Regioselectivity and mechanism of transpurination reactions in the guanine nucleosides series

Article abstract of DOI:10.1039/a602021f

The transpurination reaction of the fully acetylated derivatives of guanosine and its 7-β-D-ribofuranosyl regioisomer with 2-acetoxyethyl acetoxymethyl ether has been studied using high performance liquid chromatography (HPLC). The regioselectivity of glycosyl exchange observed in the early stages of the reaction suggests that the unsubstituted nitrogen atoms of the imidazole portion (N7 and N9) are, exclusively, sites of direct glycosylation in the case of guanine derivatives. The results lead to the conclusion that the mechanism of the glycosylation reaction of guanine is different to that of adenine.

Full text of DOI:10.1039/a602021f

View Article Online / Journal Homepage / Table of Contents for this issue
Regioselectivity and mechanism of transpurination reactions in the
guanine nucleosides series
Jerzy Boryski
Institute of Bioorganic Chemistry, Polish Academy of Sciences, Noskowskiego 12/14,
PL-61704 Poznan, Poland
The transpurination reaction of the fully acetylated derivatives of guanosine and its 7-â-D -ribofuranosyl
regioisomer with 2-acetoxyethyl acetoxymethyl ether has been studied using high performance liquid
chromatography (H PLC). The regioselectivity of glycosyl exchange observed in the early stages of the
reaction suggests that the unsubstituted nitrogen atoms of the imidazole portion (N 7 and N 9) are,
exclusively, sites of direct glycosylation in the case of guanine derivatives. The results lead to the
conclusion that the mechanism of the glycosylation reaction of guanine is different to that of adenine.
Transpurination is a glycosyl exchange reaction, in which the
purine moiety of the nucleoside is transferred from one sugar to
another. This process was originally noticed by Lichtenhaller
and Kitahara, who described the migration of hypoxanthine
from ribose to glucose when 2Ј,3Ј-O-isopropylideneinosine
was reacting with 1-bromo-2,3,4,6-tetra-O-acetyl-α--gluco-
pyranose in the presence of mercury() cyanide.1 More recently,
the transpurination reaction was applied as a convenient
method for the synthesis of acyclonucleosides of pronounced
antiviral activity, e.g. acyclovir and ganciclovir.2–4
Transpurination, and its better known counter reaction,
transglycosylation,2,3,5–13 are examples of glycosyl migration
reactions and play an important role in nucleoside chemistry.
For example, transglycosylation is the second step of N-
glycosyl linkage formation, in which the kinetic glycosylation
product, obtained directly from an appropriate heterocyclic
base and a protected sugar derivative, undergoes a rearrange-
ment to the thermodynamically more stable nucleoside. This
isomerisation process appears to be an intermolecular reaction
catalysed by Lewis acids.12,13
In the purine series, the mechanism of glycosyl exchange
seems to be well documented in the case of adenine and its
derivatives. According to the proposed mechanism, adenine
undergoes initial glycosylation at position 3. The unstable 3-
glycosyladenine derivative then reacts with another sugar cat-
ion to form a 3,9-diglycosylpurine quaternary salt intermediate.
Its decomposition gives the final 9-substituted product. The
3→9 sequence of glycosylation and transglycosylation has also
been suggested for purine bases other than adenine.12,13
derivatives react in line with the Baker’s rule,16 producing
almost exclusively β-ribosides. Thirdly, the ribo- and acyclo-
nucleosides can be readily distinguished by chromatographic
means.
Results and discussion
In the transpurination experiments tetraacetylguanosine 1 or its
7-β-regioisomer 6 (obtained from 1 by using the thermal 7
9
isomerisation method)3,11 were subjected to the action of a 10-
fold molar excess of 2-acetoxyethyl acetoxymethyl ether.17 The
reactions were performed in dry chlorobenzene at 120 ЊC for
2 h, in the presence of toluene-p-sulfonic acid (10 mol%) as
a catalyst. The reaction mixtures were analysed by HPLC
(see Experimental).
Interestingly, after a prolonged reaction time (2 h) both
experiments gave an almost identical distribution of products:
the regioisomeric mixture of acyclonucleosides 3 and 8 as major
products (in a ratio of ca. 1:1), as well as smaller amounts of
ribonucleosides 1 and 6. The pathways to reach that dynamic
equilibrium state, however, were different in both cases, as
judged by HPLC (Figs. 1 and 2).
However, as shown by Dudycz and Wright in a more recent
study,14 the respective 7-ribonucleosides (not 3-ribo derivatives)
were the kinetic products in ribosylation of persilylated 2-
bromohypoxanthine, N 2-acetylguanine and N 2-(p-butyl-
phenylguanine). In addition, they isolated the corresponding
7,9-diribofuranosylpurines as unstable reaction intermediates.
Those findings may suggest that the transglycosylation of 6-
oxopurines, i.e. guanine, hypoxanthine and their derivatives,
proceeds according to a different mechanism to that of adenine.
In order to elucidate the mechanism of glycosyl exchange in
the case of guanine, the previously reported 2,3 conversion of
guanosine into acyclovir has been investigated as a model reac-
tion.15 In this approach, two different substituents, i.e. 2,3,5-tri-
O-acetyl-β--ribofuranosyl and (2-acetoxyethoxy)methyl, were
present in the reaction medium. That pair of substituents has
been chosen for three reasons. First, both groups are of similar
reactivity in glycosyl exchange reactions.11,19 Secondly, applica-
tion of these groups allows us to avoid anomerization: the
(2-acetoxyethoxy)methyl substituent does not introduce a new
centre of asymmetry, whereas tri-O-acetyl-β--ribofuranosyl
Fig. 1 Time-dependent product distribution for the reaction of 9-
(2Ј,3Ј,5Ј-tri-O-acetyl-β--ribofuranosyl)-N 2-acetylguanine (compound
1) with 2-acetoxyethyl acetoxymethyl ether (see Experimental): 1 ᭿;
6 ᮀ; 8 ᭹; 3 ᭺. The mole fractions are based on HPLC data.
J. Chem. Soc., Perkin Trans. 2, 1997
649

View Article Online
Scheme 2
for the 7-riboside 6. The resulting unstable 7,9-diglycosyl-
guanine intermediates then undergo decomposition: the inter-
mediate 2 decomposes to 7-acyclonucleoside 3 (Scheme 2),
whereas compound 7 gives 9-acyclonucleoside 8. Additionally,
both exchange reactions produce the acyloxonium cation 10
(the structure of the respective carboxonium ion can be taken
into consideration as well),13 and therefore the reactions are
reversible. These transpurination reactions are shown in
Scheme 1 as horizontal arrows.
Fig. 2 Time-dependent product distribution for the reaction of 7-
(2Ј,3Ј,5Ј-tri-O-acetyl-β--ribofuranosyl)-N 2-acetylguanine (6) with 2-
acetoxyethyl acetoxymethyl ether: 1 ᭿; 6 ᮀ; 8 ᭹; 3 ᭺
When the peracetylated 9-riboside 1 was used as the sub-
strate, the 7-acyclonucleoside 3 was found as a single reaction
product after 2 min (Fig. 1), whereas in the transpurination
reaction of the 7-riboside 6 the 9-substituted acyclonucleoside 8
(diacetylacyclovir) was apparently the kinetic reaction product
(Fig. 2). The results of the HPLC analysis were also confirmed
by thin-layer chromatography (TLC) and 1H NMR spectro-
scopy. Any putative compound with the structure of a 3-sub-
stituted N2-acetylguanine was not detected in transpurination
experiments.
However, the liberated ribosugar cations 10 may in turn react
with the starting nucleoside, which leads to the ‘intra-
molecular’18
7
9 isomerization. This also applies to 7
9
isomerization of acyclonucleosides 3 and 8 in the presence of
acyclosugar cations 9. The vertical arrows in Scheme 1 denote
the reversible 7 9 transglycosylation reactions.
As a result, the reaction mixtures in both cases contain all
possible 9- and 7-regioisomers: two pairs of guanine ribo- and
acyclo-nucleosides (compounds 1, 6 and 8, 3, respectively).
Under the conditions described, equilibrium is achieved after
ca. 2 h. The acyclonucleosides 3 and 8 are predominant
products because an excess of 2-acetoxyethyl acetoxymethyl
ether has been applied (Figs. 1 and 2).
The mechanism and sequence of these glycosyl exchange
reactions may be depicted as shown in Schemes 1 and 2.
It can also be deduced from Scheme 1 that any direct ‘diag-
onal’ conversion is not possible, e.g. diacetylacyclovir 8 is pro-
duced from tetraacetylguanosine 1 either via 7-ribonucleoside
6, or via 7-acyclonucleoside 3. In both cases the conversion
requires not just one, but two reactions: transpurination and
7
9 transglycosylation. The respective 7,9-disubstituted
derivatives (2, 4, 5, 7), not isolated in this study, are postulated
as reaction intermediates for each single process of glycosyl
exchange.
Furthermore, a detailed comparison of Figs. 1 and 2 suggests
that 7-riboside 6 reacts more readily with the acyclosugar cation
9 than its 9-regioisomer 1. It has been previously shown for the
irreversible 2→1 transglycosylation of riboindazoles19 that
kinetic products of glycosylation undergo a fast reaction with
other sugar cations, whereas thermodynamic products do not.
Therefore, in the case of guanine nucleosides the 9-regioisomers
may be considered as the ‘more thermodynamic’ products than
the 7-substituted compounds. However, that difference is not
significant and all glycosyl exchange reactions in the guanine
series are reversible.
The results obtained in this study lead to the following
conclusions.
(i) In the guanine series only the nitrogen atoms of the imida-
zole portion (i.e. N7 and N9) participate in glycosyl exchange
reactions. This perhaps applies also to other 6-oxopurine
nucleosides.
(ii) There is no experimental evidence that N3 of guanine
takes part in glycosylation or transglycosylation reactions.
Scheme 1
Reaction of 2-acetoxyethyl acetoxymethyl ether with toluene-p-
sulfonic acid generates carboxonium cation 9, which is then
attacked by the unsubstituted nitrogen atom of the imidazole
ring, i.e. N7 in the case of the 9-riboside 1 (Scheme 2), and N9
650
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 1 1H NMR chemical shifts [δ (ppm)]a
Compound N1H
N2H
8-H
1ЈH b
2ЈH
3ЈH
4ЈH
5ЈaH b
5ЈbH b N-Acetyl
O-Acetyl
1
6
8
3
12.11
s, 1
12.25
s, 1
12.07
s, 1
12.17
s, 1
11.62
s, 1
11.69
s, 1
11.79
s, 1
11.63
s, 1
8.28
s, 1
8.53
s, 1
8.14
s, 1
8.38
s, 1
6.11
5.84
t, 1
5.81
t, 1
—
5.50
dd, 1
5.47
dd, 1
—
4.39
m, 2
4.33
dd, 1
4.25
dd, 1
2.21
s, 3
2.18
s, 3
2.18
s, 3
2.17
s, 3
2.14, 2.07, 2.05
3 × s, 9
2.11, 2.04, 2.03
3 × s, 9
1.95
s, 3
1.95
s, 3
d, 1; 6.6c
6.31
4.33
q, 1
4.07
m, 2
4.06
m, 2
4.43
dd, 1
d, 1; 5.7c
5.48
s, 2
5.69
s, 2
3, 70
m, 2
3.72
m, 2
—
—
a
Recorded at 300 MHz in (CD3)2SO; ref. SiMe4. Figures following the observed multiplicities are numbers of protons as estimated by integration.
b Or the respective methylene groups for acyclonucleosides. c Coupling constant J1Ј2Ј in Hz.
Table 2 13C NMR chemical shifts [δ (ppm)]a
Compound NCOCH3 OCOCH3 C-6
C-4
C-2
C-8
C-5
C-1Јb C-4Јb C-2Ј C-3Ј C-5Јb NCOCH3 OCOCH3
1
6
173.50
173.39
170.03
169.39
169.19
169.98
169.30
169.10
170.12
170.11
154.68 148.58 148.22 137.79 120.38 84.57 79.83 72.16 70.30 63.05 23.82
158.37 152.17 147.51 144.12 110.56 87.54 79.46 73.06 69.62 62.94 23.66
20.47
20.33
20.10
20.43
20.26
20.08
20.44
20.45
8
3
173.44
173.30
154.77 148.76 148.03 139.96 120.08 73.32 66.58
157.35 152.40 147.16 145.02 111.05 74.84 66.29
—
—
—
—
62.57 23.69
62.63 23.62
a
Recorded at 75.43 MHz, ref. SiMe4. b Or the respective carbon atoms of the acyclic chain.
Table 3 UV spectral data, TLC chromatography, and elemental analyses
Rf -value in system a
λmax(MeOH)/nm
Found
%C
Requires
Compound
(ε/dm3 molϪ1 cmϪ1
)
A
B
Formula
%H
%N
%C
%H
%N
1
6
8
3
259 (16 100), 279 (11 400)
263 (13 400), 280 (sh,b 10 300)
259 (16 200), 278 (11 600)
263 (13 500), 280 (sh,b 10 100)
59
65
35
53
37
53
23
42
C18H21N5O9
C18H21N5O9
C12H15N5O5
C12H15N5O5
48.1
48.0
46.7
46.7
4.7
4.8
4.9
4.8
15.3
15.3
22.2
22.4
47.9
47.9
46.6
46.6
4.7
4.7
4.9
4.9
15.5
15.5
22.6
22.6
a
See Experimental section. b Shoulder.
(iii) The unsubstituted nitrogen (i.e. of hybridization sp2) is
the nucleophilic centre that reacts directly with sugar cations.
Therefore, the structure of kinetic glycosylation products in the
guanine series depends on the substitution of substrates: the
9-substituted derivatives of guanine give 7-glycosylated kinetic
products, and vice versa, the 7-substituted substrates undergo
direct glycosylation at N9.
(iv) The formation of 7,9-diglycosyl-6-oxopurine quaternary
salts, postulated as the glycosylation intermediates by Dudycz
and Wright 14, allows us to rationalise the regioselectivity
observed in this study.
Division of Millipore: Nova Pak C18 column (8 × 100 mm
Radial-Pak Cartridge), 600E Multisolvent Delivery System
with U6K Universal Liquid Chromatography Injector, 486
Tunable Absorbance Detector and 746 Data Module. 9-(2Ј,3Ј,-
5Ј-tri-O-acetyl-β--ribofuranosyl)-N2-acetylguanine 1 was pre-
pared according to Reese and Saffhill,21 and then precipitated
from chloroform to diethyl ether to get 1 as a white powder.
9-[(2-Acetoxyethoxy)methyl]-N2-acetylguanine and its 7-
regioisomer (compounds 8 and 3, respectively), applied as the
HPLC and TLC standards, were synthesised as described
previously.3
(v) The sequence and mechanism of glycosylation reaction in
the guanine series is different to that documented for aden-
ine.12,13 While adenine is initially glycosylated at N3 and the
resulting kinetic product undergoes irreversible rearrangement
to the 9-substituted adenine, guanine is directly glycosylated at
either N7 or N9 (see point ii), and the reversible 7 9 transgly-
cosylation accounts for the final distribution of regioisomers.10
7-(2Ј,3Ј,5Ј-Tri-O-acetyl-â-D-ribofuranosyl)-N2-acetylguanine 6
Tetraacetylguanosine 1 (2.0 g, 4.43 mmol) was heated in an
open flask without solvents at 220 ЊC (temperature of an oil
bath) for 10 min. The resulting dark melt was dissolved in
toluene–ethanol (6:1, v/v) and chromatographed on a silica gel
short column (5 × 9 cm) in this solvent system. Evaporation of
fractions 20–25 (ca. 20 cm3) yielded chromatographically pure 6
as an oil. The oil was redissolved in chloroform (5 cm3), and
added dropwise to a stirred mixture of diethyl ether–hexane
(300 cm3; 1:1, v/v). The white precipitate was collected by fil-
tration, washed with hexane and dried under reduced pressure
over P2O5 to yield compound 6 as a white powder (0.762 g,
38%). The spectroscopic and analytical data are presented in
Tables 1–3.
Experimental
UV spectra were measured in methanol on a Beckman DU-65
1
spectrophotometer. H NMR spectra were recorded on a Var-
ian Unity 300 FT NMR spectrometer with SiMe4 as an internal
standard. Elemental analyses were performed on a Perkin-
Elmer 240 Elemental Analyzer. TLC was conducted on Merck
silica gel F254 60 plates using the following solvent systems
(measured by volume): A, chloroform–methanol (9:1); B,
toluene–ethanol (4:1). For preparative short-column chroma-
tography Merck TLC gel HF254 60 was used.
Fractions 38–52 contained pure 1 and were evaporated to
recover the unreacted substrate as an oil (1.180 g, 59%).
Reactions of 9- and 7-(2Ј,3Ј,5Ј-tri-O-acetyl-â-D-ribofuranosyl)-
N2-acetylguanines with 2-acetoxyethyl acetoxymethyl ether
General procedure. The peracetylated ribonucleoside (1 or 6,
Analytical high performance liquid chromatography (HPLC)
was performed using the following components from Waters
J. Chem. Soc., Perkin Trans. 2, 1997
651

View Article Online
22.6 mg, 0.05 mmol) and 2-acetoxyethyl acetoxymethyl ether 17
(88 mg, 0.5 mmol) were dissolved in dry chlorobenzene (2 cm3).
Reactions were initiated by injecting a solution of toluene-p-
sulfonic acid monohydrate (0.95 mg, 0.005 mmol) in
acetonitrile (50 mm3) into the pre-thermostatted reaction
medium. The transpurination experiments were carried out at
120 ЊC with stirring in stoppered flasks immersed in a thermo-
statted oil bath for 2 h. Aliquots of 10 mm3 (ca. 25 nmol) were
withdrawn at suitable time intervals (see Figs. 1 and 2). The
samples of reaction mixtures were then evaporated to dryness
under diminished pressure at room temp. The residues resulting
from evaporation of each sample were dissolved in 50% aq.
methanol (200 mm3) and the solutions obtained were used for a
preliminary qualitative TLC analysis in solvents A and B, and
for the HPLC quantitative analysis (see below).
Results of the HPLC method were also compared to those
obtained in the 1H NMR study. Thus, the transpurination
experiments of compounds 1 and 6 were repeated on a 0.05
mmol scale for chosen time intervals (20 and 60 min). The reac-
tion mixtures were then evaporated to dryness, the resulting oils
were redissolved in (CD3)2SO and directly analysed by 1H
NMR spectroscopy. The mole fractions of compounds 1, 6, 8
and 3 (estimated by integration) were in good agreement (±4%)
with the values calculated from the HPLC method for the same
time intervals.
HPLC Analysis. Distribution of products was quantitatively
determined by HPLC, applying the samples of reaction mix-
tures described above. Elution with a methanol–water reversed
gradient (from 25–60% of aqueous methanol after 20 min, then
60% aq. methanol for the next 10 min) gave a good separation
for compounds 1, 3, 6, 8. The flow rate was set at 1 cm3 minϪ1
and UV absorption was measured at 260 nm. The following
average ε260 values were taken for calculation of mole frac-
tions: 16 000 dm3 molϪ1 cmϪ1 for the 9-isomers (compounds 1
and 8) and 13 300 dm3 molϪ1 cmϪ1 for the 7-isomers (com-
pounds 6 and 3). The assignment of the peaks was performed
by comparing the retention times and UV spectra with those
of the original samples. Retention times (in order of elu-
tion): compound 8, 9.2; 3, 10.6; 1, 23.3 and 6, 24.9 min. The
results are presented in Figs. 1 and 2. In addition to those
four compounds, the reaction mixtures contained up to 0.5%
of N2-acetylguanine (retention time 4.3 min), not shown in
Figs. 1 or 2.
Acknowledgements
This study was supported by the Komitet Badan Naukowych,
Project No 2 2671 92 03.
References
1 F. W. Lichtenhaller and K. Kitahara, Angew. Chem., 1975, 87, 815.
2 J. Boryski and B. Golankiewicz, Nucleosides, Nucleotides, 1987, 6,
385.
3 J. Boryski and B. Golankiewicz, Nucleosides, Nucleotides, 1989, 8,
529.
4 H. Shiragami, Y. Koguchi, Y. Tanaka, S. Takamatsu, Y. Uchida,
T. Ineyama and K. Izawa, Nucleosides, Nucleotides, 1995, 14, 337.
5 T. Azuma and K. Isono, Chem. Pharm. Bull., 1977, 25, 3347.
6 M. Imazawa and F. Eckstein, J. Org. Chem., 1978, 43, 3044.
7 M. Imazawa and F. Eckstein, J. Org. Chem., 1979, 44, 2039.
8 T. Uematsu and R. J. Suhadolnik, Nucleic Acids Chemistry (Part 1),
eds. L. B. Townsend and R. S. Tipson, Wiley, New York, 1978,
p. 405.
9 Y. Mizuno, K. Tsuchida and H. Tampo, Nucleic Acids Res. Symp.
Ser. 1982, 45.
10 Y. Mizuno, K. Tsuchida and H. Tampo, Chem. Pharm. Bull., 1984,
32, 2915.
11 J. Boryski and B. Golankiewicz, Nucleic Acids Res. Symp. Ser., 1987,
18, 45.
12 M. Miyaki and B. Shimidzu, Chem. Pharm. Bull., 1970, 18, 1446.
13 K. A. Watanabe, D. H. Hollenberg and J. J. Fox, J. Carbohydr.
Nucleosides Nucleotides, 1974, 1, 1, and references cited therein.
14 L. W. Dudycz and G. E. Wright, Nucleosides, Nucleotides, 1984, 3, 33.
15 Preliminary communication: J. Boryski, Ninth Symposium on the
¯
Chemistry of Nucleic Acids Components, Tre3t Castle, The Czech
Republic, September 13–17, 1993; Collect. Czech. Chem. Commun.
Special Issue, 1993, 58, 5.
16 B. R. Baker, R. E. Schaub and H. M. Kissman, J. Am. Chem. Soc.,
1955, 77, 5911.
17 A. Rosovsky, S. H. Kim and M. Wick, J. Med. Chem., 1984, 24, 1177.
18 The term ‘intramolecular’ refers to the apparent results of the
reversible 7 9 transglycosylation, not to its mechanism, which is
intermolecular.
19 J. Boryski, Nucleosides, Nucleotides, 1995, 14, 287.
20 J. Boryski and A. Manikowski, Nucleosides, Nucleotides, 1995, 14,
287.
21 C. B. Reese and R. J. Saffhill, J. Chem. Soc., Perkin Trans. 2, 1972,
2937.
Paper 6/02021F
Received 22nd March 1996
Accepted 5th November 1996
652
J. Chem. Soc., Perkin Trans. 2, 1997