Crystal structure and photochemical behavior in solution of the 3′-N-sulfamate analogue of thymidylyl(3′-5′)thymidine

Article abstract of DOI:10.1021/jo061488a

The 3′-N-sulfamate analogue of thymidylyl(3′-5′)thymidine (TnsoT, 1) exhibits a preference for a C3′-endo conformation in the solution and solid states. Its photochemical behavior in solution is compared to that of its natural counterpart, thymidylyl(3′-5

Full text of DOI:10.1021/jo061488a

Crystal Structure and Photochemical Behavior in Solution of the
3′-N-Sulfamate Analogue of Thymidylyl(3′-5′)thymidine
Ce´line Moriou,† Martial Thomas,† Marie-The´re`se Adeline,† Marie-The´re`se Martin,†
Ange`le Chiaroni,† Sylvie Pochet,‡ Jean-Louis Fourrey,† Alain Favre,§ and Pascale Clivio*,†
Institut de Chimie des Substances Naturelles, CNRS, 1 AVenue de la Terrasse,
91190 Gif sur YVette, France, Unite´ de Chimie Organique, URA 2128, Institut Pasteur, 25 rue du Docteur
Roux, 75724 Paris cedex, France, and Laboratoire de Photobiologie Mole´culaire, Institut Jacques Monod,
CNRS, UniVersite´s Paris 6 et Paris 7, 2 Place Jussieu, 75251 Paris cedex 05, France
pascale.cliVio@uniV-reims.fr
ReceiVed July 18, 2006
The 3′-N-sulfamate analogue of thymidylyl(3′-5′)thymidine (TnsoT, 1) exhibits a preference for a C3′-
endo conformation in the solution and solid states. Its photochemical behavior in solution is compared
to that of its natural counterpart, thymidylyl(3′-5′)thymidine (TpT, 2), to get further insight into the
significance of the C3′-endo conformation on the photoproduct formation at the single-stranded dinucleotide
level. Irradiation at 254 nm of 1 led to the same type of photoproducts as observed with 2. However, 1
was significantly more photoreactive than 2, and accordingly, the initial rate of photoproduct formation
was enhanced in accordance with its propensity to base stack compared to 2. The corresponding quantum
yields were determined and showed that the enhancement factor (1 compared to 2) is moderate for the
cyclobutane pyrimidine dimer (CPD) (1.26) and much higher for the (6-4) photoproduct (1.8). These
data strongly suggest that the CPD and (6-4) photoproduct arise from distinct minor stacked conformations.
Introduction
in denatured than in native DNA at the same temperature as
well as in film and frozen native DNA;3 and (2) this tendency
is reversed for the pyrimidine(6-4)pyrimidone photoproduct ((6-
4) PP) formation.3 Such an observation suggests either that CPD
formation requires more conformational freedom than (6-4) PP
formation or that the geometry of adjacent thymines in the solid
state (generally associated with the A-DNA conformation) is
more prone to undergo a Paterno´-Bu¨chi reaction (leading to
(6-4) formation) than the one achieved in solution. However,
recently, CPD formation was shown to be reduced in single-
and double-stranded DNA upon a temperature increase whereas
(6-4) PP formation was slightly increased.4 Finally, the A
The interplay between DNA conformation and its photo-
chemistry is among the most important issues to address with
regard to the respective contribution of each kind of DNA
photoproduct (PP) in photoimmune suppression, photocarcino-
genesis, apoptosis, or cell-cycle arrest.1,2 Previous studies
performed on the DNA double helix have evidenced two
interesting photochemical features at dithymine sites: (1) the
formation of the cyclobutane pyrimidine dimer (CPD) is higher
* Current address: FRE 2715 CNRS, Universite´ de Reims Champagne-
Ardenne, 51 rue Cognacq Jay, 51 096 Reims cedex, France.
† Institut de Chimie des Substances Naturelles.
‡ Unite´ de Chimie Organique, De´partement de Biologie Structurale et Chimie.
§ Laboratoire de Photobiologie Mole´culaire.
(3) Patrick, M. H.; Rahn, R. O. In Photochemistry and Photobiology of
Nucleic Acids: Photochemistry of DNA and Polynucleotides: Photoprod-
ucts; Wang, S. Y., Ed.: Academic Press: New York, 1976; Vol. 2, pp
35-95.
(1) (a) Ullrich, S. E. Front. Biosci. 2002, 7, d684-703. (b) Matsumura,
Y.; Ananthaswamy, H. N. Front. Biosci. 2002, 7, d765-783.
(2) Lo, H.-L.; Nakajima, S.; Ma, L.; Walter, B.; Yasui, A.; Ethell, D.
W.; Owen, L. B. BMC Cancer 2005, 5, 135-142.
(4) Douki, T. J. Photochem. Photobiol. B: Biol. 2006, 82, 45-52.
10.1021/jo061488a CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/07/2006
J. Org. Chem. 2007, 72, 43-50
43

Moriou et al.
SCHEME 1a
conformation of DNA or RNA, which is associated with more
nucleobase rigidity, reduces considerably the formation of
dithymine photoproducts with respect to the B conformation
of DNA.5 Despite these observations, no precise relationship
linking PP formation and DNA conformation has ever been
determined. An in-depth understanding of these relationships
at the single-stranded dinucleotide level could be helpful to
understand the phenomenon occurring at the more complex
double-stranded DNA level.
The sugar pucker is one of the major conformational
parameters that distinguishes B- from A-DNA.6a At the dinucle-
otide level, the lack of base-pairing and stacking interactions
with neighboring bases provides accurate information on the
intrinsic puckering effects on the photochemistry. Currently, we
investigate the photochemistry of dinucleotide analogues en-
dowed with an A sugar conformation (major C3′-endo pucker-
ing) and have recently demonstrated that the 254 nm irra-
diation of 2′-O,5-dimethyluridylyl(3′-5′)-2′-O,5-dimethyluridine
(TmopTmo), an analogue of thymidylyl(3′-5′)thymidine (TpT,
2) exhibiting a major C3′-endo conformer population, is
qualitatively comparable to that of 2 but significantly faster.7
These results were rationalized by the increased propensity of
the modified dinucleotide to adopt a base-stacked conformation
geometry reminiscent of that of TpT.7
a Reagents and conditions: (a) 4 Å molecular sieves, Et3N, CH2Cl2, 68%;
(b) MeOH, Et3N, H2O (10:1:1); (c) CH2Cl2, 80% aq AcOH, 82% (two
steps); (d) 3% CF3COOH, CH2Cl2, 92%.
To probe the generality of this observation, we decided to
examine the photochemical behavior of other analogues of TpT
possessing a major C3′-endo sugar pucker and an increased base
stacking ability compared to 2. To get deeper insight into the
impact of the C3′-endo sugar pucker on the photochemistry, it
is crucial to investigate TpT analogues that are not modified at
the C2′-position. In this respect, the influence of the modification
of the phosphate backbone on the photochemistry looked
particularly attractive. The 3′-N-sulfamate analogue of thymi-
dylyl(3′-5′)thymidine (TnsoT, 1), in which the phosphodiester
linkage is replaced by a sulfamate group, has been reported to
possess the two above conformational criteria in solution.8
Because the X-ray crystallographic study that we carried out
confirmed this propensity, we decided to study its photochemical
behavior at 254 nm. We herein report our results and compare
them with those of the reference compound TpT (2).
FIGURE 1. ORTEP drawing of the X-ray structure of molecule
(AB) 1.
4-Nitrophenyl 5′-O-(4,4′-dimethoxytrityl)-3′-deoxythymidine-
3′-sulfamate 38 was reacted with 4 to yield the orthogonally
diprotected dimer intermediate 5 (68%). Subsequent deacety-
lation (MeOH, Et3N, H2O) and detritylation (80% aq acetic acid)
afforded 1 in 82% yield (Scheme 1).
X-Ray Crystallographic Structure of 1. Upon standing in
methanol, slowly occurring transesterification of 6 afforded 1
which crystallized in the chiral space group P1 of the triclinic
system, with two independent TnsoT molecules, AB and CD,
in the unit cell. One of them, AB, appears in Figure 1, giving
the atomic labeling, with the known absolute configuration.
These two molecules are hydrogen bonded together through the
3′OH-group of the first molecule (AB) and the C2-oxygen of
Results
Synthesis of 3′-Deoxythymidin-3′-ylsulfamoyl-[3′(N)->5′-
(O)]-thymidine (1). Compound 1 was prepared following a
slightly modified literature procedure (3′-O-acetylthymidine 4
was used in place of 3′-O-(tert-butyldimethylsilyl)thymidine).8
(5) (a) Becker, M. M.; Wang, Z. J. Biol. Chem. 1989, 264, 4163-4167.
(b) Kundu, L. M.; Linne, U.; Marahiel, M.; Carell, T. Chem.-Eur. J. 2004,
10, 5697-5705.
(6) (a) Saenger, W. In Principles of Nucleic Acid Structure; Cantor, C.
R., Ed.; Springer-Verlag: New York, 1984. (b) Altona, C.; Sundaralingam,
M. J. Am. Chem. Soc. 1972, 94, 8205-8212.
(8) (a) Fettes, K. J.; Howard, N.; Hickman, D. T.; Adah, S. A.; Player,
M. R.; Torrence, P. F.; Micklefield, J. Chem. Commun. 2000, 765-766.
(b) Fettes, K. J.; Howard, N.; Hickman, D. T.; Adah, S.; Player, M. R.;
Torrence, P. F.; Micklefield, J. J. Chem. Soc., Perkin Trans. I 2002, 485-
495.
(7) Ostrowski, T.; Maurizot, J.-C.; Adeline, M.-T.; Fourrey, J.-L.; Clivio,
P. J. Org. Chem. 2003, 68, 6502-6510.
44 J. Org. Chem., Vol. 72, No. 1, 2007

3′-N-Sulfamate Analogue of Thymidylyl(3′-5′)thymidine
1
8) as evidenced by HPLC and H NMR analysis of the crude
irradiation mixture. Photoproduct 7 showed an UV maximum
absorption at 327 nm in accordance with a (6-4) PP structure,10
whereas photoproduct 8 lacked UV absorbance above 240 nm
1
as expected for a CPD PP. H NMR analysis of the crude
irradiation mixture confirmed that the two PPs indeed belonged
to the (6-4) and CPD PP classes, respectively.10,11 The presence
of two H6 protons at δ 8.08 and 5.22 ppm and of two methyl
protons at δ 2.35 and 1.77 ppm in the 1H NMR was the signature
of the C5 (R) and C6 (S) stereochemistry of 7 as observed for
11, the (6-4) adduct of TpT.10 In contrast, the stereochemistry
of 8 was more ambiguous to assign from the chemical shifts of
its two shielded methyl signals (δ 1.59 and 1.48).11 Therefore,
we carried out structural analysis studies to unambiguously
establish the stereochemistry of the CPD derived from TnsoT.
Assignment of the Stereochemistry of the Photoproducts.
Preparative irradiation was conducted on the 3′-O-acetyl deriva-
tive 6, obtained in 92% yield upon TFA treatment of 5 (Scheme
1), to ease the purification step. Irradiation of an aqueous
solution of 6 at 254 nm led to the formation of a (6-4) PP 9
and a CPD PP 10 isolated by reversed-phase HPLC in 19 and
10% yield, respectively, together with 28% of recovered 6.
Attribution of the stereochemistry of the asymmetric centers of
10 was performed using 2D NMR techniques. The cis-syn (c,s)
stereochemistry of 10 was deduced from the NOEs between
H6 Tnso- and H3′ Tnso-, between CH3 Tnso- and H6 -nsoT,
between H6 -nsoT and H2′ -nsoT, and between CH3 -nsoT and
CH3 Tnso-. NOEs observed for 9 confirmed the C5 (R) and C6
(S) stereochemistry.
Thus, 9 and 10 arise from cycloaddition reactions between
the two thymine bases of 6 in an anti glycosidic bond confor-
mation as in the case of the natural dinucleotide 2. To demon-
strate that the same stereochemical course of the photoreactions
had occurred upon irradiation of 1, we irradiated a mixture of
1 and 6 and observed, upon alkaline treatment (see Experimental
Section), the coalescence of the peaks corresponding to the CPD
and (6-4) PPs of 6 with those of 1 (photoproducts 7 and 8).
Therefore, the stereochemistry of PPs 7 and 8 is identical to
that of PPs 11 and 12 derived from TpT (Scheme 2).
Comparative Photoreactivity of 1 and 2. From the data
above, it can be concluded that 1 and 2 generate the same major
photoproducts. To closely compare their photochemical behav-
ior, an aqueous equimolar mixture of 1 and 2 was irradiated at
254 nm and the crude irradiation mixture was analyzed by RP
HPLC at various times of irradiation (see Supporting Informa-
tion). As the reactants and photoproducts were well resolved
on the HPLC chromatograms, each individual peak (230 nm)
could be accurately integrated, yielding the corresponding ∆Si
values. In the following, 1 and 2 will be abbreviated R1 and
R2 and their major photoproducts will be abbreviated C1, P1
and C2, P2 (C for CPD and P for (6-4) PP), respectively.
Kinetics of Photoproduct Formation. The fractional amounts
ZR1 and ZR2 of TnsoT (1) and TpT (2), respectively, at time t
of the reaction are given by eq 1
FIGURE 2. View of the geometry of molecule (AB) nucleobases,
projected on the A thymine plane.
TABLE 1. Backbone Torsion Angles r(i), â(i), γ(i), δ(i), E(i), and ú(i)
of the Sugar of Residue ia in the X-ray Structure of Molecule (AB) 1
γA
δA
ꢀA
úA
RB
âB
γB
δB
1 (AB) 51.9 81.8 -170.5 -78.2 -60.2 -179.9 62.9 148.9
a i ) A corresponds to the 5′-residue, and B corresponds to the 3′-residue.
the second one (CD) according to the scheme: O3′B-H‚‚‚O2C
) 3.010(3) Å, H‚‚‚O2C ) 2.35 Å, angle O-H‚‚‚O ) 138.4°.
Molecule CD can be fitted on molecule AB by a rotation of
180° around the a cell axis direction followed by a translation
of 4.117 Å along this axis. The distance between their respective
centroids is 8.781 Å. These two molecules display an extremely
similar conformation as shown by the root mean square (rms)
difference between their atomic coordinates (excluding H atoms)
of 0.207 Å, leading to the respective bond and angle rmsd of
0.010 Å and 0.77°. So, only the conformational parameters of
molecule AB will be discussed (a complete description of the
two molecules is reported in the Supporting Information).9
In this molecule (AB), the two thymine bases are in an anti
glycosidic bond conformation where øA ) -140.1° and øB )
-105.7°. The 5′-sugar exhibits a C3′-endo conformation (P )
12.6° and τm ) 33.6°), whereas the 3′-sugar displays a C2′-
endo conformation (P ) 179.7° and τm ) 35.7°).6 These two
bases are nearly parallel as revealed by the dihedral angle of
their plane (8.2°). The angle of vectors along the two C5-C6
bonds of 8.75° indicates that the corresponding bonds are nearly
parallel. Consequently, no twist exists between the bases, as in
A- or B-DNA. The distance Cg(A)-Cg(B) between the ring
centroids of the two bases is 5.603(7) Å. With the perpendicular
distance of Cg(A) on ring B being 4.020 Å and that of
Cg(B) on ring A being 4.388 Å, the two mean planes can be
considered distant from 4.204 Å. However, the two thymine
bases are not superimposed and only the methyl C7B and the
carbon atom C4B nearly overlap with atoms N1A and O2A,
leading to a relative gliding of 3.702 Å between the rings, as
shown in Figure 2.
The backbone torsion angles are reported in Table 1. The
torsion angle ú (C3′-N-S-O) is in the g- conformation
(-78.2°).
Photochemical Behavior of 1. Irradiation of an aqueous
solution of 1 at 254 nm led to the formation of two PPs (7 and
∆SR1
∆S01
∆SR2
∆S02
ZR1
)
ZR2
)
(1)
(9) CCDC 612018 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge at www.ccdc.ca-
m.ac.uk/conts/retrieving.html [or from the Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB 1EZ, U.K.; fax/(internat.) +44-
1223/336-033; E-mail deposit@ccdc. cam. ac.uk].
(10) Rycyna, R. E.; Alderfer, J. L. Nucleic Acids Res. 1985, 13, 5949-
5963.
(11) Koning, T. M. G.; van Soest, J. J. G.; Kaptein, R. Eur. J. Biochem.
1991, 195, 29-40.
J. Org. Chem, Vol. 72, No. 1, 2007 45

Moriou et al.
SCHEME 2
TABLE 2. Molar Extinction Coefficient E (M-1 cm-1) of TpT and
for the formation of the primary TpT PPs can be written as
Its Major PPsa
P2 7Φ R2 {ΦΦC2} C2
P2
TpT (2)
(6-4) PP (11)
c,s CPD (12)
2-
ꢀ230
5300
7300
4000
a Derived from Johns et al. data.12
where ΦP2 and ΦC2 (ΦC2 ) ΦC2c,s + ΦC2t,s I) are, respectively,
the quantum yields for the formation of P2 (the (6-4) PP) and
C2 (C2 ) C2c,s + C2 trans-syn I (t,s I); the c,s CPD of 2 and
the t,s I CPD isomer of 2 (13),13 respectively). This reaction
scheme is justified by the fact that Φ2- (the CPD reversion yield)
is identical for both the c,s and t,s I CPD isomers with the
consequence that their molar ratio t,s I over c,s (whether
originating from 1 (for C1t,s I, see structure 14)13 or 2) remains
practically constant (between 0.05 and 0.1) during the course
of the photoreaction as initially observed by Johns et al.12 The
scheme above, however, is valid only for short irradiation times
(t e 5 min) as secondary photoproducts such as the Dewar
isomer of P2 will then accumulate.
where ∆S01 (∆S02) and ∆SR1 (∆SR2) are the peak areas (at 230
nm) corresponding to R1 (R2) at, respectively, time 0 and t of
the photoreaction. Practically at t ) 0, ∆S01 ≈ ∆S02.
Similarly, the fractional amounts ZC1 and ZC2 of the major
c,s CPD of 1 and 2 are
R1
C1 230
R2
C2 230
∆S
∆S
ꢀ
∆S
∆S
ꢀ
ZC1
)
ZC2
)
(2)
C1
C2
01 ꢀ230
02 ꢀ230
and the fractional amounts ZP1 and ZP2 of the (6-4) PP of 1 and
2 are
R1
P1 230
R2
P2 230
∆S
∆S
ꢀ
∆S
∆S
ꢀ
ZP1
)
ZP2
)
(3)
P1
P2
01 ꢀ230
02 ꢀ230
In the following, we assumed that the ꢀ at 230 nm of R1 and
R2 and of their corresponding PPs within the same class are
identical (ꢀX1 ) ꢀ2X320). Such an approximation is justified, at
230
least for R1 and R2, by the fact that their absorption spectra
are superimposable between 210 and 300 nm. The ꢀ values used
are derived from Johns et al. data (Table 2).12
The fractional amount of 1 and 2 and of their respective PPs
as a function of the irradiation time is shown in Figure 3. The
decay of TpT and the formation of its PPs in our experiment
follow kinetics closely similar to the one reported by Johns et
al.12 upon 254 nm irradiation of TpT under comparable
The rate of formation, dP2/dt (mol min-1), of P2 is given by
eq 4:
DR2
conditions. In particular, the c,s CPD PP increases up to ZC2
)
dP2
dt
) ΦP2(1 - 10-D)IoS
(4)
0.225 (Figure 3), compared to 0.210.12 This fully justifies the
use of TpT as an internal control to evaluate the photoreactivity
of its 3′-N-sulfamate analogue TnsoT. Obviously, from Figure
3, it appears that R1 (TnsoT, 1) is more photoreactive than R2
(TpT, 2) as it decays faster and the initial rates of formation of
the major R1-derived PPs are higher.
D
where D is the total optical density of the solution at 254 nm
and DR2 is the corresponding optical density due to R2. The
term IoS(1 - 10-D)DR2/D thus represents the 254 nm light dose
rate (einsteins min-1) absorbed by R2 at time t. For t e 5 min
of the photoreaction, D remains sufficiently large (D g 2) so
that the product F ) (1 - 10-D)IoS can be considered as a
constant. This leads to eq 4′:
Quantum Yield Determination. The method that we de-
veloped determines the quantum yield of PP formation with
respect to those obtained for TpT.12 From extensive studies on
TpT photochemistry, the simplest reaction scheme accounting
(13) t,s I CPDs were tentatively identified from their retention time and
UV spectrum. As they were minor compounds, they were not investigated
further.
(12) Johns, H. E.; Pearson, M. L.; LeBlanc, J. C.; Helleiner, C. W. J.
Mol. Biol. 1964, 9, 503-524.
46 J. Org. Chem., Vol. 72, No. 1, 2007

3′-N-Sulfamate Analogue of Thymidylyl(3′-5′)thymidine
FIGURE 3. Fractional amount of the 3′-N-sulfamates TnsoT (1) and TpT (2) and of their respective photoproducts as a function of irradiation
time.
D*R2 DC2
DR2
dP2
dt
DR2
D
dC2
dt
) ΦP2F
(4′)
) FΦC2
1 -
(5′′)
D
*
[
DR2
DC2
]
Taking into account the fact that DR1 and DR2 are the main
components of D, eq 4′ can be integrated leading to a first
approximation of eq 4′′:
It should be noticed that because Φ2- is identical for CPD
isomers eqs 5, 5′, and 5′′ remain valid for each CPD isomer.
The ratio ΦC1c,s/ΦC2c,s is then obtained as described in the
Experimental Section. The ratio of quantum yields and their
absolute values taking the Φ values determined by Johns et al.12
as reference yields are listed in Tables 3 and 4, respectively.
Consequently, the formation of both (6-4) PP and CPD is
increased in the 3′-N-sulfamate series compared to TpT.
Interestingly, the observed enhancement is significantly higher
for the (6-4) PP than for the CPD PP (1.5 times).
DR2
P2 ) tFΦP2
(4′′)
D
where DR2 and Dh are the mean values of DR2 and D in the
interval 0,t. Practically, DR2 and Dh are the values of DR2 and D
at t/2. The ratio of the quantum yields of formation of P2 and
P1 is given by eq 4′′′:
Discussion
In dilute solutions, TnsoT (1) and TpT (2) are expected to
adopt a large variety of conformations, whereas in their
respective crystal structures, only one of these conformations
is captured. Nevertheless, the data obtained so far indicate that
characteristics observed in the solid state are generally repre-
sentative of the major features observed in solution. For
example, in the crystal structure of TnsoT reported here, the
5′- and 3′-sugars adopt a C3′-endo and C2′-endo puckering,
respectively, in agreement with the NMR conformational
analysis of 1 indicating that the 5′- and 3′-sugars exist in a major
C3′-endo (69%) and a major C2′-endo (56%) conformation,
respectively.8b Similarly, in the pTpT crystal structure,14 both
the 5′- and 3′-sugars are C2′-endo, the conformations that largely
predominate at both ends in TpT in solution.15 The observed
ΦP1
DR2
DR1
P1
)
(4′′′)
ΦP2 P2
The rate dC2/dt of formation of the CPDs of R2 is depicted by
eq 5:
D
DC2
D
) F ΦC2 R2 - Φ2-
(5)
dC2
dt
[
]
D
where DC2 refers to the 254 nm absorbance due to CPDs. At
time t* ≈ 10-15 min, C2 reaches its maximal amount, C2*
(Figure 3), and thus eq 5′ can be written as
ΦC2D*R2 ) Φ2-D*C2
leading finally to eq 5′′:
(5′)
(14) Camerman, N.; Fawcett, J. K.; Camerman, A. J. Mol. Biol. 1976,
107, 601-621.
(15) Cheng, D. M.; Sarma, R. H. J. Am. Chem. Soc. 1977, 99, 7333-
7348.
J. Org. Chem, Vol. 72, No. 1, 2007 47

Moriou et al.
TABLE 3. Quantum Yield Ratios of PP Formation and CPD
crystal structure of 1 reported here. Examination of its stacking
geometry (Figure 2) indicates that none of the sets of bonds
involved in PP formation (C5-C6 of A and C5-C6 of B or
C5-C6 of A and O4-C4 of B for the CPD or (6-4),
respectively) are adequately positioned for a [2+2] cycloaddition
to occur (even though these bonds are nearly parallel, they are
too far in space (ca. 5.5 Å) for bond formation). This structure
is thus typical of a stacked nonreactive conformation. Accord-
ingly, the stacked fraction of 1 or 2 will comprise both
nonproductive and minor photoreactive conformers. It can be
concluded that the predominant C3′-endo sugar pucker at the
5′-end favors stacking but increases (6-4) PP formation more
significantly than CPD formation. These data strongly suggest
that distinct minor conformers lead, respectively, to the forma-
tion of (6-4) PPs and CPDs. The results obtained by the present
study demonstrate that photochemistry is an efficient method
to evidence the presence of minor but highly reactive conforma-
tions, whereas circular dichroism data reflect only average
conformation properties.
The question that needs to be addressed at this stage is
whether the ground-state minor active conformers are directly
involved in the formation of the corresponding photoproducts.
Excitation is fast (10-15 s),17 and the first excited singlet state
of thymine thought to be at the origin of CPD and (6-4) PP
formation has a relatively short lifetime (10-12 s)18 compared
to the motion required to interconvert an active stacked
conformer into either the corresponding unstacked form or
another productive or nonproductive stacked form. Therefore,
CPD and (6-4) PP formation is highly likely to derive essentially
from the excitation of their corresponding minor active con-
formers. Clearly, the study of the dynamics of the stacked forms
of 1 and 2 will help to clarify this point.
Reversion between the 3′-N-Sulfamate and the Natural Series
Φ
C1c,s/ΦC2c,s
Φ
P1/ΦP2
Φ
1- /Φ2-
1.26 ( 0.08
1.8 ( 0.2
0.9 ( 0.1
TABLE 4. Quantum Yields of PP Formation and CPD Reversion
in the Natural12 and the 3′-N-Sulfamate Series
ΦP
ΦCc,s
Φ-
0.75
0.67 ( 0.08
TpT11
TnsoT
10-3
(1.8 ( 0.2) 10-3
1.1 ( 10-2
(1.38 ( 0.06) 10-2
conformational shift of the 5′-sugar toward C3′-endo puckering
of 1 relative to 2 (∆ ) 33%) has been attributed to a reduction
in the gauche effect between the 3′-substituent and the O4′ atom,
given that the 3′-amino group in 1 is less electronegative than
the corresponding 3′-oxygen atom of 2.8b On the other hand,
circular dichroism data indicate that the estimated fraction of 1
present in solution in the stacked form is two times higher than
that observed with 2.8b In their respective crystal structures,
intramolecular base stacking is clearly seen in 1 whereas in 2
the two bases are too far apart.
Thus, both crystal and solution data suggest that increasing
the population of C3′-endo sugar puckering favors base stacking
as was previously reported in the case of TmopTmo.7 The
stacking process could be mediated by the torsion angle ú
(C3′-N-S-O in the present case) which is driven by the sugar
pucker-dependent torsion angle δA.7 Hence, a C3′-endo con-
formation would lead to stacking whereas a C2′-endo conforma-
tion would lead to destacking. The crystal structure of 1 shows
that both the C3′-endo conformation of the 5′-sugar and the
g- conformation of ú is associated with the stacked conforma-
tion. Interestingly, in the case of the pTpT crystal structure,14
the reverse phenomenon is observed: the extended form, as a
result of the trans conformation of the O3′-P bond, is associated
with a C2′-endo puckering of the 5′-sugar. Consequently, it is
likely that, in the phosphate and 3′-N-sulfamate series at least,
only the 5′-sugar conformation is important for the stacking/
destacking process.
The comparative irradiation experiment (Figure 3) shows that
the initial rate of TnsoT (1) decay is roughly 1.35 times faster
than that observed with TpT (2). This increased photoreactivity
correlates with the higher propensity of 1 for base stacking than
2. This is in line with our observations indicating that increasing
the C3′-endo population of the sugar pucker of TpT by a
chemical modification (2′-R-methoxylation or 3′-N-sulfamate
analogue) results in an enhancement of base stacking and hence
of the photoreactivity. In addition, the study herein indicates
that the modification of the 5′-sugar conformation is sufficient
to achieve this goal.
The relationship between the fraction of stacked species in 1
and 2 evaluated by circular dichroism (to be 2 times higher for
1 than for 2) and the increased photoreactivity (1.35) is however
not direct and should be more closely examined. Both 1 and 2
yielded the same types of photoproducts in solution, the CPD
and (6-4) PP. Determination of their quantum yields of formation
using 2 as a reference shows that both products form more
efficiently from 1 than from 2 (Table 2). These results fully
corroborate previous observations linking stacking ability and
PP formation when temperature and solvent were varied.16 In
addition, the enhancement factor is significantly higher for the
(6-4) PP than for the CPD (Table 3). A clue to further explore
the relationship between stacking and photoreactivity is the
Interestingly, when the TpT sequence is embedded in an RNA
duplex, its photoreactivity is dramatically reduced as compared
to the one observed in B-DNA helix.5b This has been attributed
to an increased rigidity of the double helix in the A conformation
with respect to the B conformation which would preclude the
prerequisite orientation for PP formation to occur. Indeed, in
our context study, base pairing is absent as is the influence of
neighboring nucleobases and these differences are likely to
explain the different photochemical results.
Conclusion
Our results demonstrate that an increase of the C3′-endo sugar
pucker of thymine dinucleotide analogues, through enhancement
of the base stacking process, provokes an increased photore-
activity and an enhancement of PP quantum yields. The
modification of the 5′-residue only appears to be sufficient to
obtain the modification of the stacking status. Interestingly, the
3′-N-sulfamate backbone represents the first type of chemically
modified dinucleotide analogue, reported so far, able to favor
the (6-4) PP formation compared to the natural phosphate
backbone. A search for modified thymine dinucleotide analogues
leading to the selective and efficient (6-4) PP formation would
(16) (a) Hosszu, J. L.; Rahn, R. O. Biochem. Biophys. Res. Commun.
1967, 29, 327-330. (b) Wacker, A.; Lodemann, E. Angew. Chem. 1965,
77, 133. (c) Tramer, Z.; Wierzchowski, K. L.; Shugar, D. Acta Biochim.
Pol. 1969, 16, 83-107.
(17) Turro, N. J. In Molecular Photochemistry; Breslow, R., Karplus,
M., Eds.; Benjamin, W. A. Inc.: New York, 1967.
(18) Ruzsicska, B. P.; Lemaire, D. G. E. In CRC Handbook of Organic
Photochemistry and Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC
Press: Boca Raton, 1995; pp 1289-1317.
48 J. Org. Chem., Vol. 72, No. 1, 2007

3′-N-Sulfamate Analogue of Thymidylyl(3′-5′)thymidine
3′-Deoxythymidin-3′-ylsulfamoyl-[3′(N)->5′(O)]-thymidine 1.
4-Nitrophenyl 5′-O-(4,4′-dimethoxytrityl)-3′-deoxythymidine-3′-
sulfamate 38b (697 mg, 0.93 mmol) was condensed to 3′-O-
acetylthymidine 4 (500 mg, 1.76 mmol) according to the pub-
lished procedure.8b The dimer intermediate 5 (566 mg) was ob-
tained in 68% yield. This intermediate (176 mg, 0.19 mmol) was
deacetylated in a mixture of MeOH/TEA/H2O (8:1:1) (10 mL)
for 20 h at room temperature and then for 4 h at 50 °C and
then concentrated to dryness. The residue was solubilized in
CH2Cl2 (3 mL). Then 80% aqueous acetic acid (5 mL) was added,
and the mixture was stirred at room temperature for 3 h. The
reaction mixture was diluted with H2O (30 mL). The aqueous
phase was washed with CH2Cl2 (30 mL), concentrated, and purified
by reversed-phase chromatography (LiChroprep RP-18) using a
gradient of CH3CN in H2O (0-20%) to give 18b in 82% yield (two
steps).
greatly facilitate the chemical synthesis of oligonucleotides
containing (6-4) PP surrogates. Such compounds are steadily
required for biological studies. In addition, 1 belongs to the
promising 3′-N-sulfamate class of potential antisense and
antigene agents. The report of its X-ray crystallographic structure
should be helpful to the antisense and antigene chemist
community.
Experimental Section
Irradiation Conditions. For analytical studies, an aqueous
solution (1500 µL, HPLC grade) of compound 1 (ODmax ) 6.3) or
a 1:1 mixture of compounds 1 and 2 (ODmax ) 6.3 for each solution)
was prepared in a quartz cuvette (0.5 cm optical path). The oxygen
of the solution was removed by argon bubbling for 30 min, and
then the solution was exposed to the 254 nm light source (2 × 15
W, VL 215C Lamp, Vilber Lourmat, Marne la Valle´e, France).
An aliquot of the solution was sampled at t ) 0, 0.75, 2.5, 5, 10,
20, and 30 min and analyzed by reversed-phase HPLC. For
preparative purposes, an aqueous solution of 6 (115 mg in 230 mL)
was exposed for 2 h to the 254 nm light (12 × 8 W; T8C UVC 7H
8W Lamps, Vilber Lourmat, Marne la Valle´e, France). The
photolysate was concentrated to dryness and purified by reversed-
phase HPLC.
HPLC Conditions. Analytic HPLC. A portion of 25 µL of the
irradiation mixture was injected on a SYMMETRY C18 (5 µm,
4.6 × 250 mm) column using a 50 min, 1 mL/min gradient of
0-15%, then 5 min of 15-20% CH3CN in 0.05 M aqueous
ammonium acetate. A photodiode array detector was used. Peak
areas were measured at 230 nm. Retention time (min): TnsoT
series, (6-4) PP 7 17.6, CPD PP 8 21.4, TnsoT 1 42.1; TpT series,
CPD PP 12 7.1, (6-4) PP 11 10.1, TpT 2 28.1.
Preparative HPLC. Preparative purification was performed on
an Hypersil HSC18 (5 µm, 100 Å, 21.2 × 250 mm) column using
a 45 min, 12 mL/min linear gradient of 0-25% CH3CN in H2O.
The detection was set at 230 nm. The (6-4) PP 9 (22 mg, 19%),
the CPD PP 10 (12 mg, 10%), and the recovered starting material
6 (32 mg, 22%) were obtained after evaporation of the appropriate
fractions. Retention time (min): (6-4) PP 9, 31; CPD PP 10, 46;
compound 6, 50.
Chemical Correlation between PPs of 6 and 2. An aqueous
solution of a 1:1 mixture of compounds 6 and 2 (ODmax ) 6.3 for
each solution) was exposed for 10 min to the 254 nm light source.
An aliquot of the solution was analyzed by HPLC. Then TEA was
added to the photolysate, and the mixture was stirred for 5 min.
An aliquot of the solution was analyzed by HPLC.
3′-Deoxythymidin-3′-ylsulfamoyl-[3′(N)->5′(O)]-3′-O-acetylth-
ymidine 6. To the dimer intermediate 5 (440 mg, 0.49 mmol) was
added a solution of 3% CF3COOH in CH2Cl2 (40 mL). The reaction
was stirred for 10 min at room temperature, and then MeOH was
added. The solvants were evaporated to give a residue that was
purified by silica gel chromatography using a gradient of MeOH
in CH2Cl2 (0-10%) give 6 in 92% yield (266 mg). 1H NMR (600
MHz, D2O): δ 1.91 (3H, s, CH3-5 pT), 1.92 (3H, s, CH3-5 Tp),
2.19 (3H, s, COCH3), 2.47-2.56 (2H, m, H-2′ and H-2′′Tp), 2.50-
2.60 (2H, m, H-2′ and H-2′′ pT), 3.84 (1H, dd, J ) 13.0, 3.7 Hz,
H-5′ Tp), 3.96 (1H, dd, J ) 13.0, 2.5 Hz, H-5′′ Tp), 4.04-4.10
(2H, m, H-3′ and H-4′ Tp), 4.45-4.51 (3H, m, H-4′, H-5′ and H-5′′
pT), 5.43 (1H, ddd, J ) 6.8, 3.4, 3.4 Hz, H-3′ pT), 6.13 (1H, dd,
J ) 6.8, 4.9 Hz, H-1′ Tp), 6.37 (1H, dd, J ) 7.0, 7.0 Hz, H-1′ pT),
7.62 (1H, s, H-6 pT), 7.75 (1H, s, H-6 Tp). IR (film, cm-1) 3410,
3207, 3066, 1692, 1467, 1361, 1238, 1172, 1097, 1044, 957. HRMS
(ES) (M + Na)+ calcd for C22H29N5O12NaS, 610.1431; found,
610.1456.
1
Cyclobutane Photoproduct 10. H NMR (600 MHz, D2O): δ
1.48 (3H, s, CH3-5 pT), 1.59 (3H, s, CH3-5 Tp), 2.12 (3H, s,
COCH3), 2.25 (1H, br ddd, J ) 14.4, 7.8, 7.2 Hz, H-2′′ Tp), 2.29
(1H, ddd, J ) 13.9, 4.8, 1.6 Hz, H-2′′ pT), 2.51 (1H, ddd, J )
13.9, 10.3, 7.3 Hz, H-2′ pT), 3.00 (1H, ddd, J ) 14.4, 7.2, 5.2 Hz,
H-2′ Tp), 3.75 (1H, dd, J ) 13.6, 2.9 Hz, H-5′ Tp), 3.85 (1H, ddd,
J ) 13.6, 3.6 Hz, H-5′′ Tp), 3.94 (1H, ddd, J ) 6.2, 3.5, 2.9 Hz,
H-4′ Tp), 4.00 (1H, ddd, J ) 7.8, 7.2, 6.2 Hz, H-3′ Tp), 4.30 (1H,
m, H-4′ pT), 4.37 (1H, d, J ) 6.4 Hz, H-6 Tp), 4.44-4.49 (2H, m,
H-5′ and H-5” pT), 4.57 (1H, d, J ) 6.4 Hz, H-6 pT), 5.29 (1H,
ddd, J ) 7.3, 3.2, 1.6 Hz, H-3′ pT), 5.93 (1H, dd, J ) 10.3, 4.8
Hz, H-1′ pT), 5.96 (1H, br dd, J ) 7.2, 5.2 Hz, H-1′ Tp). 13C NMR
(150 MHz, D2O): δ 17.8 (CH3 pT), 18.1 (CH3 Tp), 21.5 (COCH3),
34.0 (C-2′ Tp), 35.0 (C-2′ pT), 47.0 and 54.2 (C-5 Tp and pT),
53.5 (C-3′ Tp), 56.5 (C-6 pT), 60.6 (C-6 Tp), 61.1 (C-5′ Tp), 70.6
(C-5′ pT), 73.6 (C-3′ pT), 80.7 (C-4′ pT), 82.4 (C-4′ Tp), 86.9 (C-
1′ Tp), 87.5 (C-1′ pT), 155.5 (C-2 Tp), 156.1 (C-2 pT), 172.8 (C-4
Tp), 174.8 (COCH3), 175.3 (C-4 pT). IR (film, cm-1) 3427, 3233,
3075, 1705, 1454, 1366, 1282, 1207, 1177, 1089, 1036, 952. HRMS
(ES) (M + Na)+ calcd for C22H29N5O12NaS, 610.1431; found,
610.1426.
Quantum Yield Determination. Quantum yields were deter-
mined using the assumption that ꢀX1 ) ꢀX2 at both 230 and 254
nm. The ratios can then be expressed directly from experimental
data.
ΦP1 ∆SP1 ∆SR2
)
(1)
(2)
ΦP2 ∆SP2
∆SR1
ΦC1 ∆SC1 ∆SR2 F2
∆S*R2 ∆S
∆S*
)
with F2 ) 1 -
C2 and
(6-4) Photoproduct 9. 1H NMR (600 MHz, D2O): δ 1.49 (1H,
ddd, J ) 13.9, 6.7, 1.0 Hz, H-2′ Tp), 1.77 (3H, s, CH3-5 Tp), 2.08
(1H, ddd, J ) 13.9, 13.1, 8.4 Hz, H-2′′ Tp), 2.16 (3H, s, COCH3),
2.35 (3H, s, CH3-5 pT), 2.86 (1H, ddd, J ) 15.4, 8.0, 7.2 Hz, H-2′′
pT), 3.16 (1H, ddd, J ) 15.4, 7.4, 2.5 Hz, H-2′ pT), 3.24 (1H,
ddd, J ) 13.1, 9.9, 6.7 Hz, H-3′ Tp), 3.30 (1H, ddd, J ) 9.9, 3.2,
2.3 Hz, H-4′ Tp), 3.82 (1H, dd, J ) 13.1, 3.2 Hz, H-5′ Tp), 4.01
(1H, dd, J ) 13.1, 2.3 Hz, H-5′ Tp), 4.04 (1H, dd, J ) 11.3, 2.0
Hz, H-5′ pT), 4.47 (1H, dd, J ) 11.3, 1.4 Hz, H-5′ pT), 4.49 (1H,
m, H-4′ pT), 5.22 (1H, s, H-6 Tp), 5.67 (1H, ddd, J ) 7.4, 7.2, 4.2
Hz, H-3′ pT), 6.23 (1H, dd, J ) 8.4, 1.0 Hz, H-1′ Tp), 6.59 (1H,
dd, J ) 8.0, 2.5 Hz, H-1′ pT), 8.08 (1H, s, H-6 pT). 13C NMR
ΦC2 ∆SC2
F1
∆SR1
C2 ∆SR2
∆S*R1 ∆SC1
∆S*
F1 ) 1 -
C1 ∆SR1
Φ1-
Φ2-
Φ
Φ
C1 ∆S*R1 ∆S*
)
C2 where ∆SR2 ) ∆SR2 measured at t/2 (3)
*
*
C2 ∆SC1 ∆SR2
As discussed in the Results, eqs 2 and 3 can be applied to c,s CPD,
t,s I CPD, or their sum.
J. Org. Chem, Vol. 72, No. 1, 2007 49

Moriou et al.
(150 MHz, D2O): δ 14.9 (CH3 pT), 21.4 (COCH3), 26.3 (CH3
Tp), 33.7 (C-2′ pT), 35.2 (C-2′ Tp), 51.0 (C-3′ Tp), 59.0 (C-5′ Tp),
59.3 (C-6 Tp), 71.5 (C-5′ pT), 73.7 (C-3′ pT), 74.4 (C-5 Tp), 83.4
(C-1′ Tp and C-4′ Tp), 83.9 (C-4′ pT), 89.5 (C-1′ pT), 117.9 (C-5
pT), 155.1 (C-2 Tp), 158.4 (C-2 pT), 174.9 (COCH3), 175.2 (C-4
Tp), 176.0 (C-4 pT). IR (film, cm-1) 3401, 3242, 3110, 1723, 1687,
1652, 1507, 1450, 1366, 1243, 1177, 1102, 1053, 939. HRMS
(ES) (M + Na)+ calcd for C22H29N5O12NaS, 610.1431; found,
610.1449.
Acknowledgment. We thank Dr. Daniel Royer for recording
IR spectra.
Supporting Information Available: General information, chro-
matograms of the irradiation of 1 + 2, NMR spectra of compound
1 irradiation, compounds 6, 9, and 10, and X-ray crystallographic
data for 1 (in CIF format). This material is available free of charge
via the Internet at http://pubs.acs.org.
JO061488A
50 J. Org. Chem., Vol. 72, No. 1, 2007