5616 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Saladino et al.
reaction conditions; the reactivity ratio Cs/As is 0.15, and the
reactivity of Ts is close to 0. The variations on the basic
protocols used to solve these biases have been described.8-11
In order to simplify the formamide sequencing procedure and
to abolish the possibility of errors in the determination of Gs
and As residues, we have found the use of purine derivatives
to be quite promising. These could (i) be incorporated into the
DNAs to be analyzed by either the polymerase chain reaction
(PCR)12 or primer extension (PE)7 procedures and (ii) be
degraded at different rates, in order to increase the sensitivity
intervals.
Knowledge of the degradation pathway of purine nucleosides
with formamide is necessary to select analogues to be incor-
porated and then selectively degraded. Although the degradation
pathway of purine nucleosides and nucleotides by the action of
strong bases (usually HO-) is widely described in literature,13
the reaction with a weak base such as formamide has not been
previously studied. In this paper we describe the mechanism
of degradation of 2′-deoxyadenosine (1) and 2′-deoxyguanosine
(2) with formamide. The degradation pattern of purine nucleo-
sides, as defined by this analysis, allows us to predict the
positions of the purine ring whose structural modifications may
modify the sensitivity toward formamide, therefore allowing
substantial improvement of the DNA chemical sequencing
procedure.
previously described for efficient nucleophilic agents, such as
hydrazine and hydroxylamine.16
According to the degradation pathway hypothesized (Scheme
1) the initial C(8)-adduct between formamide and 2′-deoxy-
adenosine (which could not be isolated, probably because of
its instability) may easily undergo the C(8) imidazole ring-
opening to yield the derivative 3. Compound 3 is probably the
first to form and it might be slowly transformed to 4 by loss of
the initially acquired molecule of formamide. Under sufficiently
vigorous conditions (i.e, removal of the formamide by distillation
under vacuum) the 2′-deoxyribosyl residue may be hydrolyti-
cally detached to yield compound 5.
It is noteworthy that in the reaction of 1 with formamide
products of possible alternative degradation pathways (namely,
nucleophilic additions at the C(6) and/or C(2)-positions on the
purine ring) are not recovered in the reaction mixture.
2′-Deoxyguanosine (2) reacts faster than 1 with formamide
at 110 °C, in agreement with the order of sensitivity observed
for the purine bases in DNA.8-11 The reaction was complete in
only 3 h resulting in a complex mixture that was not character-
ized further. When the reaction was performed at 90 °C a minor
conversion (15%) of the substrate was observed. In the latter
case, 2,6-diamino-4-oxo-5-formamido-6-[N-(2′-deoxy-â-D-ri-
bofuranosyl)]pyrimidine 6 and 2,6-diamino-4-oxo-5-formami-
dopyrimidine 7 were obtained (Scheme 2) after purification by
TLC in reverse phase (butanol saturated by water) or by flash-
chromatography (after distillation of the excess formamide).
Formamidopyrimidine derivatives (FAPyr) 4, 5, 6, and 7 are
usually referred to as products of imidazole C(8)-ring cleavage
of purine nucleosides and nucleotides by ionizing radiation17-20
or by basic treatment after heterocyclic nitrogen alkylation.21
On the basis of the structural analogies among the isolated
degradation products it is reasonable to suggest that the
degradation pathway of 2′-deoxyguanosine is similar to that
described for 2′-deoxyadenosine, even if, in the first case, other
degradation pathways cannot be completely excluded because
of the lack of characterization of all possible intermediates.
Results
In order to obtain information on the degradation pathway
of purine nucleosides by formamide, we have analyzed the
reactions of 2′-deoxyadenosine (1) and 2′-deoxyguanosine (2)
using the experimental conditions described for the formamide
DNA sequencing method.11 2′-Deoxyadenosine (1) (538 mg,
2 mmol) was added to formamide (10 mL) in the presence of
low amount of water (0.1 mL), and the mixture was refluxed at
110 °C until the disappearance of the substrate. Small amounts
of the formamidopyrimidine nucleoside 3 and 6-amino-5-
formamido-4-[N-(2′-deoxy-â-D-ribofuranosyl)]pyrimidine 4
(Scheme 1) were obtained in isolated form by analytical TLC
purification in reverse phase (butanol saturated with water) and
characterized by capillary gas chromatography-mass spectrom-
etry (CG-MS) after silylation with N,O-bis(trimethylsilyl)-
trifluoro acetamide (BSTFA),14 1H-NMR, and elemental mi-
croanalysis. Compound 3 was found to be unstable even if
recovered under nitrogen atmosphere, and it was easily con-
verted to 4 by mild warming or by treatment with acidic SiO2.
The remaining reaction mixture was distilled under high vacuum
to eliminate the excess formamide, and the crude product was
purified by flash-chromatography (chloroform:methanol ) 9.5:
0.5 as eluent) to give 4,6-diamino-5-formamidopyrimidine 5 as
the only recovered product. These data suggest that formamide
reacts with 1 cleaving the adenine base through nucleophilic
attack on the C(8) position of the purine ring. The attack on
this position is in agreement with its electrophilic character15
and with the chemical behavior toward DNA purine components
The indication that formamide degrades purine nucleosides
by C(8)-imidazole ring-opening may provide the chemical clues
to predict the variation of reactivity for various purine analogues.
In fact, structural modifications that reduce the electrophilic
character of the C(8)-position of the purine ring might reduce
its reactivity toward a weak nucleophile such as formamide.
Available 7-deaza-2′-deoxypurine nucleosides, which lack the
electron-withdrawing nitrogen at the adjacent C(7) position,
satisfy this condition and might support nucleophilic attack less
efficiently. In order to test this hypothesis and to apply this
rationale to the chemical DNA sequencing procedure, we have
analyzed the sensitivity to formamide of 7-deazaanalogues of
purine nucleosides (both adenosine, guanosine and inosine)
when inserted into polynucleotides.
Figure 1 shows in panel a the molecular construct used to
synthesize double stranded DNA molecules carrying 7-deaza-
purine residues. The indicated 34mer template segment (lower
strand) was annealed to the 16mer oligo (upper strand) and used
as a template-primer substrate for a Sequenase-driven polym-
erization. The resulting upper strand contains (outside the oligo)
(12) Erlich, H. A. PCR Technology: Principles and application for DNA
amplification; Stockton Press: 1989.
(13) (a) Mian, A. M.; Walker, R. T. J. Chem. Soc. 1968, (C), 2577. (b)
Price, C. C.; Gache, G. M.; Koneru, P.; Shibakawa, R.; Sawa, J.R.;
Yamaguchi, M. Biochim. Biophys. Acta 1968, 166, 327. (c) Lawley, P. D.;
Brookes, P. J. Mol. Biol. 1967, 25, 143. (d) Townsed, L. B.; Robins, R. K.
J. Am. Chem. Soc. 1963, 85, 247. In Synthetic procedures in nucleic acid
chemistry; Zorbach, W. W., Tpson, R. S., Eds.; Wiley: New York-London,
1968; p 315. (e) Magrath, D. I.; Brown, G. B. J. Am. Chem. Soc. 1957, 79,
3245. (f) Gordon, M. P.; Weliky, V. S.; Brown, G. B. J. Am. Chem. Soc.
1957, 79, 3245. (g) Lawley, P. D. Proc. Chem. Soc. 1957, 290.
(14) Ali, S. L. Chromatographia 1975, 8, 33-39.
(16) Verwoerd, D. W.; Zillig, W.; Kohlage, H. Z. Physiol. Chem. 1963,
332, 184-193.
(17) Fornace, A. J.; Little, J. B. Biochim. Biophys. Acta 1977, 477, 343-
355.
(18) Ward, J. F. Isr. J. Chem. 1972, 10, 1123-1138.
(19) Phillips, G. O.; Criddle, W. J. J. Chem. Soc. 1962, 2740-2744.
(20) Chetsanga, C.; Grigorian, C. Proc. Natl. Acad. Sci. U.S.A. 1985,
82, 633-637.
(15) Pullman, B.; Jortner J. Molecular basis of specificity in nucleic
acids-drug interactions; Kluwer Academic: Dortrecht, 1961.
(21) Tabor, S.; Richardson, C. C. J. Biol. Chem. 1989, 264, 6447-6450.