Dynamics of alkene radical cation/phosphate anion pair formation from nucleotide C4′ radicals. The DNA/RNA paradox revisited

Article abstract of DOI:10.1021/ja010841l

The fragmentation of nucleotide C4′ radicals generated by thiyl radical addition to C4′C5′ exocyclic glycals has been re-examined and found to be a function of the thiol and, probably, the initiating system employed. It has been demonstrated that C4′ radi

Full text of DOI:10.1021/ja010841l

J. Am. Chem. Soc. 2001, 123, 9239-9245
9239
Dynamics of Alkene Radical Cation/Phosphate Anion Pair Formation
from Nucleotide C4′ Radicals. The DNA/RNA Paradox Revisited
David Crich* and Wenhua Huang
Contribution from the Department of Chemistry, UniVersity of Illinois at Chicago, 845 West Taylor Street,
Chicago, Illinois 60607-7061
ReceiVed March 30, 2001
Abstract: The fragmentation of nucleotide C4′ radicals generated by thiyl radical addition to C4′C5′ exocyclic
glycals has been re-examined and found to be a function of the thiol and, probably, the initiating system
employed. It has been demonstrated that C4′ radicals of DNA and RNA models fragment even in the very
nonpolar benzene solution if the correct (aliphatic) thiol is employed. 17O-Labeling experiments are used to
demonstrate that the fragmentation of nucleotide C4′ radicals (2-deoxyribo- and ribo-) to contact ion pairs is
either irreversible or so rapidly reversible as to preclude prior reorganization of the contact ion pair. Formation
of the solvent-separated ion pair is an irreversible step, with all such ion pairs proceeding to product formation.
Introduction
Scheme 1
In collaboration with Newcomb and co-workers, we have
recently put forward a model for the chemistry of â-(phos-
phatoxy)alkyl radicals in which rate-determining fragmentation
leads to the formation of a contact ion pair consisting of an
alkene radical cation and a phosphate anion. Collapse of the
ion pair leads to the rearranged products. Equilibration of the
contact ion pair with solvent provides a solvent-separated ion
pair and eventually observable free ions (Scheme 1). The
outcome of a particular reaction, rearrangement or fragmenta-
tion, is then seen to be the result of a complex interplay between
collapse of the contact ion pair and its equilibration with solvent-
separated and free ions. This interplay, in turn, is a function of
the ability of substituents and solvents to stabilize the various
charged species.1,2 Support for this model derives from the
closely related pre-exponential factors in the Arrhenius functions
for either the rearrangement or fragmentation of a broad series
of â-(phosphatoxy)alkyl and â-(acyloxy)alkyl radicals in a range
of solvents of widely differing polarity, which suggests a
common rate-determining step, namely that of fragmentation
to the contact ion pair. Further support comes from the high
linear correlation of rate constants, rearrangement or fragmenta-
tion, with the ET(30) solvent polarity scale in a range of solvents
spanning benzene and aqueous acetonitrile.1,2 Earlier trapping
experiments with stereochemically labeled probes3 are now best
interpreted as involving nucleophilic attack on the contact ion
pair.
potentials in acetonitrile,4 the model should apply to nucleotide
C4′ radicals; indeed, chemically induced dynamic nuclear
polarization experiments with 3-phosphatoxy-2-tetrahydrofura-
nyl radicals provided very strong support for the formation of
enol ether radical cations.5 More recently, using a method in
which enol ether radical cations oxidize triarylamines to the
highly chromophoric triarylaminium radical cations, we have
established that simple R-alkoxy-â-(phosphatoxy)alkyl radicals
(1 and 2) do, indeed, undergo fragmentation to diffusively free
alkene radical cations (3) in acetonitrile solution.6,7 From the
studies conducted so far, there appears to be no reason to suspect
that the reactions of R-alkoxy-â-(phosphatoxy)alkyl radicals do
not conform to the general mechanistic picture of Scheme 1,
involving equilibria between a series of ion pairs and free ions.
Indeed, the approximately 100-fold rate difference for the
fragmentation of radicals 2 and 4, giving 3 and 5, respectively,
noted by two different groups in solvents of similar polarity is
readily explained by this model and the different kinetic methods
used (Scheme 2). Thus, the classical competition kinetic method
employed for 4 by the Giese group5 probably functions at the
Due to the need for strongly UV-absorbing radicals and
radical cations in the laser flash photolytic method used to
establish the above model, all of our studies were carried out
with systems leading to benzyl radicals and styrene-type radical
cations. However, it seems very reasonable to extrapolate the
model to any system capable of fragmenting to give a radical
cation that is at least as stable as that derived by oxidation of
styrene. As ethyl vinyl ether and styrene have the same oxidation
(4) Katz, M.; Riemenschneider, P.; Wendt, H. Electrochim. Acta 1972,
17, 1595-1607.
(5) Gugger, A.; Batra, R.; Rzadek, P.; Rist, G.; Giese, B. J. Am. Chem.
Soc. 1997, 119, 8740-8741.
(1) Newcomb, M.; Horner, J. H.; Whitted, P. O.; Crich, D.; Huang, X.;
Yao, Q.; Zipse, H. J. Am. Chem. Soc. 1999, 121, 10685-10694.
(2) Whitted, P. O.; Horner, J. H.; Newcomb, M.; Huang, X.; Crich, D.
Org. Lett. 1999, 1, 153-156.
(6) Newcomb, M.; Miranda, N.; Huang, X.; Crich, D. J. Am. Chem. Soc.
2000, 122, 6128-6129.
(7) Bales, B. C.; Horner, J. H.; Huang, X.; Newcomb, M.; Crich, D.;
Greenberg, M. M. J. Am. Chem. Soc. 2001, 123, 3623-3629.
(3) Crich, D.; Gastaldi, S. Tetrahedron Lett. 1998, 39, 9377-9380.
10.1021/ja010841l CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/31/2001

9240 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Crich and Huang
Scheme 2
Scheme 4
Scheme 3
oxoruthenium(IV) complex is retarded by the presence of
electron-withdrawing substituents at the 2′-position, and which
is attributed to destabilization of the nascent 1′-carbocation.16-18
In support of our hypothesis, we found with simple model
systems that the 2-deoxyribo system fragments significantly
faster than the ribo. More precisely, the 2-deoxynucleotide 6
was found to undergo fragmentation, in CD3OD/D2O (10:1) at
40 °C in the presence of excess thiophenol and di-tert-butyl
peroxalate, to the glycal 12 in 87% yield after only 10 min,
whereas the ribo analogue 7 was recovered unchanged after 24
h under the same conditions, despite repeated addition of initiator
(Scheme 4). Given that γ-C-O bonds have only a very minimal
effect on the stability of free radicals,19-24 we interpreted this
difference of reactivity in terms of a reversible addition of PhS•,
with comparable forward and reverse rate constants in both
systems, but with the fragmentation of the ribo radical 9 being
very significantly retarded over that of its deoxy analogue 8
due to the presence of the inductively withdrawing 2-C-O bond.
After our initial report, Giese and co-workers described a series
of experiments in which an RNA C4′ radical in an undecamer
was found to cleave more slowly than the comparable single-
stranded DNA C4′ radical, but only by a factor of 3.25
Here we present the results of experiments designed to probe
the reversibility of initial fragmentation to the contact ion pair
and the effect of a 2′-C-O bond on the fragmentation and ion
pair equilibria. The reasons underlying the very different results
previously reported by Giese and ourselves on the effect of the
2′-C-O bond on nucleotide C4′ radical fragmentation have also
been determined.
level of the contact ion pair, whereas only free radical cations
are reported by oxidation of triarylamines in the laser flash
photolytic method developed with 2.6 The smaller rate constant
determined by the latter method for the fragmentation of 2 is
therefore a global one for fragmentation and cage escape and
necessarily factors in the equilibrium constants for distribution
between the various ion pairs.6
A number of interesting questions are raised by the general
model, one of which pertains to the dynamics of the contact
ion pair: does it always proceed in the forward direction to the
rearranged radical or the solvent-separated ion pair, or does it
partition with the initial radical, i.e., is the initial fragmentation
reversible? A second point of interest to us revolves around
the so-called DNA/RNA paradox,8 in which the antitumor
antibiotic iron-bleomycin (BLM) degrades “tDNAHis ” more
rapidly than the corresponding “tRNAHis ”,9-11 through the
formation and cleavage of C4′ radicals,12-14 yet binds the RNA
more tightly. By comparison with carbohydrate chemistry,
wherein it is widely appreciated that 2-deoxy glycosides are
solvolyzed much more rapidly than the analogous glycosides
(Scheme 3),15 a phenomenon attributed to destabilization of the
anomeric cation by the adjacent C-O bond, we proposed that
fragmentation of the ribonucleotide C4′ radical is retarded by
the 2′-C-O bond, which destabilizes the 3′4′ radical cation.8
Thorp and co-workers have described a very similar concept in
which hydride abstraction from the 1′-site of nucleotides by an
Results and Discussion
Nucleotide C4′ radical generation by the addition of phen-
ylthiyl radicals to exocyclic glycals, as employed in our initial
(16) Farrer, B. T.; Pickett, J. S.; Thorp, H. H. J. Am. Chem. Soc. 2000,
122, 549-553.
(17) Neyhart, G. A.; Cheng, C.-C.; Thorp, H. H. J. Am. Chem. Soc. 1995,
117, 1463-1471.
(8) Crich, D.; Mo, X.-S. J. Am. Chem. Soc. 1997, 119, 249-250.
(9) Holmes, C. E.; Hecht, S. M. J. Biol. Chem. 1993, 268, 25909-25913.
(10) Hecht, S. M. Bioconjugate Chem. 1994, 5, 513-526.
(11) Hecht, S. M. In DNA and RNA CleaVers and Chemotherapy of
Cancer and Viral Diseases; Meunier, B., Ed.; Kluwer Academic: Dordrecht,
1996; pp 77-89.
(12) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107-1136.
(13) Stubbe, J.; Kozarich, J. W.; Wu, W.; Vanderwall, D. E. Acc. Chem.
Res. 1996, 29, 322-330.
(14) Hecht, S. M. Acc. Chem. Res. 1986, 19, 383-391.
(15) Capon, B. Chem. ReV. 1969, 69, 407-498. Additionally, 2-azido-
1,2-dimethoxypropane is solvolyzed 600 times more slowly than 2-azido-
2-methoxypropane in aqueous media: Amyes, T. L.; Jencks, W. P. J. Am.
Chem. Soc. 1989, 111, 7888-7900.
(18) Thorp, H. H. Chem. Biol. 2000, 7, R33-R36.
(19) Busfield, W. K.; Grice, I. D.; Jenkins, I. D.; Monteiro, M. J. J.
Chem. Soc., Perkin Trans. 2 1994, 1071-1077.
(20) Busfield, W. K.; Grice, I. D.; Jenkins, I. D. J. Chem. Soc., Perkin
Trans. 2 1994, 1079-1086.
(21) Beckwith, A. L. J.; Norman, R. O. C. J. Chem. Soc. (B) 1969, 400-
403.
(22) Smith, P.; Karukstis, K. K. J. Magn. Reson. 1980, 39, 137-140.
(23) Gilbert, B. C.; Norman, R. O. C.; Williams, P. S. J. Chem. Soc.,
Perkin Trans. 2 1981, 1401-1405.
(24) Edge, D. J.; Kochi, J. K. J. Am. Chem. Soc. 1972, 94, 7695-7702.
(25) Strittmatter, H.; Dussy, A.; Schwitter, U.; Giese, B. Angew. Chem.,
Int. Ed. Engl. 1999, 38, 135-137.

Alkene Radical Cation/Phosphate Anion Pair Formation
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9241
study (Scheme 4), was first described by the Giese group,26,27
who found inter alia that cleavage of the dialkyl phosphates 6
and 16, leading to the fragmentation product 12, was in
competition with reduction to 14 and 17 on irradiation at 30 °C
with 1.0-2.4 M thiophenol in both toluene and ethanol/water
(4/1). The very high differences in conversion of 6 and 7
isotopic dilution. It can therefore be concluded that the same
was true for 17O and that the extent of enrichment of 18 and 19
in 17O approached the 10% in the original source water.28,29
Direct determination of the 17O incorporation by integration of
the 31P NMR spectra is precluded by the quadrupolar nature of
5
17O (I ) /2), which results in the 17OdP resonance being
dispersed in a complex multiplet.28,29
Exocyclic glycal 18 was allowed to react with 2.1 M
thiophenol, initiated by DBPO at 40 °C in 10/1 CD3OD/D2O
for 15 min, after which preparative TLC enabled the isolation
of 12 and recovered 18 in 33 and 35% yields, respectively.
Inspection of the recovered substrate (18) by 17O NMR
spectroscopy revealed the presence of a single peak at δ 76.3
which, as the PdO and PsO resonances are expected to be
very readily resolved,30-32 indicates no scrambling of the label.
Under the same conditions 19 was recovered quantitatively, also
with no scrambling of the label. The very considerable differ-
ences in reactivity of 18 and 19 mimic those in our original
experiments with the label-free substrates 6 and 7. To observe
fragmentation with 19, it was necessary to operate in 2/1
CD3OD/D2O when 13 was isolated in 26% yield after 1 h, with
18% of 19 being recovered, again without detectable scrambling
of the label.33 Although we have principally relied on 17O NMR
spectroscopy as a means of detecting scrambling of stereo-
chemistry at phosphorus, the results were always fully cor-
roborated by analysis of the 31P NMR spectra.34 In effect, the
use of water enriched in both 17O and 18O for the synthesis of
observed earlier in this laboratory, using 1.5 M thiophenol in
CD3OD/D2O 10/1 at 40 °C, with initiation by di-tert-butyl
peroxalate (DBPO), coupled with the very minor influence of
γ-C-O bonds on radical stability,19-24 prompted us to modify
the original Giese mechanism to include the reversibility of the
initial addition of the phenylthiyl radical.8 Were the initial
addition of PhS• not reversible, the slower fragmentation of 9
would have been expected to manifest itself in terms of the
formation of a greater amount of reduction product 15, rather
than an apparent lack of reactivity, which was not the case.
We now suggest that the rapid reversibility of addition of
RS• to exocyclic glycals coupled with the use of a regiospe-
cifically 17O-labeled phosphate provides a method for probing
the contact ion pair formed on expulsion of the phosphate group.
Specifically, if a monolabeled phosphate 18 or 19 is allowed to
react with RSH and a radical initiator and the reaction is stopped
after significant conversion to the fragmentation products 12
and 13, respectively, then scrambling of the label between the
PdO and PsO in the recovered nucleoside must indicate
reversibility of phosphate fragmentation. Alternatively, if it can
be demonstrated that phosphate cleavage is taking place but
scrambling of the label is not observed in recovered 18 and/or
19, then the inference must be that the contact ion pair
recombines before reorganization or not at all. The latter scenario
implies that once the contact ion pair becomes solvent-separated,
recombination cannot compete with other modes of decomposi-
tion. Toward this end, 20 and 21 were allowed to react with
18 and 19 enables the isotope shifts due to the presence of 18
O
to be employed as a probe of scrambling. The 31P NMR spectra
of 18 and 19 as initially prepared showed two singlets resulting
from the 16O and 18O isotopomers. Scrambling of the 18O
between the PdO and PsO positions would be expected to
lead to 31P NMR spectra containing three resonances: one each
for the unlabeled, the original Pd18O labeled, and the inverted
Ps18O substances.
We next turned to the use of intramolecular nucleophiles with
the expectation that cyclization might be competitive with
collapse of the contact ion pair. As illustrated in Scheme 5,
experiments of this type were conducted with 2-mercaptoetha-
nol, which served the dual function of thiol and nucleophile.
When 6 was exposed to 0.15 M 2-mercaptoethanol, with
1
initiation by DBPO, in C6D6 at 40 °C, H NMR spectroscopy
after 45 min revealed consumption of the substrate and the
formation of a rather complex reaction mixture. One nucleoside
predominated in this reaction mixture to the extent of ap-
proximately 50% as judged from the intensity of the “anomeric”
signals. Isolation of this product from the complex reaction
mixture was not possible owing to the small scale of the
experiment and the instability of the product. The experiment
was therefore repeated on a somewhat larger scale in benzene,
when 1H NMR spectroscopy of the crude reaction mixture again
diethyl chlorophosphite in THF and pyridine, followed by
oxidation with a mixture of iodine and 10% 17O-labeled water.
In this manner, 18 and 19 were prepared in 37 and 57% isolated
yields, respectively, and the label was shown to reside only in
the PdO oxygen by the presence of a unique peak in the 17O
NMR spectra at δ 76.4 and 76.1, respectively. Determination
of the exact extent of 17O incorporation by mass spectrometry
was complicated by the further 57% enrichment of the water
employed in 18O. Inspection of the 31P NMR spectra of 18 and
19, however, revealed that 18O had been incorporated with little
(28) Lowe, G.; Potter, B. V. L.; Sproat, B. S.; Hull, W. E. J. Chem.
Soc., Chem. Commun. 1979, 733-735.
(29) Tsai, M.-D. Biochemistry 1979, 18, 1468-1472.
(30) Kintzinger, J. P. In NMR of Newly Accessible Nuclei; Laszlo P.,
Ed.; Academic Press: New York, 1979; Vol. 2, pp 79-104.
(31) Evans, S. A. In 17O NMR Spectroscopy in Organic Chemistry;
Boykin, D. W., Ed.; CRC Press: Boca Raton, FL, 1990; pp 263-320.
(32) Tsai, M.-D.; Bruzik, K. In Biological Magnetic Resonance; Berliner,
L. J., Reuben, J., Eds.; Plenum Press: New York, 1983; Vol. 5, pp 129-
181.
(33) The low mass balances in most of the experiments described here
are explained in large part, as verified by control experiments, by the
instability of most of the substrates and products under the reaction
conditions, with diethylphosphoric acid being a necessary byproduct.
(34) Seela, F.; Ott, J.; Potter, B. V. L. J. Am. Chem. Soc. 1983, 105,
5879-5886.
(26) Giese, B.; Beyrich-Graf, X.; Burger, J.; Kesselheim, C.; Senn, M.;
Schafer, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1742-1743.
(27) Giese, B.; Burger, J.; Kang, T. W.; Kesselheim, C.; Wittmer, T. J.
Am. Chem. Soc. 1992, 114, 7322-7324.

9242 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Crich and Huang
Scheme 5
indicated formation of the same major product. Repeated
chromatography on silica gel eventually enabled isolation of
this product in ∼5% yield, which reflects the extensive
decomposition in the course of the isolation. The new product
decomposed rapidly in CDCl3 solution but was stable for
extended periods of time in C6D6 and CD3OD, undergoing only
a slow isomerization to a second diastereoisomer. The general
instability toward chromatography and CDCl3, coupled with the
slow isomerization, point to a spiroacetal 32. The spiroacetal
structure was supported by the high-resolution mass spectrum,
which was consistent with a molecular formula of C19H20N5O3S
for the protonated species. In the 13C spectrum of a mixture of
the two equilibrating diastereomers, weak signals at δ 104.2
and 105.1 are attributable to the acetal carbon. None of the
of the reduction product indicates that the fragmentation of the
ribo radical 23 is significantly slower than that of its 2-deoxy
counterpart 22. Second, the formation of 27 rather than the
spiroacetal 33 suggests that the ribo radical ion pair 25 is more
reactive than its deoxy counterpart 24 and is reduced to 27 in
competition with nucleophilic trapping. The actual mechanism
of this reduction probably involves a competing deprotonation
within the contact ion pair to give an allyl radical, followed by
hydrogen transfer from the thiol. The 2′-C-O bond in the ribo
series both destabilizes the radical cation and stabilizes the allyl
radical, which accounts for the higher rate of deprotonation as
compared to that in the deoxy ribo series.36 Third and most
important, the very formation of 27 indicates that ion pair 25
can be formed, even in benzene. This is in contrast to the earlier
experiments (Scheme 4) when 7 was recovered unchanged even
after 24 h in the far more polar methanol/water mixtures, leading
to the original suggestion that fragmentation of radical 9 to ion
pair 10 was severely retarded by the 2-C-O bond.
1
resonances assigned to the H5′ hydrogens in the H spectrum
had any significant 3J couplings which, along with the absence
of a signal attributable to H4,′ further confirmed the spiroacetal
structure.
In an attempt to stabilize the spiroacetal structure, the reaction
was repeated using 2-mercapto-1,1-dimethylethanol but, unfor-
tunately, with a similar result, namely a high conversion to the
anticipated product (36) but a low isolated yield. Again, the
spectroscopic data for 36 fully support the structure assigned.
The obvious difference between Schemes 4 and 5 is the
choice of thiol and its effect on the initial equilibrium between
the exocyclic glycals and the adduct C4′ radicals. It is well
known that analysis of the kinetics of addition of thiyl radicals
to alkenes is complicated by the reversibility of the reaction.
Nevertheless, following earlier work by Davies and Roberts,37
Griller and co-workers established an approximate value for
n-BuS• addition to 1-octene of 1.9 × 106 M-1 s-1 in isooctane
at 25 °C,38 whereas Ito and Matsuda determined a value of 1.8
× 105 M-1 s-1 for the addition of p-ClPhS•, which is more
reactive than simple PhS•, to the more reactive isobutyl vinyl
ether.39 Wagner and co-workers, on the other hand, determined
the relative rate constants for elimination of n-BuS• and PhS•
from secondary â-mercaptoalkyl radicals to be 1 and 687,
respectively.40,41 Although these numbers are only approximate
1
In particular, the H NMR spectrum is a simplified version of
that of 32 with the absence of resonances between δ 3.1 and
6.0 signaling the lack of a C3′-O bond and of a C4′ H, thereby
excluding any structure arising from ring closure at C3′ rather
than C4′.
(36) It was originally suggested by Giese that the endocyclic glycal
products derive from thiol reduction of the radical cation.26 However, it
has now been demonstrated that deprotonation to give allyl radicals can
take place within the contact ion pair for styrene radical cations,1,7 and
likely also does so for enol ether radical cations: Horner, J. H.; Newcomb,
M. J. Am. Chem. Soc. 2001, 123, 4364-4365. It is therefore highly likely,
at least in nonpolar solvents such as benzene, that the products arise from
deprotonation in the contact ion pair followed by subsequent reduction. A
reviewer has pointed out, correctly, that this deprotonation/reduction
mechanism cannot be operative in the experiments conducted in CD3OD,
as there is no incorporation from deuterium in the product from the
exchanged thiol.
The experiment was repeated, using 2-mercaptoethanol and
the ribonucleotide 7 in benzene. Unlike the above reactions with
6, no evidence for the formation of a spiroacetal (33) was found.
The two major isolated products were the fragmentation product
27 (24%) and the addition product 37 (44%),35 arising from
quenching of radical 23 (Scheme 5). Several conclusions can
be drawn from this experiment. First, the isolation in high yield
(35) The stereochemistry of 37, formed as a single diastereomer, was
not assigned rigorously. On the basis that quenching of the C4′ radical by
the thiol will take place preferentially from the less hindered face, it is
likely that C5′ is cis to the phosphate.
(37) Davies, A. G.; Roberts, B. P. J. Chem. Soc., B 1971, 1830-1837.
(38) McPhee, D. J.; Campredon, M.; Lesage, M.; Griller, D. J. Am. Chem.
Soc. 1989, 111, 7563-7567.
(39) Ito, O.; Matsuda, M. J. Am. Chem. Soc. 1979, 101, 1815-1819.

Alkene Radical Cation/Phosphate Anion Pair Formation
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9243
Scheme 6
and are not measured for vinyl ethers, it is evident that the
reversible addition of thiyl radicals to alkenes favors the adduct
radical to a much greater extent for alkylthiyl radicals than for
arylthiyl radicals. This leads directly to the conclusion that
radical 23 (Scheme 5) has a longer lifetime than radical 9
(Scheme 4) and that the increase in lifetime of 23 is sufficient
to permit fragmentation to 25 to compete with the back reaction.
A revised general reaction scheme for the fragmentation should
therefore be as presented in Scheme 6.
The fragmentation of 6 and 7 in benzene with 2-mercapto-
ethanol but not with thiophenol is not a function of the hydroxy
group on 2-mercaptoethanol. The reaction of 6 with ethanethiol
in benzene at 40 °C on initiation with DBPO resulted in
extensive degradation, and it was not possible to isolate the
anticipated product 38. However, reaction of 7 with ethanethiol
under the same conditions resulted in the formation of the
fragmentation product 39 and the reduction product 40, in 21
and 34% yields respectively, together with recovered 7,
demonstrating again that fragmentation of the C4′ radical is a
function of thiol. Reaction of 6 with ethanethiol and of 7 with
tert-butanethiol in CD3OD resulted in the formation and isolation
of the fragmentation products 38 and 41 in 79 and 24% yields,
respectively.
and it is therefore concluded that all solvent-separated ion pairs
proceed in the forward direction toward products. This last result
is in full agreement with earlier failed crossover experiments.42,43
A further conundrum concerns the discrepancy between the
original work of the Giese group,26 who observed fragmentation
product 12 and addition products 17 on addition of PhS• radicals
to exocyclic glycals 6 and 16 in toluene at 30 °C, and our own
experiments with PhS• and 6 in C6D6, which resulted in very
little reaction at 40 °C. The answer, we believe, is again to be
found in the initial equilibrium between the thiyl radical and
its adduct with the glycal. In effect, the position of this
equilibrium is affected not only by the nature of the thiol but
also by the concentration of thiyl radicals in the reaction mixture.
This, in turn, is a function of the radical initiation method. In
the work described here, as in our previous work, we have
preferred chemical initiation with DBPO, as this allows us to
constantly monitor experiments in the probe of the NMR
spectrometer. Under these conditions, the flux of PhS• radicals
is determined by the half-life for the decomposition of the
initiator at the temperature employed (t1/2(DBPO) at 40 °C ≈ 2
h).44 Giese, on the other hand, employed photolytic initiation
which, presumably, provided a higher flux of radicals and so
shifted the equilibrium in the forward direction.
Conclusion
It has been demonstrated that the overall rate of fragmentation
of nucleotide C4′ radicals, when generated by addition of thiyl
radicals to exocyclic glycals, is a function of the position of
the equilibrium between the thiyl radical and alkene and their
adduct. This, in turn, means that the observed rate of fragmenta-
tion will be a function of the thiol used and of the initiating
system. These complications suffice to explain the discrepancies
between our earlier work8 and that of the Giese group25 on the
differing rates of fragmentation between nucleotide C4′ radicals
and their 2-deoxy analogues. While the results reported earlier
were correct and were shown to be reproducible, the results
from the Giese group, using a more straightforward method of
radical generation, probably reflect more accurately the true
extent of the phenomenon. 17O-Labeling experiments unambigu-
ously demonstrate that the fragmentation of nucleotide C4′
radicals is either not reversible or so rapid as to preclude
reorganization of the contact ion pair. All solvent-separated ion
pairs proceed to product formation.
The reaction of ethanethiol, DBPO, and 7 in benzene was
repeated using the 17O-labeled isotopomer 19 and C6D6 as
solvent. The fragmentation product 39 was isolated in 19% yield,
the reduced product 17O-40 in 38% yield, and the substrate 19
in 33% yield. Examination of recovered 19 and 17O-40 by 17
O
NMR spectroscopy revealed that no scrambling of the label had
taken place in either case; i.e., each substance exhibited a single
peak in the 17O NMR spectrum corresponding to the PdO
labeled substance. It is therefore established beyond reasonable
doubt that C4′ radicals in both the 2-deoxy and ribo series do
fragment to give contact ion pairs, even in benzene solution. It
is likewise established that if the formation of the contact ion
pair is reversible, the reverse reaction is more rapid than
reorganization of the contact ion pair. After equilibration with
the solvent, collapse back to the initial radical is not observed,
Experimental Section
General. 1H and 13C NMR spectra were recorded in CDCl3 solutions
unless otherwise stated. 17O NMR spectra were recorded in CDCl3
(40) Wagner, P. J.; Sedon, J. H.; Lindstrom, M., J. J. Am. Chem. Soc.
1978, 100, 2579-2580.
(41) Griller, D.; Simo˜es, J. A. M. In Sulfur-Centered ReactiVe Intermedi-
ates in Chemistry and Biology; Chatgilialoglu, C.; Asmus, K.-D., Eds.;
Plenum: New York, 1989; pp 327-340.
(42) Crich, D.; Escalante, J.; Jiao, X.-Y. J. Chem. Soc., Perkin Trans. 2
1997, 627-630.
(43) Beckwith, A. L. J.; Duggan, P. J. J. Am. Chem. Soc. 1996, 118,
12838-12839.
(44) Walling, C. Tetrahedron 1985, 41, 3887-3900.

9244 J. Am. Chem. Soc., Vol. 123, No. 38, 2001
Crich and Huang
Table 1. Reaction of Exocyclic Glycals 6, 7, 18, and 19 with Thiols
fragmentation
product
(isolated yield)
recovered
substrate
(isolated yield)
31P (ppm) of
recovered
substrate
17O (ppm) of
recovered
substrate
DBPO
(mol %)
time
(min)
substrate
thiol
solvent (v/v)
1
2
3
4
5
6
7
8
9
18
18
18
18
19
19
7
6
19
7
PhSH
PhSH
PhSH
PhSH
PhSH
PhSH
PhSH
EtSH
EtSH
EtSH
ButSH
CD3OD/D2O (10/1)
CD3CN
CD3CN/CD3OD (10/1)
CD3OD
CD3OD/D2O (10/1)
CD3OD/D2O (2/1)
CD3OD/D2O (2/1)
CD3OD
40
246
465
100
40
158
120
100
50
15
195
75
45
45
105
105
30
30
45
12 (33)
(35)
(28)
(71)
(24)
(100)
(18)
nd
nd
(33)
nd
-3.40, -3.36
-3.37, -3.40
-3.37, -3.41
-3.36, -3.40
-3.25, -3.21
-3.26, -3.22
76.29
76.13
75.99
76.42
75.12
75.12
12 (62)
13 (26)
13 (41)
38 (79)
39 (19)
39 (26)
41 (24)
C6D6
CD3OD
CD3OD
-3.25, -3.22
75.78
10
11
150
150
7
45
nd
solution on a Bruker Avance spectrometer at 68 MHz. 17O chemical
shifts are given with respect to external H2O. High-resolution FAB
and ESI mass spectra were recorded by the University of Minnesota
Mass Spec Laboratory and the UIC Research Resources Center,
respectively. Deuterated solvents and O-labeled water (10% 17O, 55%
18O) were purchased from Cambridge Isotope Laboratories, Inc., and
used directly. Other solvents was dried and distilled by standard
procedures before use. All reactions were performed under argon.
Thiophenol, ethanethiol, and 2-methyl-2-propanethiol were purchased
from ACROS and used directly. 2-Mercaptoethanol was distilled before
use. 2-Hydroxy-2-methylpropanethiol,45 di-tert-butylperoxyoxalate,46
and substrates8 6 and 7 were prepared according to literature procedures.
17O-Labeled 6-N-Benzoyl-9-(2,5-dideoxy-3-O-diethylphosphoryl-
â-D-glycero-pent-4-enofuranosyl)adenine (18). Diethyl chlorophos-
phite (0.34 mL, 2.36 mmol) was added dropwise to a solution of 6-N-
benzoyl-9-(2,5-dideoxy-â-D-glycero-pent-4-enofuranosyl)adenine8 (200
mg, 0.59 mmol) and pyridine (1.0 mL) in THF (10.0 mL) at 0 °C.
After 30 min, a solution of iodine (1.27 g, 5.0 mmol) in a mixture of
THF, pyridine, and O-labeled water (5.0:1.0:0.2 mL) was added
dropwise until the red color of iodine persisted. The reaction mixture
was then treated with 10% aqueous Na2S2O3 (10.0 mL) and phosphate
buffer solution (10.0 mL, pH ) 7.0). The aqueous layer was extracted
with EtOAc (2 × 50 mL), and the combined organic layer was dried
over Na2SO4 and concentrated in vacuo. Preparative TLC on silica gel
(eluent: 20/1 CHCl3/MeOH) yielded the regioselectively labeled
nucleotide 18 (103 mg, 37%) as a white solid. 1H NMR δ: 1.38-1.41
(6 H, m), 1.81-1.87 (1 H, m), 2.12-2.20 (1 H, m), 4.16-4.20 (4 H,
m), 4.60 (1 H, d, J ) 2.6 Hz), 4.09 (1 H, d, J ) 2.4 Hz), 5.60-5.64
(1 H, m), 6.71 (1 H, t, J ) 6.6 Hz), 7.50-7.55 (2 H, m), 7.60-7.63 (1
H, m), 8.02-8.05 (2 H, m), 8.14 (1 H, s), 8.80 (1 H, s), 9.16 (1 H, br
s). 13C NMR δ: 16.5 (2 C), 38.7, 64.7 (2 C), 75.6, 85.3, 88.8, 124.0,
128.3 (2 C), 129.3 (2 C), 133.3, 133.9, 141.4, 150.2, 152.0, 153.4,
158.9, 165.0. 31P NMR δ: -3.44, -3.40. 17O NMR δ: 76.42. ESI-MS
(m/z, relative intensity): 473 (M, 71), 474 (M + 1, 32), 475 (M + 2,
100), 476 (M + 3, 24).
19 (Table 1). A solution of glycal (0.06 mmol) and thiol (30 equiv) in
the appropriate solvent (0.6 mL) was degassed by sparging with Ar at
0 °C for 5 min and then warmed to 40 °C. DBPO (Table 1) in the
reaction solvent (0.25 mL) was added, and the reaction was monitored
by TLC or NMR as appropriate. In the cases when little or no reaction
took place, further DBPO was added periodically over the course of
the reaction up to the total amount given in Table 1. The products
were isolated by preparative TLC (eluent: 20/1 CHCl3/MeOH) fol-
lowing direct deposition of the reaction mixture onto the plate. The
results are presented in Table 1, and the spectral data of the products
are recorded below.
6-N-Benzoyl-9-(2,3,5-trideoxy-5-phenylthio-â-D-glycero-pent-3-
enofuranosyl)adenine (12).8 1H NMR δ: 2.93 (1 H, d, J ) 17.1 Hz),
3.32-3.45 (1 H, m), 3.63 (1 H, d, J ) 15.9 Hz), 3.72 (1 H, d, J )
15.0 Hz), 5.02-5.06 (1 H, m), 6.90 (1 H, dd, J ) 3.2, 9.2 Hz), 7.20-
7.68 (8 H, m), 8.00-8.05 (2 H, m), 8.11 (1 H, s), 8.80 (1 H, s), 9.04
(1 H, br s).
6-N-Benzoyl-9-(3,5-dideoxy-2-O-methyl-5-phenylthio-â-D-glycero-
pent-3-enofuranosyl)adenine (13).8 1H NMR δ: 3.44 (3 H, s), 3.69
(1 H, d, J ) 15.2 Hz), 3.81 (1 H, d, J ) 15.2 Hz), 4.76 (1 H, s), 5.28
(1 H, d, J ) 1.8 Hz), 6.63 (1 H, s), 7.20-7.65 (8 H, m), 8.00-8.05 (2
H, m), 8.06 (1 H, s), 8.83 (1 H, s), 9.17 (1 H, br s). IR (film) ν: 1695,
1655 cm-1
.
6-N-Benzoyl-9-(2,3,5-trideoxy-5-ethylthio-â-D-glycero-pent-3-eno-
1
furanosyl)adenine (38). H NMR (CD3OD) δ: 1.21 (3 H, t, J ) 7.4
Hz), 2.57 (2 H, q, J ) 7.4 Hz), 3.15 (1 H, m), 3.30 (2 H, m), 3.45 (1
H, m), 5.16 (1 H, m), 6.94 (1 H, dd, J ) 3.9, 9.0 Hz), 7.55 (2 H, m),
7.65 (1 H, m), 8.07 (2 H, m), 8.55 (1 H, s), 8.73 (1 H, s). 13C NMR
(CD3OD) δ: 16.7, 25.0, 26.7, 36.1, 84.7, 96.3, 128.0, 128.3, 132.5,
141.7, 152.0, 154.0. HRFAB-MS: calcd for C19H20N5O2S, 382.1338;
found, 382.1357 (M + H)+.
6-N-Benzoyl-9-(3,5-dideoxy-5-ethylthio-2-O-methyl-â-D-glycero-
pent-3-enofuranosyl)adenine (39). 1H NMR δ: 1.30 (3 H, t, J ) 7.4
Hz), 2.62-2.67 (2 H, m), 3.35 (1 H, d, J ) 15.0 Hz), 3.42 (1 H, d, J
) 15.0 Hz), 3.54 (3 H, s), 4.81 (1 H, s), 5.34 (1 H, d, J ) 2.4 Hz),
6.65 (1 H, d, J ) 1.4 Hz), 7.55-7.60 (2 H, m), 7.62-7.67 (1 H, m),
17O-Labeled 6-N-Benzoyl-9-(5-deoxy-3-O-diethylphosphoryl-2-O-
methyl-â-D-glycero-pent-4-enofuranosyl)adenine (19). In a manner
similar to that described above for 18, the ribonucleotide 19 was
prepared as a colorless gum in 57% yield from 6-N-benzoyl-9-(5-de-
oxy-2-O-methyl-â-D-glycero-pent-4-enofuranosyl)adenine.8 1H NMR
δ: 1.35-1.40 (6 H, m), 3.49 (3 H, s), 4.14-4.28 (4 H, m), 4.66 (1 H,
d, J ) 2.7 Hz), 4.73 (1 H, d, J ) 2.6 Hz), 4.95-5.00 (1 H, m),
5.47-5.52 (1 H, m), 6.33 (1 H, d, J ) 7.6 Hz), 7.50-7.55 (2 H, m),
7.60-7.65 (1 H, m), 8.00-8.05 (2 H, m), 8.13 (1 H, s), 8.81 (1 H, s),
9.20 (1 H, br s). 13C NMR δ: 16.5 (2 C), 59.1, 64.6, 64.8, 73.4, 80.8,
87.3, 91.0, 124.2, 128.3 (2 C), 129.3 (2 C), 133.3, 133.9, 142.2, 150.2,
152.4, 153.5, 156.9, 165.0. 31P NMR δ: -3.29, -3.25. 17O NMR δ:
76.10. ESI-MS (m/z, relative intensity): 503 (M, 65), 504 (M + 1,
35), 505 (M + 2, 100), 506 (M + 3, 24).
8.05-8.08 (2 H, m), 8.28 (1 H, s), 8.85 (1 H, s), 9.10 (1 H, br s). 13
C
NMR δ: 14.4, 26.6, 28.1, 52.4, 88.3, 89.4, 98.3, 123.0, 128.1 (2 C),
129.1 (2 C), 133.1, 133.6, 140.3, 149.6, 152.6, 158.8, 160.8, 183.6.
6-N-Benzoyl 5′-Deoxy-5′-ethylthio-3′-O-diethylphosphoryl-2′-O-
methyl-D-adenosine (40) and Its 4′-Epimer. 40. 1H NMR δ: 1.25 (3
H, t, J ) 7.4 Hz), 1.35-1.45 (6 H, m), 2.61 (2 H, q, J ) 7.4 Hz), 2.96
(1 H, dd, J ) 5.4, 14.2 Hz), 3.08 (1 H, dd, J ) 5.6, 14.3 Hz), 3.51 (3
H, s), 4.15-4.25 (4 H, m), 4.50-4.55 (1 H, m), 4.80-4.85 (1 H, m),
5.05-5.10 (1 H, m), 6.11 (1 H, d, J ) 5.6 Hz), 7.50-7.55 (2 H, m),
7.60-7.65 (1 H, m), 8.03-8.08 (2 H, m), 8.31 (1 H, s), 8.81 (1 H, s)
9.20 (1 H, br s). 13C NMR δ: 14.7, 16.1, 27.2, 33.5, 58.8, 64.3, 75.5,
80.9, 83.2, 87.2, 120.8, 128.0, 128.9, 133.0, 133.4, 142.3, 149.5, 151.7,
152.2, 164.6. 31P NMR δ: -3.03, -2.99. 17O NMR δ: -77.35. HRESI-
MS: calcd for C24H32N5O7PSNa, 588.1658; found, 588.1616 (M +
Na)+. 4′-Epi-40. 1H NMR δ: 1.27 (3 H, t, J ) 7.4 Hz), 1.35-1.40 (6
H, m), 2.60 (2 H, q, J ) 7.5 Hz), 2.85 (1 H, dd, J ) 6.8, 13.8 Hz),
2.94 (1 H, dd, J ) 7.0, 13.8 Hz), 3.47 (3 H, s), 4.20-4.30 (4 H, m),
4.95-5.00 (1 H, m), 5.10-5.15 (1 H, m), 5.30-5.35 (1 H, m), 5.96(1
H, d, J ) 7.1 Hz), 7.50-7.55 (2 H, m), 7.60-7.65 (1 H, m), 8.02-
General Protocol for the Reaction of Monofunctional Thiols with
Exocyclic Glycals 6 and 7 and Their 17O-Labeled Analogues 18 and
(45) Singh, R.; Whitesides, G. M. J. Am. Chem. Soc. 1990, 112, 1190-
1197.
(46) Bartlett, P. G.; Benzing, E. P.; Pincock, R. E. J. Am. Chem. Soc.
1960, 82, 1762-1768.

Alkene Radical Cation/Phosphate Anion Pair Formation
J. Am. Chem. Soc., Vol. 123, No. 38, 2001 9245
Table 2. Intramolecular Capture of Radical Cation
fragmentation
(NMR/isolated
yield)
recovered
substrate
DBPO
(mol %)
thiol adduct
spirocycle
substrate
solvent
thiol
(NMR/isolated) (NMR/isolated) (NMR/isolated)
7
7
7
6
6
CD3OD
CD3CN
C6D6
C6D6
C6H6
150
150
150
100
100
2-mercaptoethanol
2-mercaptoethanol
2-mercaptoethanol
2-hydroxy-2-methylpropanethiol
2-mercaptoethanol
27 (95/76)
27 (12/7.3)
27 (24/2.4)
37 (15/3.5)
37 (44/1.8)
(53/12)
nd/nd
nd/nd
nd/nd
36 (nda/5.2)
32 (nda/3.3)
a The yields of 32 and 36 in the reaction could not be accurately determined because of the complexity of the mixtures; nevertheless, both 32
and 36 were the major products before attempted isolation.
1
8.07 (2 H, m), 8.10 (1 H, s), 8.80 (1 H, s), 9.25 (1 H, br s). 13C NMR
δ: 15.1, 16.6, 27.0, 30.7, 59.6, 64.7, 76.2, 82.1, 83.3, 88.7, 124.3, 128.4,
129.4, 133.5, 133.7, 143.6, 150.0, 152.0, 152.4, 164.9. 31P NMR δ:
-2.77, -2.73. 17O NMR δ: -77.35. HRESI-MS: calcd for C24H32N5O7-
PSNa, 588.1658; found, 588.1668 (M + Na)+.
Isomer 1. H NMR (CD3OD) δ: 2.20-2.25 (1 H, m), 2.32-2.40 (1
H, m), 2.50-2.60 (2 H, m), 2.74-2.78 (1 H, m), 2.80-2.90 (2 H, m),
3.16 (1 H, d, J ) 13.7 Hz), 3.90-3.95 (1 H, m), 4.25-4.28 (1 H, m),
6.50-6.55 (1 H, m), 7.55-7.60 (2 H, m), 7.63-7.67 (1 H, m), 8.05-
1
8.10 (2 H, m), 8.53 (1 H, s), 8.73 (1 H, s). Isomer 2. H NMR (CD3-
6-N-Benzoyl-9-(5-tert-butylthio-3,5-dideoxy-2-O-methyl-â-D-glyc-
ero-pent-3-enofuranosyl)adenine (41). 1H NMR δ: 1.38 (9 H, s), 3.38
(1 H, d, J ) 14.6 Hz), 3.47 (1 H, d, J ) 14.6 Hz), 3.52 (3 H, s), 4.77
(1 H, s), 5.38 (1 H, d, J ) 2.4 Hz), 6.66 (1 H, d, J ) 1.3 Hz), 7.54-
7.58 (2 H, m), 7.60-7.65 (1 H, m), 8.05-8.10 (2 H, m), 8.29 (1 H, s),
8.86 (1 H, s), 9.15 (1 H, br s). HRESI-MS: calcd for C22H25N5O3SNa,
462.1576; found, 462.1580 (M + Na)+.
OD) δ: 2.20-2.25 (1 H, m), 2.32-2.40 (1 H, m), 2.40-2.48 (2 H,
m), 2.74-2.78 (1 H, m), 2.80-2.90 (2 H, m), 3.13 (1 H, d, J ) 13.6
Hz), 3.90-3.95 (1 H, m), 4.25-4.28 (1 H, m), 6.35-6.45 (1 H, m),
7.43-7.45 (2 H, m), 7.55-7.60 (1 H, m), 7.95-8.00 (2 H, m), 8.20 (1
H, s), 8.25 (1 H, s). Unassigned Isomers 1 and 2. 13C NMR (CD3-
OD) δ: 25.6, 30.4, 30.5, 32.3, 35.9, 36.0, 63.1, 85.1, 85.6, 104.2, 105.1,
128.4, 128.6, 128.8, 129.5, 132.9, 140.1, 152.2, 152.9. HRFAB-MS:
calcd for C19H20N5O3S, 398.1287; found, 398.1280 (M + H)+.
2-(N-Benzoyl-9-adeninyl)-7,7-dimethyl-1,6-dioxa-9-thia-spiro[4,5]-
decane (36). The compound 36 was isolated in 5.2% yield by
preparative TLC on silica gel (eluent: 100/1/1 EtOAc/MeOH/Et3N).
1H NMR (CD3OD) δ: 1.59 (6 H, s), 2.05-2.15 (1 H, m), 2.42-2.48
(3 H, m), 2.77 (1 H, d, J ) 13.7 Hz), 2.70-2.90 (2 H, m), 3.03 (1 H,
d, J ) 13.7 Hz), 6.52-6.55 (1 H, m), 7.55-7.60 (2 H, m), 7.62-7.66
(1 H, m), 8.05-8.10 (2 H, m), 8.52 (1 H, s), 8.73 (1 H, s). HRFAB-
MS: calcd for C21H23N5O3SNa, 448.1419; found, 448.1417 (M + Na+).
6-N-Benzoyl-9-[5-deoxy-5-(2-hydroxyethylthio)-2-O-methyl-3-O-
diethylphosphoryl-r-L-lyxo-furanosyl]adenine (37). The product was
isolated by preparative TLC on silica gel (eluent: 20/1 CHCl3/MeOH).
1H NMR δ: 1.40-1.45 (6 H, m), 2.90-2.93 (2 H, m), 3.03 (1 H, dd,
J ) 5.4, 14.4 Hz), 3.19 (1 H, dd, J ) 6.7, 14.4 Hz), 3.53 (3 H, s),
3.80-3.95 (2 H, m), 4.20-4.30 (4 H, m), 4.58-4.63 (1 H, m),
4.87-4.92 (1 H, m), 5.18-5.23 (1 H, m), 6.14 (1 H, d, J ) 5.6 Hz),
7.55-7.60 (2 H, m), 7.64-7.68 (1 H, m), 8.05-8.10 (2 H, m), 8.40 (1
H, s), 8.85 (1 H, s). 31P NMR δ: -2.95. HRESI-MS: calcd for
C24H32N5O8PSNa, 604.1607; found, 604.1643 (M + Na+).
General Protocol for the Reaction of 2-Hydroxyalkanethiols with
Exocyclic Glycals 6 and 7 (Table 2). A solution of 6 or 7 (ca. 0.01
mmol) and hydroxythiol (30 equiv) in CD3OD or C6D6 (ca. 0.6 mL)
was degassed by sparging Ar at 0 °C for 5 min and then allowed to
warm to 40 °C under argon. DBPO (100 and 150 mol % for 6 and 7,
respectively) was then added, and the reaction mixture was stirred for
1
45 min at 40 °C. The resulting mixture was analyzed directly by H
NMR spectroscopy with 1,3,5-tri-tert-butylbenzene as internal standard.
6-N-Benzoyl-9-[3,5-dideoxy-2-O-methyl-5-(2-hydroxyethylthio)-
â-D-glycero-pent-3-enofuranosyl]adenine (27). The product was
isolated by preparative TLC on silica gel (eluent: 20/1 CHCl3/MeOH).
1H NMR δ: 2.81 (2 H, t, J ) 5.8 Hz), 3.35 (1 H, d, J ) 15.0 Hz), 3.42
(1 H, d, J ) 15.0 Hz), 3.50 (3 H, s), 3.79 (2 H, m), 4.82 (1 H, s), 5.36
(1 H, d, J ) 1.9 Hz), 6.61 (1 H, d, J ) 0.9 Hz), 7.50-7.55 (2 H, m),
7.60-7.63 (1 H, m), 8.02-8.07 (2 H, m), 8.34 (1 H, s), 8.83 (1 H, s),
9.08 (1 H, br s). 13C NMR δ: 28.2, 35.4, 56.6, 60.7, 88.8, 89.5, 98.7,
122.6, 128.0 (2 C), 128.9 (2 C), 133.0, 133.3, 140.3, 149.3, 151.1,
152.7, 160.3, 164.7. IR (film) ν: 2360, 2339, 1698, 1669 cm-1. HRESI-
MS: calcd for C20H21N5O4S, 427.1314; found, 427.1311 (M+).
2-(N-Benzoyl-9-adeninyl)-1,6-dioxa-9-thia-spiro[4,5]decane (32).
6 (31.7 mg, 0.067 mmol), 2-mercaptoethanol (141 µL, 2.01 mmol),
and DBPO (15.7 mg, 0.067 mmol) were dissolved in benzene (13.4
mL) at 40 °C. After 45 min, the resulting reaction mixture was separated
by preparative TLC on silica gel (eluent: 100/1/1 EtOAc/MeOH/Et3N),
resulting in the isolation of 32 in 3.3% yield as a mixture of isomers.
Acknowledgment. We thank the NIH (CA 60500) and NSF
(CHE 9986200) for partial support of this work.
JA010841L