Thio effects on the departure of the 3′-linked ribonucleoside from diribonucleoside 3′,3′-phosphorodithioate diesters and triribonucleoside 3′,3′,5′-phosphoromonothioate triesters: Implications for ribozyme catalysis

Article abstract of DOI:10.1002/chem.200601835

To provide a solid chemical basis for the mechanistic interpretations of the thio effects observed for large ribozymes, the cleavage of triribonucleoside 3′,3′,5′-phosphoromonothioate triesters and diribonucleoside 3′,3′-phosphorodithioate diesters has be

Full text of DOI:10.1002/chem.200601835

DOI: 10.1002/chem.200601835
Thio Effects on the Departure of the 3’-Linked Ribonucleoside from
Diribonucleoside 3’,3’-Phosphorodithioate Diesters and Triribonucleoside
3’,3’,5’-Phosphoromonothioate Triesters: Implications for Ribozyme Catalysis
Tuomas Lçnnberg,* Mikko Ora,* Sami Virtanen, and Harri Lçnnberg[a]
Abstract: To provide a solid chemical
basis for the mechanistic interpreta-
tions of the thio effects observed for
large ribozymes,the cleavage of triri-
bonucleoside 3’,3’,5’-phosphoromono-
thioate triesters and diribonucleoside
3’,3’-phosphorodithioate diesters has
been studied. To elucidate the role of
the neighboring hydroxy group of the
departing 3’-linked nucleoside,hydroly-
sis of 2’,3’-O-methyleneadenosin-5’-yl
bis[5’-O-methyluridin-3’-yl] phosphoro-
monothioate (1a) has been compared
to the hydrolysis of 2’,3’-O-methylenea-
a wide pH range. The phosphoromono-
thioate triesters 1a,b undergo two com-
peting reactions: the starting material
is cleaved to a mixture of 3’,3’- and
3’,5’-diesters,and isomerized to 2 ’,3’,5’-
and 2’,2’,5’-triesters. With phosphorodi-
thioate diesters 2a,b,hydroxide-ion-
droxide-ion-catalyzed isomerization of
the 3’,3’,5’-triester to 2’,3’,5’- and
2’,2’,5’-triesters with 1a is 11 times as
fast as that compared with 1b. These
accelerations have been accounted for
by stabilization of the anionic phos-
phorane intermediate by hydrogen
bonding with the 2’-hydroxy function.
Thio substitution of the nonbridging
oxygens has an almost negligible influ-
ence on the cleavage of 3’,3’-diesters
2a,b,but the hydrolysis of phosphoro-
monothioate triesters 1a,b exhibits a
sizable thio effect, kPO/kPS =19. The ef-
fects of metal ions on the rate of the
ꢀ
catalyzed cleavage of the P O3’ bond
is the only reaction detected at pH >6,
but under more acidic conditions desul-
furization starts to compete with the
cleavage. The 3’,3’-diesters do not un-
R
dergo isomerization. The hydroxide-
ion-catalyzed cleavage reaction with
both 1a and 2a is 27 times as fast as
that compared with their 2’-O-methy-
lated counterparts 1b and 2b. The hy-
denosin-5’-yl
2’,5’-di-O-methyluridin-3’-yl phospho-
romonothioate (1b) and the hydrolysis
5’-O-methyluridin-3’-yl
cleavage of diesters and triesters have
A
G
been studied and discussed in terms of
the suggested hydrogen-bond stabiliza-
tion of the thiophosphorane intermedi-
ates derived from 1a and 2a.
of bis[uridin-3’-yl] phosphorodithioate
(2a) to the hydrolysis of uridin-3’-yl
Keywords: hydrolysis
· kinetics ·
2’,5’-di-O-methyluridin-3’-yl
phos-
phosphoesters · phosphorothioates ·
ribozymes · thio effect
A
ACHTREUNG
have been followed by RP HPLC over
Introduction
bridging phosphoryl oxygen atom,either pro- R or pro-S
oxygen,as a coordination site of metal ions playing a cata-
lytic role.[2–4] The catalytic activity of large ribozymes de-
pends on the Mg2+ ion,which prefers oxygen over sulfur as
a donor atom. Accordingly,replacement of an oxygen atom
that participates in binding of the Mg2+ ion with sulfur ex-
pectedly results in rate-deceleration,which may then be at
least partially restored by using a more thiophilic Mn2+ or
Cd2+ ion instead of Mg2+ as a cofactor. In addition,replace-
ment of the bridging 3’-oxygen atom with sulfur has been
exploited in attempting to elucidate the importance of this
atom as a donor of an electron pair for hydrogen bonding or
metal-ion binding.[4,5] Thio-substitution may,however,simul-
taneously affect several structural parameters,such as chain
folding,hydrogen bonding,metal-ion binding,solvation,and
van der Waals interactions,and,hence,interpretation of the
Oligonucleotides containing a chiral phosphoromonothioate
linkage have been extensively used as substrates in mecha-
nistic studies of large ribozymes.[1] An oxygen atom assumed
to be catalytically important is substituted with a sulfur
atom and the accompanying change in the reaction rate,
known as a “thio effect”,is quantified. Usually this ap-
proach has been applied to the identification of the non-
[a] Dr. T. Lçnnberg,Dr. M. Ora,S. Virtanen,Prof. Dr. H. Lçnnberg
Department of Chemistry,University of Turku
20014 Turku (Finland)
Fax : (+358)2-333-6700
E-mail: tuanlo@utu.fi
mikora@utu.fi
4614
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

FULL PAPER
[21]
ꢀ
observed changes in reaction rate is not always unambigu-
ous. Comparative studies with small molecule models usual-
ly facilitate evaluation of the inherent influence of a given
structural modification and help to distinguish between al-
ternative mechanistic interpretations.
The thio effects of phosphoesters have previously been
studied in some detail both experimentally[6–13] and by theo-
retical calculations.[14–16] Generally speaking,the thio effect
(kPO/kPS) resulting from replacement of a nonbridging phos-
phoryl oxygen atom of a phos-
the P O3’ bond. To quantify the influence of this kind of
intramolecular hydrogen bonding on the thio effect and to
provide a solid chemical basis for mechanistic inferences
concerning the action of large ribozymes,we now report on
thio effects determined for the cleavage and isomerization
of simple chemical models of their cleavage site. For this
purpose,cleavage and isomerization of a 3 ’,3’,5’-phosphoro-
monothioate triester,2 ’,3’-O-methyleneadenosin-5’-yl bis[5’-
O-methyluridin-3’-yl] phosphoromonothioate (1a),which
phodiester with sulfur usually
falls in the range 1–10,while
phosphotriesters exhibit consid-
erably higher and monoesters
lower,that is,inverse thio ef-
fects.[6] In other words,this kind
of thio-substitution appears to
favor a dissociative mechanism
via a monoanionic metaphos-
phate-like intermediate (mono-
ester cleavage) and disfavor an
associative mechanism via
a
pentacoordinated monoanionic
phosphorane intermediate (triester cleavage). The magni-
tude of the thio-effect,hence,allows mechanistic conclu-
sions even in cases for which metal-ion binding does not
play a role. A complicating factor,however,is that the elec-
tronic and solvation effects of the thio-substitution are op-
posite. While the soft sulfur atoms stabilize anionic phos-
phorane-like transition states by more facile delocalization
of the negative charge,this rate-accelerating effect is largely
compensated by less efficient solvation of the thiophosphor-
ane intermediates compared to their oxygen counterparts.[15]
For this reason,the cleavage of phosphodiester via a di-
gives a tri-O-alkyl monoanionic phosphorane upon the
attack of hydroxide ion,and a 3 ’,3’-phosphorodithioate di-
A
O-alkyl dianionic phosphorane,have been studied over a
wide pH range and the results are compared to those report-
ed previously for their oxygen analogues 3a,b[19,22] and
4a,b.[20] The reasoning behind the choice of 1a as a model
compound is that the 2’-hydroxy group of the leaving 3’-
linked nucleoside may donate a hydrogen bond either to the
departing 3’-oxyanion or to the oxyanion ligand of the
monoanionic phosphorane intermediate obtained by an in-
G
A
tramolecular attack of the other 3’-linked nucleoside
(Scheme 1). The relative importance of these two processes
is estimated by comparing the rate and product distribution
(departure of 3’-linked versus 5’-linked nucleoside) to the
corresponding quantities obtained with 1b,containing the
effect.[7,10,12] Bearing the importance of solvation in mind,
one might expect the thio effects of associative phosphoester
cleavage reactions via a phosphorane-like transition state to
be sensitive to hydrogen-bonding interactions. In other
words,stabilization of a phosphorane intermediate or leav-
ing group by intramolecular hydrogen bonding could well
be reflected in the magnitude of the thio effect.
With large ribozymes,the cleavage of 3 ’,5’-phosphodiester
bonds proceeds by an attack of a 3’-hydroxy group of an ex-
ternal nucleoside on the phosphorus atom with concomitant
departure of the 3’-linked nucleoside.[1] The neighboring 2’-
hydroxy group of the departing nucleoside is rate accelerat-
[2,6,17]
ing,in particular with group I introns,
and may,in prin-
ciple,donate a hydrogen bond either to an oxyanion of the
phosphorane intermediate or to the 3’-oxyanion of the de-
parting nucleoside. We have previously preferred the former
alternative for the reason that the 2’-hydroxy group appears
to accelerate the departure of the 3’- and 5’-linked nucleo-
sides as efficiently,[18–20] but the opposite view has also re-
ceived support from model studies,namely from the obser-
vation that methoxide-ion-catalyzed methanolysis of ribonu-
cleoside 3’(2’)-dimethylphosphate proceeds by cleavage of
Scheme 1. Thiophosphorane intermediates for the hydrolysis of trinucleo-
side 3’,3’,5’-phosphorothioate 1a; ap=apical.
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4615

T. Lçnnberg,M. Ora et al.
methylated 2’-hydroxy group of the leaving nucleoside. Ac-
Product distributions and pH rate profiles for the reactions
of phosphorothioate triesters 1a,b: The hydrolysis of 1a and
1b was followed over a wide pH range (from H0 =ꢀ0.1 to
pH 9) at 258C by analyzing the composition of the aliquots
withdrawn from the reaction mixture at appropriate time in-
tervals by RP HPLC. The nucleosidic products (6, 14, 20 in
Scheme 4) were identified by spiking with authentic samples.
The diester products were identified either as di(uridin-3’-
yl)esters (15a, 16a, 17 from 1a; 15b, 16b from 1b) or
uridin-3’-yl adenosine-5’-yl esters 18 and 19 by ESI-HPLC/
MS. The triester isomers of the starting materials were also
identified by ESI-HPLC/MS. Assignment of 12a and 13 as
2’,3’,5’- and 2’,2’,5’-isomers,respectively,was based on the
fact that 12a appeared as an intermediate in the formation
of 13.
ꢀ
celeration of the P O5’ cleavage by intramolecular hydro-
ꢀ
gen bonding to the 5’-oxyanion is less probable,as the O2 ’
ꢀ
H O5’ hydrogen bond,if ever formed,must be cleaved
upon release of the 5’-oxyanion as an immediate product.
The influence of the hydrogen-bonding on the thio effect is
evaluated by comparing the cleavage rates of 1a,b to the
cleavage rates of their oxygen analogues 3a,b. Studies with
diesters 2a,b and 4a,b elucidate,in turn,the thio effect re-
lated to the cleavage reaction via a dianionic phosphorane.
Results
Preparation of the trinucleoside 3’,3’,5’-phosphoromono-
thioates 1a,b and dinucleoside 3’,3’-phosphorodithioates
2a,b: Trinucleoside 3’,3’,5’-phosphoromonothioates were ob-
Over the pH range studied,two reactions expectedly took
place: isomerization of the 3’,3’,5’-triesters 1a,b to 2’,3’,5’-
and 2’,2’,5’-triesters 12a,b and 13,respectively,and cleavage
of the triesters to a mixture of diesters. Two sets of cleavage
products were obtained (Scheme 4). Release of 2’,3’-O-
tained in a 4,4’-dimethoxytrityl (DMTr)-protected form,
[19]
8a,b,essentially as previously
described for the corre-
sponding trinucleoside monophosphates,namely by tetra-
zole-promoted stepwise displacement of the dimethylamino
groups from tris(dimethylamino)phosphine with the appro-
priately protected nucleosides,followed by oxidation with
elemental sulfur (Scheme 2).
ꢀ
methyleneadenosine (14) by the P O5’ bond fission (route
A) resulted in formation of bis(5’-O-methyluridin-3’-yl)
phosphorothioate (15a) and its 2’,3’- and 2’,2’-isomers 16a
Preparation of the protected nu-
cleosides,5
’-O-methyl-2’-O-
(4,4’-dimethoxytrityl)uridine
(5),[22] 2’,5’-di-O-methyluridine
(6),[23] and N6-benzoyl-2,3’-O-
methyleneadenosine (7)[23] has
been described earlier. The di-
methoxytrityl protections were
removed only prior to each ki-
netic run with hydrogen chlo-
ride in 1,4-dioxane.
Bis(uridin-3’-yl) phosphorodi-
thioate (2a) and 2’,5’-di-O-
methyluridin-3’-yl
uridin-3’-yl
phosphorodithioate (2b) were
synthesized by the hydrogen
phosphonate methodology de-
scribed by Stawinski et al.[24]
(Scheme 3). Accordingly,2 ’,5’-
bis-O-(tert-butyldimethylsilyl)ur-
Scheme 2. a) P
A
zole,MeCN; d) S ,CH Cl2; e) NH3,MeOH; f) HCl/14, -dioxane.
idine
3’-hydrogenphosphono-
thioate (10) prepared from
2’,5’-bis-O-(tert-butyldimethylsi-
lyl)uridine (9) was allowed to
react with this nucleoside or
2’,5’-di-O-methyluridine (6) in
the presence of diphenylphos-
phorochloridate and the prod-
uct was oxidized with elemental
sulfur to obtain 11a,b. Desilyla-
tion of the latter products gave
the desired phosphorodithioate
diesters 2a,b.
ꢀ
Scheme 3. a) H2PO2 HN+Et3,PivCl,Py; b) 1) 9 (to obtain 11a) or 6 (to obtain 11b),diphenylphosphorochlo-
AHCTREUNG
4616
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
The pH rate profiles for the
cleavage (routes A and B in
Scheme 4) and isomerization
(route C) of 1a and 1b are de-
picted in Figure 1. The rate con-
stants indicated are those ex-
trapolated to buffer concentra-
tion zero. With both com-
pounds,the cleavage is pH in-
dependent
conditions. The cleavage of 1a
becomes hydroxide-ion-cata-
under
acidic
lyzed at pH 4–5,while with 1b,
this change takes place at
pH 6–7. Between pH 4 and 6,
the cleavage of 1b is still pH in-
dependent,although slightly
faster than under more acidic
conditions. Compound 1b ac-
tually is a mixture of two dia-
stereomers,which can be sepa-
rated by RP HPLC. Both dia-
stereomers are,however,hydro-
lyzed as rapidly. With both 1a
ꢀ
ꢀ
and 1b,the P O3’ and P O5’
bonds are cleaved almost equal-
Scheme 4. Products for the cleavage and isomerization of trinucleoside 3’,3’,5’-phosphorothioate triesters 1a
and 1b.
and 17,respectively,from
1a,and 5 ’-O-methyluridin-3’-yl
2’,5’-di-O-methyluridin-3’-yl phosphorothioate (15b) and its
2’,3’-isomer 16b from 1b. Release of 5’-O-methyluridine
ꢀ
(20) from 1a or 2’,5’-di-O-methyluridine (6) from 1b by P
O3’ bond fission (route B),in turn,resulted in formation of
2’,3’-O-methyleneadenosin-5’-yl
5’-O-methyluridin-3’-yl
phosphoromonothioate (18) and its 2’,5’-isomer 19. No de-
sulfurization was observed to take place.
The isomerization and cleavage reactions most likely pro-
ceed via a common thiophosphorane intermediate indicated
in Scheme 1. While endocyclic bond cleavage of this inter-
mediate leads to isomerization,departure of either the 5 ’- or
3’-linked nucleoside leads to the cleavage products by routes
A and B,respectively. The cleavage reactions,in fact,give a
2’,3’-cyclic phosphorothioate triester as the immediate prod-
uct (Scheme 5). Attack of water on this highly unstable
triester gives another thiophosphorane intermediate that by
endocyclic cleavage gives the diester products 15–19. In
principle,the cyclic thiophosphorane may also undergo exo-
cyclic cleavage to a nucleoside 2’,3’-cyclic phosphate. Previ-
ous studies[9] have,however,shown that the cyclic triester
undergoes only endocyclic cleavage,except at low pH
(pH <3),for which exocyclic cleavage starts to compete as
a side reaction (see the discussion on diester cleavage
below). Consistent with this,the monomeric nucleosides 14
or 6/20 and the respective isomeric phosphorothioate di-
Scheme 5. Breakdown of 2’,3’-cyclic phosphorothioate triester intermedi-
ates.
ly rapidly under acidic conditions (pH <6),but on going to
ꢀ
more alkaline conditions,cleavage of the P O3’ bond gradu-
ally becomes the predominant reaction.
The isomerization exhibits a similar dependence of rate
on acidity as the cleavage,except that the pH-independent
region is not as wide: the reaction becomes first order in hy-
droxide-ion concentration already around pH 2 with 1a and
around pH 3 with the diastereomers of 1b. Over the whole
pH range studied,isomerization proceeds at essentially
equal rates in both directions. In contrast to the cleavage re-
action,the diastereomers of 1b are isomerized at a different
rate,the faster eluting diastereomer somewhat less rapidly
than the more slowly eluting one.
A
ratio at pH >2.
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4617

T. Lçnnberg,M. Ora et al.
independent reactions (kWcl and kWcl) are given in Tables 1
and 2.
The buffer-independent rate constant, kWis ,for the isomeri-
zation of 1a and 1b may be expressed by Equation (1),in
which ki0s and kOH are the rate constants for the pH-inde-
is
pendent and hydroxide-ion-catalyzed reactions. A similar
equation,Equation (2),applies to the buffer-independent
rate constant, kWcl,of the cleavage of 1a. Kw is the ionic prod-
uct of water under the experimental conditions (1.87”
10ꢀ14 mol2 Lꢀ1).[25] The observed first-order rate constant for
the cleavage of 1b may,in turn,be expressed by Equa-
tion (3),in which kXcl is the rate constant for the additional
pH-independent reaction observed between pH 4 and 6,and
Ka is the equilibrium constant referring to the change in the
ionic form of the reactants which results in the observed in-
crement in the rate of the otherwise pH-independent cleav-
age on passing pH 4. The pKa value,4.2,coincides with the
pKa value of N1 in the adenine moiety.[26]
Figure 1. pH rate profiles for the cleavage and isomerization of phospho-
triesters 1a and 1b at 258C, I
(NaNO3)=1.0 molLꢀ1. All the rate con-
AHCTREUNG
stants refer to buffer concentration zero. The curves are obtained by
least-squares fitting by Equations (1)–(3). The dashed and dotted lines
refer to the cleavage of triesters 3a[19] and 3b,[22] respectively. Notation:
kWis ¼ ki0s þ kOH ðKw=½HþŠÞ
ð1Þ
ð2Þ
ð3Þ
is
OH
kWcl ¼ kc0l þ k ðKw=½HþŠÞ
~
!
&
( ) cleavage of 1a (kclA +kclB),( and ) cleavage of the slowly and fast-
eluting diastereomer of 1b,respectively ( kclA +kclB),( &) isomerization of
cl
!
~
1a (kis),( and ) isomerization of the fast and slowly eluting diastereo-
mer of 1b (kis). For the rate constants,see Scheme 4. kis stands for the
sum of the first-order rate constants for the forward and reverse reaction
of 1a or 1b to a mixture of other isomers.
kWcl ¼ fkc0l½HþŠ þ kXclKa þ kOcHl KaðKw=½HþŠÞg=ð½HþŠ þ KaÞ
The rate constants for the overall cleavage, kWcl =kclA +kclB
(see Scheme 4),and isomerization, kWis ,have been obtained
by nonlinear least-squares fitting of the experimental data
to Equations (1)–(3). To take into account the fact that in
1a there are two potential 2’-OH nucleophiles to initiate the
reaction,the rate constants corresponding to the cleavage
and isomerization of 1a have been divided by two. The
overall rate constants have then be broken down to contri-
butions of kclA and kclB on the basis of observed product dis-
tribution. The rate constants obtained in this manner are
listed in Table 3.
As mentioned above,buffer catalysis was observed in car-
boxylic acid,MES and HEPES buffers for the cleavage,and
to a lesser extent,for the isomerization reactions. Measure-
ments at various concentration ratios of the buffer constitu-
ents revealed both reactions to be susceptible to general
base catalysis. The rate constants obtained at different
obs
cl
obs
is
buf
buffer concentrations (k and k ) and the rate constants
buf
for the general base catalysis (k and k ) and the buffer-
cl
is
buf
cl
buf
is
Table 1. Rate constants for the buffer-catalyzed and -independent cleavage (k and kWcl) and isomerization (k and kWis ) of 1a at T=258C and
I
A
.
[HA]/[Aꢀ]
[HA]
[mol Lꢀ1
[Aꢀ]
[mol Lꢀ1
k
”105
]
k
”104
kWcl ”105
[sꢀ1
k
”104
]
k
”104
kWis ”104
[sꢀ1
obs
cl
buf
cl
obs
is
buf
is
Buffer acid
formic
N
]
A
]
A
[Lmolꢀ1 sꢀ1
]
U
]
A
[Lmolꢀ1 sꢀ1
]
E
]
3:1
1:3
1:1
1:3
0.033
0.069
0.139
0.011
0.023
0.046
5.30
7.58
11.7
1.90
2.49
3.07
4.53ꢁ0.07[a]
10.0ꢁ0.8
20ꢁ3
3.32ꢁ0.09[b]
5ꢁ1
8ꢁ1[a]
32[c]
18
1.6ꢁ0.2[b]
0.011
0.023
0.046
0.033
0.069
0.139
9.71
13.4
23.6
15.7
18.7
13[c]
acetic
0.024
0.048
0.095
0.024
0.048
0.095
15.5
20.2
43.2
32.5
75.2
76.9
4ꢁ4
74
0.012
0.024
0.048
0.036
0.071
0.143
17.5
27.3
48.3
21.6ꢁ0.3
7.0ꢁ0.4
[a] Statistical error of the slope of the linear regression of the rate constants obtained at various buffer concentrations. [b] Statistical error of the rate
constant extrapolated to buffer concentration zero. [c] The values based on rate constants obtained at two buffer concentrations only.
4618
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
buf
cl
buf
is
Table 2. Rate constants for the buffer-catalyzed and -independent cleavage (k and kWcl) and isomerization (k and kWis ) of the fast-eluting diastereomer
of 1b at T=258C and I .
(NaNO3)=1.0 molLꢀ1
G
Buffer acid
formic
[HA]/[Aꢀ]
3:1
[HA]
[molLꢀ1
[Aꢀ]
[molLꢀ1
k
”10
]
k
”104
kWcl ”105
[sꢀ1
k
”104
]
k
”104
kWis ”104
[sꢀ1
obs
cl
buf
cl
obs
is
[sꢀ1
buf
is
U
]
G
]
A
[Lmolꢀ1 sꢀ1
]
G
]
A
[Lmolꢀ1 sꢀ1
]
H
]
0.033
0.069
0.139
0.011
0.023
0.046
1.87
2.33
3.14
3.65
5.49
8.91
0.90ꢁ0.02[a]
2.5ꢁ0.3
1.48ꢁ0.03[b]
2.1ꢁ0.3
3.73ꢁ0.03[a]
14.6ꢁ0.3
40ꢁ10
0.20ꢁ0.01[b]
0.50ꢁ0.03
1ꢁ1
1:3
1:1
1:3
3:1
1:1
3:1
0.011
0.023
0.046
0.033
0.069
0.139
3.06
4.63
6.64
1.13
1.87
3.19
acetic
0.024
0.048
0.095
0.024
0.048
0.095
3.56
4.84
7.00
2.76
6.37
9.33
2.4ꢁ0.1
2.5ꢁ0.1
0.012
0.024
0.048
0.036
0.071
0.143
4.25
6.00
9.35
11.6
15.1
3.56ꢁ0.04
0.6ꢁ0.2
2.58ꢁ0.05
2.7ꢁ0.2
37[c]
8.1[c]
MES
0.057
0.086
0.143
0.019
0.029
0.048
3.09
3.51
3.80
0.024
0.048
0.095
0.024
0.048
0.095
3.60
3.78
4.48
0.63ꢁ0.09
3.3ꢁ0.1
HEPES
0.071
0.143
0.024
0.048
6.76
8.24
1.56[c]
5.28[c]
[a] Statistical error of the slope of the linear regression of the rate constants obtained at various buffer concentrations. [b] Statistical error of the rate
constant extrapolated to buffer concentration zero. [c] The values based on rate constants obtained at two buffer concentrations only.
Table 3. Rate constants for the pH-independent isomerization (ki0s),hydroxide-ion-catalyzed isomerization
sulted in accumulation of uri-
dine 2’,3’-cyclic phosphorodi-
thioate (21). In addition,a
small amount of uracil (2–5%)
was formed at pH 7–9 parallel
to 21. Under acidic conditions
(pH <6),desulfurization (route
B) competed with the forma-
tion of 21. At pH <4, 21 was
not any more accumulated,but
2a and 2b were converted to a
mixture of 2’,3’- (23a,b),3 ’,3’-
OH
(kOH),pH-independent cleavage ( kc0l),and hydroxide-ion-catalyzed cleavage ( k ) of phosphorothioate tri-
is
cl
A
N
(see Scheme 4) and kXcl to the pH-independent cleavage of 1b at pH 4–6. For comparison,the corresponding
data for the oxygen–phosphate counterparts 3a[19] and 3b[22] are given.
1[a]
1b[b]
1b[c]
3a[a]
3b
kc0l ”105 [sꢀ1
]
1.7ꢁ0.1
1.4ꢁ0.1
14ꢁ2
9.0ꢁ0.9
6.8ꢁ0.7
2.3ꢁ0.2
0
kclA ”105 [sꢀ1
]
1.05ꢁ0.06
0.65ꢁ0.04
0.83ꢁ0.03
0.56ꢁ0.02
9ꢁ1
5ꢁ1
kcl0B ”105 [sꢀ1
]
kXcl ”105 [sꢀ1
]
2.7ꢁ0.2
2.4ꢁ0.6
ki0s ”105 [sꢀ1
]
2.4ꢁ0.4
10.5ꢁ0.9
1.3ꢁ0.1
9.2ꢁ0.8
58ꢁ6
1.8ꢁ0.2
0.39ꢁ0.07
0.02ꢁ0.004
0.37ꢁ0.07
5.5ꢁ0.7
OH
k
k
k
”10ꢀ3 [Lmolꢀ1 sꢀ1
”10ꢀ3 [Lmolꢀ1 sꢀ1
”10ꢀ3 [Lmolꢀ1 sꢀ1
]
]
]
]
200ꢁ10
22ꢁ1
7.7ꢁ0.9
0.9ꢁ0.1
6.6ꢁ0.8
500ꢁ100
cl
OH
clA
OH
clB
178ꢁ9
kOH ”10ꢀ5 [Lmolꢀ1 sꢀ1
24ꢁ5
(24a,b),and
thioate diesters (25),probably
via 2’,3’-cyclic phosphoro-
monothioate triester (not accu-
2
’,2’-phosphoro-
is
[a] The rate constants referring to compounds 1a and 3a have been subjected to a statistical correction (divi-
sion by a factor of two) for the reason that these molecules contain two attacking 2’-OH groups,while 1b and
3b contain only one 2’-OH. [b] Fast-eluting diastereomer. [c] Slowly eluting diastereomer.
a
AHCTREUNG
mulated). Consistent with pre-
vious studies,[11] the latter inter-
Product distributions and pH rate profiles for the reactions
of phosphoroditioate diesters 2a,b: The cleavage of phos-
phorodithioate diesters is much slower than the hydrolysis
of triesters. The reactions of 2a and 2b were,hence,fol-
lowed at an elevated temperature,namely 90 8C. Under
basic and neutral conditions (pH >6),the only reaction de-
mediate underwent only endocyclic fission to acyclic phos-
phoromonothioate diesters (23–25) at pH >4. On going to
more acidic solutions,exocyclic fission of the 2 ’,3’-cyclic
phosphoromonothioate triester to 2’,3’-cyclic phosphoromo-
A
nothioate (26a),however,started to compete and these re-
actions became significant side reactions at pH <2. The
phosphoromonothioate diesters containing two dissimilar
nucleosides 23a,b and 24b were expectedly obtained as
pairs of two diastereomeric forms. All the phosphoromono-
ꢀ
tected was the hydroxide-ion-catalyzed cleavage of the P
O3’ bond of 2a (route A in Scheme 6). Release of uridine
(22) from 2a and 2’,5’-di-O-methyluridine (6) from 2b re-
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4619

T. Lçnnberg,M. Ora et al.
&
*
Figure 2. pH rate profiles for the cleavage (kcl) of 2a ( ) and 2b ( ) and
*
desulfurization (kds) of 2a ( ) and 2b (&) at 908C and I
A
0.1 molLꢀ1. The dashed and dotted lines show the corresponding curves
[20]
for 4a and 4b,respectively.
kcl ¼ kHcl½HþŠ þ kc0l þ kOcHl ðKw=½HþŠÞ
kds ¼ kdHs½HþŠ þ kd0s
ð4Þ
ð5Þ
OH
In these equations,the rate constants kHcl, k0c,and k refer
cl
to oxonium-ion-catalyzed,pH-independent and hydroxide-
ion-catalyzed cleavage (route A in Scheme 6),respectively,
and kdHs and kd0 are the rate constants of the oxonium-ion-cat-
alyzed and pH independent desulfurization (route B). A
value of 6.17”10ꢀ13 mol2 Lꢀ2 has been used for the ionic
product of water.[25] To take into account the fact that in 2a
there are two intramolecular 2’-OH nucleophiles to initiate
the reaction,the rate constants corresponding to the cleav-
age and desulfurization of 2a have been divided by two. The
values of these constants are listed in Table 4.
Scheme 6. Products for the cleavage and desulfurization of dinucleoside
3’,3’-phosphorodithioate diesters 2a and 2b.
thioate diesters 23–25 were subsequently desulfurized to a
mixture of the corresponding oxygen diesters,as previous-
ly[11] described in detail. It is worth noting that no isomeriza-
tion of the starting material 2a,b took place under any con-
ditions. No sign of participation of the 5’-OH as a nucleo-
phile was observed.
Table 4. Rate constants for the oxonium-ion-catalyzed cleavage (kHcl) and
desulfurization (kdHs),pH-independent cleavage ( kc0l) and desulfurization
OH
(kd0s),and hydroxide-ion-catalyzed cleavage ( k ) of phosphorodithioate
cl
diesters 2a and 2b. T=908C and I
A
The pH rate profiles for the cleavage and desulfurization
of 2a and 2b are depicted in Figure 2. The rate constants in-
dicated refer to buffer concentration zero. In fact,the rate
constants were independent of the buffer concentration in
the low concentration region (<0.1 molLꢀ1) used in the ki-
netic measurements. The cleavage is oxonium-ion-catalyzed
at pH <4,hydroxide-ion-catalyzed at pH >6,and nearly
pH independent between these two limiting pH values. Both
the oxonium- and hydroxide-ion-catalyzed cleavage is first-
son,the corresponding data for their oxygen–phosphate analogues 4a
and 4b[20] are given.
2a[a]
2b
90ꢁ30
4a[a]
4b
kHcl ”105 [Lmolꢀ1 sꢀ1
]
170ꢁ5
130ꢁ60
230ꢁ70
kc0l ”105 [sꢀ1
]
0.15ꢁ0.06 0.05ꢁ0.02 0.15ꢁ0.05 0.06ꢁ0.02
OH
cl
k
[Lmolꢀ1 sꢀ1
]
10.1ꢁ3.1
36ꢁ13
0.37ꢁ0.08 14.9ꢁ3.1
55ꢁ21
0.46ꢁ0.12
kdHs ”105 [Lmolꢀ1 sꢀ1
kd0s ”105 [sꢀ1
]
]
0.09ꢁ0.02 0.08ꢁ0.02
[a] The rate constants referring to compounds 2a and 4a have been sub-
jected to a statistical correction (division by a factor of two) for the
reason that these molecules contain two attacking 2’-OH groups,while
2b and 4b contain only one 2’-OH.
order in the catalyst concentration,H O+ and OHꢀ,respec-
3
tively. The desulfurization is first order in [H3O+] at pH <4,
and pH independent under less acidic conditions. To illus-
trate the effects that replacement of both of the nonbridging
oxygens of diribonucleoside 3’,3’-monophosphates with
sulfur atoms has on the cleavage rate,the pH rate profiles
obtained previously[20] with 4a and 4b are also included in
Figure 2. The curves indicated have been obtained by non-
linear least-squares fitting based on Equations (4) and (5).
Metal-ion catalysis: The cleav-
age of phosphorothioate tri-
AHCTREUNG
4620
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
chelate of 1,5,9,-triazacyclododecane ([12]aneN3; 27) at
and pH 5.6 83 and 230 times as fast as in the absence of the
metal ion,respectively (Table 5). The catalytic activity of
Mg2+,a hard Lewis acid,was negligible,whereas the mark-
pH 7.28 [T=258C; I
(NaNO3)=1.0 molLꢀ1]. The Zn2+ che-
A
late was used instead of the free metal ion,as the purpose
was to study the effect of metal ions on the cleavage rate
under conditions in which this reaction is hydroxide-ion-cat-
alyzed,that is,at pH >7. Uncomplexed Zn2+ precipitates at
such a high pH. Pseudo-first-order rate constants were mea-
sured as a function of the concentration of Zn2+[12]aneN3.
Table 5. First-order rate constants for the cleavage of phosphorodithioate
diesters 2a and 2b at 908C and I .
(NaNO3)=0.1 molLꢀ1
G
Mz+
U
]
pH
kcl ”103 [sꢀ1
2b
]
]
2a
0.33
1.40
3.58
1.27
0.52
Zn2+
1.0
5.0
5.60
5.60
5.60
5.38
5.08
4.68
5.60
5.60
0.0169
0.0477
0.328
1.08
0.27
0.134
0.051
Cd2+
Mg2+
none
55.9
24.4
0.0012
0.021
5.60
0.0014
edly thiophilic Cd2+ ion efficiently accelerated both the
cleavage and desulfurization. The cleavage reactions of 2a
and 2b were accelerated by a factor of 3300 and 21000,re-
spectively,at [Cd 2+]=5 mmolLꢀ1 and pH 5.6. Desulfuriza-
tion of 2b was 13000-fold faster than in the absence of the
metal ion.
!
&
Figure 3. Acceleration of the cleavage of 1a ( ) and the fast ( ) and
Discussion
slowly eluting ( ) diastereomer of 1b by Zn2+[12]aneN3 at pH 7.28, T=
~
258C and I .
(NaNO3)=1.0 mol Lꢀ1
A
Reactions of phosphorothioate triesters 1a and 1b: The
overall cleavage of 1a and 1b becomes first-order in hydrox-
ide-ion concentration at pH 4–5 and 6–7,respectively.
Under these conditions,the reaction most probably pro-
ceeds by rapid initial deprotonation of the 2’-hydroxy group,
attack of the resulting oxyanion on the phosphorus atom
with concomitant formation of a thiophosphorane inter-
mediate and rate-limiting departure of the 3’- or 5’-linked
nucleoside (Scheme 7).[15] The fact that breakdown of the
monoanionic thiophosphorane rather than its formation is
rate limiting is evidenced by much faster isomerization of
the starting material compared to its cleavage. Methylation
of the 2’-hydroxy function of one of the 3’-linked nucleo-
sides markedly retards the cleavage: 1a is cleaved 27 times
Figure 3 shows the observed pseudo-first-order rate con-
stants (kobs) divided by values obtained in the absence of the
metal complex (ko) plotted as a function of the concentra-
tion of Zn2+[12]aneN3. As seen,both compounds are subject
to catalysis by the complex,but the cleavage of 1b is cata-
lyzed more efficiently than the cleavage of 1a. The rate ac-
celeration, kobs/ko,levels off to a value of 46 with the fast-
eluting diastereomer of 1b and a value of 10 with 1a.
The effect of the Zn2+ ion on the cleavage of phosphoro-
dithioate diesters 2a and 2b was followed as a function of
pH (5.1–5.6) and metal-ion concentration (1–10 mmolLꢀ1 at
pH 5.6). Cleavage was the only
reaction observed. Over the
narrow pH and [Zn2+] range
studied,the reaction was first-
order in both the hydroxide
and metal-ion concentration.
Cleavage was the only reaction
detected at [Zn2+]=5 mmolLꢀ1
and pH 5.6. In other words,the
desulfurization was less suscep-
tible to the Zn2+ ion catalysis
than the cleavage. The Zn2+
-
promoted cleavage of 2a and
Scheme 7. Mechanism of the hydroxide-ion-catalyzed cleavage of phosphorothioate triesters 1a and 1b; eq=
2b was at [Zn2+]=5 mmolLꢀ1 equatorial.
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4621

T. Lçnnberg,M. Ora et al.
as rapidly as 1b. This value includes the statistical correction
arising from the fact that 1a contains two and 1b only one
hydroxy function serving as an intramolecular nucleophile.
In other words,the observed rate acceleration is 54-fold.
ꢀ
The cleavage of the P O3’ bond (route B in Scheme 7) is fa-
ꢀ
vored over the cleavage of the P O5’ bond (route A),reach-
ing a maximum proportion of 88% with 1a and 95% with
1b under sufficiently alkaline conditions. As described pre-
viously for the oxygen–phosphate counterparts 4a[19] and
4b,[22] the fact that 2’-OMe substitution does not specifically
ꢀ
retard the P O3’ bond cleavage may be taken as an evi-
dence for hydrogen-bond stabilization of the phosphorane
intermediate (compare I1 and I2 in Scheme 1) rather than
the leaving group.
Isomerization of the 3’,3’-5’-phophorothioate triester (1a)
to a mixture of the 2’,3’,5’- and 2’,2’,5’-triesters 12a and 13,
respectively,is also facilitated by the 2 ’-OH function of the
3’-linked nucleoside,albeit somewhat less markedly: 1a is
isomerized 11 times as rapidly as the fast-eluting diastereo-
mer of 1b and 2.4 times as fast as the slowly eluting one.
The only viable hydrogen bond acceptors are the nonbridg-
ing sulfur and the departing 3’-oxygen atom. Hydrogen
bonding of the 2’-OH to the latter is not expected to acceler-
ate isomerization. Accordingly,the observed marked rate
acceleration of isomerization by the 2’-OH function may be
taken as compelling evidence for hydrogen-bond stabiliza-
tion of the thiophosphorane intermediate. The fact that the
acceleration,however,is less prominent than the accelera-
tion of cleavage,argues against the suggestion that accelera-
tion of the cleavage by leaving group stabilization is of
minor importance. Possibly,intramolecular hydrogen bond-
ing,while stabilizing the phosphorane intermediate,simulta-
neously increases the pseudorotation barrier,argued to limit
at least partially the rate of isomerization of monoanionic
phosphorane intermediate (though not a thiophosphor-
ane).[27]
Over a wide range from H0 =ꢀ0.1 to pH 4,the overall
cleavage of both 1a and 1b is pH independent. Evidently
nucleophilic attack of the 2’-oxygen on the phosphorus atom
concerted with transfer of the 2’-hydroxy proton to the
sulfur atom (either direct or water-mediated transfer) gives
a thiophosphorane intermediate which undergoes rate-limit-
ing expulsion of the leaving group concerted with water-
mediated proton transfer from the mercapto ligand to the
departing oxyanion (Scheme 8).[9,18] The rate-retarding effect
of the 2’-OMe substitution is again very similar to that ob-
served previously[19,22,28] for the
Scheme 8. Mechanism of the pH- and buffer-independent cleavage of
phosphorothioate triesters 1a and 1b.
group departs as an alcohol,the neighboring 2 ’-hydroxy
group does not markedly stabilize the phosphorane inter-
mediate or the leaving group by hydrogen bonding. With
1b,another slightly faster pH independent reaction is ob-
served between pH 4 and 6. The origin of this modest (ap-
proximately twofold) rate difference between the two pH-
independent reactions remains obscure,but it possibly is re-
lated to the protolytic equilibrium at the N1 site of the ade-
nine moiety at around pH 4.[26]
Between pH 3 and 5,the overall cleavage of 1a and 1b is
susceptible to general base catalysis by carboxylate anions.
The reaction may be interpreted as a sequential specific
base/general acid catalysis,in which the attacking 2 ’-OH is
deprotonated in a pre-equilibrium step and proton transfer
from the general acid to the departing oxyanion occurs con-
ꢀ
certed with the rate-limiting P O bond fission
(Scheme 9).[18] The rate-retarding effect of the 2’-OMe sub-
stitution is more marked than with the pH- and buffer-inde-
pendent reaction,but not as prominent as in the hydroxide-
ion-catalyzed reaction: 1a is cleaved by carboxylic acid buf-
fers 4–6 times as fast as 1b (the statistical correction made).
The fact that the effects of 2’-OMe substitution on the
cleavage rate of 1a and its oxygen analogue 3a are almost
equal is somewhat unexpected. The sulfur atom is less elec-
tronegative than oxygen and,hence,a weaker hydrogen-
bond acceptor. Accordingly,one might expect the 2 ’-OMe
substitution to be less rate-retarding with 1a than with 3a,
assuming that stabilization of the monoanionic phosphorane
oxygen analogues 3a,b. Com-
pound 1a is cleaved 2.3 times
as rapidly as 1b,and cleavage
ꢀ
ꢀ
of the P O5’ (route A) and P
O3’ (route B) bonds is equally
rapid. With 3a,the rate acceler-
ation has been reported to be
1.6 compared to 3b. As the thi-
ophosphorane intermediate is
now neutral and the leaving Scheme 9. Mechanism of the general-base-catalyzed cleavage of phosphorothioate trimesters 1a and 1b.
4622
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
intermediate by intramolecular hydrogen bonding is the
main source of the reactivity difference between the 2’-OH
and 2’-OMe substituted compounds. One should,however,
bear in mind that the sulfur atom is considerably larger than
the oxygen atom and this may contribute to the strength of
intramolecular hydrogen bonding. The cleavage of ribonu-
cleoside 3’-thioesters,which contain a bridging sulfur atom,
serves as an example of the importance of geometry upon
the formation of a cyclic phosphorane intermediate. The 3’-
thioesters are converted to phosphorane intermediates up to
100 times more readily than their oxygen counterparts,[8] al-
though the inductive effect of thio substitution on nucleo-
philic attack on phosphorus is rather more rate retarding
than rate accelerating.
The hydroxide-ion-catalyzed cleavage of 1a and 1b exhib-
its a thio effect of kPO/kPS =19. A considerably smaller
effect,namely kPO/kPS =4,has been reported for the corre-
sponding reaction of uridine 3’-dimethylphosphorothioates,[9]
whereas the thio effects for the hydroxide-ion-catalyzed hy-
drolysis of triethyl phosphate[29] and dialkyl 4-nitrophenyl
phosphate[30] are of this magnitude,12 and 30,respectively.
Stabilization of the monoanionic phosphorane intermediate
by intramolecular hydrogen bonding,suggested above to be
the reason of 27-fold faster cleavage of 1a compared to 1b,
is not reflected in the thio effect: both compounds exhibit
exactly the same effect. The thio effects for the pH-inde-
pendent cleavage of 1a and 1b,proceeding via a neutral
phosphorane,are considerably smaller,4.2 and 6.4,respec-
tively,but still higher than obtained with uridine 3 ’-di-
Scheme 10. Mechanism for the hydroxide-ion-catalyzed cleavage of phos-
phorodithioate diesters 2a and 2b.
weaker hydrogen-bond acceptor than oxyanion. According-
ly,the difference between the cleavage rates of the di-
thioesters 2a and 2b could be expected to be smaller than
the difference between the cleavage rates of the oxygen
esters, 4a.
The thio effects are small: the 3’-O-linked nucleoside de-
parts 1.5 times more slowly from 2a than from 4a,[20] and
the corresponding reactivity ratio for 2b and 4b is 1.2. In
other words,stabilization of the dianionc phosphorane by
intramolecular hydrogen bonding is not appreciably reflect-
ed to the thio effect. Previously,a small thio-effect has been
reported for the cleavage of 3’,5’-phosphoromono-
thioates[10,11] and their dithioate analogues.[7] Previous theo-
retical calculations suggest that the small values of the thio
effect result from opposite electronic and solvation ef-
fects.[14,15] While higher polarizability of sulfur compared to
oxygen stabilize the dianionic phosphorane intermediate,
poor solvation of the intermediate cancel the resulting rate
acceleration.
Desulfurization starts to compete with cleavage only
under mildly acidic conditions (pH <5),that is,under con-
ditions in which the reactions turn acid-catalyzed and pro-
ceed via a neutral phosphorane intermediate. Under these
conditions,methylation of the 2 ’-OH of the departing alco-
hol has only a minor effect on the rate of the desufurization
and cleavage reactions. In other words,stabilization of the
phosphorane intermediate or leaving group by intramolecu-
lar hydrogen bonding does not play a role.
A
that the thio effect of the hydroxide-ion-catalyzed reaction
is three- to fourfold compared to that of the pH-independ-
ent cleavage means that the thioester is less prone to hy-
droxide-ion catalysis than the oxyester.
The thio effect for the hydroxide-ion-catalyzed isomeriza-
tion of 1a,b appears to be even larger than that of the cleav-
age reaction,of the order of 100. Again,ribonucleoside 3 ’-
dimethylphosphorothioates show a much weaker effect, kPO
/
kPS =4. Estimation of the rate constant for the very rapid hy-
droxide ion-catalyzed isomerization of the oxygen triesters,
3a,b,is,however,susceptible to errors.
Metal-ion-catalyzed cleavage and desulfurization: Zn2+
[12]aneN3 promotes the cleavage of phosphorothioate tri-
Reactions of phosphorodithioate diesters 2a and 2b: In con-
trast to phosphorothioate triesters, 1a,b,which are hydro-
lyzed via a monoanionic thiophosphorane intermediate
under alkaline conditions,the hydroxide-ion-catalyzed
cleavage of phosphorodithioate diesters, 2a,b,proceeds via
a dianionic dithiophosphorane (Scheme 10). Methylation of
the 2’-OH (and 5’-OH) of the departing nucleoside retards
this reaction by a factor of 27 (the statistical correction re-
sulting from the presence of two attacking 2’-OH groups has
been taken into account). The corresponding value for the
rate-retarding influence of the 2’-O-methylation of guanylyl-
3’,3’-uridine (4a versus 4b) has been reported to be 33 and
attributed,at least predominantly,to stabilization of the
phosphorane intermediate by intramolecular hydrogen
bonding.[20] As discussed above,sulfide ion expectedly is a
C
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4623

T. Lçnnberg,M. Ora et al.
centrations of Zn2+[12]aneN3 is five. The situation is qualita-
tively similar with the phosphorodithioate diesters 2a and
2b: the 2’-O-methylated 2b is more susceptible to catalysis
by the Zn2+ aquo ion. The mechanistic explanation for the
difference may well be the same as described for 1a,b. It
should be noted,however,that the reaction now proceeds
pendent of the 2’-substitution of the leaving group. In other
words,stabilization of the thiophosphorane intermediate by
2’-OH hydrogen bonding is not reflected in the value of the
thio effect. The thio effect for the hydroxide-ion-catalyzed
isomerization is of the order of 100. Consistent with the pro-
posed stabilization of the thiophosphorane intermediate by
hydrogen bonding with the 2’-OH of the departing nucleo-
side,the difference between the cleavage rates of the 2 ’-OH
(1a) and 2’-OMe-substituted (1b) triesters is markedly di-
minished at saturating concentrations of Zn2+[12]aneN3.
Evidently binding of Zn2+ to the sulfur ligand of the thio-
phosphorane intermediate does not allow any extra stabili-
zation by hydrogen-bonding and,hence,the cleavage rates
of 1a and 1b become comparable. Phosphorodithioate di-
via
a dianionic phosphorane and,hence,simultaneous
metal-ion binding and 2’-OH hydrogen bonding to the
anionic sulfur ligand is,in principle,still possible.
The Cd2+ ion accelerates desulfurization in addition to
cleavage. In the absence of metal ions,the negatively charg-
ed sulfur ligands initially adopt equatorial positions upon
formation of the dianionic dithiophosphorane. Evidently,
binding of Cd2+ to the sulfur atom makes the CdS+ ligand
sufficiently apicophilic to compete with the 3’-linked nucleo-
side for an apical position. Accordingly,desulfurization by
departure of CdS may compete with the cleavage
A
behave analogously. The rate-retarding effect of 2’-O-meth-
ylation of the leaving nucleoside is 80% of that observed
with the corresponding oxygen diesters,consistent with the
fact that sulfur is not as good a hydrogen-bond acceptor as
(Scheme 11). As with Zn2+,the cleavage of the 2 ’-OMe-sub-
2+
stituted diester, 2b,is accelerated by Cd
more efficiently
than its 2’-OH counterpart, 2a. The cleavage of 2b is accel-
erated slightly more markedly than the desulfurization.
oxygen. The thio effect is small,1.5 with 2a and 1.2 with 2b.
As with triesters 1a,b,thiophilic metal ions,Zn
2+
and Cd2+
,
accelerate more markedly the
cleavage of 2b than the cleav-
age of 2a. Cd2+ induces desul-
furization,which otherwise
occurs only under acidic condi-
tions,in which the reactions
proceed via a neutral dithio-
phosphorane.
Experimental Section
Materials: Nucleosides,uridine
2 ’,3’-
Scheme 11. Mechanism of the metal-ion-catalyzed desulfurization of phosphorodithoate diesters 2a and 2b.
monophosphate and uridine 2’- and 3’-
monophosphates,all used as reference
materials,were commercial products
Conclusion
of Sigma. The preparation of the diasteromeric uridine 2’,3’-cyclic phos-
phoromonothioates,also used as reference materials,has been described
earlier.[12] All the reagents employed were of reagent grade.
In summary,comparative kinetic measurements with phos-
phoromonothioate analogues of trinucleoside 3’,3’,5’-mono-
phosphates 1a and 1b and phosphorodithioate analogues of
dinucleoside 3’,3’-monophosphates 2a and 2b indicate that
2’-OH,as an intramolecular hydrogen-bond donor,acceler-
ates the cleavage of the departing 3’-linked nucleoside. The
observed 27-fold rate-acceleration effect is almost as large
as that observed previously with the oxygen analogues. In
the case of phosphoromonothioate triester 1a,stabilization
of the monoanionic thiophosphorane intermediate by hydro-
gen bonding of the 2’-OH group of a 3’-linked nucleoside
evidently is the major source of the observed rate accelera-
tions,stabilization of the leaving group being only of minor
importance: the 2’-OH accelerates as efficiently the depar-
ture of both the 3’-and 5’-linked nucleoside. The observed
11-fold rate acceleration of isomerization by the 2’-OH func-
tion may be taken as compelling evidence for hydrogen-
bond stabilization of the phosphorane intermediate. The
thio effect for the cleavage reaction, kPO/kPS,is 19 and inde-
2’,3’-O-Methyleneadenosin-5’-yl
bis[2’-O-(4,4’-dimethoxytrityl)-5’-O-
G
methyluridin-3’-yl] phosphorothioate (8a): 2’-O-(4,4’-Dimethoxytrityl)-5’-
O-methyluridine[22] (5,0.56 g,1.00 mmol) was coevaporated three times
from anhydrous pyridine. The residue was dissolved in anhydrous MeCN
(1.0 mL),and then tris(dimethylamino)phosphine (190 mL,1.01 mmol)
and 1H-tetrazole (81.9 mg,1.17 mmol) in anhydrous MeCN (2.6 mL)
were added. After being stirred at room temperature for 2 h,the reaction
mixture was divided into two portions. 2’-O-(4,4’-Dimethoxytrityl)-5’-O-
methyluridine (5,0.28 g,0.50 mmol) was coevaporated three times from
anhydrous pyridine and the residue was added to one portion of the reac-
tion mixture,followed by 1 H-tetrazole (69.3 mg,0.99 mmol) in anhydrous
MeCN (2.2 mL). The reaction mixture was stirred at room temperature
for 120 h,after which
N6-benzoyl-2’,3’-O-methyleneadenosine[23] (7,
0.38 g,0.99 mmol),coevaporated three times from anhydrous pyridine,
and 1H-tetrazole (63.0 mg,0.90 mmol) in anhydrous MeCN (2.0 mL)
were added. The reaction mixture was stirred at room temperature for
27 h,after which elemental sulfur (0.16 g,0.62 mmol) and CH
2Cl2
(1.5 mL) were added. The reaction mixture was stirred at room tempera-
ture for 24 h,after which the sulfur was separated by filtration. The fil-
trate was then neutralized by addition of Et3N and evaporated to dryness.
The residue was dissolved in CH2Cl2 and a conventional aq NaHCO3/
CH2Cl2 workup was carried out. The organic phase was evaporated to
4624
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
dryness and the residue was dissolved in saturated methanolic ammonia
(15 mL). After being stirred at room temperature for 6 h,the reaction
mixture was evaporated to dryness and the residue was purified first on a
silica-gel column eluting with a mixture of MeOH,Et 3N,and CH 2Cl2
(4:1:95 v/v) and then by HPLC on a Supelcosil LC-18 column (25 cm”
21.2 mm,12 mm) eluting with a mixture of water and MeCN (30:70,v/v).
The overall yield was 12.7% from 5 (92.8 mg). 1H NMR (500 MHz,
[D6]DMSO,25 8C,TMS): d=11.49 (d,1H, J=2.1 Hz),11.36 (d,1H, J=
then diphenylphosphorochloridate (0.91 g,3.36 mmol) was added. After
stirring for 20 min at ambient temperature,elemental sulfur (0.11 g,
3.36 mmol) was added. After an additional 2 h stirring,the reaction mix-
ture was evaporated to dryness. The crude product was isolated by a con-
ventional aqueous work up and purified on a silica-gel column,eluting
with a mixture of CH2Cl2 and MeOH (95:5%,v/v). The tert-butyldi-
A
fluoride (1 molLꢀ1) in THF (3 mL). After being stirred for 20 h,the solu-
tion was evaporated to dryness and the product was purified by reversed-
phase chromatography on a Lobar RP-18 column (37”440 mm,40–
63 mm),eluting with a mixture of water and MeCN (92:2%,v/v). Finally,
the product was desalted and passed through a Dowex 50-W Na+ (100–
200 mesh) column. 1H NMR (500 MHz,D 2O,25 8C,external TMS): d=
7.83 (d,2H, J=8.0 Hz),5.90 (d,2H, J=5.2 Hz),5.85 (d,2H, J=8.4 Hz),
2.0 Hz),8.30 (s,1H),8.17 (s,1H),7.5–6.7 (m,28H),7.16 (d,1H,
J=
8.1 Hz),7.06 (d,1H, J=8.1 Hz),6.31 (d,1H, J=8.0 Hz),6.19 (d,1H, J=
8.3 Hz),6.17 (d,1H, J=2.8 Hz),5.49 (d,1H, J1 =2.1, J2 =8.1 Hz),5.38
(dd,1H, J1 =2.0, J2 =8.1 Hz),5.33 (dd, J1 =2.7, J2 =6.5 Hz),5.17 (s,2H),
4.96 (dd, J1 =4.2, J2 =6.4 Hz),4.49 (m,1H),4.35 (m,1H),4.31 (m,1H),
4.27 (m,1H),4.20 (m,1H),4.16 (m,1H),4.00 (m,1H),3.71 (s,6H),
3.70 (s,3H),3.69 (s,3H),3.58–3.48 (m,2H),3.28–3.18 (m,2H),3.19 (s,
4.93 (m,2H),4.47 (t,2H,
12.8, J2 =3.0 Hz),3.85 ppm (dd,2H,
J=5.2 Hz),4.38 (m,1H),3.91 (dd,2H,
J1 =
31P NMR (202 MHz,
J1 =12.8, J2 =4.0 Hz); 31P NMR
3H),2.98 (s,3H),2.80–2.73 ppm (m,2H);
[D6]DMSO,25 8C,H 3PO4): d=66.45 ppm; HRMS (FAB): m/z: calcd for
(202 MHz,D O,25 8C,H PO4): d=116.7; ESI-MS: m/z: 581.0.
2
3
[M]ꢀ: 1458.4430; found: 1458.4459.
2’,5’-Di-O-methyluridin-3’-yl uridin-3’-yl phosphorodithioate (2b): 2’,5’-
Di-O-methyluridin-3’-yl uridin-3’-yl phosphorodithioate (2b) was pre-
pared in a similar manner as 2a,but by using 2 ’,5’-di-O-methyluridine[22]
(6) instead of 2’,5’-bis-O-(tert-butyldimethylsilyl)uridine (9). 1H NMR
(500 MHz,D 2O,25 8C,external TMS): d=7.82 (d,1H, J=8.0 Hz),7.75
(d,1H, J=8.0 Hz),5.93 (d,1H, J=4.5 Hz),5.86 (d,1H, J=4.5 Hz),5.81
(d,1H, J=8.0 Hz),5.80 (d,1H, J=8.0 Hz),4.98 (m,1H),4.89 (m,1H),
2’,3’-O-Methyleneadenosin-5’-yl 2’-O-(4,4’-dimethoxytrityl)-5’-O-methyl-
uridin-3’-yl 2’,5’-di-O-methyluridin-3’-yl phosphorothioate (8b): 2’,5’-Di-
O-methyluridine[23] (6,0.136 g,0.5 mmol),coevaporated three times from
anhydrous pyridine,and 1 H-tetrazole (1.00 mmol,80.0 mg) in anhydrous
MeCN (2.2 mL) were added to the remaining portion of 2’-O-(4,4’-di-
A
dite) (0.68 g,0.5 mmol) in anhydrous MeCN (1.8 mL). The reaction mix-
ture was stirred at room temperature for 120 h,after which N6-benzoyl-
2’,3’-O-methyleneadenosine (7,0.38 g,0.99 mmol),coevaporated three
times from anhydrous pyridine,and 1 H-tetrazole (63.0 mg,0.90 mmol) in
anhydrous MeCN (2.0 mL) were added. The reaction mixture was stirred
at room temperature for 27 h,after which elemental sulfur (0.16 g,
0.62 mmol) and CH2Cl2 (1.5 mL) were added. The reaction mixture was
stirred at room temperature for 24 h. The sulfur was then separated by
filtration. The filtrate was neutralized by addition of Et3N and evaporat-
ed to dryness. The residue was dissolved in CH2Cl2 and a conventional aq
NaHCO3/CH2Cl2 workup was carried out. The organic phase was evapo-
rated to dryness and the residue was dissolved in saturated methanolic
ammonia (10 mL). After being stirred at room temperature for 6 h,the
reaction mixture was evaporated to dryness and the residue was purified
first on a silica-gel column,eluting with a mixture of MeOH,Et 3N,and
CH2Cl2 (4:1:95,v/v) and then by HPLC on a Supelcosil LC-18 column
(25 cm”21.2 mm,12 mm) eluting with a mixture of water and MeCN
(47:53,v/v). Compound 8b was obtained as two diastereomers in an ap-
proximately 2:3 ratio. The overall yield was 9.3% calculated from 6
4.24 (m,1H),4.43 (m,2H),4.34 (m,1H),4.18 (t,1H),3.87 (dd,1H
13.0, J2 =3.0 Hz),3.81 (dd,1H, J1 =13.0, J2 =4.5 Hz),3.75 (dd,1H, J1 =
11.5, J2 =2.0 Hz),3.67 (dd,1H, J1 =11.5, J2 =4.5 Hz),3.44 (s,3H),
3.35 ppm (s,3H);
31P NMR (202 MHz,D 2O,25 8C,H 3PO4): d=
J1 =
116.0 ppm; ESI-MS: m/z: 609.2.
Kinetic measurements: Reactions were carried out in sealed tubes im-
mersed in a thermostated water bath,the temperature of which was ad-
justed to 258C within ꢁ0.18C with 1a and 1b,and 90 8C with 2a and 2b.
Prior to the actual kinetic runs,the 2 ’-O-(4,4’-dimethoxytrityl) groups
were removed from 8a and 8b by adding the starting material in DMSO
(3 mL) to 80 mL of 0.1 molLꢀ1 HCl in 1,4-dioxane. After 30 min at 508C,
1520 mL of the desired reaction solution (prethermostated to 258C) was
added. The aliquots were quenched by cooling to 08C and adjusting their
pH to approximately three with a formic acid buffer when necessary. The
oxonium-ion concentration of reaction solutions was adjusted with nitric
acid,hydrogen chloride,and sodium hydroxide,and with formate,ace-
tate,2-( N-morpholino)ethanesulfonic acid (MES),( N-[2-hydroxyethyl]pi-
perazine-N-[2-ethanesulfonic acid]) (HEPES) and glycine buffers. The
oxonium ion concentrations of the buffer solutions were calculated with
the aid of the pKa values of the buffer acids under the experimental con-
ditions.[32] In addition,the pH values of the solutions were checked with
a pH meter. The initial substrate concentration in the kinetic runs was
1
(54.2 mg). H NMR (500 MHz,[D 6]DMSO,25 8C,TMS; for the fast-elut-
ing diastereomer): d=11.46 (brs,2H),8.31 (s,1H),8.18 (s,1H),7.69 (d,
1H, J=8.1 Hz),7.38 (brs,2H),7.33–7.15 (m,9H),7.21 (d,1H,
J=
8.1 Hz),6.77 (m,4H),6.27 (d,1H, J=8.2 Hz),6.18 (d,1H, J=2.7 Hz),
approximately 10ꢀ4 molLꢀ1
.
5.87 (d,1H, J=6.1 Hz),5.74 (d,1H, J=8.1 Hz),5.49 (d,1H, J=8.1 Hz),
5.36 (dd,1H, J1 =2.7, J2 =6.5 Hz),5.19 (s,1H),5.18 (s,1H),5.03 (dd,
1H, J1 =4.0, J2 =6.4 Hz),4.95 (m,1H),4.29 (m,3H),4.25–4.15 (m,2H),
4.09 (m,1H),3.70 (s,3H),3.69 (s,3H),3.46–3.37 (m,2H),3.29 (s,3H),
3.26 (s,3H),3.19 (s,3H),3.22–3.10 (m,2H),2.75–2.67 ppm (m,2H);
1H NMR (500 MHz,[D 6]DMSO,25 8C,TMS; for the slowly eluting dia-
stereomer): d=11.45 (brs,2H),8.31 (s,1H),8.15 (s,1H),7.68 (d,1H,
The composition of the samples withdrawn at appropriate time intervals
was analyzed by HPLC on a Hypersil-Keystone Aquasil C18 column (4”
150 mm,5 mm) by using a mixture of 0.06 molLꢀ1 formic acid buffer and
MeCN as an eluent. The amount of MeCN was 7% for the first 10 min,
after which it was increased linearly to 27% during 20 min. The observed
retention times (tR/min) for the products of 1a and 1b (the flow rate was
1.0 mLminꢀ1) were as follows: 35.2 (both diastereomers of 1b); 32.3
(both diastereomers of 12b); 32.3 (1a); 30.0,29.4 (diastereomers of 12a);
27.2 (13); 21.1,18.8,17.9,15.8 ( 15a,diastereomers of 16a, 17); 23.7,22.8,
21.4,20.9 (diastereomers of 15b and 16b); 21.7,19.5,18.7,16.8 (diaste-
reomers of 18 and 19); 7.8 (6); 6.8 (14); 3.9 (20). The nucleosidic prod-
ucts (6, 14, 20) were identified by spiking with authentic samples. The
diester products were identified either as di(uridin-3’-yl)esters (15a, 16a,
17 from 1a; 15b, 16b from 1b) or uridin-3’-yl adenosine-5’-yl esters (18,
19) by ESI-HPLC/MS. The triester isomers of the starting materials were
also identified by ESI-HPLC/MS. Assignment of 12a and 13 as 2’,3’,5’-
and 2’,2’,5’-isomers,respectively,was based on the fact that 12a appeared
as an intermediate of the formation of 13. In calculations,the total con-
centrations of di(uridin-3’-yl) esters ([15a]+[16a]+[17] or [15b]+[16b]),
uridin-3’-yl adenosine-5’-yl esters ([18]+[19]),and isomerized triesters
([12a]+[13]) were used.
J=8.1 Hz),7.35 (brs,2H),7.42–7.15 (m,9H),7.16 (d,1H,
J=8.1 Hz),
6.82 (m,4H),6.23 (d,1H, J=2.5 Hz),6.20 (d,1H, J=8.2 Hz),5.90 (d,
1H, J=6.1 Hz),5.76 (d,1H, J=8.1 Hz),5.46 (d,1H, J=8.1 Hz),5.36
(dd,1H, J1 =2.5, J2 =6.5 Hz),5.16 (s,1H),5.15 (s,1H),5.04 (dd,1H,
J1 =3.8, J2 =6.4 Hz),4.95 (m,1H),4.34–4.17 (m,4H),4.15–4.04 (m,2H),
3.72 (s,3H),3.71 (s,3H),3.46–3.37 (m,2H),3.25 (s,3H),3.19 (s,3H),
3.18 (s,3H),3.22–3.10 (m,2H),2.75–2.67 ppm (m,2H);
31P NMR
(202 MHz,[D 6]DMSO,25 8C,H 3PO4; for the fast-eluting diastereomer):
d=66.75 ppm; 31P NMR (202 MHz,[D6]DMSO,25 8C,H 3PO4; for the
slowly eluting diastereomer): d=66.13 ppm; HRMS (FAB): m/z: calcd
for [M]ꢀ: 1170.3280; found: 1170.3283.
Bis(uridin-3’-yl) phosphorodithioate (2a): 2’,5’-Bis-O-(tert-butyldimethyl-
silyl)uridin3 3’-hydrogenphosphonothioate[24] (10,0.25 g,0.57 mmol) and
2’,5’-bis-O-(tert-butyldimethylsilyl)uridine[24] (9,0.58 g,1.21 mmol) were
dissolved in 10 mL of a mixture of MeCN and pyridine (1:4,v/v),and
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4625

T. Lçnnberg,M. Ora et al.
A good separation of the product mixture of 2a was obtained,when an
8 min isocratic elution with a acetic acid/sodium acetate buffer (0.045/
0.015 molLꢀ1) containing 0.1 molLꢀ1 ammonium chloride and 1% aceto-
nitrile was followed by a linear gradient (32 min) up to 6.0% MeCN. The
observed retention times (tR/min) for the products of 2a on a Hypersil
ODS 5 column (4”250 mm,5 mm) at flow rate 1 mLminꢀ1 were: 27.0,
21.8,19.8 ( 23a, 24a, 25); 22.2,21.8,18.6,17.6 (fully desulfurized di-
2425–2438; e) J. M. Warnecke,J. P. Fꢁrste,W.-D. Hardt,V. A. Erd-
mann,R. K. Hartmann, Proc. Natl. Acad. Sci. USA 1996, 93,8924–
8928; f) T. S. McConnell,T. R. Cech, Biochemistry 1995, 34,4056–
4067; g) M. Podar,P. S. Perlman,R. A. Padgett,
Mol. Cell. Biol.
1995, 15,4466–4478; h) R. A. Padgett,M. Podar,S. C. Boulanger,
P. S. Perlman, Science 1994, 266,1685–1688; i) D. Herschlag,M.
Khosla, Biochemistry 1994, 33,5291–5297; j) D. Herschlag,F. Eck-
stein,T. R. Cech, Biochemistry 1993, 32,8312–8321; k) D. Her-
esters); 14.2 (21); 11.7 (RP-26a); 7.1 (22); 6.1 (SP-26a) and 4.5 (26b). The
observed retention time for the starting material 2a was 38.5 min. With
2b,a sufficient separation of the product mixture was obtained,when an
8 min isocratic elution with buffer was followed by a linear gradient
(57 min) up to 14.0% acetonitrile. The observed retention times (tR/min)
for the products of 2b were as follows: 39.7,37.8,31.2 ( RP-23b, SP-23b,
RP/SP-24b); 28.3,27.2 (desulfurized diesters); 18.1 ( 21); 16.0 (RP-26); 9.5
(22); 7.2 (SP-26); 7.0 (2’-UMP); 5.3 (26b); 5.8 (3’-UMP); and 4.5 (uracil).
The observed retention time for starting materials 4 was 49.7. The prod-
ucts were either characterized by spiking with authentic samples mono-
meric nucleoside/nucleotides or identified as phosphoromonothioate di-
A
l) J. Rajagopal,J. A. Doudna,J. W. Szostak, Science 1989, 244,692–
694.
[4] a) A. Yoshida,S. Sun,J. A. Piccirilli, Nat. Sruct. Biol. 1999, 6,318–
321; b) J. A. Picirilli,J. S. Vyle,M. H. Caruthers,T. R. Cech, Nature
1993, 361,85–88.
[5] a) P. M. Gordon,E. J. Sontheimer,J. A. Piccirilli, Biochemistry 2000,
39,12939–12952; b) J. M. Warnecke,E. J. Sontheimer,J.A. Picciril-
li,R. K. Hartmann, Nucleic AcidsRe.s 2000, 28,720–727; c) E. J.
Sontheimer,P. M. Gordon,J. A. Piccirilli,
1729–1741; d) L. B. Weinstein,B. C. M. N. Jones,R. Cosstick,T. R.
Cech, Nature 1997, 388,805–808.
A
furized diesters by mass spectrometric analysis (HPLC/ESI-MS) by using
a mixture of MeCN and aqueous fomic acid (0.4%) as an eluent. In cal-
culations,the total concentration of phosphoromonothioate diesters
([23a]+[24a]+[25] or [23b]+[24b]) was used.
[6] a) J. Purcell,A. C. Hengge, J. Org. Chem. 2005, 70,8437–8442;
b) D. Herschlag, J. Am. Chem. Soc. 1994, 116,11631–11635,and
references therein.
[7] M. Ora,J. Jꢂrvi,M. Oivanen,H. Lçnnberg, J. Org. Chem. 2000, 65,
2651–2657.
[8] M. I. Elzagheid,M. Oivanen,K. D. Klika,B. C. N. M. Jones,R. Cos-
stick,H. Lçnnberg, Nucleosides Nucleotides 1999, 18,2093–2108.
[9] M. Ora,M. Oivanen,H. Lçnnberg, J. Org. Chem. 1997, 62,3246–
3253.
[10] a) H. Almer,R. Strçmberg, J. Am. Chem. Soc. 1996, 118,7921–
7928; b) H. Almer,R. Strçmberg, Tetrahedron Lett. 1991, 32,3723–
3726.
[11] M. Ora,M. Oivanen,H. Lçnnberg, J. Org. Chem. 1996, 61,3951–
3955.
The signal areas of the nucleotide analogues,obtained as products from
hydrolysis of 2a and 2b,were assumed to be proportional to concentra-
tions,as the base moiety of all the compounds was the same. With the di-
nucleoside monophosphates and their thioate analogues,the molar ab-
sorptivities of the two base moieties were assumed to be additive. In the
case of 1a and 1b,the peak areas were converted to relative concentra-
tions by calibrating the system with uridine and adenosine solutions of
known concentrations. The contribution of the buffer catalysis to the ob-
served rate constants of 2a and 2b was insignificant even at the higher
buffer concentration employed (0.1 molLꢀ1).
Calculations of the rate constants: The pseudo-first-order rate constants
for the decomposition of starting materials were obtained by applying
the integrated first-order rate equation to the time-dependent concentra-
tion of the starting material. In cases for which desulfurization competed
with the cleavage,the first-order rate constants ( kds) for the desulfuriza-
tion of 2a were calculated by applying the kinetics of consecutive parallel
first-order reactions [Eq. (6)). In Equation (6), kd is the first-order rate
constant for the disappearance of 2a, [2a]0 stands for the initial concen-
tration of 2a and [23–25]t denotes the sum concentration of all diastereo-
meric forms of 23a, 24a,and 25 at moment t. k1 is the first-order rate
constant for the disappearance of the latter compounds.
[12] M. Oivanen,M. Ora,H. Almer,R. Strçmberg,H. Lçnnberg,
J. Org.
Chem. 1995, 60,5620–5627.
[13] a) K. Sorensen-Stowell,A. C. Hengge, J. Org. Chem. 2006, 71,
7180–7184; b) I. E. Catrina,A. C. Hengge, J. Am. Chem. Soc. 2003,
125,7546–7552; c) M. I. Elzagheid,E. Mꢂki,U. Kaukinen,M. Oiva-
nen,H. Lçnnberg, Nucleosides Nucleotides Nucleic Acids 2000, 19,
827–838; d) I. E. Catrina,A. C. Hengge, J. Am. Chem. Soc. 1999,
121,2156–2163; e) M. Ora,M. Peltomꢂki,M. Oivanen,H. Lçnn-
berg, J. Org. Chem. 1998, 63,2939–2947; f) M. Ora,M. Oivanen,H.
Lçnnberg, J. Chem. Soc. Perkin Trans. 2 1996,771–774.
[14] Y. Liu,B. A. Gregersen,A. C. Hengge,D. M. York,
Biochemistry
2006, 45,10043–10053.
[15] a) Y. Liu,X. Lopez,D. M. York, J. Phys. Chem. B 2005, 109,19987–
½23 ꢀ 25Št=½2aŠ ¼ ½kds=ðk1ꢀkdފ ꢃ ½expðꢀkdtÞꢀexpðꢀk1tފ
ð6Þ
0
20003; b) B. A. Gregersen,X. Lopez,D. M. York,
J. Am. Chem.
Soc. 2004, 126,7504–7513.
In the case of 2b,the first-order rate constants ( kds) for the desulfuriza-
tion were obtained by breakdown of the rate constant of the disappear-
ance of the starting material (kd) to the rate constants of the two parallel
first-order reactions (desulfurization and cleavage) on the basis of the
product distribution at the early stages of the reaction. The same method
was applied to obtain the rate constants for the parallel reactions of
triesters 1a and 1b. The resulting diesters did not undergo further de-
composition during the time used to follow the disappearance of the
triesters. The isomeric triesters,the RP- and SP-diasteromers 1a, 12a and
13 on one hand,and the RP- and SP-diasteromers of 1b and 12b on the
other hand,were assumed to react at equal rates.
[16] a) Y. Liu,X. Lopez,D. M. York, Chem. Commun. 2005,3909–3911;
b) C. S. Lopez,O. N. Faza,B. A. Gregersen,X. Lopez,A. R.
de Lera,D. M. York, ChemPhysChem 2004, 5,1045–1049.
[17] a) S. A. Strobel,L. Ortoleva-Donnelly, Chem. Biol. 1999, 6,153–
165; b) S.-O. Shan,A. Yoshida,S. Sun,J. A. Piccirilli,D. Herschlag,
Proc. Natl. Acad. Sci. USA 1999, 96,12299–12304.
[18] a) T. Lçnnberg,H. Lçnnberg, Curr. Opin. Chem. Biol. 2005, 9,665–
673; b) M. Ora,H. Linjalahti,H. Lçnnberg, J. Am. Chem. Soc. 2005,
127,1826–1832.
[19] T. Lçnnberg,J. Korhonen, J. Am. Chem. Soc. 2005, 127,7752–7758.
[20] A. Kiviniemi,T. Lçnnberg,M. Ora, J. Am. Chem. Soc. 2004, 126,
11040–11045.
[21] S. Bakalova,W. Siebrand,A. Fernandez-Ramos,Z. Smedarchina,
D. D. Petkov, J. Phys. Chem. B 2002, 106,1476–1480.
[1] Y. Takagi,Y. Ikeda,K. Taira, Top. Curr. Chem. 2004, 232,213–251.
[2] a) S. Shan,A. V. Kravchuk,J. A. Piccirilli,D. Herschlag,
Biochem-
[22] T. Lçnnberg,J. Kiiski,S. Mikkola,
1089–1096.
Org. Biomol. Chem. 2005, 3,
istry 2001, 40,5161–5171; b) A. Yoshida,S. Shan,D. Herschlag,
J. A. Piccirilli, Chem. Biol. 2000, 7,85–96.
[23] T. Lçnnberg,S. Mikkola, J. Org. Chem. 2004, 69,802–810.
[24] a) J. Stawinski,M. Thelin,E. Westman,R. Zain, J. Org. Chem. 1993,
[3] a) S. Shan,D. Herschlag, RNA 2000, 6,795–813; b) J. M. Warnecke,
R. Held,S. Busch,R. K. Hartmann, J. Mol. Biol. 1999, 290,433–
445; c) M. Podar,P. S. Perlman,R. A. Padgett, RNA 1998, 4,890–
58,3503–3506; b) R. Zain,J. Stawinski,
J. Org. Chem. 1996, 61,
900; d) Y. Chen,X. Li,P. Gegenheimer,
Biochemistry 1997, 36,
6617–6622.
4626
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
Chem. Eur. J. 2007, 13,4614 – 4627

Thio Effects
FULL PAPER
[25] H. S. Harned,G. E. Mannweiler, J. Am. Chem. Soc. 1935, 57,1873–
[30] J. A. A. Ketelaar,H. R.; K. Koopmans, Recl. Trav. Chim. Pay-Bas
1952, 71,1253–1258.
1876.
[26] R. M. Izatt,J. J. Christensen,J. H. Rytting,
439–481.
[27] E. Mꢂki,M. Oivanen,P. Poijꢂrvi,H. Lçnnberg, J. Chem. Soc. Perkin
Trans. 2 1999,2493–2499.
[28] S. Mikkola,M. Kosonen,H. Lçnnberg, Curr. Org. Chem. 2002, 6,
523–538.
Chem. Rev. 1971, 71,
[31] S. Mikkola,E. Stenman,K. Nurmi,E. Yousefi-Salakdeh,R. Strçm-
berg,H. Lçnnberg, J. Chem. Soc. Perkin Trans. 2 1999,1619–1625.
[32] E. P. Serjeant,B. Dempsey, Ionization Constants of Organic Acids in
AqueousSolution ,IUPAC Chemical Data Series No. 23,Pergamon,
Oxford, 1979.
[29] T. Fanni,K. Taira,D. G. Gorenstein,R. Vaidyanathaswamy,J. G.
Verkade, J. Am. Chem. Soc. 1986, 108,6311–6314.
Received: December 20,2006
Published online: March 1,2007
Chem. Eur. J. 2007, 13,4614 – 4627
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
4627