RNA photolabeling mechanistic studies: X-ray crystal structure of the photoproduct formed between 4-thiothymidine and adenosine upon near UV irradiation

Full text of DOI:10.1021/ja961329g

8142
J. Am. Chem. Soc. 1996, 118, 8142-8143
Scheme 1. Photochemical Reaction between
4-Thiothymidine (1) and Adenosine (2)
RNA Photolabeling Mechanistic Studies: X-ray
Crystal Structure of the Photoproduct Formed
between 4-Thiothymidine and Adenosine upon Near
UV Irradiation
Carole Saintome´,†,‡ Pascale Clivio,† Alain Favre,‡
Jean-Louis Fourrey,*,† and Claude Riche†
Institut de Chimie des Substances Naturelles
CNRS, 91198 Gif-sur-YVette Cedex, France
Institut Jacques Monod, CNRS-UniVersite´ Paris VII
2 Place Jussieu, 75251 Paris Cedex O5, France
ReceiVed April 22, 1996
In recent years, reliable photolabeling methods have been
developed to probe RNA tertiary structure in solution.1 One
of these exploits the remarkable photochemical properties of
the sulfur analogs of the current nucleic acid bases 4-thiouracil,
4-thiothymine, and 6-mercaptopurine to gain informative data
on tertiary interactions within such biomolecules.2 Indeed, the
latter bases in which an oxygen has been replaced by sulfur are
stable in the dark and can be selectively photoactivated to give
highly reactive excited states manifesting the capacity to undergo
covalent bonding with any nucleic acid base.3 Hence, when a
thio-substituted nucleobase is introduced either ramdomly or
at a defined position in such a system (by application of
appropriate enzymatic2b or chemical4 methods), subsequent
irradiation usually results in the formation of informative cross-
links. Obviously, only residues located at bonding distance with
the thio-modified nucleobase can be involved in the cross-links
which can be mapped by suitable sequencing procedures.5
Interestingly, the spatial relationships which are evidenced in
this manner can serve as constraints for the reconstruction of
the three-dimensional structure of the system and the analysis
of its conformational flexibility by molecular modeling.6
Furthermore, the respective orientation of the two residues
leading to a cross-link can be more precisely defined when the
corresponding photochemical pathways leading to the products
have been elucidated. In general, these pathways can be
reasonably deduced from the structural analysis of the photo-
products. So far, detailed structural analysis of photoproducts
resulting from thio-substituted nucleobases has been accom-
plished with pyrimidine only.7 As photo-cross-links involving
purine residues are frequently encountered in the analysis of
RNA folding,5 we describe here the X-ray crystal structure of
the unique photoproduct8 which was formed upon irradiation
(ca. 360 nm, inert atmosphere) of an equimolecular mixture
(8.6 mM) of 4-thiothymidine (1) and adenosine (2) in water.
The reaction, monitored by the disappearance on the UV
spectrum of the thiocarbonyl absorption maximum of 1 at 335
nm, yielded 3 (mp 167-168 °C) in 90% yield after purification
(based on reacted 2 after ca. 70% consumption) (Scheme 1).9
Structure 3 was supported by the spectroscopic data. Exact mass
measurement of the quasimolecular ion (m/z 532.1790, M +
Na+) in the positive mode FAB mass spectrum indicated that 3
resulted from the condensation of 1 with 2 and replacement of
the sulfur atom by an oxygen atom. The UV absorption peak
of 3 (λmax 293 nm, ꢀ 13 840 mol-1 cm-1 dm3, H2O, pH 7) was
in good agreement with that of a 5-methylcytosine. Key NMR
arguments for structure 3 are the following: opening of the
imidazole ring of adenosine to give a 6-N-formyl-4,5,6-
triaminopyrimidine is supported by the presence on the HMBC
spectrum of the penta-O-acetyl derivative 4,10 of a methine
carbon at 164.2 ppm that correlates with H1′ of ribose. This
long-range C-H coupling (3J) allowed this carbon to be
† Institut de Chimie des Substances Naturelles.
‡ Institut Jacques Monod.
(1) (a) Branch, A. D.; Benenfeld, B. J.; Paul, C. P.; Robertson H. D.
Methods Enzymol. 1989, 180, 418-442. (b) Jaeger, J. A.; Santa Lucia, J.,
Jr.; Tinoco, I., Jr. Annu. ReV. Biochem. 1993, 62, 255-287.
(2) (a) Favre, A. In Bioorganic Photochemistry: Photochemistry and
the Nucleic Acids; Morrison, H., Ed.; J. Wiley & Sons: New York, 1990;
Vol. 1, pp 317-378. (b) Sontheimer, E. J. Mol. Biol. Rep. 1994, 20, 35-
44. (c) Favre, A.; Fourrey, J.-L. Acc. Chem. Res. 1995, 28, 375-382.
(3) Favre, A.; Lemaigre Dubreuil, Y.; Fourrey, J.-L. New J. Chem. 1991,
15, 593-599.
(4) (a) Beaucage, S.; Iyer, R. P. Tetrahedron 1993, 49, 6123-6194. (b)
Beaucage, S.; Iyer, R. P. Ibid. 1992, 48, 2223-2311.
(7) (a) Fourrey, J.-L.; Gasche, J.; Fontaine, C.; Guittet, E.; Favre, A. J.
Chem. Soc., Chem. Commun. 1989, 1334-1336. (b) Clivio, P.; Fourrey,
J.-L.; Gasche, J.; Favre, A. J. Am. Chem. Soc. 1991, 113, 5481-5483. (c)
Woisard, A.; Favre, A.; Clivio, P.; Fourrey, J.-L. J. Am. Chem. Soc. 1992,
114, 10072-10074. (d) Clivio, P.; Favre, A.; Fontaine, C.; Fourrey, J.-L.;
Gasche, J.; Guittet, E.; Laugaˆa, P. Tetrahedron 1992, 48, 1605-1616. (e)
Clivio, P.; Fourrey, J.-L.; Gasche, J.; Favre, A. Tetrahedron Lett. 1992,
33, 1615-1618. (f) Blazek, E. R.; Alderfer, J. L.; Tabaczynski, W. A.;
Stamoudis, V. C.; Churchill, M. E.; Peak, J. G.; Peak, M. J. Photochem.
Photobiol. 1993, 57, 255-265. (g) Clivio, P.; Fourrey, J.-L.; Szabo, T.;
Stawinski, J. J. Org. Chem. 1994, 59, 7273-7283.
(5) (a) Wollenzien, P.; Expert-Bezanc¸on, A.; Favre, A. Biochemistry
1991, 30, 1788-1795. (b) Lemaigre Dubreuil, Y.; Expert-Bezanc¸on, A.;
Favre, A. Nucleic Acids Res. 1991, 19, 3653-3660. (c) Dos Santos, V. D.;
Vianna, A.; Fourrey, J.-L.; Favre, A. Nucleic Acids Res. 1993, 21, 201-
207. (d) Dos Santos, V. D.; Fourrey, J.-L.; Favre, A. Biochem. Biophys.
Res. Commun. 1993, 190, 377-385. (e) Sontheimer, E. J.; Steitz, J. A.
Science 1993, 262, 1989-1996. (f) Woisard, A.; Fourrey, J.-L.; Favre, A.
J. Mol. Biol. 1994, 239, 366-370. (g) Rinke-Appel, J.; Ju¨nke, N.;
Brimacombe, R.; Lavrik, I.; Dokudovskaya, S.; Dontsova, O.; Bogdanov,
A. Nucleic Acids Res. 1994, 22, 3018-3025. (h) Bravo, C.; Lescure, F.;
Laugaˆa, P.; Fourrey, J.-L.; Favre, A. Nucleic Acids Res. 1996, 24, 1351-
1359. (i) Saintome´, C. Unpublished results.
(8) From inspection of the 1H NMR spectrum of the crude irradiated
mixture.
(9) Purification was achieved on a medium-pressure (400 mbar) RP 18
column using a gradient of acetonitrile in water. Nucleosides 1 and 2 were
recovered in 24% and 33% yield, respectively.
(6) Laugaˆa, P.; Woisard, A.; Fourrey, J.-L.; Favre, A. C. R. Acad. Sci.,
Life Sci. 1995, 318, 307-313.
S0002-7863(96)01329-7 CCC: $12.00 © 1996 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 8143
identified as being originally C8 of adenine. Its deshielded
chemical shift (δ∆ 22 ppm), compared to the corresponding
one of the parent adenosine (δ 142 ppm), is typical of a
formamido group. Analysis of the NMR chemical shifts of the
carbons of the former thymine base11 favors substitution of
carbon C4 by a nitrogen (the chemical shifts of C4, C5, and
C6 are very similar to the corresponding ones of 5-methylcy-
tosine12) rather than by a carbon as in the case of the
pyrimidinone of the [6-4] photoproduct of TpT.13 Upon
standing in aqueous solution, photoproduct 3 was slowly
transformed into 5, this transformation being accelerated and
brought to completion upon mild heating.14 The positive FAB
mass spectrum of 5 exhibited a pseudomolecular ion at m/z 350
(M + H+), suggesting the loss of the ribose unit of 3 and
deformylation. This was evident from the examination of the
1H NMR spectrum of 5, recorded in a D2O/DMSO-d6 mixture,
which showed only the signals of the deoxyribose moiety
together with one methyl group (δ 2.08 ppm) and two vinylic
protons (δ 7.86 and 7.77 ppm). The most significant difference
between the 13C NMR spectra of 5 and 3 was the shielding of
carbon C5 (δ∆ ca. 17 ppm) of the 4,5,6-triaminopyrimidine
unit. This was in accordance with the replacement of the
formamido group at position C6 by an amino group, inducing
shielding in the ortho position. Interestingly, all the carbon
chemical shifts which derived from the adenine part are very
close to those of 4,6-diamino-5-(formylamino)pyrimidine (Fapy-
Ade),15 in agreement with N-5-substitution. This interpretation
is in accordance with the data derived from the HMBC spectrum
recorded in DMSO-d6. Indeed, this spectrum showed an
exchangeable signal at 7.92 ppm integrating for one proton and
correlating with C2 (4J), C4 (2J), and C5 (3J) of the 5-methyl-
cytosine unit and C5 (2J), magnetically equivalent C4 and C6
(3J) of the 4,5,6-triaminopyrimidine unit of 5, respectively.
The structure of photoproduct 3 (Figure 1) was definitively
established by X-ray diffraction studies carried out on crystals
grown from a concentrated aqueous solution (1 M).16 In this
structure, the deoxyribose unit shows a C4′-endo/C3′-exo
conformation (South-type puckered conformation) with a pseu-
dorotation phase angle P ) 218.3°,17 and a maximum puckering
amplitude, τm ) 34.0°, as in the case of 5-methyl-2′-deoxycy-
tidine.18 The 5-methylcytosine base retains an anti conformation
(ø ) -178.6°).19 On the other hand, the ribose unit adopts a
South-type puckered geometry (C2′-endo, P ) 160.3° and τm
Figure 1. ORTEP drawing of 3. Ellipsoids are at the 30% level. The
two water molecules of crystallization are depicted.
) 40.8°) which differs from the one found in the crystalline
structure of adenosine (C3′-endo).20 Finally, in 3 the conforma-
tion about the ribose glycosidic bond (ø ) -108.5°) is anti as
in the case of the parent adenosine. Molecules, in the crystal
structure, are linked in a three-dimensional lattice through eight
independent intermolecular hydrogen bonds. In particular, this
lattice implicates two water molecules which bridge four
molecules of photoproduct.
The most probable mechanism to account for the formation
of 3 (Scheme 1) is a [2 + 2] cycloaddition in which the
thiocarbonyl of 1 adds across the N7-C8 double bond of 2.
This should lead to an unstable four-membered heterocyclic
intermediate (thiaazetidine) prone to undergo ring opening to
give an imidazoline, yielding 3 after further hydrolysis.21
In summary, 3 is the first characterized purine-pyrimidine
photoadduct, arising by saturating the N7-C8 double bond of
adenine, by means of a remarkably high yielding bimolecular
reaction. Indeed, a photoaddition reaction involving the N7-
C8 double bond of adenine has already been reported to occur
in the case of d(ApA) between the two adjacent adenine bases.22
But, to the best of our knowledge, such a N7-C8 regioselective
reaction has never been observed to be so efficient with a
pyrimidine unit.23 The herein reported X-ray structural data of
3 might hopefully provide important information to be used for
the reconstitution of tertiary interactions based on photolabeling
experiments. Finally, the facile cleavage of the N-glycosidic
bond of the parent adenosine might be potentially used for
designing artificial specific endonucleases.
(10) Attempts to obtain correlations between C and H of the N-6-formyl-
N-6-ribosyl-4,5,6-triaminopyrimidine part on the HMBC spectrum of 3,
recorded either in D2O or in DMSO-d6 at room temperature or below, failed
due to line broadening. Although recording at 90 °C in D2O led to a well-
resolved 1H NMR spectrum of 3 (see the Supporting Information), the
transformation of 3 to 5 precluded a HMBC experiment at this temperature.
Finally, this was circumvented by inspection of the HMBC spectrum of 4
(m/z 742, M + Na+) in CD3OD whose 13C NMR chemical shifts of the
aglycon nuclei are almost identical to those of 3.
Supporting Information Available: X-ray crystallographic data,
1H NMR spectra of 3, and 13C NMR chemical shifts of 3, 4, and 5 (10
pages). See any current masthead page for ordering and Internet access
instructions.
(11) Attributed from the HMBC spectrum.
JA961329G
(12) Shaw, A. A.; Shetlar, M. D. J. Am. Chem. Soc. 1990, 112, 7736-
7742.
(19) Saenger, W. In Principles of Nucleic Acid Structure; Cantor, C. R.,
Ed.; Springer-Verlag: New York, 1984.
(13) Rycyna, R. E.; Alderfer, J. L. Nucleic Acids Res. 1985, 13, 5949-
5863.
(20) Lai, T. F.; Marsh, R. E. Acta Crystallogr. 1972, B28, 1982-1989.
(21) The preferred N7-C8 imidazole ring fission of adenosine leading
to a an N-9-formyl derivative is probably governed by the electron deficient
nature of the N-7-substituent. This is inferred from analogous ring opening
of 9-alkylated adenines undergoing N-7-acylation (Altman, J.; Ben-Ishai,
D. J. Heterocycl. Chem. 1968, 5, 679-682. Leonard, N. J.; McDonald, J.
J.; Henderson, R. E. L.; Reichmann, M. E. Biochemistry 1971, 10, 3335-
3342).
(14) Purification was performed by HPLC on a semipreparative NOVA-
PAK C18, 6µ radial pak cartridge, using a gradient of acetonitrile in 0.05
M sodium acetate.
(15) Raoul, S.; Bardet, M.; Cadet, J. Chem. Res. Toxicol. 1995, 8, 924-
933.
(16) Crystal data: C20H27N7O9‚2H2O, monoclinic, space group P21,
colorless, a ) 7.381 (7) Å, b ) 21.298 (12) Å, c ) 8.414 (1) Å, â )
108.92 (8)°, Z ) 2, R ) 0.067, Rw) 0.086, goodness of fit 1.04, number
of observed reflexions 1945.
(22) Kumar, S.; Joshi, P. C.; Sharma, N. D.; Bose, S. N.; Davies, R. J.
H.; Takeda, N.; McCloskey, J. A. Nucleic Acids Res. 1991, 19, 2841-
2847.
(23) For a precedent case see: Paszyc, S.; Skalski, B.; Wenska, G.
Tetrahedron Lett. 1976, 449-450.
(17) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94, 8205-
8212.
(18) Sato, T. Acta Crystallogr. 1988, C44, 870-872.