Kinetic and Theoretical Studies on the Mechanism of Alkaline Hydrolysis of DNA

Article abstract of DOI:10.1021/jo000323d

The reaction mechanism of alkaline hydrolysis of DNA has been investigated by kinetic analysis and density-functional-theory calculation. The rates of hydrolysis of thymidine 3′-monophosphate esters (including thymidylyl(3′-5′)thymidine (Tp-OT)) monotonically decrease as the leaving groups get poorer. According to the theoretical calculation in which the solvent effects are incorporated, no intermediate is formed in the course of the reaction. In the alkaline hydrolysis of the activated Tp-OT analogues having good leaving groups, the 3′,5′-cyclic monophosphate of thymidine is concurrently formed through the intramolecular attack by the 5′-alkoxide ion. In the hydrolysis of the native dinucleotide, however, this side reaction does not occur, since the transition state leading to the departure of its poor leaving group cannot be formed due to conformational restraint. These arguments are supported by the theoretical analysis on the hydrolysis of both dimethyl phosphate and its O(bridging)→S substituted analogue.

Full text of DOI:10.1021/jo000323d

J . Org. Chem. 2000, 65, 4391-4396
4391
Kin etic a n d Th eor etica l Stu d ies on th e Mech a n ism of Alk a lin e
Hyd r olysis of DNA
Naoya Takeda, Masahiko Shibata,† Nobuo Tajima,† Kimihiko Hirao,† and Makoto Komiyama*
Research Center for Advanced Science and Technology, The University of Tokyo,
Komaba, Tokyo 153-8904, J apan, and Department of Applied Chemistry, Graduate School of
Engineering, The University of Tokyo, Hongo, Tokyo 113-8656, J apan
Received March 7, 2000
The reaction mechanism of alkaline hydrolysis of DNA has been investigated by kinetic analysis
and density-functional-theory calculation. The rates of hydrolysis of thymidine 3′-monophosphate
esters (including thymidylyl(3′-5′)thymidine (Tp-OT)) monotonically decrease as the leaving groups
get poorer. According to the theoretical calculation in which the solvent effects are incorporated,
no intermediate is formed in the course of the reaction. In the alkaline hydrolysis of the activated
Tp-OT analogues having good leaving groups, the 3′,5′-cyclic monophosphate of thymidine is
concurrently formed through the intramolecular attack by the 5′-alkoxide ion. In the hydrolysis of
the native dinucleotide, however, this side reaction does not occur, since the transition state leading
to the departure of its poor leaving group cannot be formed due to conformational restraint. These
arguments are supported by the theoretical analysis on the hydrolysis of both dimethyl phosphate
and its O(bridging)fS substituted analogue.
In tr od u ction
atom, and a pentacoordinated intermediate (or transition
state) is formed. In the following stage, the bond between
the P atom and the 5′-O (or 3′-O) atom of 2′-deoxyribo-
nucleotide is cleaved. A pioneering quantum-chemical
study on alkaline hydrolysis of dimethyl phosphate was
achieved by Karplus et al.12,13 However, correlation of the
results with kinetic results has not been attempted.
In the present paper, both kinetic and theoretical
approaches are adopted to investigate the mechanism of
alkaline hydrolysis of DNA. Thymidylyl(3′-5′)thymidine
(Tp-OT) and its analogues containing various leaving
groups are hydrolyzed under alkaline conditions, and the
dependence of hydrolysis rate on the basicity of leaving
group is determined. The rates of intramolecular trans-
esterification in these analogues, in which the P atom is
attacked by the 5′-OH, are compared with the rates of
alkaline hydrolysis. Theoretical studies are also achieved
in the density-functional-theory calculation and the effect
of substitution of the bridging O atom to an S atom is
analyzed in detail. The structure of the transition-state
is clarified, showing how it resembles the reactants or
the products.
Nonenzymatic hydrolysis of DNA and RNA is one of
the most important and attractive themes in nucleic acid
chemistry. So far, many studies on the mechanism of
RNA hydrolysis have been achieved.1-4 A number of
kinetic data have been accumulated and sufficiently
rationalized by theoretical studies. However, the infor-
mation on the mechanism of DNA hydrolysis has been
scarce, mainly because spontaneous hydrolysis of DNA
is virtually nil under physiological conditions. Recently,
outstanding activities of lanthanide ions and other metal
ions were found, and DNA was hydrolyzed at reasonable
rates.5-9 A new era has been opened in this field. The
information on the mechanism of DNA hydrolysis10,11 is
crucially important for precise understanding of the
catalytic mechanisms and also for design of active
catalysts.
DNA hydrolysis consists of two stages (Figure 1). In
the first stage, hydroxide ion attacks the phosphorus
* To whom correspondence should be addressed. Tel and Fax:
+81-3-5452-5209. E-mail: komiyama@mkomi.rcast.u-tokyo.ac.jp.
† Graduate School of Engineering.
(1) Perreault, D. M.; Anslyn, E. V. Angew. Chem., Int. Ed. Engl.
1997, 36, 432-450.
(2) Oivanen, M.; Kuusela, S.; Lo¨nnberg, H. Chem. Rev. 1998, 98,
961-990.
(3) Zhou, D.-M.; Taira, K. Chem. Rev. 1998, 98, 991-1026.
(4) Li, Y.; Breaker, R. R. J . Am. Chem. Soc. 1999, 121, 5364-5372.
(5) Shiiba, T.; Yonezawa, K.; Takeda, N.; Matsumoto, Y.; Yashiro,
M.; Komiyama, M. J . Mol. Catal. 1993, 84, L21-L25.
(6) Takasaki, B. K.; Chin, J . J . Am. Chem. Soc. 1994, 116, 1121-
1122.
Resu lts a n d Discu ssion
Alk a lin e Hyd r olysis of Tp -OT a n d Its An a logu es.
(1) Kin etic Stu d ies. At pH 14 and 80 °C, the pseudo-
first-order rate-constant for alkaline hydrolysis (khyd) of
Tp-OT is 2.0 × 10-3 h-1. The products are thymidine
(Thd) and its 3′- or 5′-monophosphate (Tp and pT,
respectively), as expected.
(7) Komiyama, M.; Takeda, N.; Takahashi, Y.; Uchida, H.; Shiiba,
T.; Kodama, T.; Yashiro, M. J . Chem. Soc., Perkin Trans. 2 1995, 269-
274.
To shed light on the reaction mechanism, the leaving
group in Tp-OT was systematically changed, with the Tp
moiety in the 5′-side kept constant (see Figure 2). The
rate constants (khyd) for their hydrolyses are listed in
(8) Irisawa, M.; Komiyama, M. J . Biochem. 1995, 117, 465-466.
(9) Takeda, N.; Imai, T.; Irisawa, M.; Sumaoka, J .; Yashiro, M.;
Shigekawa, H.; Komiyama, M. Chem. Lett. 1996, 599-600.
(10) Sumaoka, J .; Takeda, N.; Okada, Y.; Takahashi, H.; Shigekawa,
H.; Komiyama, M. Nucleic Acids Res. Symp. Ser. 1998, 39, 137-138.
(11) The mechanism of Ce(IV) catalysis has been recently discussed.
Komiyama, M.; Takeda, N.; Shigekawa, H. J . Chem. Soc., Chem.
Commun. 1999, 1443-1451.
(12) Dejaegere, A.; Lim, C.; Karplus, M. J . Am. Chem. Soc. 1991,
113, 4353-4355.
(13) Dejaegere, A.; Liang, X.; Karplus, M. J . Chem. Soc., Faraday
Trans. 1994, 90, 1763-1770.
10.1021/jo000323d CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/21/2000

4392 J . Org. Chem., Vol. 65, No. 14, 2000
Takeda et al.
F igu r e 1. Scheme of DNA hydrolysis.
F igu r e 3. Brønsted-type correlation for the alkaline hydroly-
sis of Tp-OT and its analogues at pH 14 and 80 °C. All the
data are from Table 1. The slope of the fitting line is -0.27.
F igu r e 2. Structures of the substrates utilized in the present
kinetic studies (a) and in the theoretical analyses (b).
conjugate acids of leaving groups, (pKa,lg) are around 10.6
and 14.8.14 This argument is strongly supported by the
fact that Tp-OCH2CF3 was hydrolyzed slightly faster
than was Tp-OT, as expected from its smaller pKa,lg value
(Table 1). Here, the perturbation around the P atom is
marginal. The analogues involving aryl leaving groups
were also hydrolyzed more promptly than Tp-OT.
As shown in Figure 3, the reaction rate monotonically
Ta ble 1. P seu d o-F ir st-Or d er Ra te Con sta n ts for th e
Alk a lin e Hyd r olysis of P h osp h od iester Lin k a ge (kh yd
)
a n d for th e In tr a m olecu la r Tr a n sester ifica tion by th e
5′-OH Atta ck (kin tr a ) a t p H 14 a n d 80 °Ca
b
substrate
Tp-OT
Tp-OCH2CF3
Tp-ST
Tp-OPh(o-Cl)
Tp-OPh(p-NO2)
Tp-(S)T
pKa,lg
khyd/h-1
kintra/h-1
14.8
12.4
10.6
8.1
7.2
14.8
2.0 × 10-3
3 × 10-3d
2.3 × 10-2
4.2 × 10-2
3.2 × 10-1
2.3 × 10-3
1.2 × 10-1
0c
2 × 10-3d
6.6 × 10-2
4.2 × 10-2
4.8 × 10-1
decreases with increasing pKa,lg
: Tp-OPh(p-NO2) >
Tp-OPh(o-Cl) > Tp-ST > Tp-OCH2CF3 > Tp-OT. The
removal of the leaving group from the P atom is, at least
partially, rate limiting. The slope of this Brønsted-type
plot is -0.27. This value is considerably smaller than the
value (-0.76) reported by Chin et al.15 This is probably
because, in the present study, the esters of Tp were used
as the substrates, to minimize undesirable perturbation
cTMP
a
The error in the rate constants is around 10%, unless noted
b
otherwise. The pKa values of the conjugate acids of the leaving
groups.14 c No cTMP was detected by the HPLC (see the ref 24).
d
Precise determination of the khyd and kintra values is difficult,
since the reaction is very slow (the error is around 30%).
(14) pKa values. (a) Tp-OT; its pKa was taken as identical with the
value of CH3OCH2CH2OH (Thompson, J . E.; Raines, R. T. J . Am. Chem.
Soc. 1994, 116, 5467-5468; Ballinger, P.; Long, F. A. J . Am. Chem.
Soc. 1960, 82, 795-798). (b) Tp-OCH2CF3 (Murto, J . The Chemistry of
Functional Groups, The Chemistry of the Hydroxyl Group, Part 2; Patai,
S., Ed; J ohn Wiley & Sons: London, 1971; pp 1087-1115). (c) Tp-ST;
estimated from the value of ethyl mercaptan (Kagaku-binran Kiso-
hen; The Chemical Society of J apan, Ed.; Maruzen: Tokyo, 1984; II-
339). (d) Tp-OPh(o-Cl) and Tp-OPh(p-NO2) (Kortu¨m, G.; Vogel, W.;
Andrussow, K. Dissociation Constants of Organic Acids in Aqueous
Solution; Butterworth: London, 1961; pp 444-451). (e) H2O (Sille´n,
L. G.; Martell, A. E. Stability Constants of Metal-Ion Complexes,
Supplement No 1; The Chemical Society: London, 1971; pp 14-15).
(15) Chin, J .; Banaszczyk, M.; J ubian, V.; Zou, X. J . Am. Chem. Soc.
1989, 111, 186-190.
Table 1 (these values were determined by using eqs 1
and 2, as described in the Experimental Section). When
the bridging O atom in Tp-OT was replaced with S
(Tp-ST), the hydrolysis rate increased by 11-fold (khyd
)
2.3 × 10-2 h-1). In contrast, Tp-(S)T, in which one of the
two nonbridging O atoms in Tp-OT is substituted with
S, was hydrolyzed at almost the same rate as Tp-OT.
Apparently, the OfS substitution around the P atom
does not significantly alter the electrophilicity of the P
atom. Tp-ST is hydrolyzed faster than Tp-OT, since its
leaving group is better; the pKa’s of R-SH and R-OH, the

Mechanism of Alkaline Hydrolysis of DNA
J . Org. Chem., Vol. 65, No. 14, 2000 4393
bond (0.17324). The incomplete formation of this bond is
further substantiated.
All these arguments were supported by the natural
population analysis, in which the charge of each atom
(i.e., atomic charge) is accurately evaluated.18 The charges
for the O atoms of OH- (nucleophile) and of the OH
residue in the product (CH3OP(dO)(OH)(O-)) are -1.47
and -1.08, respectively. The corresponding value of the
O atom of the P-OH(axial) group in the transition state
is -1.25. According to the Leffler’s approach,19 the extent
of the formation of the P-OH bond in the transition state
(R) is estimated to be 56.5% () [-1.47 - (-1.25)]/[-1.47
- (-1.08)]). Similarly, the extent of the P-OCH3 bond
cleavage in the transition state (â) is estimated as 14.2%
(the charges of the O atoms of the leaving OCH3 group
in DMP-O, of the P-OCH3(axial) in the transition state
and of the product (CH3O-) are -0.89, -0.92, and -1.13,
respectively). The transition state resembles the initial
state, rather than the products (vide post).20 These results
fairly agree with the structural analysis using P-O bond
length (described above) and also with the small slope
(-0.27) in the Brønsted-type plot in Figure 3.
When the quantum-chemical calculation was made on
the DMP-O hydrolysis in a gas phase (without the
contribution of the solvent effects), the energy barrier for
the second stage (relating to the scission of the P-OCH3
linkage) was increased, and an intermediate emerged
between the two transition states (the OH- attack and
the P-OCH3 scission). However, the former energy
barrier is higher than the latter by only 0.99 kcal mol-1
and the energy-well is too shallow to be kinetically
significant (0.01 kcal mol-1 depth with respect to the
second transition state). The energy surface between the
two transition states would be rather flat.
For the purpose of comparison, one of the two bridging
O atoms in DMP-O was replaced with S (DMP-S). The
hydrolysis products are CH3OP(dO)(OH)(O-) and CH3S-.
As was the case in the hydrolysis of DMP-O, the solvation
effects were incorporated in the calculation, and both the
attacking OH- and -SCH3 were placed at the axial
positions at the start of the calculation.21 It was found
that the P-S bond is cleaved in one step without any
intermediate (the broken line in Figure 4).22 The activa-
tion free energy (24.48 kcal mol-1) is by 9.16 kcal mol-1
smaller than that in the DMP-O hydrolysis. In the
transition state, the length of the P-OH(axial) bond is
2.950 Å. This bond length is significantly longer than the
corresponding value in the DMP-O hydrolysis (2.125 Å).
In the DMP-S hydrolysis, the P-OH bond formation is
less complete in the transition state than that in the
DMP-O hydrolysis. The DMP-S has a better leaving
group (CH3S-), so that the transition state is readily
formed when the attacking OH- binds to the P atom to
F igu r e 4. Energy-diagrams for the alkaline hydrolysis of
DMP-O (the solid line) and DMP-S (the broken line). At the
top, the structures of the transition states, optimized at RHF/
6-31+G* level, are presented. The calculation was made at
the B3LYP/6-31+G*//RHF/6-31+G* level, where the solvation
effects were incorporated by the PCM method. The sum of the
energy potentials of the reactants was taken as 0 kcal mol-1
.
around the P atom. In ref 15, the substrates were diaryl
phosphates and dialkyl phosphates.
(2) Qu a n tu m -Ch em ica l Stu d ies. In the ab initio
calculation, dimethyl phosphate (DMP-O) was used as a
model compound of DNA. By using the polarized con-
tinuum model (PCM), the solvation effects by water mole-
cules were sufficiently incorporated in the calculation of
the Gibbs free energies. The charge of each atom in
various species on the reaction-coordinate was evaluated
by the natural population analysis. To examine the
extents of the bond formation [PrO(attacking)] and the
bond cleavage [P- - -O(leaving)] in the transition state,
the imaginary frequencies were calculated by normal
coordinate analysis of force constants. On optimizing the
structures of phosphoranes, the calculations were started
by placing a hydroxide ion and one of the OCH3 groups
in the DMP-O at the axial positions of the pentacoordi-
nated P atom. The phosphoranes, formed from (CH3O)2P-
(dO)(O-) and OH-, are dianionic (the two OH residues
at the equatorial positions are dissociated), consistently
with the corresponding pKa values (pKa1 ) 6.5-11 and
pKa2 ) 11.3-15).1,16,17 The leaving OCH3 group from the
axial position departed as an anion.
According to this calculation, the alkaline hydrolysis
proceeds in one step (the solid line in Figure 4). The
activation free energy is 33.64 kcal mol-1. No stable
intermediate is formed along the reaction coordinate. In
the transition state, the bond length between the O atom
of the attacking OH residue (in the axial position) and
the P atom is 2.125 Å. This bond is considerably longer
than the bond (1.795 Å) between the O atom of the
leaving OCH3 and the P atom. The P-OH bond in the
product (CH3OP(dO)(OH)(O-)) is 1.627 Å. In the transi-
tion state, the P-OH(axial) bond is partially formed, and
the P-OCH3(axial) bond is being kept still rather intact.
Consistently, the imaginary frequency of P-OH(axial)
bond (0.87564) is larger than that of the P-OCH3(axial)
(18) On the RHF/6-31+G* level employed in the present study, the
natural population analysis gives more reasonable values of the charge
than the Mulliken population analysis. J ensen, F. Introduction to
Computational Chemistry; J ohn Wiley & Sons: Chichester, 1999; pp
230-234.
(19) Leffler, J .; Grunwald, E. Rates and Equilibria of Organic
Reactions; J ohn Wiley & Sons: New York, 1963; pp 156-161.
(20) Karplus et al. pointed out that the alkaline hydrolysis of DMP-O
has the early transition state, based on its Mulliken charge distribution
(in ref 13).
(21) When the SCH3 group was located at the equatorial position
in the beginning of calculation, the intermediate was generated.
However, it was not normally hydrolyzed but unreasonably released
the SCH3 group during the pseudo-rotation process.
(22) In a gas phase, no intermediate was formed in the alkaline
hydrolysis of DMP-S.
(16) Kluger, R.; Colvitz, F.; Dennis, E.; Williams, D.; Westheimer,
F. H. J . Am. Chem. Soc. 1969, 91, 6066-6072.
(17) Guthrie, J . P. J . Am. Chem. Soc. 1977, 99, 3991-4001.

4394 J . Org. Chem., Vol. 65, No. 14, 2000
Takeda et al.
Furthermore, the intramolecular reaction is much more
appropriate for the analysis, since the position of the
nucleophile, with respect to the P atom, is greatly
restricted therein.
(1) Kin etic Stu d ies. The rate constants for the
intramolecular transesterification by the 5′-OH attack
(kintra) are listed in Table 1, together with the correspond-
ing khyd values (see eqs 1 and 2 in the Experimental
Section). For Tp-OPh(p-NO2), Tp-OPh(o-Cl), and Tp-ST,
the transesterification is faster than (or comparable with)
the phosphodiester hydrolysis. In Tp-OCH2CF3, however,
the transesterification is slower than the hydrolysis. With
Tp-OT, the transesterification was not perceived at all
and only the hydrolysis occurred.24 Although the kintra
decreases with increasing pKa,lg as does the khyd, its
F igu r e 5. HPLC profile in the gradient elution for the
cleavage of Tp-OPh(o-Cl) in 1 M NaOH (aq) at 80 °C and t )
6 h. By the intramolecular attack of the 5′-OH group, cTMP
is formed. Tp is generated by the hydrolyses of both Tp-OPh-
(o-Cl) and cTMP.24 The gradient conditions are presented in
the Experimental Section.
dependence is far more drastic than that of khyd
.
(2) Estim a tion of th e Dista n ce betw een th e Nu -
cleop h ile a n d th e Electr op h ile for th e In tr a m olecu -
la r Tr a n sester ifica tion . The minimal distance between
the 5′-O atom and the P-atom in Tp-OT and its analogues
was evaluated by using HFMP (in Figure 2b) as a model
compound. First, its structure was fully optimized by the
quantum-chemical calculation. Then the C3(furanose)-
OP bond and the C4(furanose)-CH2OH bond were ro-
tated, keeping other structural parameters unchanged.
It has been found that the minimal distance between
these two atoms is 2.524 Å. To shorten this distance
further, enormous energy is required.
(3) Releva n ce betw een th e In tr a m olecu la r Tr a n s-
ester ifica tion a n d th e P h osp h od iester Hyd r olysis.
According to the calculation on the alkaline hydrolysis
of DMP-O (Figure 4), the distance between the attacking
OH- and the P atom in the transition state is 2.125 Å. It
is reasonable to assume that the corresponding value in
the transesterification of Tp-OT should be equal to (or
smaller than) this value, since the 5′-alkoxide ion is less
nucleophilic than is hydroxide ion.14 However, the mini-
mal distance between the 5′-O atom and the P atom in
Tp-OT is 2.524 Å, as estimated above. Thus, the trans-
esterification cannot proceed to a measurable extent. In
Tp-ST hydrolysis, however, the distance between the
attacking OH- and the P in the transition state (2.950
Å) is greater than the minimal distance in the reactant.
Accordingly, the transesterification efficiently takes place.
The intramolecular reactions are also evident in Tp-OPh-
(p-NO2) and Tp-OPh(o-Cl),25 whereas it is inefficient in
Tp-OCH2CF3. The kinetic analysis and the theoretical
study are completely consistent with each other.
Ta ble 2. Ch a r ge of Ea ch Atom of th e P -O(ester ) Bon d
in DMP -O a n d of th e P -S(th ioester ) Bon d in DMP -S
charge
substrate
P
O or S
DMP-O
DMP-S
2.58
2.28
-0.89
-0.25
some extent. According to the natural population analy-
sis, the extent of the formation of the P-OH(axial) bond
in the transition state (R) is 20.3%, which is considerably
smaller than the value (56.5%) in the DMP-O hydrolysis.
The leaving group (CH3S-) is removed there in a greater
extent (â ) 66.4%). The P-OH(axial) bond and the
P-SCH3(axial) bond have similar imaginary frequencies
(0.74001 and 0.53765, respectively), and both bonds are
weak in the transition state. In the DMP-O hydrolysis,
however, the OH- attack must be more complete in order
to remove the poorer leaving group (CH3O-) from the P
atom. Thus, the hydrolysis rate of the Tp-OT derivatives
monotonically increases with decreasing pKa,lg (Figure 3).
The acceleration of hydrolysis on the OfS substitution
is never ascribed to the change in the electrophilicity of
P atom. As shown in Table 2, the P atom of DMP-S (the
charge ) 2.28) is less positive than that of DMP-O (2.58).
The Pauling’s electronegativity of S atom (2.5) is smaller
than that of O atom (3.5).23 Accordingly, DMP-S is less
susceptible to the nucleophilic attack by OH- ion than
DMP-O. The acidity of leaving group predominantly
governs the reactivity of alkaline hydrolysis.
In tr a m olecu la r Atta ck by th e 5′-OH in th e Acti-
va ted Tp -OT An a logu es for th e Tr a n sester ifica tion .
When the Tp-OT analogues having good leaving groups
were incubated at pH 14, significant amounts of 3′,5′-
cyclic monophosphate of thymidine (cTMP) were gener-
ated (together with the hydrolysis products). The typical
HPLC profile is presented in Figure 5. Quite interest-
ingly, Tp-OT itself did not produce cTMP at all under
the same conditions. Only in these activated phospho-
esters, the intramolecular attack by the 5′-OH to the P
atom (path 1 in Figure 6) occurred concurrently with
alkaline hydrolysis of the phosphodiester linkage (path
2). Important information on the mechanism of DNA
hydrolysis is obtained by analyzing this transesterifica-
tion, since the reaction mechanism is similar (nucleo-
philic attack by hydroxide ion vs that by alkoxide ion).
Con clu sion s
By combining the kinetic studies and the theoretical
studies, the mechanism of alkaline hydrolysis of DNA has
(24) Considering the detection limit for cTMP, the kintra value of Tp-
OT is at least smaller than 3 × 10-4 h-1. In this reaction, Tp and pT
were formed in comparable amounts, and no cTMP was perceived. This
shows that the intramolecular transesterification by the 5′-OH is not
occurring. Otherwise, significant amount of cTMP should be ac-
cumulated, and, furthermore, Tp should be predominantly produced
(according to control experiments, cTMP is mostly hydrolyzed to Tp
under alkaline conditions).
(25) As reported previously (Borden, R. K.; Smith, M. J . Org. Chem.
1966, 31, 3247-3253), cTMP was also generated by the 3′-OH attack,
when p-nitrophenyl ester of thymidine 5′-monophosphate was incu-
bated at pH 14. The rate constant k3′-OH ) 0.5 × 10-1 h-1. This value
was significantly smaller than the kintra of Tp-OPh(p-NO2) (4.8 × 10-1
h-1), although the khyd’s of both substrates were almost the same. The
mutual conformation between the nucleophile and the target P atom
is overwhelmingly important for the intramolecular transesterification.
(23) Kagaku-binran Kiso-hen; The Chemical Society of J apan, Ed.;
Maruzen: Tokyo, 1984; II-589.

Mechanism of Alkaline Hydrolysis of DNA
J . Org. Chem., Vol. 65, No. 14, 2000 4395
F igu r e 6. Two parallel reactions in the cleavage of Tp-X having good leaving groups. Path 1: the intramolecular transesterification
by the 5′-OH. Path 2: alkaline hydrolysis of the phosphodiester linkage.
ture for 3 h.27 After dithiothreitol (218 mg, 1.4 mmol) in MeOH
been clarified. It proceeds in one step without any
(4 mL) was added,27 the turbid solution was vigorously stirred
for 1 h. The precipitate was filtered off, and the solvent was
evaporated. The silica gel column chromatography with diethyl
ether provided 3′-O-benzoyl-5′-mercapto-5′-deoxythymidine, Rf
) 0.25 (diethyl ether).
intermediate. In the transition state, the attack by OH-
ion as a nucleophile toward the P atom is only partially
complete, whereas the leaving group (alkoxide ion) is still
bound to the P atom rather tightly. The lengths of the
corresponding P-O bonds are 2.125 and 1.795 Å, respec-
tively. Consistently, the slope of the Brønsted-type plot
(the logarithm of hydrolysis rate vs the pKa of leaving
group) is small (-0.27). All these analyses indicate that
efficient acid catalysts, which facilitate the removal of
the leaving group from the P atom, are the primary
requisites for prompt DNA hydrolysis under alkaline
conditions. The studies on the mechanism of spontaneous
and metal-catalyzed DNA hydrolysis under neutral con-
ditions are currently under way in our laboratory.
In dry acetonitrile (2 mL), the 3′-O-benzoyl-5′-mercapto-5′-
deoxythymidine (49.1 mg, 0.14 mmol) was treated with 5′-O-
dimethoxytritylthymidine-3′-O-(2-cyanoethyl)-N,N′-diisopro-
pylphosphoroamidite (the phosphoramidite reagent for thy-
midine from PerSeptive; 203 mg, 0.27 mmol) and 1H-tetrazole
(38.1 mg, 0.54 mmol) under nitrogen for 20 min. Then, 2,6-
lutidine (73.6 mg, 0.68 mmol) and n-tetrabutylammonium
periodate in CH2Cl2 (2 mL) were added.28 The fully protected
Tp-ST was purified by the silica gel column chromatography,
Rf ) 0.44 (5/95 EtOH/AcOEt). The protecting residues were
removed by the mixture of NH4OH and MeOH (3:1) at 50 °C
for 14 h and then by the deblocking solution (2% CHCl2COOH
in CH2Cl2) for 30 min. The final product Tp-ST was extracted
into water, washed by CH2Cl2 three times, and then purified
by the reversed-phase HPLC (95/5 H2O/acetonitrile mixture
containing 50 mM ammonium formate, tR ) 61.0 min): 1H
NMR (D2O) δ 7.61 (s, 1H), 7.54 (s, 1H), 6.25 (t, J ) 6.6 Hz,
1H), 6.22 (t, J ) 6.9 Hz, 1H), 4.85 (m, 1H), 4.44 (m, 1H), 4.20
Exp er im en ta l Section
Ma t er ia ls. The activated Tp-OT analogues (Tp-ST,
Tp-OCH2CF3, and Tp-OPh(o-Cl)) and Tp-(S)T were synthesized
as described below. Pyridine was dried by KOH. Dry acetoni-
trile (graded for DNA synthesis) was purchased from Dojindo.
The characterization of the products and reaction inter-
mediates was made by NMR spectroscopy (1H NMR at 270
MHz, 31P NMR at 109.4 MHz), mass spectroscopy (TOF-MS,
ESI-MS), and reversed-phase gradient HPLC equipped with
an ODS column (Merck, LiChroCART 250-4, LiChrospher 100
RP-18(e)). For the TLC analysis, silica gel 60 F254 (Merck) was
used. Tp-OT, Tp-OPh(p-NO2), and cTMP were purchased from
Sigma and used without further purification. Solvent water
was purified by a Milli-Q Labo system (Millipore) and auto-
claved. All the implements were sterilized. Great care was
taken to avoid contamination by natural nucleases throughout
the experiments.
(1) Syn th esis of Tp -ST. In dry pyridine (5 mL), 5′-(S-
trityl)mercapto-5′-deoxythymidine (370 mg, 0.74 mmol), syn-
thesized as reported previously,26 was dissolved. To this
solution was added benzoyl chloride (155 mg, 1.1 mmol), and
the mixture was stirred at room temperature under nitrogen
for 20 h. After the reaction was quenched by water on an ice
bath, the product was purified by the silica gel column
chromatography (the eluent was gradually changed from
CH2Cl2 to 5/95 EtOH/CH2Cl2), Rf ) 0.27 (5/95 EtOH/CH2Cl2).
The 3′-O-benzoyl-5′-(S-trityl)mercapto-5′-deoxythymidine ob-
tained (280 mg, 0.47 mmol) was detritylated by AgNO3 (160
mg, 0.94 mmol) and pyridine (86.5 mg, 1.1 mmol) in the 1:1
mixture of acetonitrile and MeOH (40 mL) at room tempera-
2
3
(q, J ) 3.6 Hz, 1H), 4.11 (m, 1H), 3.81 (q, J ) 12.4 Hz, J )
3.6 Hz, 1H), 3.75 (q, 2J ) 12.5 Hz, 3J ) 4.6 Hz, 1H), 2.95-
3.21 (m, 2H), 2.60 (oc, 1H), 2.31-2.44 (m, 3H), 1.86 (s, 6H);
31P NMR (D2O) δ 22.96; ESI-MS calcd for C20H26N4O11PS [(M
- H)-] 561.5, found 561.4.
(2) Tp -OCH2CF 3. 2,2,2-Trifluoroethanol (15.0 mg, 0.15
mmol) was reacted in dry acetonitrile (6 mL) with the
phosphoroamidite reagent for thymidine (225 mg, 0.30 mmol)
and 1H-tetrazole (42.0 mg, 0.60 mmol). After being kept at
room temperature under nitrogen for 4 h, the mixture was
treated with a 0.1 M I2 solution of THF/2,6-lutidine/H2O (4/
1/0.1) for 30 min. The fully protected Tp-OCH2CF3 was purified
by the silica gel column chromatography (Rf ) 0.83 (2/1 AcOEt/
EtOH)) and treated with the deblocking solution (4 mL) and
then with NH4OH (20 mL). The product was extracted into
water and washed with CH2Cl2 four times. The final purifica-
tion of Tp-OCH2CF3 was achieved by the reversed-phase HPLC
(90/10 H2O/acetonitrile mixture containing 50 mM ammonium
formate, tR ) 28.9 min): 1H NMR (D2O) δ 7.64 (s, 1H), 6.31 (t,
J ) 6.9 Hz, 1H), 4.81 (m, 1H), 4.32 (quin, J ) 8.6 Hz, 2H),
2
3
4.20 (q, J ) 3.3 Hz, 1H), 3.85 (q, J ) 12.5 Hz, J ) 3.6 Hz,
1H), 3.78 (q, 2J ) 12.5 Hz, 3J ) 4.9 Hz, 1H), 2.55 (oc, 1H),
2.44 (quin, 1H), 1.89 (s, 3H); 31P NMR (D2O) δ 2.55; TOF-MS
calcd for C12H15F3N2O8P [(M - H)-] 403.2, found 403.6.
(26) Mag, M.; Lu¨king, S.; Engels, J . W. Nucleic Acids Res. 1991,
19, 1437-1441.
(27) Connolly, B.; Rider, P. Nucleic Acids Res. 1985, 13, 4485-4502.
(28) Cosstick, R.; Vyle, J . Nucleic Acids Res. 1990, 18, 829-835.

4396 J . Org. Chem., Vol. 65, No. 14, 2000
Takeda et al.
rate constants presented here are the averages of the results
of two or three runs.
Sch em e 1
(1) Deter m in a tion of kh yd a n d kin tr a in Sch em e 1. In the
cleavage of the activated Tp-OT analogues (Tp-X), both the
hydrolysis of the phosphodiester linkage and the intramolecu-
lar transesterification by the 5′-OH (formation of cTMP)
concurrently proceed, as shown in Scheme 1 (see also Figures
5 and 6). The rate constant of the disappearance of Tp-X (kobs
)
is the sum of the rate constants (khyd and kintra) of these two
reactions (eq 1).
(3) Tp -OP h (o-Cl). In the deblocking solution (15 mL), 5′-
O-(4,4′-dimethoxytrityl)thymidine 3′-(2-chlorophenyl)phos-
phate (100 mg, 0.12 mmol; from Sigma) was incubated at room
temperature for 15 min. The product was purified by the
reversed-phase HPLC (H2O/acetonitrile mixture containing 50
mM ammonium formate, linear gradient from 96/4 to 60/40
at t ) 40 min, tR ) 29.1 min): 1H NMR (D2O) δ 7.62 (s, 1H),
7.49 (d, J ) 7.6 Hz, 1H), 7.37 (d, J ) 7.9 Hz, 1H), 7.31 (t, J )
7.3 Hz, 1H), 7.15 (t, J ) 7.3 Hz, 1H), 6.28 (t, J ) 7.3 Hz, 1H),
kobs ) khyd + kintra
(1)
On the other hand, kintra can be expressed by eq 2.
kintra ) [[cTMP]t(kcTMP - kobs)] × 1/[[Tp-X]0[exp(-kobst) -
exp(-kcTMPt)]] (2)
Here, [cTMP]t is the concentration of cTMP, which is
accumulated in the reaction mixture at the reaction time t.
The rate constant kcTMP for alkaline hydrolysis of cTMP was
determined independently by using its authentic sample. By
fitting the kobs and [cTMP]t values to eqs 1 and 2, khyd and kintra
were determined.
Qu a n t u m -Ch em ica l Ca lcu la t ion s. For the calculation,
the Gaussian 98 program was used.30 The geometries of all
the substrates, intermediates, transition states, and products
were fully optimized in the gas phase at the RHF/6-31+G*
level with the frequency calculations. The Gibbs free energies
were obtained by the density-functional-theory calculation
with B3LYP/6-31+G* level. The solvation effects were incor-
porated by PCM under the recommended conditions.31 The
natural population analysis was conducted by the NBO
Version 3.1.32
2
4.90 (sep, 1H), 4.18 (q, J ) 3.6 Hz, 1H), 3.78 (q, J ) 12.5 Hz,
2
3
3J ) 3.3 Hz, 1H), 3.71 (q, J ) 12.5 Hz, J ) 4.9 Hz, 1H), 2.52
(oc, 1H), 2.37 (quin, 1H), 1.86 (s, 3H).
(4) Tp -(S)T. To 3′-O-benzoylthymidine (17.3 mg, 0.050
mmol; from Sigma) in dry acetonitrile (2 mL) were added the
phosphoroamidite reagent for thymidine (74.5 mg, 0.10 mmol)
and 1H-tetrazole (14.0 mg, 0.20 mmol) in dry acetonitrile
(1 mL each). The reaction mixture was stirred at room
temperature under nitrogen for 1.5 h. Then, dry acetonitrile
solution (2 mL) of tetraethylthiuram disulfide (594 mg, 2.0
mmol) was poured.29 The mixture was incubated under
nitrogen for 30 min and subjected to the silica gel column
chromatography. The product (Rf ) 0.29 (1/2 AcOEt/CH2Cl2))
was treated with a 3:1 mixture of NH4OH and MeOH at 60
°C overnight and then with the deblocking solution for 15 min.
The Tp-(S)T was purified by the reversed-phase HPLC (94/6
H2O/acetonitrile mixture containing 50 mM ammonium for-
mate, tR ) 77.6 min): 1H NMR (D2O) δ 7.70 (s, 1H), 7.64 (s,
1H), 6.29 (t, J ) 6.6 Hz, 1H), 6.18 (t, J ) 6.6 Hz, 1H), 4.94 (m,
1H), 4.56 (m, 1H), 4.08-4.21 (m, 4H), 3.84 (q, 2J ) 12.5 Hz, 3J
) 3.3 Hz, 1H), 3.77 (q, 2J ) 12.5 Hz, 3J ) 4.6 Hz, 1H), 2.53
(oc, 1H), 2.28-2.38 (m, 3H), 1.90 (s, 3H), 1.85 (s, 3H); 31P NMR
(D2O) δ 60.1; ESI-MS calcd for C20H26N4O11PS [(M - H)-]
561.5, found 561.
Ack n ow led gm en t. The authors should like to thank
Prof. Kimitsuna Watanabe and Dr. Tsutomu Suzuki for
kind assistance with the ESI-MS measurements. This
work was supported by a Grant-in-Aid for Scientific
Research from the Ministry of Education, Science,
Sports and Culture, J apan, and J SPS Research Fellow-
ships for Young Scientists (for N.T.).
Kin etic An a lyses. Small aliquots of reaction mixtures were
collected at appropriate intervals and injected to the reversed-
phase HPLC after neutralization by 1 M HCl. All the products
were characterized by co-injection with authentic samples
under two different eluting conditions: the isocratic system
(8/92 acetonitrile/water) and the linear gradient system (0/100
at t ) 0 min f 50/50 at t ) 30 min). The ꢀ260 values of T-base
in all the substrates and the products were determined by
using authentic samples. The ꢀ260’s of Tp-OPh(p-NO2) and
Tp-OPh(o-Cl) were the sum of the values of those T-bases and
the corresponding phenyl derivatives. With these treatments,
the mass balance of each kinetic runs were satisfactory. In
addition, ESI-MS and/or LC-MS were also utilized for the
characterization of the products. The initial concentration of
the substrate was fixed at 0.1 mM. All the pseudo-first-order
J O000323D
(30) Gaussian 98 (Revision A.7): Frisch, M. J .; Trucks, G. W.;
Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J . R.;
Zakrzewski, V. G.; Montgomery, J . A.; Stratmann, R. E.; Burant, J .
C.; Dapprich, S.; Millam, J . M.; Daniels, A. D.; Kudin, K. N.; Strain,
M. C.; Farkas, O.; Tomasi, J .; Barone, V.; Cossi, M.; Cammi, R.;
Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J .;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.; Cioslowski, J .; Ortiz,
J . V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Keith, T.;
Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.;
Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen, W.; Wong, M.
W.; Andres, J . L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople,
J . A. Gaussian, Inc., Pittsburgh, PA, 1998.
(31) Solvent, water; Model, PCM/UAHF; Icomp ) 4; Version, Matrix
inversion; Cavity, Pentakisdodecahedra with 60 initial tesserae.
Barone, V.; Cossi, M.; Tomasi, J . Chem. Phys. 1997, 107, 3210-3221.
(32) NBO Version 3.1, Glendening, E. D.; Reed, A. E.; Carpenter,
J . E.; Weinhold, F.
(29) Zon, G.; Stec, W. J . In Oligonucleotides and Analogues, A
Practical Approach; Eckstein, F., Ed.; Oxford University Press: New
York, 1991; pp 87-108.