e 3. Hyd r olysis P r od u cts of a m Up sU a t p H 6 a n d 10
deoxythymidine (Table 1) and compound 12, which
subsequently degraded as indicated in Scheme 3.
The dinucleotides TpsT, TpnT, and amUpT9 were
completely stable toward hydrolysis under these condi-
tions, emphasizing the requirement of an internal 2′-
hydroxyl nucleophile for hydrolysis.
The reaction products of dinucleotide hydrolysis reac-
tions were identified by comparison with authentic
samples on HPLC and by 31P NMR spectroscopy (Table
1). The hydrolysis products 9, 10, 11, 14, and 15 were
identified by HPLC and, where possible, 31P NMR
spectroscopic (pH 10) comparison with authentic samples.
To the best of our knowledge there is no data available
for cyclic 1,2,3-oxazaphospholidin-2-ones, such as 12, with
F igu r e 2. log kobs (s-1) versus pH for the hydrolysis of UpsU
([), amUpsU (]), UpnU (1), amUpnT (3) and UpU (9).
previous studies having been carried out with the alky-
lated nitrogen derivatives.20,21 Compound 12 was as-
signed on the basis of its 31P NMR multiplet at 28 ppm,
which is similar to the chemical shift of 20 ppm observed
by J ohnson et al.22 for an N-methylated oxazaphospho-
lidin-2-one ester. Compound 13 was assigned on the

6276 J . Org. Chem., Vol. 61, No. 18, 1996
Thomson et al.
Ta ble 2. Dep en d en ce of th e Ra te of Up sU, Up n U, a n d
a m Up sU Hyd r olysis w ith Meta l Ion s (10 m M, p H 7.5)a
be determined due to its instability at pH 7.5, where its
rate of hydrolysis is approximately 1 order of magnitude
faster than those of UpsU, amUpsU, and UpnU.
none
Mg2+
Zn2+
Cd2+
UpsU
UpnU
amUpsU
97
140
89
227
197
93
1300
1491
223
803
1682
3083
Discu ssion
Polyribonucleotides and oligoribonucleotides containing
a 2′-amino group as opposed to a 2′-hydroxy group are
stable to alkaline, RNaseA, and hammerhead ribozyme-
induced hydrolyses.9-12 This amino group should how-
ever be sufficiently nucleophilic to attack the phosphorus
of the internucleotidic linkage in a transesterification
reaction to yield the cyclic phosphoramidate. The pKa
of the 2′-amino group is 6.2,17,18 and therefore protonation
cannot be the cause for the observed stability above pH
7. We considered the properties of the leaving group to
be important for a successful displacement reaction
involving the 2′-amino group, and therefore the uridine
dinucleotide analogues UpsU, amUpsU, UpnU, and
amUpnT were prepared to systematically investigate
their hydrolytic properties.
The hydrolysis of uridine dinucleotide UpU has been
extensively studied in the past.24-27 It is extremely stable
toward acid-base hydrolysis with a kobs ) 9.8 × 10-9 s-1
at pH 8.36 and 41 °C.28 As a result of this stability,
elevated temperatures are required to obtain reliable
kinetic data for hydrolysis at near neutral pH values.26
In order to assess the influence of the leaving group on
the hydrolysis/transesterification of UpU, we have changed
the leaving group from a 5′-hydroxyl to a thiol and a
primary amine. The hydrolysis rates of UpU and of these
analogues are compared (Figure 2). The log kobs vs pH
profile for UpsU, in the range 6-9, indicates that UpsU
hydrolysis is base-catalyzed, and this is indicative of the
deprotonation of the 2′-hydroxyl group being rate-limit-
ing. This is consistent to that observed for UpU at 90
°C26 and uridine 3′-(o-carboxyphenyl) ester,29 which were
also base-catalyzed in this region and only became pH
independent as the pH values neared the pKa of the 2′-
hydroxyl group, pKa ) 13.30 Additionally, the hydrolysis
of UpU, at pH 8,28 is at least 5 orders of magnitude slower
than that of UpsU, which is consistent with the differ-
ences between the rate of hydrolysis of alkyl S-phospho-
rothioate and phosphate esters.31,32 Hydrolysis of
amUpsU, in contrast to the stability of amUpT, was only
pH dependent in the region from pH 5 to 7, above which
the hydrolysis rate plateaus to its maximum value at pH
a
kobs (10-6 s-1); 15% error between experiments.
Ta ble 3. Mich a elis-Men ten P a r a m eter s for Up U, Up sU,
a n d Up n U a s Su bstr a tes for P a n cr ea tic Ribon u clea se A
kcat
KM
(mM)
kcat/ KM
c
(s-1
5.3
195
6.5
)
(mM-1 s-1
)
UpUa
8
7.3
4
0.66
26.7
1.6
UpnUa
UpsUb
a
b
c
HEPES (50 mM), NaCl (200 mM). HEPES (500 mM). kcat
15% error between experiments.
basis of its 31P NMR chemical shift of 7.5 ppm, similar
to the chemical shifts of UpnU and TpnT at 9.3 and 8.4
ppm, respectively. This is consistent with our observa-
tion that there is not a big change in chemical shift
between the phosphate monoester, uridine-2′(3′)-mono-
phosphate (4 ppm), and the phosphate diester, UpU (0
ppm), at pH 10. The intermediacy of the cyclic phosphor-
amidate 12 is reasonable, particularly considering that
both TpsT and TpnT, which have no 2′-hydroxyl group,
are completely stable toward hydrolysis in the pH range
from 5 to 11.5.
Dep en d en ce of Hyd r olytic Sta bility u p on th e
P r esen ce of Diva len t Meta l Ca tion s. The rates of
hydrolysis of UpsU, amUpsU, and UpnU were also
determined in the presence of Mg2+, Zn2+, and Cd2+
(Table 2), but no reliable data could be obtained for
amUpnT due to its instability at pH 7.5. The hydrolysis
rates in the absence of metal ions did not change upon
addition of 10 mM EDTA, which implied that the buffers
did not contain any metal ion contamination. In the
presence of Mg2+, there is a relatively small increase in
the rate of hydrolysis with all three dinucleotides. In the
presence of Zn2+ or Cd2+, an increase in hydrolysis rate
of ca. 10-fold is observed with UpsU and UpnU. In
contrast, amUpsU hydrolysis was not significantly en-
hanced in the presence of either Mg2+ or Zn2+; however,
a 35-fold increase was observed in the presence of Cd2+
.
Su bstr a te P r op er ties w ith RNa se A. The Michae-
lis-Menten parameters for the reaction of UpU, UpsU,
and UpnU with RNase A were determined (Table 3). For
UpU degradation kcat/KM is 0.66 mM-1 s-1 and is ap-
proximately 4-fold lower than that previously published,
kcat ) 11 s-1, KM ) 3.7 mM, and kcat/KM ) 2.97 mM-1
11.5, kobs ) 138 × 10-6 s-1
. The profile observed for
amUpsU (Figure 2) correlates well with the pKa of the
2′-ammonium group of 6.2 determined for amUpT.17,18
Thus, as the pH decreases, the equilibrium shifts toward
the protonated amine, which is not nucleophilic. How-
ever, above ca. 7.2, one unit above the pKa, the amino
species predominates and hydrolysis plateaus to its
maximum rate.
s-1 23
Due to the presence of AcOH in our sample of
.
UpsU, its kinetic determination was carried out in
HEPES (pH 7.5, 500 mM), compared with HEPES (pH
7.5, 50 mM) and NaCl (200 mM) for UpnU and UpU.
Hydrolysis of UpU using both buffer conditions gave
values within experimental error. UpnU has a 40-fold
higher catalytic efficiency as a result of the same increase
in the value of kcat. In contrast the 2-fold increase in
catalytic efficiency observed for UpsU is mainly a result
of its 2-fold lower KM value. No data for amUpnT could
(24) Butzow, J . J .; Eichhorn, G. L. Biochemistry 1971, 10, 2019-
2027.
(25) Ikenaga, H.; Inoue, Y. Biochemistry 1974, 13, 577-582.
(26) J a¨rvinen, P.; Oivanen, M.; Lo¨nnberg, H. J . Org. Chem. 1991,
56, 5396-5401.
(27) Kuusela, S.; Lo¨nnberg, H. J . Chem. Soc. Perkin Trans. 1 1994,
2301-2306.
(28) Chapman, W. H. J .; Breslow, R. J . Am. Chem. Soc. 1995, 117,
5462-5469.
(29) Chandler, A. J .; Kirby, A. J . J . Chem. Soc., Chem. Commun.
1992, 1769-1770.
(21) Brown, C.; Boudreau, J . A.; Hewitson, B.; Hudson, R. F. J .
Chem. Soc. Perkin Trans. 1 1976, 888-895.
(30) Narlikar, G. J .; Gopalakrishnan, V.; McConnell, T. S.; Usman,
N.; Herschlag, D. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 3668-3672.
(31) Hudson, R. F.; Keay, L. J . Chem. Soc. Perkin Trans. 1 1956,
3269-3271.
(22) J ohnson, C. R.; Elliott, R. C.; Penning, T. D. J . Am. Chem. Soc.
1984, 106, 5019-5020.
(23) Richards, F. M.; Wyckoff, H. W. In The Enzyme; Boyer, P. D.,
Ed.; Academic Press: New York 1971; Vol. IV, p 647-806.
(32) Thain, E. M. J . Chem. Soc. Perkin Trans. 1 1957, 4694-4699.

Diuridine Phosphate Analogues
J . Org. Chem., Vol. 61, No. 18, 1996 6277
In light of the fact that nucleophilic displacement
reactions with primary amines on phosphate diesters
have been reported for the methyl 2,4-dinitrophenyl
diesters at pH 10, a contributing factor in the hydrolysis
of both UpsU and amUpsU is the pKa of the leaving
group. Such phosphoryl transfer reactions typically occur
with rates of 8 × 10-2 M-1 min-1, where the leaving group
is 2,4-dinitrophenol, which has an extremely low pKa of
4.33 The dependence of rates on the pKa of the leaving
group has been demonstrated for reactions of nicotina-
mide with phosphate monoester monoanions, for which
the nucleophilic substitution reaction is the same as for
diester monoanions.34 This dependence is also reported
for the reaction of quinuclidine and pyridine nucleophiles
with phosphate ester monoanions.35 Thus, the hydrolysis
of UpsU and amUpsU, compared to UpU and amUpT, is
clearly more facile due to the differences in pKa between
the 5′-thio (pKa ) 11) and the 5′-oxo (pKa ) 16) leaving
group.36 This difference in hydrolytic activity can also
be attributed to the much lower P-S bond strength, ca.
50 kcal mol-1, compared to the P-O bond strength of ca.
91 kcal mol-1 and the P-N bond strength of ca. 70 kcal
group than uridine, and hydrolysis is facilitate by the
ready protonation of the 5′-imino function. Introduction
of 2′-amino groups clearly stabilizes phosphodiester bond
hydrolysis with both 5′-thio- and 5′-amino leaving groups,
but only above pH’s 7 and 10, respectively. This study
demonstrates that the stability of 2′-amino containing
poly- and oligoribonucleotides is due to the poor leaving
characteristics of the 5′-nucleoside and the greater nu-
cleophilicity of a 2′-oxo compared to a 2′-amine.
Met a l Ion Dep en d en ce. The rate of hydrolysis of
UpU24,25,27,39 and phosphate diesters29,39-42 is dramatically
increased in the presence of divalent metal ions and the
metal ion is thought to co-ordinate to the phosphate
group. Zn2+ and Cd2+ were used in this study since their
metal-aquo complexes have similar pKa values, but Cd2+
is considered “softer” and therefore it should enable
differentiation between activation of hydrolysis through
general base catalysis and co-ordination to the 5′-thio
leaving group. The observed rate enhancements in UpsU
hydrolysis do not correlate with what would be expected
from increased general-base catalysis, since the pKa
values are 11.4, 8.8, and 9.0 for the aquo/hydroxy
complexes of Mg2+, Zn2+, and Cd2+, respectively.43 Ad-
ditionally, on the basis of the “hardness” and “softness”
of Zn2+ and Cd2+, respectively, one would have expected
an enhancement in UpsU hydrolysis in the presence of
Cd2+. The ca. 2-, 11-, and 12-fold rate enhancements
observed for UpsU and UpnU hydrolysis in the presence
of Mg2+, Zn2+, and Cd2+, respectively, is comparable to
that observed for UpU hydrolysis27 and does not correlate
with increased general-base catalysis.
mol-1 37
.
Hydrolysis of UpnU below pH 6 is acid-catalyzed but
the rate of hydrolysis between pH 6 and 11.5 is constant,
ca. 140 × 10-6 s-1 (Figure 2). A log linear increase in
the hydrolysis of UpU26 and uridine 3′-(o-carboxyphenyl)
ester29 was also observed in the acidic pH range down to
pH 1, which was attributed to protonation of the phos-
phate diester with a pKa of ca. 1.5.38 With UpnU there
exists the additional possibility for rate acceleration
through protonation of the 5′-imino group, for which the
pKa is not known. The leveling off of the hydrolysis rate
of UpnU above pH 6 is not clearly understood. One
would have expected a log linear increase in the rate of
hydrolysis as the pKa of the 2′-hydroxyl group is ap-
proached, since this increases the concentration of the
2′-oxo anion species, which is more nucleophilic than a
2′-hydroxyl. However, at higher pH one would also
expect a decrease in the rate as the protonation of the
5′-imino becomes more difficult, thus possibly canceling
out any increase derived from the greater concentration
of the 2′-oxo anion. The profile observed for amUpnT
hydrolyis gives a straight line of gradient -0.7 (Figure
2), which supports the idea that protonation of the 5′-
imino group is rate-limiting. It is more readily hydro-
lyzed than UpnU and it is not until above pH 10 that
amUpnT exhibits greater hydrolytic stability than UpnU.
Thus, the 2′-oxo anion in UpnU more readily displaces
the 5′-amine than does the 2′-amino group in amUpnT,
as expected from the difference in P-O and P-N bond
strengths discussed above.
Metal ion dependent hydrolysis of amUpsU is poten-
tially the most interesting, since the general-base proper-
ties of the metal ion-aquo complex should not be
important because the 2′-amino group is mainly depro-
tonated at pH 7.5 and does not require activation by a
metal ion to enable nucleophilic attack on the phospho-
rus. Thus activation of the 5′-thio leaving group by a
“soft” metal ion might be most easily visible. The
increase in rate of hydrolysis from Mg2+ to Zn2+ is 2-fold
and is another 15-fold to Cd2+. This is the effect one
would expect if there was rate acceleration by the soft
metal ion co-ordinating to the soft thio leaving group.
There are no numbers available for the preference of co-
ordination of metal ions to oxygen vs sulfur except for
the difference in interaction with ATP and ATPâS.44
There it is concluded that the ratio of oxygen vs sulfur
coordination is 31 000 for Mg2+, 3100 for Ca2+ , 160 for
Mn2+, and 0.018 for Cd2+
.
It is very difficult to establish exactly how metal ions
promote hydrolysis of dinucleotides, as several effects
may contribute to the rate enhancement. (1) The metal
ion aquo-hydroxy complex can enhance deprotonation
or protonation by general-acid-base catalysis. (2) The
metal ion could co-ordinate to the phosphate acting as a
Lewis acid, lowering the negative charge on the phos-
phate group, making it more susceptible to nucleophilic
Clearly 5′-thiouridine is a better leaving group than
uridine and two factors, the P-S bond strength and the
lower pKa of 5′-thiouridine of approximately 11 vs uridine
of approximately 16, enable it to be displaced by a
primary amine as a nucleophile in an intramolecular
reaction. 5′-Aminouridine is also a much better leaving
(39) Breslow, R.; Huang, D.-L. Proc. Natl. Acad. Sci. U.S.A. 1991,
88, 4080-4083.
(33) Kirby, A. J .; Younas, M. J . Chem. Soc. (B) 1970, 510-513.
(34) Kirby, A. J .; Varvoglis, A. G. J . Chem. Soc. (B) 1968, 135-141.
(35) Herschlag, D.; J encks, W. P. J . Am. Chem. Soc. 1989, 111,
7587-7596.
(36) March, J . Advanced Organic Chemistry, 4th ed.; J ohn Wiley
and Sons: New York, 1992.
(37) Hudson, R. F. Structure and Mechanisms in Organo-Phosphorus
Chemistry; Academic Press: London, 1965.
(38) Kumler, W. D.; Eiler, J . J . J . Am. Chem. Soc. 1943, 65, 2355-
2361.
(40) Breslow, R.; Berger, D.; Huang, D.-L. J . Am. Chem. Soc. 1990,
112, 3686-3687.
(41) Browne, K. A.; Bruice, T. C. J . Am. Chem. Soc. 1992, 114, 4951-
4958.
(42) Dempcy, R. O.; Bruice, T. C. J . Am. Chem. Soc. 1994, 116,
4511-4512.
(43) Basolo, F.; Pearson, R. Mechanismen in der Anorganischen
Chemie; Georg Thieme Verlag: Stuttgart, 1973.
(44) Pecoraro, V. L.; Hermes, J . D.; Cleland, W. W. Biochemistry
1984, 23, 5262-5271.

6278 J . Org. Chem., Vol. 61, No. 18, 1996
Thomson et al.
attack from the 2′-hydroxyl. This is similar to protona-
tion of the phosphate diester, which has been previously
attributed to the increase in hydrolysis of UpU26 and
uridine 3′-(o-carboxyphenyl) ester29 under acidic condi-
tions. (3) Direct co-ordination of the metal ion to the
leaving group would stabilize the bond cleavage step
without requiring protonation, as proposed for the ham-
merhead45 and tetrahymena46 ribozymes. Care must also
be taken when extrapolating data obtained with activated
esters, since clearly they can hydrolyze by a different
mechanism as a result of the lower pKa of the leaving
group. In a study on the cleavage of a hammerhead
ribozyme substrate containing a 5′-thio leaving group, it
was concluded that no metal ion was bound to the leaving
group in the transition state. This does not however
imply that the native hammerhead ribozyme, which
contains a 5′-oxo leaving group, does not bind a metal
ion to stabilize the leaving group, since the rate-
determining step for the hydrolysis reaction may be
different for the two substrates on account of the different
pKa’s of the leaving group. Thus, the 5′-thio leaving
group may not need to be stabilized by protonation or
metal ion co-ordination. This difficulty in attributing rate
enhancements to a particular effect is exemplified by the
linkage into the cleavage site of ribozymes would be a
suitable modification to probe the sensitivity of the
leaving group toward protonation. If protonation does
operate in the cleavage mechanism of ribozymes, one
would expect to see an increase in the rate of cleavage
at pH 7 and a change in the pH vs rate profile.
It is interesting to note that amUpsU is not a substrate
for RNaseA whereas UpsU is, indicating that P-S bond
cleavage is not a problem. Although it is not possible at
present to identify the reason for this stability, one might
speculate that the 2′-amino group is not placed in the
correct orientation for the reaction to proceed, as the 2′-
amino nucleoside is predominantly in the 2′-endo con-
formation and not the 3′-endo conformation usually
adopted by the ribonucleosides.
Con clu sion s
We have undertaken the synthesis of the uridine
dinucleotides UpsU, amUpsU, UpnU, and amUpnT and
studied their hydrolytic properties under various condi-
tions. These dinucleotides are readily hydrolyzed in the
pH range from 5 to 11.5. with the most important finding
being that amUpsU and amUpnT are hydrolyzed in
contrast to amUpT, indicating the importance of the
leaving group when the 2′-amino group is the nucleophile.
The metal ion effect of Cd2+ on the hydrolysis of amUpsU
is a clear example for metal ion interaction with the
leaving group. UpnU demonstrated a marked rate
enhancement, relative to UpU and UpsU, when cleaved
by RNase A, consistent with the postulated protonation
of the 5′-amino leaving group in the hydrolysis of UpnU
and amUpnT.
hydrolysis of various XpXp(2′) dinucleotides with Zn2+
,
Cu2+, Co2+, and Ni2+, the four metal ions for which data
is available.24 The rates are all within 1 order of magni-
tude of each other and therefore it seems that the
structure of the metal ion/nucleotide complex is more
important for the rate of reaction than enhanced depro-
tonation of the 2′-hydroxyl group. Thus, the metals
enhance hydrolysis irrespective of the leaving group for
UpsU and UpnU, indicating a similar mode of enhance-
ment for both.
Exp er im en ta l Section
The results presented here for amUpsU hydrolysis are
consistent with the concept that Cd2+ can stabilize the
developing negative charge on the 5′-thio- leaving group;
however, it is not clear why this enhancement does not
also manifest itself in UpsU hydrolysis. Thus, the data
presented here should caution the interpretion of results
in terms of particular interactions as at present a clear
understanding of metal ion-catalyzed hydrolysis of RNA
or oligoribonucleotides is far from complete.
R Na se-Ca t a lyzed H yd r olysis. The above com-
pounds were also tested as substrates for pancreatic
ribonuclease (RNase A) which catalyses RNA phosphodi-
ester bond cleavage 3′ to pyrimidine nucleosides forming
2′,3′-cyclic phosphate and 5′-hydroxy termini.23 Of the
five dinucleotides investigated only UpU, UpnU, and
UpsU were cleaved, but amUpsU and amUpnT were not.
The three cleavable dinucleotides had only marginal
differences in their KM values (Table 3), which suggests
that association of the substrate with the enzyme in the
ground state is approximately the same for all of them.
The kcat values for UpU and UpsU were also similar.
However, UpnU displays a kcat value 35-fold higher than
that observed for UpsU and UpU. This argues that
UpnU degradation by RNase A is enhanced, compared
to UpU and UpsU, because it is more easily protonated
in the active site of the enzyme by the general acid His-
119. In contrast to UpU, which is only acid-catalyzed
below pH 526 the hydrolysis of UpnU is acid-catalyzed
below ca. pH 7 and incorporation of a phosphoramidate
Ma ter ia ls a n d Meth od s. Ribonuclease inhibitor (110
units/mL) was purchased from United States Biochemicals and
bovine pancreatic ribonuclease (RNase A) (DNase free) was
obtained from Boehringer Mannheim. 5′-O-(4,4′-Dimethoxy-
trityl)-2′-O-(tert-butyldimethylsilyl)uridine 3′-O-(â-cyanoethyl
N,N′-diisopropylphosphoramidite) and 5′-O-(4,4′-dimethoxytri-
tyl)thymidine 3′-O-(â-cyanoethyl N,N′-diisopropylphosphor-
amidite) were obtained from Applied Biosystems (Dreieich,
Germany) and Milliore BioSyntech (Hamburg, Germany),
respectively. 5′-Iodo-5′-deoxyuridine,47 5′-azido-5′-deoxyuri-
dine,48 5′-amino-5′-deoxyuridine,49 5′-thio-5′-deoxyuridine,50 5′-
O-(4,4′-dimethoxytrityl)-2′-deoxy-2′-[(trifluoroacetyl)amino]-
uridine 3′-O-(â-cyanoethyl N,N′-diisopropylphosphoramidite),9
thymidylyl (3′-5′) 5′-amino-5′-deoxythymidine (TpnT),51 and
2′-amino-2′-deoxyuridylyl (5′-3′)-thymidine17 were prepared
as previously published. All other reagents were purchased
commercially and used as received except for dry pyridine, 1,4-
dioxane, and tetrahydrofuran, which were stored over 4 Å
molecular sieves. CH2Cl2 used in phosphitylation reactions
was stored over sodium-lead alloy (10% Na, Merck) and
passed down a column of basic alumina prior to use. Analyti-
cal thin layer chromatography (TLC) was carried out using
silicagel 60 F254 plates (Merck), which were developed with
one of the following solvent systems: solvent A, CH2Cl2/EtOAc/
Et3N (47.5:47.5:5 v/v); solvent B, CH2Cl2/EtOAc/Et3N (5:90:5
v/v); solvent C, CH2Cl2/MeOH (90:10 v/v); solvent D, CH2Cl2/
(47) Tsuji, T.; Takenaka, K. Nucleosides Nucleotides 1987, 6, 575-
580.
(48) Yamamoto, I.; Sekine, M.; Hata, T. J . Chem. Soc. Perkin Trans.
1 1980, 306-310.
(49) Saneyoshi, M.; Fujii, T.; Kawaguchi, T.; Sawai, K.; Kimura, S.
In Nucleoside Chemistry (part 4); Townsend, L. B. a. T., L. S., Ed.; J .
Wiley: New York, 1987.
(50) Reist, E. J .; Benitez, A.; Goodman, L. J . Org. Chem. 1964, 29,
554-558.
(51) Mag, M.; Engels, J . W. Tetrahedron 1994, 50, 10225-10234.
(45) Sawata, S.; Komiyama, M.; Taira, K. J . Am. Chem. Soc. 1995,
117, 2357-2358.
(46) Piccirilli, J . A.; Vyle, J . S.; Caruthers, M. H.; Cech, T. R. Nature
1993, 361, 85-88.

Diuridine Phosphate Analogues
J . Org. Chem., Vol. 61, No. 18, 1996 6279
MeOH (80:20 v/v); solvent E, i-PrOH/NH3/H2O (55:35:10 v/v).
Column chromatography was performed on silica gel 60
(Merck) of particle size 0.063-0.200 mm.
2′-O-(ter t-Bu tyldim eth ylsilyl)u r idin e 3′-ph osph or oth io-
a te (1). 5′-O-(4,4′-Dimethoxytrityl)-2′-O-(tert-butyldimethyl-
silyl)uridine 3′-O-(â-cyanoethyl N,N-diisopropylphosphoramid-
ite) (500 mg, 580 µmol) was dissolved in anhydrous CH3CN (3
mL) under an argon atmosphere. A solution of tetrazole, in
CH3CN (0.5 M, 1.5 mL, 1.3 equiv), and 3-hydroxypropionitrile
(44 µL, 1.2 equiv) was then added, and the mixture stirred at
room temperature for 1.5 h, Rf (starting material) 0.8 (solvent
A). The reaction mixture was then added, under argon, to 3H-
1,2-benzodithiol-3-one 1,1-dioxide (Beaucage reagent)54 dis-
solved in CH3CN, (0.2 M, 15 mL), and stirring continued for a
further 2.5 h. The solution was concentrated in vacuo and
the residue was suspended in H2O and extracted with ether
(3 × 60 mL), and the organic phase was dried (Na2SO4),
filtered, and evaporated to dryness. The residue was then
suspended in MeOH (15 mL)/ammonia (32%, 56 mL) and the
mixture heated at 55 °C for 2 h. After cooling on ice for 15
min, this mixture was concentrated in vacuo, AcOH (80%) was
added, and the mixture was kept at room temperature for 30
min before being evaporated to dryness. The residue was then
dissolved in H2O (40 mL) and washed with ether (2 × 40 mL),
and the aqueous phase was evaporated to dryness. The white
crude product was then dissolved in a small volume of H2O
and purified by chromatography using DEAE-Sephadex A-25
eluted with a linear gradient of TEAB (0 to 0.5M, total volume
3 L). This yielded the disulfide-bridged nucleotide dimer,
which after concentration in vacuo was taken up in H2O (2.8
mL), treated with an aqueous solution of dithiothreitol (1 M,
100 µL), and maintained at room temperature for 4.5 h. The
mixture was subsequently evaporated to dryness, dissolved in
TEAB (100 mM, 1 mL), and purified by chromatography over
DEAE-Sephadex A-25 eluted with a linear gradient of TEAB
(0-0.5 M, total volume 1 l) to yield 1 (1170 A260 units, 117
µmol, 20%): 31P NMR (D2O) 56 ppm; MS m/ z 453.093 483,
calcd for C15H26N2O8S1P1Si1 453.091 680.
2′-O-(ter t-Bu tyld im eth ylsilyl)u r id ylyl (3′-5′) 5′-Th io-
5′-d eoxyu r id in e. Compound 1 (1170 A260 units, 117 µmol)
was loaded onto a Merck I ion exchange resin (pyridinium
form) column (2 × 16 cm) and eluted with H2O (400 mL) into
a solution of tri-n-butylamine (254 µmol, 3 equiv, 90 µL) in
MeOH (3 mL). This solution was then evaporated to dryness
and the residue redissolved in MeOH (5 mL). The tri-n-
butylammonium salt of 2′-(O-tert-butyldimethylsilyl)uridine 3′-
monophosphate (500 A260 units, 50 µmol, 2.12 mL) was then
evaporated to dryness, the residue dried by twice evaporation
from anhydrous N,N′-dimethylformamide (2 mL) and then
dissolved in anhydrous N,N′-dimethylformamide (1 mL). 5′-
Iodo-5′-deoxyuridine (35.3 mg, 2 equiv) and tri-n-butylamine
(36 µL, 3 equiv) were then added. The reaction was then
maintained at room temperature for 48 h, after which conver-
sion to product was quantitative, as observed by 31P NMR. The
reaction mixture was evaporated to dryness, dissolved in H2O
(5 mL), and purified by chromatography over DEAE-Sephadex
A-25 eluted with a linear gradient of TEAB (0-0.5 M, total
volume 1 l) to yield the product (598 A260 units, 27 µmol,
50%): HPLC tR (buffer system 1, gradient 2) 27.5 min; 31P
Mass spectral52 and NMR53 analyses were carried out as
previously described. Analytical and preperative HPLC were
carried out using the following buffer systems. Buffer system
1: buffer A, TEAA (50 mM); buffer B, TEAA (50 mM)/CH3CN
(70% v/v). Buffer system 2: buffer A, TEAB (100 mM); buffer
B, TEAB (100 mM)/CH3CN (70% v/v). Gradient 1, 20 min
linear gradient from 100% to 80% buffer A; gradient 2, 40 min
linear gradient from 100% to 50% buffer A; gradient 3, 25 min
linear gradient from 80% to 0% buffer A.
Kin et ic An a lysis for t h e Hyd r olyt ic Degr a d a t ion of
Din u cleotid es. The rate of dinucleotide hydrolysis was
established by fitting the fraction of remaining substrate,
relative to an internal standard, to a single exponential decay.
Buffers were as follows, with pH adjusted at 37 °C: sodium
acetate (pH 5, 100 mM); Na2HPO4-citric acid (pH 6, 100 mM);
HEPES (pH 7 to 8, 100 mM); MES (pH 9, 100 mM); glycine-
NaOH (pH 9, 100 mM); carbonate-bicarbonate (pH 10, 100
mM) and disodium phosphate (pH 11.5, 100 mM), NaCl
solution was used to adjusted ionic strength to 1 M. Reaction
mixtures were prewarmed at 37 °C before the reaction was
initiated by addition of the dimer solution containing the
internal standard nucleoside. Concentrations of dimer and
internal standard were UpnU (400 µM) and ddT (800 µM),
UpsU (200 µM) and ddA (800 µM), amUpsU (400 µM) and dA
(800 µM), or amUpnT (200 µM) and dA (400 µM). With the
exception of amUpnT hydrolysis, aliquots (5 µL) were removed
at appropriate time points, diluted with TEAA buffers (50 mM,
40 µL), frozen at liquid nitrogen temperature, and maintained
at -80 °C until required for analysis and the fraction of
remaining dinucleotide established by HPLC analysis (buffer
system 1, gradient 1). For amUpnT the aliquots were diluted
with glycine-NaOH buffer (pH 12, 500 mM, 40 µL) and
analysed as above. Dependence of the hydrolysis rates upon
the buffer concentration was assessed in the above manner
as described above using HEPES (pH 7.5), which was varied
from 50 to 500 mM. Dependence of hydrolysis rates upon the
presence of metal ions was also carried out as described above
using HEPES (pH 7.5, 100 mM) in the presence of the relevant
metal ion (10 mM) and reaction was quenched with EDTA (50
mM, pH 7.5., 10 µL), frozen in liquid nitrogen, and maintained
at -80 °C until required for HPLC. Duplication of the
hydrolysis reactions gave an error of 15% for the values of kobs
.
Mich a elis-Men ten P a r a m eter s for RNa se A Clea va ge
of th e Din u cleotid es. Kinetic constants kcat and KM were
determined from Eadie-Hofstee plots obtained from initial
velocities, at various substrate concentrations, under multiple
turnover conditions. Nucleotide dimer concentrations of 0.5-6
mM were used and the RNase A concentration was varied from
18.25 to 365 nM, depending upon the rate of the reaction. The
concentration of internal standard, ddT for hydrolysis with
UpnU and UpU, ddA for hydrolysis with UpsU, and dA for
hydrolysis with amUpsU, was from 2 to 8 mM. For UpU and
UpnU a solution containing the nucleotide dimer, internal
standard, HEPES (pH 7.5, 50 mM), and NaCl (200 mM) was
incubated for 10 min at 25 °C and the reaction initiated by
addition of RNase A. For UpsU a solution containing the
nucleotide dimer, internal standard, and HEPES (pH 7.5, 500
mM) was incubated for 10 min at 25 °C and the reaction
initiated by addition of RNase A. Aliquots (5 µL) were removed
at appropriate time points, quenched by addition to a solution
of RNAse inhibitor (1.6 units/mL), HEPES (pH 7.5, 20 mM),
KCl (50 mM), dithiothreitol (5 mM), and glycerol (50%), and
maintained at -80 °C. Initial rates were established by
quantifying the amount of starting material remaining, with
respect to the internal standard, by HPLC analysis (buffer
system 1, gradient 1) of the collected time points. Duplication
of the RNase A-induced degradation reactions gave an error
NMR (D2O) 20.4 ppm; MS m/z 679.148 956, calcd for C24H36
-
N4O13S1P1Si1 679.150 651.
Ur id ylyl (3′-5′) 5′-Th io-5′-d eoxyu r id in e (Up sU). 2′-O-
(tert-Butyldimethylsilyl)uridylyl (3′-5′) 5′-thio-5′-deoxyuridine
(300 A260 units, 15 µmol) in H2O (1.3 mL) was evaporated to
dryness, dried by twice evaporation from tetrahydrofuran (3
mL), dissolved in triethylamine trihydrogen fluoride (1 mL),
and maintained at room temperature for 22 h. The product
was then purified by preparative HPLC (buffer system 2,
gradient 2) followed by lyophilization to yield a white solid
(135 A260 units, 6.8 µmol, 45%): HPLC tR (buffer system 1,
gradient 2) 9.6 min; 31P NMR (D2O) 18.8 ppm; MS m/z
565.064 941, calcd for C18H22N4O13S1P1 565.064 172.
2′-Am in o-2′-deoxyu r idin e 3′-P h osph or oth ioate (2). This
was prepared by the same method described for compound 1
but using 5′-O-(4,4′-dimethoxytrityl)-2′-(trifluoroacetamido)-
2′-deoxyuridine 3′-O-(â-cyanoethyl N,N-diisopropylphosphor-
of 15% for the values of kobs
.
(52) Eckart, K.; Spiess, J . Am. Soc. Mass Spectrom. 1995, 6, 912-
919.
(53) Ludwig, J .; Eckstein, F. J . Org. Chem. 1991, 56, 1777-1783.
(54) Iyer, R. P.; Phillips, L. R.; Egan, W.; Regan, J . B.; Beaucage, S.
L. J . Org. Chem. 1990, 55, 4693-4699.

6280 J . Org. Chem., Vol. 61, No. 18, 1996
Thomson et al.
amidite) (420 mg, 500 µmol) and S8 dissolved in CS2/pyridine
(1:1) for oxidation to give 2 (2540 A260 units, 254 µmol, 51%):
31P NMR (D2O) 44.45 ppm; MS m/z 338.020 500, calcd for
C9H13N3O7P1S1 338.021 185.
2′-Am in o-2′-d eoxyu r id ylyl (3′-5′) 5′-Th io-5′-d eoxyu r i-
d in e (a m Up sU). This was prepared by the same method
described for 2′-O-(tert-butyldimethylsilyl)uridylyl (3′-5′) 5′-
thio-5′-deoxyuridine but using compound 2 (500 A260 units, 50
µmol), tri-n-butylamine (36 µL, 3 equiv), and 5′-iodo-5′-
deoxyuridine (35 mg, 2 equiv, 100 µmol) to give amUpsU (496
A260 units, 24.8 µmol, 49.6%): 31P NMR (D2O) 20.16 ppm; MS
m/z 564.079 086, calcd for C18H23N5O12S1P1 564.080 157.
Th ym id in e (3′-5′) 5′-Th io-5′-d eoxyth ym id in e (Tp sT).
This was prepared by the same method described for amUpsU
but using thymidine 3′-thiomonophosphate (125 A267 units,
12.5 µmol), tri-n-butylamine (9 µl, 3 equiv), and 5′-iodo-5′-
deoxythymidine (9 mg, 2 equiv 25 µmol) to give TpsT (125 A267
units, 6.8 µmol, 50%): 31P NMR (D2O) 19.80 ppm; MS m/z 561
(M - H+).
indicated that the reaction had gone to 90% completion [(buffer
system 1, gradient 2,) tR (UpnU) 8 min, tR (5), 25 min].
Purification was carried out by preparative HPLC (buffer
system 2, gradient 1), and the appropriate fractions were
collected and frozen before being combined and lyophilized to
yield UpnU as a white solid containing ca. 10% 2′,3′cUMP.
The reaction product was dissolved in H2O (2 mL) and stored
at -20 °C (225 A260 units, 11.3 µmol, 75%): 31P NMR (D2O)
9.29 (s), [cUMP 20.83 (s)]; MS m/z 548.101 456, calcd for
C18H23N5O13P1 548.103 000.
(Rp /Sp )-5′-O-(4,4′-Dim eth oxytr ityl)-2′-d eoxy-2′-(tr iflu o-
r oa cet a m id o)u r id ylyl (3′-5′) 5′-Am in o-5′-d eoxyt h ym i-
d in e Meth yl Ester (6). This was prepared by the same
method described for compound
3 but using 5′-O-(4,4′-
dimethoxytrityl)-2′-deoxy-2′-(trifluoroacetamido)uridine (1.364
g, 2.12 mmol) dissolved in CH2Cl2:CH3CN (1:1, v/v, 40 mL),
tetrazole (0.5 M, 6.36 mL, 1.5 eq), (N,N′-diisopropylamino)-
dimethoxyphosphane (0.613 g, 3.18 mmol, 1.5 eq), LiCl (0.445
g, 10.6 mmol, 5 eq), and 5′-azido-5′deoxythymidine (0.566 g,
2.12 mmol, 1 eq). The reaction mixture was concentrated in
vacuo twice from toluene (2 × 10 mL), and the residue was
dissolved in EtOAc (300 mL) and subsequently washed with
5% NaHCO3 (100 mL) and NaCl (saturated 100 mL), dried
(Na2SO4), filtered, and evaporated to dryness. The crude
product was purified by silica gel chromatography followed by
preperative HPLC to give 6 (8060 A260 unit 403 µmol, 19%):
31P NMR (CD3OD) 12.58 (s) and 12.66 (s), 1:1; MS m/z
957.268 193, calcd for C43H46N6O14P1F3 957.268 349.
5′-O-(4,4′-Dim et h oxyt r it yl)-2′-d eoxy-2′-(t r iflu or oa cet -
a m id o)u r id ylyl (3′-5′) 5′-Am in o-5′-d eoxyth ym id in e (7).
This was prepared by the same method as described for
compound 4 but starting with compound 6 to give 7 (89%):
31P NMR (CD3OD) 8.5; MS m/z 943.255 030, calcd for
C42H44N6O14P1F3 943.2526 992.
(Rp /Sp )-5′-O-(4,4′-Dim et h oxyt r it yl)-2′-O-(ter t-b u t yld i-
m et h ylsilyl)u r id ylyl (3′-5′)-5′-Am in o-5′-d eoxyu r id in e
Meth yl Ester (3). 5′-O-(4,4′-Dimethoxytrityl)-2′-O-(tert-bu-
tyldimethylsilyl)uridine (1.336 g, 2 mmol) was dissolved in dry
CH2Cl2/CH3CN (10:7, v/v, 17 mL), and the flask was sealed
with a septum and flushed with argon. To this was then added
a solution of tetrazole dissolved in CH3CN (0.5 M, 3 mL, 0.75
equiv) and (N,N′-diisopropylamino)dimethoxyphosphane (0.404
g, 2.09 mmol, 1.05 equiv), and the solution was stirred for 4 h
at room temperature. With maintenance of the argon atmo-
sphere, anhydrous ether (50 mL) was then added and the
resultant white precipitate removed by filtration. The super-
natant was concentrated in vacuo and the residue dissolved
in anhydrous pyridine (5 mL). To this solution was added LiCl
(0.420 g, 10 mmol, 5 equiv) followed by 5′-azido-5′-deoxyuridine
(0.875 g, 3.25 mmol, 1.62 equiv), and the reaction mixture was
stirred for 20 h at room temperature, Rf (product) 0.25 (solvent
C). The crude product was then concentrated, in vacuo, and
purified by silica gel column chromatography, eluting with
CH2Cl2 (0%-10% MeOH) containing triethylamine (1%), fol-
lowed by preparative HPLC [(buffer system 2, gradient 3) tR
(3) 16 min, tR (5′-azido-5′-deoxyuridine) 6 min] to yield 3 (5560
A260 units, 278 µmol, 14%). 31P NMR (CD3OD) 12.30 (s) and
12.66 (s), 1:1; m/z 978.336 304, calcd for C46H57N5O15P1Si1
978.335 809.
5′-O-(4,4′-Dim eth oxytr ityl)-2′-O-(ter t-bu tyld im eth ylsi-
lyl)u r id yl (3′-5′) 5′-Am in o-5′-d eoxyu r id in e (4). Compound
3 (1000 A260 units, 50 µmol), in MeOH (1.8 mL), was evapo-
rated to dryness and evaporated twice from dioxane (0.5 mL).
To this residue was then added a solution of thiophenol/
triethylamine/dioxane (1:2:3, v/v, 6 mL) and the resulting
solution stirred at room temperature for 24 h; TLC (solvent
D) Rf (4) 0.4, Rf (3) 0.68. The reaction mixture was concen-
trated in vacuo, the residue was twice evaporated from MeOH
(2 mL), and the crude product was purified by silica gel
chromatography, eluting with CH2Cl2 (0%-10% MeOH) con-
taining triethylamine (1%), to yield 4 (871 A260 units, 43.5
µmol, 87%): 31P NMR (CD3OD) 8.4; MS m/z 964.320 159, calcd
for C45H55N5O15P1Si1 964.320 159.
2′-O-(ter t-Bu tyld im eth ylsilyl)u r id ylyl (3′-5′) 5′-Am in o-
5′-d eoxyu r id in e (5). To 4 (600 A260 units, 30 µmol) was added
acetic acid (80%, 200 µL) and the solution was thoroughly
shaken by vortexing. After 6 min, removal of DMTr was
quantitative by HPLC [(buffer system 1, gradient 2) tR (5) 17
min, tR (4) 25 min] and the reaction mixture was cooled in an
ice bath and brought to pH 7 by addition of NaOH (1.2 mL, 2
M). The reaction mixture was purified by chromatography
over DEAE-Sephadex A-25 eluted with a linear gradient of
TEAB (0 to 0.5M, total volume 1 l) to yield 5 (340 A260 units,
17 µmol, 57%): 31P NMR (CD3OD) 9.74; MS m/z 662.188 828,
calcd for C24H37N5O113P1Si1 662.189 479.
Ur id ylyl (3′-5′) 5′-Am in o-5′-d eoxyu r id in e Up n U. Com-
pound 5, dissolved in H2O (0.9 mL) (300 A260 units, 15 µmol),
was evaporated to dryness, the residue evaporated twice from
THF (500 µL), TBAF (500 µL, 1.1 M) added, and the solution
incubated for 16 h at rt. Subsequent analytical HPLC
2′-Deoxy-2′-(t r iflu or oa cet a m id o)u r id ylyl (3′-5′) 5′-
Am in o-5′-d eoxyth ym id in e (8). This was prepared by the
same method as described for compound 5 but starting with
compound 7 to give 8 (93%): 31P NMR (CD3OD) 9.71; MS m/z
641.122 494, calcd for C21H26N6O12P1F3 641.122 019
2′-Am in o-2′-d eoxyu r id ylyl (3′-5′) 5′-Am in o-5′-d eoxy-
th ym id in e (a m Up n T). To a solution of compound 8 in MeOH
(600A260 units, 30 µmol, 1 mL) was added Na2CO3 (10%, H2O/
MeOH 8:2, v/v, 6 mL), and the mixture was stirred for 8 h at
rt. Subsequent analytical HPLC indicated that the reaction
had gone to 80% completion; (buffer system 1, gradient 1,) tR
(amUpnT) 10.55 min, tR (8), 15.2 min. Purification was carried
out by preparative HPLC (buffer system 2, gradient 1), and
the appropriate fractions were collected and frozen before being
combined and lyophilized to yield amUpnT as a white solid
contaminated with ca. 12% of cyclic phosphoramidate (12). The
product was dissolved in H2O (2 mL) and stored at -20 °C
(380 A260 units, 19 µmol, 63%): 31P NMR (D2O + H2O, pH 12)
9.04 (s), [28.8, s, cUMnP; 7.69, s, 2′-amino-2′-deoxyuridine 2′-
monophosphoramidate].
2′-Am in o-2′-d eoxyu r id in e 3′-Mon op h osp h a te (9). 5′-
(4,4′-Dimethoxytrityl)-2′-deoxy-2′-(trifluoroacetamido)-2′-
deoxyuridine 3′-O-(â-cyanoethyl N,N-diisopropylphosphoram-
idite) (210 mg, 250 µmol) was dissolved in anhydrous CH3CN
(3 mL) under an argon atmosphere. A solution of tetrazole in
acetonitrile (0.5 M, 0.75 mL, 1.25 equiv) and 3-hydroxypropi-
onitrile (20.38 µL, 1.2 eq) was then added and the mixture
stirred at room temperature for 1.5 h. A solution of iodine
(0.5 M, ca. 0.8 mL) in THF/pyridine/H2O (7:2:1) was added
dropwise to the reaction mixture until the color persisted, and
stirring was continued for a further 30 min. NaHSO3 (5%,
1.2 mL) was then added and the solution concentrated in
vacuo. The residue was then dissolved in H2O (50 mL) and
extracted with CH2Cl2 (2 × 50 mL), and the organic phase was
dried (Na2SO4), filtered, and evaporated to dryness. Removal
of the protecting groups was performed as described for the
synthesis of 2 to give compound 9 (1035 A260 units, 103 µmol,
41%): 31P NMR (D2O) 4.16 (d) J PH ) 8.24 Hz; MS m/z
322.043 045 (M - H)- calcd for C9H14N3O8.P1 322.044 028.
Abbr evia tion s: HEPES, N-2-hydroxyethylpiperazine-N′-
2-ethanesulfonic acid; MOPS, 3-(N-morpholino)propanesulfonic

Diuridine Phosphate Analogues
J . Org. Chem., Vol. 61, No. 18, 1996 6281
acid; TBAF, tetrabutylammonium fluoride; PEG, polyethylene
glycol; PPG polypropylene glycol; TBDMSCl, tetrabutyldimeth-
ylsilyl chloride; tetrahydrofuran, THF; TEAA, triethylammo-
nium acetate; TEAB, triethyl ammonium bicarbonate; UpnU,
uridylyl (3′-5′) 5′-amino-5′-deoxyuridine; UpsU, uridylyl (3′-
5′) 5′-thio-5′-deoxyuridine; TpnT, thymidylyl (3′-5′) 5′-amino-
5′-deoxythymidine; amUpsU, 2′-amino-2′-deoxyuridylyl (3′-
5′) 5′-thio-5′-deoxyuridine; amUpT, 2′-amino-2′-deoxyuridylyl
(3′-5′) thymidine; TpsT, thymidylyl (3′-5′) 5′-thio-5′-deoxy-
thymidine; tR, retention time; ddT, 2′,3′-dideoxythymidine;
ddA, 2′,3′-dideoxyadenosine; dA, 2′-deoxyadenosine.
Ack n ow led gm en t. We are very grateful to S. Th.
Sigurdsson and P. A. Heaton for the critical reading of
the manuscript and many stimulating discussions. We
also thank U. Kutzke for valuable and expert technical
assistance and B. Seeger for NMR spectroscopy. Sup-
port by the Deutsche Forschungsgemeinschaft and
Fonds der Chemischen Industrie is acknowledged. V.
J . thanks the Deutscher Akademischer Austauschdienst
for a short term fellowship.
J O960795L