Reactivity of a 2'-thio nucleotide analog

Article abstract of DOI:10.1021/ja962265c

The chemical reactivity of ribonucleotide analog 2'-deoxy-2'-thiouridine 3'-(p-nitrophenyl phosphate) (1), in which the 2' hydroxyl is replaced with a 2'-thiol group, has been characterized. The major reaction pathway for 1, as monitored by ³¹P NMR spectroscopy, is transphosphorylation to afford 2',3'-cyclic phosphorothioate 3, followed by hydrolysis of 3 to produce 2'-deoxy-2'-thiouridine 2'-phosphorothioate (4). Thus, the reaction pathway of 1 is similar to that of the hydrolysis of ribonucleotides, yet there are significant differences. The pH-rate profile for transphosphorylation of 1 was determined by monitoring the formation of p-nitrophenol or p-nitrophenolate by UV-visible spectroscopy. Analysis of the profile reveals the attacking nucleophile to be thiolate, and the pK(a) of the 2'-thiol was determined to be 8.3 ± 0.1. At pH 7.4, the thiol-containing ribonucleotide analog 1 is hydrolyzed at an observed rate 27-fold slower than its 2'-hydroxyl counterpart. These results indicate that the rate of thiolate attack on the adjacent phosphodiester bond is 10⁷-fold slower than that of the corresponding alkoxide. Thiolate nucleophiles, therefore, are remarkably reticent toward attack at electrophilic phosphate centers. In addition to providing new information about the reactivity of phosphodiester bonds, our studies highlight the potential of 2'-thiol-containing nucleotides for the study of an array of RNA processes, especially those in which the 2'-substituent plays a critical role.

Full text of DOI:10.1021/ja962265c

J. Am. Chem. Soc. 1996, 118, 11715-11719
11715
Reactivity of a 2′-Thio Nucleotide Analog
Cathy L. Dantzman and Laura L. Kiessling*
Contribution from the Department of Chemistry, UniVersity of WisconsinsMadison,
Madison, Wisconsin 53706
ReceiVed July 3, 1996X
Abstract: The chemical reactivity of ribonucleotide analog 2′-deoxy-2′-thiouridine 3′-(p-nitrophenyl phosphate) (1),
in which the 2′ hydroxyl is replaced with a 2′-thiol group, has been characterized. The major reaction pathway for
1, as monitored by 31P NMR spectroscopy, is transphosphorylation to afford 2′,3′-cyclic phosphorothioate 3, followed
by hydrolysis of 3 to produce 2′-deoxy-2′-thiouridine 2′-phosphorothioate (4). Thus, the reaction pathway of 1 is
similar to that of the hydrolysis of ribonucleotides, yet there are significant differences. The pH-rate profile for
transphosphorylation of 1 was determined by monitoring the formation of p-nitrophenol or p-nitrophenolate by UV-
visible spectroscopy. Analysis of the profile reveals the attacking nucleophile to be thiolate, and the pKa of the
2′-thiol was determined to be 8.3 ( 0.1. At pH 7.4, the thiol-containing ribonucleotide analog 1 is hydrolyzed at an
observed rate 27-fold slower than its 2′-hydroxyl counterpart. These results indicate that the rate of thiolate attack
on the adjacent phosphodiester bond is 107-fold slower than that of the corresponding alkoxide. Thiolate nucleophiles,
therefore, are remarkably reticent toward attack at electrophilic phosphate centers. In addition to providing new
information about the reactivity of phosphodiester bonds, our studies highlight the potential of 2′-thiol-containing
nucleotides for the study of an array of RNA processes, especially those in which the 2′-substituent plays a critical
role.
Introduction
acids. In analogy to cysteine residues in proteins, nucleotides
containing thiols can participate in disulfide bond formation.
This property may be exploited to study RNA tertiary structure4
and RNA-protein interactions through the formation of covalent
cross-links. In addition, the introduction of such thiol groups
into a ribozyme could increase the diversity of functionality
available for catalysis.5 Another feature of sulfur substituents,
their ability to coordinate metal ions, could be exploited to
illuminate the roles of metals in nucleic acid catalysis6 or to
engineer new metal-dependent catalysts. For instance, the
double-metal-ion mechanism, proposed for some enzyme- and
ribozyme-mediated RNA hydrolysis reactions, can be explored
using sulfur substituents to localize metals at specific sites.7
The utility of 2′-thio-substituted nucleotides for such RNA
structure-function studies is contingent on their chemical
reactivity, about which little is known. To provide insight into
the reactivity of phosphodiesters adjacent to 2′-thiol groups, we
synthesized ribonucleotide 18 (Figure 1) and determined the rates
of its transphosphorylation over a broad pH range. Previous
reports suggested that thiols in the 2′-position of nucleotide
analogs would react not through a transphosphorylation pathway
Impetus for studying RNA analogs has arisen from many
directions. Investigations of the reactivities of these analogs
have been propelled by a desire to understand the mechanisms
by which enzymes and ribozymes facilitate RNA cleavage.1 The
possibility of using RNA derivatives as therapeutic agents has
further fueled investigations into their properties.2 Because of
the critical role of the 2′-hydroxyl group in RNA degradation,
nucleoside analogs with modifications at this site have been
used to provide insight into the mechanism of hydrolysis and
to increase the stability of nucleic acid-derived therapeutics.1,2
One functional group substitution that could be used to
investigate RNA chemistry and biology is the replacement of
the 2′-hydroxyl with a 2′-thiol.3 Because of the unique reactivity
and metal binding properties of thiols, this functional group
could be used to endow nucleic acids with tailored properties.
2′-Thiol groups in oligonucleotides could play significant
roles in determining structure and promoting catalysis of nucleic
X Abstract published in AdVance ACS Abstracts, November 1, 1996.
(1) For leading references describing the application of analogs with 2′
modifications to probe reactions of RNA, see the following reviews: (a)
Usman, N.; Cedergren, R. Trends Biochem. Sci. 1992, 17, 334-339. (b)
Heidenreich, O.; Pieken, W.; Eckstein, F. FASEB J. 1993, 7, 90-96. (c)
Grasby, J. A.; Pritchard, C.; Gait, M. J. Proc. Indian Acad. Sci. (Chem.
Sci.) 1994, 106, 1003-1022. For recent specific examples, see: (d) Moore,
M. J.; Sharp, P. A. Science 1992, 256, 992-997. (e) Herschlag, D.; Eckstein,
F.; Cech, T. R. Biochemistry 1993, 32, 8312-8321. (f) Heidenreich, O.;
Benseler, F.; Fahrenholz, A.; Eckstein, F. J. Biol. Chem. 1994, 269, 2131-
2138. (g) Beigelman, L.; McSwiggen, J. A.; Draper, K. G.; Gonzalez, C.;
Jensen, K.; Karpeisky, A. M.; Modak, A. S.; Matulic-Adamic, J.; DiRenzo,
A. B.; Haeberli, P.; Sweedler, D.; Tracz, D.; Grimm, S.; Wincott, F. E.;
Thackray, V. G.; Usman, N. J. Biol. Chem. 1995, 270, 25702-25708.
(2) For recent reviews, see: (a) Antisense Research and Applications;
Crooke, S. T., Lebleu, B., Eds.; CRC Press, Inc.: Boca Raton, FL, 1993.
(b) Milligan, J. F.; Matteucci, M. D.; Martin, J. C. J. Med. Chem. 1993,
36, 1923-1937. (c) Wagner, R. W. Nature 1994, 372, 333-335. (d)
Sanghvi, Y. S.; Cook, P. D. Carbohydrates: Synthetic Methods and
Applications in Antisense Therapeutics: An OVerView; Sanghvi, Y. S., Cook,
P. D., Ed.; American Chemical Society: Washington, DC, 1994; Vol. 580,
pp 1-22. (e) Eaton, B. E.; Pieken, W. A. Annu. ReV. Biochem. 1995, 64,
837-863. (f) De Mesmaeker, A.; Ha¨ner, R.; Martin, P.; Moser, H. E. Acc.
Chem. Res. 1995, 28, 366-374.
(3) For syntheses of 2′-thio nucleosides, see: (a) Imazawa, M.; Ueda,
T. Tetrahedron Lett. 1970, 55, 4807-4810. (b) Imazawa, M.; Ueda, T.;
Ukita, T. Chem. Pharm. Bull. 1975, 23, 604-610. (c) Lee, C. C.; Schrier,
W. H.; Nagyvary, J. Biochim. Biophys. Acta 1979, 561, 223-231. (d) Patel,
A. D.; Schrier, W. H.; Nagyvary, J. J. Org. Chem. 1980, 45, 4830-4834.
(e) Divakar, K. J.; Mottoh, A.; Reese, C. B.; Sanghvi, Y. S. J. Chem. Soc.,
Perkin Trans. 1 1990, 969-974. (f) Marriott, J. H.; Mottahedeh, M.; Reese,
C. B. Carbohydr. Res. 1991, 216, 257-269. For synthesis of a 2′-thio
dinucleotide, see: (g) Reese, C. B.; Simons, C.; Pei-Zhuo, Z. J. Chem.
Soc., Chem. Commun. 1994, 1809-1810.
(4) (a) Goodwin, J. T.; Osborne, S. E.; Scholle, E. J.; Glick, G. D. J.
Am. Chem. Soc. 1996, 118, 5207-5215. (b) Allerson, C. R.; Verdine, G.
L. Chem. Biol. 1995, 2, 667-675. (c) Goodwin, J. T.; Glick, G. D.
Tetrahedron Lett. 1994, 35, 1647-1650.
(5) Jencks, W. P. Nature 1992, 358, 543-544.
(6) Piccirilli, J. A.; Vyle, J. S.; Caruthers, M. H.; Cech, T. R. Nature
1993, 361, 85-88.
(7) (a) Steitz, T. A.; Steitz, J. A. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
6498-6502. (b) Sawata, S.; Komiyama, M.; Taira, K. J. Am. Chem. Soc.
1995, 117, 2357-2358.
(8) The synthesis of 1 will be published elsewhere.
S0002-7863(96)02265-2 CCC: $12.00 © 1996 American Chemical Society

11716 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Dantzman and Kiessling
Figure 1. Scheme showing pathway for the hydrolysis of RNA analogs 1 and 5. Abbreviations are as follows: pNP ) p-nitrophenyl; Ura ) uracil.
but by thiol attack at the glycosidic linkage to displace the
nucleotide base.9 With a goal of examining the transphospho-
rylation reaction, we equipped our 2′-thiol ribonucleotide analog
with a p-nitrophenolate phosphodiester. The p-nitrophenol
group activates the phosphate center toward nucleophiles and
provides a convenient spectroscopic handle with which to follow
the transphosphorylation reaction. A comparison of the rates
of transphosphorylation of 1 with those obtained for the
corresponding 2′-hydroxyl ribonucleotide analog 510 provides
insight into the propensity of nucleophilic thiols to attack
electrophilic phosphodiesters.
foreshadows the reactivity differences of these nucleophiles
toward phosphodiesters. Under acidic conditions (pH ) 2.3)
the 2′-thiol derivative 1 is stable (no reaction was observed over
48 h), which contrasts with the reactivity of the 2′-hydroxyl
analog 5 at low pH.
Cyclic phosphorothioate 3 is relatively stable at pH 7, but its
hydrolysis to thiophosphate 4 under basic conditions could be
monitored by NMR spectroscopy. The selective formation of
the product 4 from the cyclic intermediate 3 contrasts with that
observed for the breakdown of 6, which yields both phosphate
7 and 8. The exclusive P-O bond cleavage observed in the
breakdown of intermediate 3 is precedented in studies of the
alcoholysis of other 5-membered ring containing phosphorothio-
ates. Unless reversible conditions are applied, only P-O bond
scission occurs in such systems.15 The susceptibility of the P-O
over the P-S bond toward cleavage can be ascribed to the
stereoelectronic features of the trigonal bipyramidal phospho-
ranes, which are formed during the course of the hydrolysis
reactions (Figure 2).16 The more apicophilic oxygen atom
prefers an apical position in the transition state (or intermediate),
as shown in structure I. Under these conditions, species I breaks
down to generate only 4; none of the 3′-phosphomonoester
product that would arise through the higher energy species II
is detected. Hydrolysis of 6 affords a mixture of 2′- and 3′-
uridine monophosphates (7 and 8) since the trigonal bipyramidal
phosphorane transition states formed are of similar energy.
pH-Rate Profile for the Reactions of 1 and 5. The 31P
NMR studies reveal that the predominant pathway for reaction
of 1 is through transphosphorylation followed by hydrolysis of
the cyclic phosphorothioate 3 (Figure 1) and not through attack
Results and Discussion
Reactivity of 1 As Determined by 31P NMR. To ascertain
the course of the hydrolysis of 1, the products of its reaction in
aqueous solution were monitored at high and low pH by 31P
NMR spectroscopy.11 Under strongly basic conditions (pH )
13), thiol 1 reacts to generate a mixture of 2′-deoxy-2′-
thiouridine 2′-phosphorothioate (4) (65% by 31P NMR), which
is presumably derived from P-O scission of intermediate 3,12
and 2′-deoxy-2′-(p-nitrophenylthio)uridine 3′-phosphate, 2 (35%
by 31P NMR), the product of nucleophilic aromatic substitu-
tion.13 The reaction of thiol 1 at pH values between 7 and 10
progressed to afford the 2′,3′-cyclic phosphorothioate 314 as the
major product, with some 2 generated by the competing reaction.
The formation of product 2 reflects the high nucleophilicity of
thiolates relative to alkoxides toward carbon centers and
(9) (a) Ryan, K.; Acton, E.; Goodman, L. J. Org. Chem. 1971, 36, 2646-
2657. (b) Johnson, R.; Reese, C. B.; Pei-Zhuo, Z. Tetrahedron 1995, 51,
5093. See also refs 3a, 3b, and 3g.
(10) Davis, A. M.; Hall, A. D.; Williams, A. J. Am. Chem. Soc. 1988,
110, 5105-5108.
(11) (a) Thompson, J. E.; Venegas, F. D.; Raines, R. T. Biochemistry
1994, 33, 7408-7414. (b) Cozzone, P. J.; Jardetzky, O. FEBS Lett. 1977,
73, 77-79. (c) Weiner, L. M.; Backer, J. M.; Rezvukhin, A. I. FEBS Lett.
1974, 41, 40-42.
(15) (a) It was also observed that hydrolysis of O,S-diethyl phospho-
rothioate occurs 25 times faster than ethylene cyclic phosphate. Gay, D.
C.; Hamer, N. K. J. Chem. Soc. B 1970, 1123-1127. (b) Kudelska, W.;
Michalska, M. Tetrahedron 1986, 42, 629-636.
(16) For reviews concerning mechanisms of phosphodiester hydrolysis,
see: (a) Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70-9. (b) Benkovic,
S. J.; Schray, K. J. In The Enzymes; Boyer, P. D., Ed.; Academic Press,
Inc.: New York, 1973; Vol. 8, pp 201-238. (c) Hall, C. R.; Inch, T. D.
Tetrahedron 1980, 36, 2059-2095. (d) Thatcher, G. R. J.; Kluger, R. In
AdVances in Physical Organic Chemistry; Bethell, D., Ed.; Academic Press,
Inc.: New York, 1989; Vol. 25, pp 99-266. (e) Breslow, R.; Anslyn, E.;
Huang, D.-L. Tetrahedron 1991, 47, 2365-2376. (f) Gerlt, J. A. In The
Enzymes; Academic Press, Inc.: New York, 1992; Vol. XX, pp 95-139.
(g) Cleland, W. W.; Hengge, A. C. FASEB 1995, 9, 1585-1594. (h)
Perreault, D. M.; Anslyn, E. V. Angew. Chem., Int. Ed. Engl., in press.
(12) The 31P NMR chemical shift of 4 is 15 ppm, which is similar to
previously reported values for S-alkyl phosphorothioates, see: (a) Meade,
T. J.; Iyengar, R.; Frey, P. A. J. Org. Chem. 1985, 50, 936-940. (b) Mu¨ller,
C. E.; Roth, H. J. Tetrahedron Lett. 1990, 31, 501-502.
(13) The 31P NMR chemical shift of 2 is 4 ppm at high pH. For additional
spectroscopic data on this compound, see the Experimental Section.
(14) The 31P NMR chemical shift of 3 is 37 ppm, which is characteristic
of oxathiaphospholanes; see: Nuretdinova, O. N.; Guseva, F. F.; Arbuzov,
B. A. IzV. Akad. Nauk. SSSR, Ser. Khim 1976, 2625-2627.

ReactiVity of a 2′-Thio Nucleotide Analog
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11717
(7.4) the 2′-thiol nucleotide undergoes transphosphorylation at
an observed rate 27-fold slower than that of 5. Given the
increased concentration of the reactive nucleophilic thiolate in
1 relative to that of the corresponding alkoxide in 5 under these
conditions, this 27-fold observed difference indicates the rate
of thiolate attack on the adjacent phosphodiester bond is a
striking 107-fold slower than that of the corresponding alkoxide.
Thus, thiolate groups are remarkably reticent nucleophiles when
the target is an electrophilic phosphate center.9,21
Figure 2. Possible trigonal bipyramidal phosphorane transition states
formed during the transphosphorylation and hydrolysis reactions.
The decreased propensity of thiolate relative to alkoxide attack
on the phosphodiester bond may be attributed to stereoelectronic
features of the trigonal bipyramidal phosphoranes, which are
formed during the course of the transphosphorylation reactions.
For 1 to undergo transphosphorylation, the incoming thiolate
nucleophile must attack the adjacent phosphodiester from an
apical position (structure II, Figure 2). Sulfur, being less
apicophilic than oxygen, prefers the equatorial position (structure
I, Figure 2). The necessity for sulfur to occupy an apical
position during attack destabilizes the transition state for the
transphosphorylation of 1.15,22 In addition, the incipient P-O
bond formed in the transition state for reaction of 5 is expected
to be stronger than the corresponding P-S bond generated by
reaction of 1.23
Figure 3. pH-rate profile for the hydrolysis of 2′-thiol UpOC6H4-p-
NO2, 1 (b), and UpOC6H4-p-NO2, 5 (O), at 25 °C with [buffer] )
0.10 M and I ) 0.20 M was obtained by monitoring the formation of
p-nitrophenol or p-nitrophenolate (330 nm, pH < 7; 400 nm, pH g 7).
Another aspect of the reactivity of the 2′-thiol nucleotide 1
is its transesterification under acidic conditions, which deviates
from the behavior of a hydroxyl nucleotide, such as 5. Below
pH 5 compound 1 undergoes hydrolysis at a basal rate, which
is pH-independent.24 By analogy to mechanisms proposed for
2′-hydroxyl nucleotides, we hypothesize that the pH-independent
region under acidic conditions for 1 is due to the ionized thiol
attacking a neutral phosphodiester.25 Below pH 2, the rate of
transphosphorylation of hydroxyl derivative 5 exhibits a first-
order dependence on hydronium ion concentration attributed
to hydroxyl attack on the neutral phosphodiester group.18b In
this pH range, which is approaching the pKa of the phosphodi-
ester (ca. -0.36),26 the p-nitrophenyl group is most likely
departing as its anion from the neutral pentacoordinated
phosphorane intermediate. The increase in rate for 5 contrasts
with that for thiol analog 1, a result that suggests an equivalent
mechanism does not occur in the acidic region (pH <2). Our
data are consistent with the transphosphorylation occurring via
thiolate attacks; the sulfhydryl group has little tendency to attack
either an anionic or a neutral phosphodiester.
at the glycosidic linkage.17 Thus, the transphosphorylation of
1 is similar to that of a ribonucleotide, although there are
intriguing differences. To further characterize the reactions of
1, the pH-rate profile for its transphosphorylation was deter-
mined (Figure 3). By monitoring formation of p-nitrophenolate,
the rate for the formation of 3 could be dissected from other
processes. For comparison, the rate of transphosphorylation for
hydroxyl analog 5 under the same conditions was also deter-
mined.
From the pH-rate profile, parallels between the reaction of
1 and that of 5 under basic conditions are apparent. Specifically,
the transesterification of both nucleotides are first order with
respect to hydroxide ion concentration from pH 6 to 8.18
Moreover, the reaction rate for 1 becomes pH-independent above
pH 9; the rate of transphosphorylation of 2′-hydroxyl ribonucle-
otide analogs also reaches a maximum when the basicity of the
solution exceeds the pKa of the alkoxide nucleophile.19 Analysis
of the pH-rate profile for 1 affords a value of 8.3 ( 0.1 for
the pKa of 2′-thiol analog 1. An independent determination of
the pKa of the related compound 2′-deoxy-2′-thiouridylyl-
(3′f5′)adenosine, obtained by monitoring the chemical shift
Our quantification of the modest reactivity of thiolate
nucleophiles toward phosphodiesters illuminates previous find-
(21) For thiols reacting with phosphodiester monochloridates, see: (a)
Merckling, F. A.; Ru¨edi, P. Tetrahedron Lett. 1996, 37, 2217-2220. For
thiols reacting with phosphorochloridothioates, see: (b) Miller, B. J. Am.
Chem. Soc. 1962, 84, 403-409.
(22) Cleavage of 3′-thiouridylyl(3′f5′)uridine (pH 10, 50 °C) occurs at
a rate 200-fold faster than that of UpU. In the transition state for the former
substrate, the sulfur substituent can occupy an equatorial position; see: Liu,
X.; Reese, C. B. Tetrahedron Lett. 1996, 37, 925-928.
(23) For bond enthalpies for P-O and P-S, see: (a) Pedley, J. B.;
Marshall, E. M. J. Phys. Chem. Ref. Data 1983, 12, 967-1026. (b) Drowart,
J.; Myers, C. E.; Szwarc, R.; Auwera-Mahieu, A. V.; Uy, O. M. High Temp.
Sci. 1973, 5, 482-488.
(24) Alternatively, the hydrolysis of 1 between pH 0.8 and 5 may only
be seemingly pH-independent. Specifically, the observed rate of hydrolysis
under increasingly acidic conditions may be due to intermolecular attack
by water rather than by thiol- or thiolate-facilitated transphosphorylation,
1
of the H2′ uridine proton by H NMR as a function of pD,
afforded a similar pKa of 8.2 ( 0.1.20 These data indicate that
the transphosphorylation reactions of 1 and 5 occur by related
mechanisms, and they are consistent with thiolate serving as
the nucleophile in the transphosphorylation of 1.
A critical feature of thiol derivative 1 is its stability over the
entire pH range studied, relative to 5. Near physiological pH
(17) Only under very basic conditions (pH ) 14) did a significant amount
(30%) of glycosidic bond cleavage occur to afford uracil and p-nitrophenyl
phosphate.
(18) Others have also demostrated this first order dependence on
hydroxide for hydrolysis of 2′-hydroxyl compounds similar to 5; see: (a)
Usher, D. A.; Richardson, D. I., Jr.; Oakenfull, D. G. J. Am. Chem. Soc.
1970, 92, 4699-4712. (b) Oivanen, M.; Lo¨nnberg, H. Acta Chem. Scand.
1991, 45, 968-971.
(19) For an estimate of the pH-independent rate for transphosphorylation
of 5, see the Experimental Section and ref 18a.
(20) (a) The pKa of the 2′-hydroxyl of a ribonucleotide derivative is
similarly depressed (pKa of 13.9 obtained by stopped-flow kinetics)18a
relative to that of 2-propanol. A pKa value of 8.0 was obtained for the thiol
in 2′-deoxy-2′-thiocytidine 3′-phosphate by ultraviolet spectroscopy.3d
since the rate of hydrolysis of ethyl p-nitrophenyl phosphodiester at pH )
-7
0.8 at 25 °C with I ) 0.20 M was determined to be 6.3 × 10
s-1
.
(25) The pH-independent region observed for 2′-hydroxyl nucleotides
similar to 5 has been postulated to be the result of an ionized 2′-hydroxyl
attacking a neutral phosphodiester backbone; see: Kosonen, M.; Oivanen,
M.; Lo¨nnberg, H. J. Org. Chem. 1994, 59, 3704-3708.
(26) This pKa was obtained for 3,3-dimethylbutyl p-nitrophenyl phos-
phate. Hengge, A. C.; Cleland, W. W. J. Am. Chem. Soc. 1991, 113, 5835-
5841.

11718 J. Am. Chem. Soc., Vol. 118, No. 47, 1996
Dantzman and Kiessling
ings in which the reactions of 2′-thiol-substituted nucleotides
were examined. For example, Reese and co-workers did not
detect any products of transphosphorylation when 2′-deoxy-2′-
thiouridylyl(3′f5′)uridine was treated with acid or base; rather,
they found that the major products were the symmetrical
disulfide dimer and uracil.3g The nucleobase uracil is presum-
ably derived from reaction of the thiolate at the anomeric
position, which forms an episulfide that undergoes further
degradation to multiple products. The origin of the differences
in the observed products between this study and ours lies in
the structures of the phosphodiester substituents employed. With
the unactivated phosphodiester investigated previously, the rate
of transphosphorylation by a thiolate is expected to be even
more diminished than that of 1. p-Nitrophenylate serves as a
better leaving group than a primary alkoxide, and this substituent
renders the phosphodiester in 1 more electrophilic. Therefore,
different reaction pathways for 2′-thiol nucleotide substrates can
be promoted by changes in the electronic properties of the phos-
phodiester group. For instance, compounds such as 2′-deoxy-
2′-thiouridylyl(3′f5′)uridine may undergo transphosphorylation
in the presence of catalysts that can increase the electrophilicity
of the phosphodiester or stabilize the leaving group.
Despite our findings that thiolates exhibit limited nucleophi-
licity toward phosphodiesters, these species are used by biologi-
cal catalysts that hydrolyze phosphomonoesters. Protein ty-
rosine phosphatases (PTPases),27 a family of enzymes that
participates in signal transduction, use a cysteine-derived thiolate
as a nucleophile to attack the tyrosine phosphate moiety. In
these reactions, which proceed through thiophosphate intermedi-
ates, the catalytic cysteine residue is essential; it cannot be
replaced by serine. A distinguishing feature of the reactions
of the PTPases that require thiolate nucleophiles is that these
processes occur through dissociative mechanisms in which there
are metaphosphate-like transition states.28 In contrast, transes-
terification of phosphodiesters generally occurs through an
associative pathway.16g Interestingly, in the examples of
biological catalysts that use a thiolate to attack phosphorus, the
phosphorus species are electron deficient. Similarly, the
increased propensity for thiolate attack at an electrophilic
phosphorus center is reflected in the different results of our
studies and previous investigations.
structure or catalysis. Chemical investigations into the intrinsic
reactivity of ribonucleotide derivatives containing 2′-thiol groups
are essential prerequisites for the multiple applications of these
compounds.
Experimental Section
General Methods. 1H, 13C, and 31P NMR spectra were recorded
on a Bruker AM-300 spectrometer at 300, 75, and 121 MHz,
respectively. Chemical shifts are reported relative to external references
of sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4, 1,4-dioxane, or 85%
1
H3PO4 for H, 13C, and 31P, respectively, unless otherwise stated. 1H
NMR data are assumed to be first order with apparent doublets and
triplets reported as d and t, respectively. A Corning 320 pH meter
equipped with a standard calomel electrode was used for pH measure-
ments. Ultraviolet absorbance measurements were made on a Cary
Model 3 spectrophotometer equipped with a Cary temperature control-
ler. To prevent reaction due to contaminating ribonucleases, water used
in the preparation of buffers for 31P NMR and UV-visible kinetics
measurements was treated with diethyl pyrocarbonate then sterilized
with an autoclave. Buffers used were glycine (pH ) 2.3, 9.2-10.2),
formate (pH ) 3.3-4.0), acetate (pH ) 5.0-5.5), MES (pH ) 6.0-
6.5), HEPES (pH ) 7.0-7.6), Tris (pH ) 8.5), and triethanolamine
(pH ) 7.5-8.0), and the buffer pH was adjusted with hydrogen chloride
or sodium hydroxide. Reactions carried out at pH’s above 13 and below
1 were done in NaOH or HCl solutions with the pH of the solution
measured by a pH meter. Buffers were filtered through a 0.1 µm filter.
To prevent oxidation of thiol containing compounds to the disulfide,
solutions were deoxygenated by three successive freeze-pump-thaw
cycles, and hydrolysis reactions monitored by NMR were performed
under argon.
Spectral Data for 1, 2, and 2′-Deoxy-2′-Thiouridylyl (3′f5′)
Adenosine.8 2′-Deoxy-2′-thiouridine 3′-(p-nitrophenyl phosphate)
1
(1): H NMR (300 MHz, D2O) δ 8.29 (d, J ) 9.2 Hz, 2 H), 7.82 (d,
J ) 8.1 Hz, 1 H), 7.41 (d, J ) 8.8 Hz, 2 H), 6.03 (d, J ) 9.2 Hz, 1 H),
5.92 (d, J ) 8.1 Hz, 1 H), 4.78-4.76 (m, 1 H), 4.36-4.34 (m, 1 H),
3.78 (d, J ) 3.3 Hz, 2 H), 3.75-3.72 (m, 1H); 13C NMR (75 MHz,
D2O) δ 165.8, 157.1 (d, JCP ) 7.0 Hz), 151.9, 143.5, 141.4, 125.8,
120.5 (d, JCP ) 5.1 Hz), 103.0, 89.8, 85.1, 77.5 (d, JCP ) 5.1 Hz),
61.1, 44.0 (d, JCP ) 7.0 Hz); 31P NMR (121 MHz, D2O) δ -5.57 (d,
J ) 6.9 Hz); ultraviolet absorption λmax 266 nm (ꢀ 13 000) at pH )
4.9; mass spectrum (LSIMS, 3-NBA, negative ion mode) m/z 460.1
[M-, calcd for C15H15N3O10PS 460.1].
2′-Deoxy-2′-(p-nitrophenylthio)uridine 3′-phosphate (2): 1H NMR
(300 MHz, D2O) δ 8.00 (d, J ) 8.6 Hz, 2H), 7.53 (d, J ) 8.5 Hz, 2H),
7.50 (d, J ) 7.7 Hz, 1H), 6.19 (d, J ) 9.0 Hz, 1H), 5.57 (d, J ) 8.1
Hz, 1H), 4.78-4.83 (m, 1H), 4.29-4.31 (m, 1H), 4.04 (dd, J ) 8.6,
5.7 Hz, 1H), 3.57 (d, J ) 3.5 Hz, 2H); 13C NMR (126 MHz, D2O) δ
165.2, 150.7, 146.3, 141.3, 141.1, 132.4, 123.9, 103.1, 89.7, 86.1, 75.5
(d, JCP ) 5.4 Hz), 61.3, 53.3 (d, JCP ) 5.4 Hz); 31P NMR (121 MHz,
D2O) δ 1.13 (d, J ) 5.4 Hz); ultraviolet absorption λmax 260 nm (ꢀ
7400), λ 332 nm (ꢀ 3900), at pH ) 5.8; mass spectrum (LSIMS, 3-NBA,
negative ion mode) m/z 460.1 [M-, calcd for C15H15N3O10PS 460.1].
Conclusions
The unique reactivity of 2′-thiol-substituted ribonucleotides
is an attribute that highlights their potential as important tools
for the study of nucleic acid chemistry and biology. Our finding
that ribonucleotide analog 1 reacts through transphosphorylation
suggests that oligonucleotides bearing 2′-sulfhydryl groups can
be used to provide mechanistic insight into enzyme- and
ribozyme-promoted RNA cleavage. Catalysts that adequately
activate the leaving 5′-hydroxyl and/or polarize the scissile
phosphodiester bond may be expected to facilitate the cleavage
of 2′-thiol-containing oligonucleotides by a similar mechanism.
Moreover, nucleoside analogs endowed with a 2′-thiol can be
used to equip ribozymes with a functional group that has
widespread use in reactions catalyzed by enzymes. Finally, 2′-
thiol substituents can be used to engineer metal binding sites
into RNA molecules. Such modifications would provide the
means to dissect the contributions of specific metal ions to RNA
1
2′-Deoxy-2′-thiouridylyl (3′f5′) adenosine: H NMR (300 MHz,
D2O) δ 8.45 (s, 1 H), 8.26 (s, 1 H), 7.73 (d, J ) 8.3 Hz, 1 H), 6.12 (d,
J ) 5.0 Hz, 1 H), 5.89 (d, J ) 8.1 Hz, 1 H), 5.82 (d, J ) 8.8 Hz, 1 H),
4.87-4.85 (m, 1 H), 4.57-4.51 (m, 2 H), 4.38-4.37 (m, 1 H), 4.26-
4.11 (m, 3 H), 3.72-3.60 (m, 2 H), 3.54 (dd, J ) 7.0, 5.2 Hz, 1 H);
13C NMR (75 MHz, D2O) δ 165.8, 155.4, 152.7, 151.8, 148.9, 141.2,
139.6, 118.6, 102.8, 89.8, 87.2, 84.9, 83.4 (d, JCP ) 9.5 Hz), 76.6 (d,
JCP ) 5.1 Hz), 73.7, 70.2, 65.2 (d, JCP ) 5.1 Hz), 61.0, 44.2 (d, JCP
7.0 Hz); 31P NMR (121 MHz, D2O) δ -0.50; ultraviolet absorption
max 260 nm (ꢀ 14 000) at pH ) 5.6; mass spectrum (LSIMS, 3-NBA,
)
λ
negative ion mode) m/z 588.1 [M-, calcd for C19H24N7O11PS 588.1].
Experimental Procedures. 31P NMR Spectroscopy Studies of the
Hydrolysis of 1. A deoxygenated solution of 1 (3.6 mg, 0.0074 mmol)
in water (0.205 mL) was added to a deoxygenated solution of buffer.
The final solution was 15 mM in 1 at a buffer concentration of 150
mM with the ionic strength of the solution adjusted to 250 mM by the
addition of NaCl. Reactions were carried out at room temperature in
a 5 mm NMR tube equipped with a D2O insert. The resulting solutions
were periodically monitored by 31P NMR. For the reaction at pH 13
(0.49 mM NaOH), the solution concentration of 1 was 15 mM.
(27) (a)Denu, J. M.; Lohse, D. L.; Vijayalakshmi, J.; Saper, M. A.; Dixon,
J. E. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 2493-2498. (b) Barford, D.;
Flint, A. J.; Tonks, N. K. Science 1994, 263, 1397-1404. (c) Wo, Y. Y.
P.; Zhou, M. M.; Stevis, P.; Davis, J. P.; Zhang, Z. Y.; Etten, R. L. V.
Biochemistry 1992, 31, 1712-1721.
(28) Hengge, A. C.; Sowa, G. A.; Wu, L.; Zhang, Z. Y. Biochemistry
1995, 34, 13982-13987.

ReactiVity of a 2′-Thio Nucleotide Analog
J. Am. Chem. Soc., Vol. 118, No. 47, 1996 11719
UV-Visible Studies of the Hydrolysis of 1 and 5. The transphos-
phorylation reactions of 2′-thiol-UpOC6H4-p-NO2, 1, and UpOC6H4-
p-NO2, 5, were monitored at 330 nm for pH <7 and at 400 nm for pH
g7. Kinetic measurements were carried out at 25 °C with substrate
concentrations at 0.15 mM and buffer concentrations at 100 mM. The
ionic strength of each reaction was adjusted to 200 mM using NaCl.
Conditions for the kinetic studies were chosen to minimize buffer-
induced catalysis. These precautions include the specific choice of
buffers used (listed under General Methods) and the reaction conditions,
which were selected so that each was carried out at a pH near the pKa
for each buffer. At pH 7.5, the rates of transphosphorylation of 1 in
two different buffers, HEPES and triethanolamine, were found to be
the same.
For substrates 1 and 5, the observed rate of transphosphorylation
was calculated from the plot of absorbance versus time using Beer’s
law. The change in molar absorptivity for both compounds at each
pH was determined from the difference in absorbance of completely
hydrolyzed compound compared to the initial absorbance of starting
material at the same concentration, pH, and ionic strength. Each data
point plotted in the pH-rate profile is an average of at least two runs.
The observed rate constant for the hydrolysis of 1 was fit to the
following Henderson-Hasselbach-derived equation:
Figure 4. Graph of the chemical shift of the H2′proton in 2′-deoxy-
2′- thiouridylyl(3′f5′)adenosine as a function of pD.
strength of a solution of 2′-deoxy-2′-thiouridylyl(3′f5′)adenosine (11
mM in D2O) was adjusted to 130 mM by the addition of NaCl. The
pD of the solution was adjusted to 4.6 by addition of a solution of DCl
in D2O. After acquisition of a 1H NMR spectrum, titration of 2′-deoxy-
2′-thiouridylyl(3′f5′)adenosine with NaOD in D2O was monitored by
1H NMR and the pD of the solution was measured on a pH meter. Ten
such measurements were made up to a pD of 11.5 (Figure 4). The
pKa was calculated from the fit of the chemical shift of the H2′ proton
of uridine versus pD using the following equation
kobs ) kmaxKa/([H+] + Ka) + kmin[H+]/([H+] + Ka)
where Ka is the acid dissociation constant for the 2′-thiol, kmax is the
asymptotic maximum rate of thiolate hydrolysis, and kmin is the
asymptotic minimum. A value to 8.3 ( 0.1 was obtained for the pKa
of the 2′-thiol in 1.
δobs ) δdownδ
up [(10pK + 10pD)/(δdown10pD + δup10pK )]
a
a
The observed rate constant for the hydrolysis of 5 was expressed
using the following equation:
where δobs is the chemical shift of the H2′ in ppm, δdown is the downfield
limit for the chemical shift, and δup is the upfield limit. A value of 8.2
( 0.1 was obtained for the pKa of the 2′-thiol.
kobs ) kmaxKa/([H+] + Ka) + kmax′[H+]/([H+] + Ka′) + kind
Acknowledgment. This research was supported by the NSF
NYI Program and the Milwaukee Foundation (Shaw Scientist
Program). We thank Prof. R. T. Raines (UWsMadison) for
use of a UV-visible spectrometer and helpful discussions and
J. M. Messmore, Dr. A. C. Hengge, and Prof. W. W. Cleland
(UWsMadison) for helpful discussions. C.L.D. thanks the NIH
where Ka is the acid dissociation constant for the 2′-hydroxyl, Ka′ is
the acid dissociation constant for the phosphate group, kmax is the
asymptotic maximum rate of alkoxide mediated hydrolysis, kmax′ is the
maximum rate of hydroxide attack on the neutral phosphate, and kind
is the rate of pH-independent hydrolysis.
The asymptotic maximum rate of alkoxide-mediated transphospho-
rylation of 5 was obtained by extrapolation of the above fit past the
pKa of the 2′-hydroxyl using 1.2 × 10-14 as the acid dissociation
constant for the 2′-hydroxyl.18a By comparing the asymptotic maximum
observed rates of transphosphorylation of 1 and the extrapolated
maximum of 5, the ratio of intramolecular alkoxide versus thiolate attack
on the phosphate was calculated to be 107.
for
a Predoctoral Biotechnology Training Fellowship
(T32GM08349).
Supporting Information Available: 1H NMR spectra for
1, 2, and 2′-deoxy-2′- thiouridylyl(3′f5′)adenosine (4 pages).
See any current masthead page for ordering and Internet access
instructions.
Titration of the Thiol in 2′-deoxy-2′-thiouridylyl(3′f5′)adenosine
As Monitored by 1H NMR. Spectra are reported relative to an internal
reference of sodium 3-(trimethylsilyl)propionate-2,2,3,3,-d4. The ionic
JA962265C