Preparation of the 2′-deoxynucleosides of 2,6-diaminopurine and isoguanine by direct glycosylation

Article abstract of DOI:10.1021/jo902616s

Chemical Equetion Presentation The purine nucleoside 2,6-diaminopurine- 2′-deoxyriboside is prepared by the direct glycosylation of the 2,6-bis(tetramethylsuccinimide) derivative of the parent purine heterocycle 4 with 2-deoxy-3,5-di-O-(p-toluoyl)-α-D-erythro-pentofuranosyl chloride 5 using the sodium salt method. 2′-Deoxyisoguanosine is prepared from 2,6-diaminopurine by a five-step procedure. The purine heterocycle isoguanine is prepared by selective diazotization of 2,6-diaminopurine and then converted to the N9-trityl derivative to increase solubility. After silylation of the O ²-carbonyl with TMSCl, the N⁶-amino group is protected as the tetramethylsuccinimide (M₄SI). The O²-carbonyl is protected as the DPC derivative, and the trityl group is removed. The resulting product is glycosylated in good yield to generate fully protected 2′-deoxyisoguanosine.

Full text of DOI:10.1021/jo902616s

pubs.acs.org/joc
Preparation of the 20-Deoxynucleosides of 2,6-Diaminopurine and
Isoguanine by Direct Glycosylation
Joseph W. Arico, Amy K. Calhoun, and Larry W. McLaughlin*
Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02467
mclaughl@bc.edu
Received December 14, 2009
The purine nucleoside 2,6-diaminopurine-20-deoxyriboside is prepared by the direct glycosylation of
the 2,6-bis(tetramethylsuccinimide) derivative of the parent purine heterocycle 4 with 2-deoxy-3,5-di-
O-(p-toluoyl)-R-D-erythro-pentofuranosyl chloride 5 using the sodium salt method. 20-Deoxyiso-
guanosine is prepared from 2,6-diaminopurine by a five-step procedure. The purine heterocycle
isoguanine is prepared by selective diazotization of 2,6-diaminopurine and then converted to the
N9-trityl derivative to increase solubility. After silylation of the O2-carbonyl with TMSCl, the
N6-amino group is protected as the tetramethylsuccinimide (M4SI). The O2-carbonyl is protected as
the DPC derivative, and the trityl group is removed. The resulting product is glycosylated in good
yield to generate fully protected 20-deoxyisoguanosine.
Introduction
N1-oxide of adenosine.4 A much simpler synthesis of the
20-deoxynucleoside and corresponding phosphoramidite
was described by Gaffney and Jones;5 it involved sulfonation
of the O6 carbonyl of dG followed by treatment with
trimethylamine and then ammonia to generate the 2,6-di-
aminopurine-20-deoxynucleoside. That procedure has been
improved by preparation of the O6-pentafluorophenyl deri-
vative, which can be converted in high yield to 2,6-diamino-
purine-20-deoxyriboside by treatment with ammonia. They
also report that this synthon can be incorporated into DNA
and then converted to 2,6-diaminopurine.6 A recent report
also describes the preparation of the ribonucleoside directly
from guanosine in high yield under high pressure condi-
tions using HMDS and triflic acid.7 One drawback to all of
these procedures is that they use a nucleoside starting
material, making them less cost-effective and less flexible
if, for example, a nucleoside bearing a modified sugar were
desired.
Among their many roles, nucleoside analogues can be used
to modulate stability for DNA, RNA, or DNA/RNA du-
plexes. 2,6-Diaminopurine-20-deoxyriboside or the corre-
sponding ribonucleoside is often used in this role because
of the presence of a minor groove amino group that can
provide a third hydrogen bond for dA-dT base pairs, or
corresponding RNA A-U base pairs, and enhance overall
helix stability.1 20-Deoxyisoguanosine has also been of inter-
est for its hydrogen-bonding capability but perhaps more so
for the observation that it pairs with isoC and provides a fifth
base pair to expand the genetic code in artificial genetic
systems.2,3
An early description of the preparation of 2,6-diamino-
purine riboside was a multistep pathway beginning with the
(1) Strobel, S. A.; Cech, T. R.; Usman, N.; Beigelman, L. Biochemistry
1994, 33, 13824–13835.
(2) Switzer, C.; Moroney, S. E.; Benner, S. A. J. Am. Chem. Soc. 1989,
111, 8322–8323.
(3) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner, S. A. Nature
(London) 1990, 343, 33–7.
(4) Tohru, U.; Kazunobu, M.; Tsuguo, K. Chem. Pharm. Bull. (Tokyo)
1978, 26, 2122–2127.
(5) Gaffney, B. L.; Marky, L. A.; Jones, R. A. Tetrahedron 1984, 40, 3–13.
(6) Gao, H.; Fathi, R.; Gaffney, B. L.; Goswami, B.; Kung, P. P.;
Youngsook, R.; Jin, R.; Jones, R. A. J. Org. Chem. 1992, 57, 6954–6959.
(7) Ross, B. S.; Springer, R. H.; Ravikumar, V. T. Nucleosides, Nucleo-
tides Nucleic Acids 2008, 27, 67–69.
1360 J. Org. Chem. 2010, 75, 1360–1365
Published on Web 02/10/2010
DOI: 10.1021/jo902616s
r
2010 American Chemical Society

Arico et al.
JOCFeatured Article
The alternative approach involves glycosylation of 2,6-di-
aminopurine or related derivative using for example, 2-deoxy-
3,5-di-O-(p-toluoyl)-R-D-erythro-pentofuranosyl chloride8 (5)
to generate a precursor to the diaminonucleoside. To achieve
highly diastereomeric yields of 20-deoxynucleosides the sodium
salt method9 is often used, requiring that the purine be activated
by a base (commonly NaH). This procedure is untenable when
free amines or even amides are present, but this problem can be
eliminated by using 2,6-dichloropurine (available from
xanthine) for glycosylation.8,10,11 Treatment of chloropurines
with NaH generates the isomeric N9/N7 salt, which upon
reaction with the chlorinated carbohydrate typically results in
a mixture of N9 and N7 purine nucleosides. The major product
is the N9-β-isomer; the minor product is the N9-R-isomer.8 The
sodium salt method requires polar solvents, which can accele-
rate the anomerization of the chloro-sugar and generate some
of the diastereomeric R-nucleoside. Reaction of the product
2,6-dichloropurine nucleoside with ammonia displaces the
chlorine at the C6 position, but not the one at C2.11 The latter
site requires a harsher hydrazine treatment followed by Raney
nickel reduction. Alternatively, NaN3 can be used to displace
the chlorines at both positions; the product diazide can then be
reduced to the desired 2,6-diaminopurine nucleoside.10 One
drawback with this overall approach is the larger number of
steps needed, first to eliminate the amines and then to regene-
rate them in order to obtain the desired 2,6-diaminopurine-20-
deoxyriboside.
FIGURE 1. Protected derivatives of (a) 2,6-diaminopurine-20-
deoxyriboside and (b) 20-deoxyisoguanosine resulting from glyco-
sylation of the corresponding purines.
SCHEME 1
Syntheses of 20-deoxyisoguanosine12 and isoguanosine suf-
fer from similar issues. Reported syntheses generally employ
a nucleoside starting material. For example, after preparing
the N1-oxide of 20-deoxyadenosine a photolytic rearrange-
ment generates the desired 20-deoxyisoguanosine13,14 in
a reported 42% yield.15 A second reported synthesis requires
four steps from guanosine to generate 2-iodoadenosine,
which could then be photolyzed to isoguanosine in a reported
55% yield.16 The other common approach to obtain the
isoguanine nucleosides is to selectively deaminate 2,6-
diaminopurine riboside or the 20-deoxy derivative with
nitrous acid.14,17,18 These deaminations still require a costly
nucleoside starting material and are limited to ribose or
20-deoxyribose sugars.
noted purity results from product isolation by precipitation.
The absence of other reports for isoguanine glycosylations
likely stems from the extremely poor solubility of isoguanine,
estimated to be 2.2 mg/L of water,20 and virtual insolubility
in organic solvents.
Here we describe simple and straightforward procedures
involving glycosylation of protected aminopurine derivatives
to generate the fully protected nucleosides of 2,6-diamino-
purine (1) and isoguanine (2) (Figure 1). The use of the
tetramethylsuccinimide (M4SI) protecting group21 affords
purine derivatives that display excellent solubility and are
excellent coupling partners in the sodium salt procedure. These
derivatives can be further elaborated to DMT-protected
nucleoside phosphoramidites for DNA synthesis, or can be
fully deprotected to generate the parent nucleosides 2,6-di-
aminopurine-20-deoxyriboside and 20-deoxyisoguanosine.
Reported procedures for the glycosylation of isoguanine
are limited. To the best of our knowledge, there has been no
report of direct glycosylation of isoguanine to prepare
20-deoxyisoguanosine. Bookser and Raffaele describe19
a
€
Vorbruggen glycosylation using microwave acceleration to
generate isoguanosine in 38% yield and in 79% purity; the
(8) Wright, G. E.; Hildebrand, C.;Freese, S.; Dudycz, L. W.; Kazimierczuk,
Z. J. Org. Chem. 1987, 52, 4617–4618.
(9) Kazimierczuk, Z.; Cottam, H. B.; Revankar, G. R.; Robins, R. K.
J. Am. Chem. Soc. 1984, 106, 6379–6382.
Results and Discussion
Glycosylation reactions to form purine 20-deoxynucleo-
sides commonly use the sodium salt method9 whereby the
purine is treated with base (e.g., NaH) and the purine salt
thus formed performs an SN2 displacement of chloride from
2-deoxy-3,5-di-O-(p-toluoyl)-R-D-erythro-pentofuranosyl
chloride (5, Scheme 1); the reaction usually generates a
mixture of N9 and N7 nucleosides in the β-configuration,
(10) Naval, M.; Thibaut, M.; Vasseur, J. J.; Debart, F. Biorg. Med. Chem.
Lett. 2002, 12, 1435–1438.
(11) Christensen, L. F.; Broom, A. D. J. Med. Chem. 1972, 15, 735–739.
(12) Kazimierczuk, Z.; Mertens, R.; Seela, F. Helv. Chim. Acta 1991, 74,
1742–1752.
(13) Roberts, C.; Bandaru, R.; Switzer, C. Tetrahedron Lett. 1995, 36,
3601–3604.
(14) Seela, F.; Gabler, B. Helv. Chim. Acta 1994, 77, 8322–8323.
(15) Roberts, C.; Bandaru, R.; Switzer, C. J. Am. Chem. Soc. 1997, 119,
4640–4649.
(16) Nair, V.; Young, D. J. Org. Chem. 1985, 50, 406–408.
(17) Jurczyk, S. C.; Kodra, J. T.; Park, J. H.; Benner, S. A.; Battersby,
T. R. Helv. Chim. Acta 1999, 82, 1005–1014.
(20) Bendich, A.; Brown, G. B.; Philips, F. W.; Thiersch, J. B. J. Biol.
Chem. 1950, 267–277.
(18) Davol, J. J. J. Am. Chem. Soc. 1951, 73, 3174–3179.
(19) Bookser, B. C.; Raffaele, N. B. J. Org. Chem. 2006, 72, 173–179.
(21) Arico, J.; Calhoun, A. K.; Salandria, K.; McLaughlin, L. W. Org.
Lett. 2010, 12, 120–122.
J. Org. Chem. Vol. 75, No. 5, 2010 1361

Arico et al.
JOCFeatured Article
SCHEME 2
SCHEME 3
with R-anomers sometimes formed due to anomerization of
the sugar. Purines containing free amines or amides generally
have poor solubility and give low product yield with the
sodium salt method.
Glycosylation of isoguanine has proven difficult with a
single report noted above using high temperatures facilitated
by microwave treatment.19 2,6-Diaminopurine (Scheme 3) is
a readily available starting material for 20-deoxyisoguano-
sine, and similar to reports14,17,18 for the corresponding
nucleoside, we could selectively deaminate it by dripping a
solution of 2,6-diaminopurine in aqueous NaOH/NaNO2
into an acidic solution and then heating at 60 °C for 1 h.
Precipitation gave isoguanine (9) in excellent yield. Owing to
its poor solubility, isoguanine was treated with excess
HMDS under reflux and then treated with trityl chloride to
generate 10. The N9-trityl derivative of isoguanine exhibited
much improved solubility but the exocyclic amine could not
be protected with M4SA; the low reactivity of the amino
group of isoguanosine has also been noted by Seela.22
Transient protection of isoguanosine by treatment with
TMSCl has been reported to enhance the reactivity of
N6,22 presumably as the result of silylation of the O2-carbo-
nyl. We adopted a similar course of action and treated 10
with TMSCl and DBU followed by M4SA and obtained 11 in
67% yield. Removal of the trityl group from 11 with acetic
acid followed by glycosylation of the resulting the M4SI-
protected nucleobase generated at least four different pro-
ducts, suggesting the presence of reactive nucleophiles in
addition to N9. This observation is also consistent with
reactions described by Seela22 who noted that alkylation
could take place at N1, N3, and N9, suggesting extensive
delocalization of the isoguanyl anion.
We wished to avoid the use of the costly 2,6-dichloro-
purine glycosylation partner and examine direct glycosyl-
ation of the readily available 2,6-diaminopurine. To accom-
plish this task, we needed to protect the exocyclic amines in
such a manner that sodium hydride treatment generated only
the N7/N9 purine salt. We chose to use tetramethylsuccinic
anhydride (M4SA)21 to form the bis(tetramethylsuccin-
imide) of 2,6-diaminopurine. After refluxing in the presence
of DBU for 22 h we obtained the fully protected purine
heterocycle 4 (Scheme 1) as a mixture with approximately
10% of the corresponding amide. After partial puri-
fication, this mixture was fully converted to 4 in an overall
yield of 74% by refluxing in anhydrous pyridine an addi-
tional 8 h.
The glycosylation of 4 occurred under standard conditions
using sodium hydride to activate the fully protected hetero-
cycle and 2.1 equivalents of the glycosylating sugar 5. Under
these conditions and at ambient temperature we obtained the
β-N9 product 1 in 93% yield (Scheme 1). None of the N7
glycosylation product and neither of the possible R-anomers
were observed.
The toluoyl groups could be removed from 1 using
ammonia/methanol (Scheme 2), but during this reaction
partial deprotection of the N6-M4SI was observed (f6).
However, it was possible to convert 1 to 6 employing aq
NaOH in 2:1 pyridine/ethanol, and after purification protect
the exocyclic amine as the dimethylformamidine derivative
(7). The material obtained in this manner can be converted to
a DMT-protected nucleoside phosphoramidite using stan-
dard protocols. The coupling product 1 could also be fully
deprotected to yield the 20-deoxynucleoside 8.
As an alternative, the DPC group (one of the preferred
O2-protecting groups for monomers of isoguanine used in
DNA synthesis protocols13) was introduced at O2 (f12) and
(22) Seela, F.; Wei, C.; Kazimierczuk, Z. Helv. Chim. Acta 1995, 78, 1843–
1854.
1362 J. Org. Chem. Vol. 75, No. 5, 2010

Arico et al.
JOCFeatured Article
SCHEME 4
trend in coupling product ratios; it may be that another
factor such as aggregation of the sodium salt of 12 during
long reaction times with NaH leads to the observed effect.
No similar effects were observed either with coupling of 4 or
with other M4SI-protected purine substrates.21
The final step of the synthesis required the selective
cleavage of the toluoyl esters (Scheme 4), but this proved
infeasible under a number of conditions (NH3/MeOH,
NaOH at 4 °C or K2CO3 at ambient temperature) as the
M4SI group was the most labile protecting group. The DPC
group was stable enough that 13 could be isolated in ∼50%
yield. However, Seela has reported24 that 13 cannot be
converted to the amidine derivative (14). Therefore, the
simplest course of action was to fully deprotect 2 using
ammonia in methanol at 60° overnight to obtain 15. The
conversion of 20-deoxyisoguanosine (15) to the fully pro-
tected DMT-nucleoside phosphoramidite (Scheme 4) has
been described.15,24
Conclusions
We describe simple and straightforward procedures for
the glycosylation of protected derivatives of 2,6-diamino-
purine and isoguanine. Both procedures rely upon the use of
the bidentate protecting group M4SI. The bidentate nature
of the M4SI group eliminates any side reactions that result
from proton abstraction at the exocyclic amines (or corre-
sponding amides); it also provides enhanced solubility for
purine heterocycles particularly for isoguanine. The M4SI
protecting group exhibits exceptional directing ability such
that high ratios of the N9 vs N7 glycosylation products
result. These valuable nucleosides can be further elaborated
to generate nucleoside phosphoramidites for solid-phase
DNA syntheses using standard protocols.
the trityl group removed. We then glycosylated the fully
protected heterocycle (12) using 5. Glycosylation of 12
proceeded smoothly to generate a quantitative yield of
coupling products with 70-79% of the desired β-N9 pro-
duct, 5-15% R-N9, and 15% R/β-N7 in an approximately
1:1 ratio. The wide variation in yield of the R-N9 pro-
duct prompted us to optimize the reaction conditions to
both minimize this product and render the reaction more
reproducible.
The same amount of R-N9 (12-15%) was produced with
either acetonitrile or dichloromethane as the solvent after
allowing 1 h for formation of the purine sodium salt with
NaH and followed with an overnight (16 h) reaction time
after addition of 5. Stopping the reaction at 1, 5, or 16 h after
addition of the sugar indicated that the highest yield of
coupling products (quantitative) was obtained after 5 h;
the same amount of R-N9 (12-15%) was obtained in each
case. To our surprise, the least amount of R-N9 (4-5%) was
produced when the reaction time of 12 with NaH was
minimized (20 min) and all other conditions were kept
identical, while allowing 1 h or more for reaction with
NaH gave the most R-N9 product. Either acetonitrile or
dichloromethane could be used as the reaction solvent with
identical yields and identical ratios of products obtained.
Therefore, the reaction time after addition of the sugar was
not important except in terms of the overall yield. This result
is surprising given that the sugar 5 is reported to anomerize at
a faster rate in polar solvents23 and thus an increasing yield of
R-nucleosides should result with increasing reaction times.
This rationale does not adequately explain our observed
Experimental Section
General Procedures. All reactions were carried out in oven-
dried glassware under an inert atmosphere with dry solvents and
anhydrous conditions, unless otherwise noted. Dry tetrahydro-
furan (THF), diethyl ether (Et2O), N,N-dimethylformamide
(DMF), pyridine (pyr), acetonitrile (MeCN), and dichloro-
methane (DCM) were obtained by passing commercially avail-
able predried, oxygen-free formulations through activated
alumina columns. Dry methanol (MeOH) was obtained by
distillation from Mg(OMe)2. Yields refer to chromatographi-
cally and spectroscopically (1H NMR) homogeneous materials,
unless otherwise stated. Reagents were purchased at the highest
commercial quality and used without further purification, un-
less otherwise stated. Reactions were monitored by thin-layer
chromatography (TLC) carried out on 0.25 mm Silicycle TLG-
˚
R10011B-323 60 A plates using UV light as visualizing agent
and ceric ammonium molybdate (CAM) stain and heat as
developing agents. Dynamic Adsorbents silica gel (particle size
32-63 μm) was used for flash column chromatography. Diffi-
cult separations were carried out using “chromatospec silica gel”
˚
(E. Merck Chromatospec silica gel (60 A pore size, particle size
15-40 μm). NMR spectra were recorded on 400, 500, or 600
instruments and calibrated using residual undeuterated solvent
(CDCl3: δH = 7.26 ppm, δC = 77.16 ppm, acetone-d6: δH
=
2.05 ppm, δC=29.84, DMSO-d6: δH=2.50 ppm, δC=39.52 ppm,
methanol-d4: δH =3.34 ppm, δC =49.00 ppm) as an internal
reference. The following abbreviations are used to designate
(23) Kotick, M. P.; Szantay, C.; Bardos, T. J. J. Org. Chem. 1969, 34,
3806–3811.
(24) Seela, F.; Wei, C. Helv. Chim. Acta 1997, 80, 73–85.
J. Org. Chem. Vol. 75, No. 5, 2010 1363

Arico et al.
JOCFeatured Article
the multiplicities: s=singlet, d=doublet, t=triplet, q=quartet,
quin=quintet, m=multiplet, br=broad. High-resolution mass
spectra (HRMS) were recorded using ESI (electrospray
ionization) or DART (direct analysis in real time).
8.27 (s, 1H), 7.01 (brs, 2H), 6.41 (dd, J=8.8, 5.7 Hz, 1H), 5.00
(brs, 1H), 4.59 (m, 1H), 4.42 (brs, 1H), 4.05 (dd, J=4.2, 2.6 Hz,
1H), 3.74 (dd, J=12.3, 2.4 Hz, 1H), 3.68-3.56 (m, 1H), 2.83
(ddd, J=13.2, 8.9, 5.4 Hz, 2H), 2.35 (s, 1H), 1.26 (s, 12H); 13
C
N,N0-(9H-Purine-2,6-diyl)bis(2,2,3,3-tetramethylsuccinimide)
(4). To a flask containing 2,6-diaminopurine 3 (257 mg, 1.71
mmol) and 2,2,3,3-tetramethylsuccinic anhydride (1.604 g,
10.27 mmol) were added 13 mL of pyridine and 1.02 mL of
DBU (6.84 mmol). The mixture was heated to reflux with
stirring for 20 h. The starting material gradually dissolved as
the reaction proceeded. TLC (9:1 DCM/MeOH containing
0.7 M NH3) indicated that some starting material remained.
Another 267 mg (1.71 mmol) of anhydride was added, and
heating continued for another 10 h. TLC showed that no
starting material remained. Volatiles were removed in vacuo,
and the residue coevaporated with toluene (3 ꢀ 3 mL). The
resulting brown oil was purified by flash chromatography
eluting with 95:5 DCM/MeOH) to obtain 738 mg of white solid.
NMR (acetone-d6, 126 MHz) δ 181.1, 158.4, 150.8, 148.6, 142.4,
90.2, 87.4, 73.2, 63.7, 48.1, 41.3, 21.6, 21.4; HRMS (ESI-TOF)
calcd for C18H24N6O5Naþ [M þ Na]þ 427.1706, found
427.1709.
9-[2-Deoxy-β-D-erythro-pentofuranosyl]-6-[[1-(dimethylamino)-
ethylidene]amino]-2-(3,3,4,4-tetramethyl-2,5-dioxopyrrolidin-1-yl)-
purine (7). To a flask containing 6 (36 mg, 0.089 mmol) was added
MeOH (0.5 mL) followed by N,N-dimethylformamide dimethyl
acetal (18 μL, 0.133 mmol). The mixture was allowed to stir at rt
for 24 h, at which time TLC indicated that all starting material had
been consumed. Volatiles were removed in vacuo, and the residue
was purified by flash chromatography eluting with 9:1 DCM/
MeOH to obtain 35 mg (85%) of 7 as a white foam. 7: Rf=0.29
(silica gel, 9:1 DCM/MeOH); 1H NMR (acetone-d6, 500 MHz) δ
8.98 (s, 1H), 8.39 (s, 1H), 6.46 (dd, J=8.8, 5.7 Hz, 1H), 4.92 (brs,
1H), 4.65 - 4.56 (m, 1H), 4.46 (brs, 1H), 4.07 (dd, J=4.5, 2.7 Hz,
1H), 3.76 (dd, J=12.3, 2.6 Hz, 1H), 3.65 (d, J=11.9 Hz, 1H), 3.26
(s, 3H), 3.21 (d, J=0.6 Hz, 3H), 2.86 (ddd, J=13.2, 8.8, 5.4 Hz,
1H), 2.37 (ddd, J=13.1, 5.8, 1.9 Hz, 1H), 1.28 (d, J=1.1 Hz, 12H);
13C NMR (acetone-d6, 126 MHz) δ 181.3, 162.2, 159.9, 152.8,
148.1, 143.6, 127.3, 90.1, 87.2, 73.2, 63.7, 48.1, 41.3, 41.3, 35.1, 21.6,
21.4; HRMS (ESI-TOF) calcd for C21H29N7O5Naþ [M þ Na]þ
482.2128, found, 482.2122.
1
This material contained the amides (∼10% by H NMR) that
coeluted with the desired product. The solid was thoroughly
dried in vacuo and treated with SOCl2 (3 mL) at reflux for 2 h
with a drying tube attached or dissolved in anhydrous pyridine
(10 mL) and heated to reflux for 8 h. Volatiles were removed in
vacuo, and the residue was coevaporated with EtOAc or
toluene, respectively (3 ꢀ 3 mL). The residue was purified by
flash chromatography eluting with 95:5 DCM/MeOH to obtain
546 mg (74%) of 4 as a white amorphous powder. 4: Rf=0.7
(silica gel, 9:1 DCM/MeOH); 1H NMR (DMSO-d6, 400 MHz) δ
14.25 (brs, 1H), 8.84 (brs, 1H), 1.29 (s, 12H), 1.24 (s, 12H); 13C
NMR (DMSO-d6, 101 MHz) δ 180.5, 180.1, 146.4, 47.7, 47.2,
21.1, 20.9; HRMS (ESI-TOF) calcd for C21H26N6O4Naþ [M þ
Na]þ 449.1940, found 449.1920.
Isoguanine (6-Amino-1H-purin-2(9H)-one) (9). 2,6-Diamino-
purine (2.0 g, 13.3 mmol) was dissolved in 1.67 M NaOH
(20 mL), and NaNO2 (1.19 g, 17.3 mmol) was added. The
resulting solution was transferred to an addition funnel and
added dropwise over 1 h, with stirring, to a flask containing 78%
H2SO4 (8 mL) immersed in a water bath at 20 °C. The tempera-
ture was carefully regulated at 20 °C during this time. After the
solution of 2,6-diaminopurine had been added, the temperature
was raised to 60 °C and allowed to stir for 1 h. The mixture was
allowed to cool to room temperature and filtered to remove the
precipitate that had formed. The precipitate was washed with
10 mL H2O, air-dried, then redissolved in 0.9 M NaOH (40 mL).
To this solution was added NaHSO3 (122 mg). The solution was
heated to 90 °C with stirring and slowly neutralized to pH 7-8
by adding 50% H2SO4 dropwise. A precipitate formed, and after
the neutralized mixture was allowed to cool to room tempera-
ture it was filtered. The precipitate was washed with 50 mL of
H2O and dried overnight in a vacuum oven at 50 °C to yield 9 as
an off-white amorphous solid (1.837 g, 91%). The product was
virtually insoluble in organic solvents and H2O but readily
soluble in alkali. The physical data were in agreement with that
reported.25 9: mp >360 °C; UV λmax=283 (0.1 M NaOH) (8.0);
IR (KBr pellet) 3051, 2785, 1828, 1698, 1666, 1524, 1450, 1398,
9-[2-Deoxy-3,5-di-O-(p-toluoyl)-β-D-erythro-pentofuranosyl]-
2,6-di-(3,3,4,4-tetramethyl-2,5-dioxopyrrolidin-1-yl)purine (1).
To a flask containing 4 (200 mg, 0.467 mmol) and NaH
(45 mg, 0.984 mmol, 50% dispersion in oil) was added 20 mL
of MeCN. The mixture was allowed to stir for 1 h, and then
2-deoxy-3,5-di-O-p-toluoyl-R-D-erythro-pentofuranosyl chlo-
ride 5 (382 mg, 0.984 mmol) was added. Stirring was continued
overnight. The reaction mixture was filtered through Celite and
the filter cake washed with acetone (25 mL). The filtrate was
evaporated in vacuo, and the resulting yellow foam was purified
by flash column chromatography eluting with 98.25:1.75 DCM/
MeOH to obtain 338 mg (93%) of 1 as a white foam. 1: Rf=0.55
(silica gel, 97:3 DCM/MeOH); 1H NMR (acetone-d6, 400 MHz)
δ 8.79 (s, 1H), 7.97 (d, J=8.2 Hz, 2H), 7.87 (d, J=8.2 Hz, 2H),
7.34 (d, J=8.2 Hz, 2H), 7.26 (d, J=8.1 Hz, 2H), 6.80 (s, 1H), 5.99
(s, 1H), 4.79 (dd, J = 10.5 Hz, J=5.0 Hz, 1H), 4.74 - 4.59 (m,
2H), 3.55 - 3.42 (m, 1H), 3.01 (ddd, J=14.5, 6.6, 3.4 Hz, 1H),
2.42 (s, 3H), 2.38 (s, 3H), 1.35 (s, 12H), 1.31 (s, 12H); 13C NMR
(acetone-d6, 101 MHz) δ 180.9, 180.6, 166.5, 166.2, 155.2, 148.7,
148.3, 147.2, 145.1, 144.7, 131.7, 130.5, 130.4, 130.1, 130.0,
128.1, 127.9, 86.6, 83.7, 76.0, 65.0, 49.0, 48.4, 37.1, 21.6, 21.5,
21.5, 21.5, 21.4; HRMS (ESI-TOF) calcd for C42H46N6O9Naþ
[M þ Na]þ 801.3224, found 801.3207.
1236, 1179, 1119, 1028, 940, 852, 775 cm-1
.
6-Amino-9-trityl-1H-purin-2(9H)-one (10). To a flask con-
taining isoguanine 9 (530 mg, 3.50 mmol) and (NH4)2SO4
(35 mg, 0.26 mmol) was added HMDS (53 mL). The mixture
was heated to reflux with stirring for 18 h. The excess HMDS
was removed in vacuo and the flask purged with nitrogen. Trityl
chloride (988 mg, 3.54 mmol) and MeCN (53 mL) were added,
and the mixture was allowed to stir for 8 h at rt under inert
atmosphere. Volatiles were removed in vacuo, and the residue
was purified by flash chromatography eluting with 9:1 DCM/
MeOH to obtain 1.268 g (92%) of 10 as a white amorphous
6-Amino-9-(2-deoxy-β-D-erythro-pentofuranosyl)-2-(3,3,4,4-
tetramethyl-2,5-dioxopyrrolidin-1-yl)purine (6). Note: anhydrous
conditions not needed. To a flask containing 1 (123 mg, 0.158
mmol) dissolved in 2:1 pyr/EtOH (1.4 mL) at 0 °C was added
1 M NaOH (0.7 mL) dropwise. The mixture was allowed to stir
at 0 °C for 2 h. The reaction was quenched by adding NH4Cl
(45 mg, 0.19 mmol). Volatiles were removed in vacuo, and the
residue was redissolved in minimal MeOH. Silica gel (700 mg)
was added, and volatiles were removed in vacuo. The
thoroughly dried powder was loaded onto a flash column and
purified by elution with 9:1 DCM/MeOH then 6:1f5:1 DCM/
EtOH to obtain 45 mg (72%) of 6 as a white foam. 6: Rf=0.13
(silica gel, 9:1 DCM/MeOH); 1H NMR (acetone-d6, 500 MHz) δ
1
powder. 10: Rf=0.38 (silica gel, 9:1 DCM/MeOH); H NMR
(DMSO-d6, 400 MHz) δ 10.21 (brs, 1H), 7.39 - 7.23 (m, 12H),
7.18-7.11 (m, 6H); 13C NMR (DMSO-d6, 101 MHz) δ 141.4,
(25) Chern, J. W.; Lee, H. Y.; M., H.; Shish, F. J. Tetrahedron Lett. 1987,
28, 2151–2154.
1364 J. Org. Chem. Vol. 75, No. 5, 2010

Arico et al.
JOCFeatured Article
129.4, 127.8, 127.3, 74.2; HRMS (ESI-TOF) calcd for
C24H19N5OHþ [M þ H]þ 394.1668, found 394.1662.
21.4; HRMS (ESI-TOF) calcd for C26H24N6O4Naþ [M þ Na]þ
507.1757, found 507.1745.
N-(2-Oxo-9-trityl-2,9-dihydro-1H-purin-6-yl)-2,2,3,3-tetra-
methylsuccinimide (11). To a flask containing 10 (593 mg, 1.51
mmol) and DBU (0.9 mL, 6.03 mmol) in 16 mL of anhydrous
pyridine at 0 °C was added TMSCl (476 μL, 3.77 mmol) dropwise
by syringe. The mixture was allowed to stir for 30 min. The ice bath
was removed and the mixture allowed to stir for another 30 min.
2,2,3,3-Tetramethylsuccinic anhydride (823 mg, 5.27 mmol) was
added and the mixture heated to reflux for 12 h. Volatiles were
removed in vacuo, and the residue was coevaporated with toluene
(3 ꢀ 5 mL). The residue was redissolved in DCM (15 mL) and
washed with satd NaHCO3 (2 ꢀ 5 mL) and brine (1 ꢀ 5 mL). The
organic layer was dried with Na2SO4, filtered, and evaporated. The
residue was purified by flash chromatography eluting with
98.25:1.75 DCM/MeOH to obtain 700 mg of a white foam. The
product contained 23% of the unreacted anhydride M4SA by
weight as determined by 1H NMR; the yield of 11 was thus 539 mg
(67%). Some of the anhydride could be removed by repeated
chromatography, but 11 could be used in subsequent reactions
without further purification. An analytical sample of 11 was
obtained by recrystallization from acetone containing a small
amount of MeOH. 11: Rf=0.44 (silica gel, 97:3 DCM/MeOH);
1H NMR (CDCl3, 400 MHz) δ 10.27 (brs, 1H), 7.93 (s, 1H),
7.36-7.25 (m, 9H), 7.24-7.08 (m, 6H), 1.28 (s, 12H); 13C
NMR (CDCl3, 101 MHz) δ 180.1, 159.9, 157.1, 145.8, 145.26,
140.6, 129.9, 128.2, 128.1, 126.0, 76.4, 48.3, 21.4; HRMS (ESI-
TOF) calcd for C32H29N5O3Naþ [M þ Na]þ 554.2168, found
554.2147.
6-(3,3,4,4-Tetramethyl-2,5-dioxopyrrolidin-1-yl)-9H-purin-2-
yl Diphenylcarbamate (12). To a flask containing 11 (781 mg,
77% 11 by weight, 1.13 mmol) and diphenylcarbamoyl chloride
(289 mg, 1.25 mmol) were added pyridine (8 mL) and DIPEA
(0.4 mL, 2.26 mmol). The mixture was allowed to stir at room
temperature. Full conversion to a UV-active spot with Rf=0.32
in 99:1 DCM/MeOH was observed by TLC after 1.5 h. Water
(5 mL) was added and stirring continued for 10 min. Volatiles
were removed in vacuo, and the residue was coevaporated with
toluene (3 ꢀ 3 mL). CHCl3 (not anhydrous, 20 mL) and AcOH
(4 mL) were added, and the mixture was heated to reflux with
stirring until TLC indicated consumption of starting material
(20 h). Volatiles were removed in vacuo. The residue was
redissolved in DCM (25 mL) and washed with satd NaHCO3
(3 ꢀ 8 mL) and brine (10 mL). The organic layer was dried with
Na2SO4, filtered, and evaporated. The residue was purified by
flash chromatography eluting with 97:3/95:5 DCM/MeOH to
obtain 508 mg (93%) of 12 as a white foam. 12: Rf=0.42 (silica
gel, 95:5 DCM/MeOH); 1H NMR (CDCl3, 400 MHz) δ 12.13 (s,
1H), 7.97 (s, 1H), 7.34 (m, 8H), 7.30 - 7.15 (m, 2H), 1.32 (s,
12H); 13C NMR (101 MHz, CDCl3) δ 180.1, 155.8, 155.4, 152.0,
146.0, 145.8, 141.5 (br), 129.4, 128.0, 127.5 (br), 126.3 (br), 48.4,
9-[2-Deoxy-3,5-di-O-(p-toluoyl)-β-D-erythro-pentofuranosyl]-
6-(3,3,4,4-tetramethyl-2,5-dioxopyrrolidin-1-yl)-9H-purin-2-yl
Diphenylcarbamate (2). To a flask containing 12 (201 mg, 0.414
mmol) and NaH (41 mg, 0.871 mmol, 50% dispersion in oil) was
added 20 mL of DCM. The mixture was allowed to stir for
20 min, and then 5 (338 mg, 0.871 mmol) was added. Stirring
was continued for 5 h, after which time the reaction mixture
was filtered through Celite and the filter cake washed with
acetone (25 mL). The filtrate was evaporated in vacuo, and
the resulting yellow foam was purified by flash chromatography
(20 ꢀ 5 cm column, chromatospec silica gel) eluting with
98.25:1.75 DCM/MeOH to obtain 273 mg (79%) of 2 as a white
foam. Also obtained were 17 mg (5%) of R-N9 product and 55 mg
(16%) of R/β-N7 products. 2: Rf = 0.38 (silica gel, 97:3 DCM/
MeOH); 1H NMR (acetone-d6, 400 MHz) δ 8.66 (s, 1H), 8.00 (d,
J=8.3 Hz, 2H), 7.87 (d, J=8.3 Hz, 2H), 7.53 (d, J=7.5 Hz, 4H),
7.43-7.32 (m, 6H), 7.32-7.19 (m, 4H), 6.75 (t, J=6.8 Hz, 1H), 6.06
(dt, J=6.4, 3.2 Hz, 1H), 4.84 (dd, J=10.5, 4.5 Hz, 1H), 4.76-4.61
(m, 2H), 3.47 (dt, J=14.0, 6.8 Hz, 1H), 3.01 (ddd, J=14.4, 6.5, 3.4
Hz, 1H), 2.43 (s, 3H), 2.37 (s, 3H), 1.34 (s, 12H); 13C NMR
(acetone-d6, 101 MHz) δ 180.5, 166.58, 166.4, 156.6, 155.6, 151.9,
147.6, 147.1, 145.1, 144.7, 143.1, 130.6, 130.4, 130.1, 130.0, 128.0,
128.0, 127.8 (br), 86.4, 83.6, 76.0, 65.0, 49.0, 37.1, 21.6, 21.6, 21.5,
21.5; HRMS (ESI-TOF) calcd for C47H44N6O9Naþ [M þ Na]þ
859.3067, found 859.3039.
9-(2-Deoxy-β-D-erythro-pentofuranosyl)-6-amino-9H-purin-
2-yl Diphenylcarbamate (13). Obtained in approximately 50%
yield by deprotection of 2 in NH3/MeOH at rt for 15 h.
Purification of the residue by flash chromatography eluting
with 92.5:7.5 DCM/MeOH gave 13 as a white foam. Physical
data were in agreement with those reported.23
20-Deoxyisoguanosine (15)14,26. To a flask containing 2 (212
mg, 0.253 mmol) was added 7 M NH3/MeOH (8 mL). The
mixture was allowed to stir at 60 °C for 15 h, and volatiles were
removed in vacuo. The residue was washed with DCM (2 ꢀ
5 mL) and EtOAc (2 ꢀ 5 mL) and dried in vacuo to yield 58 mg
of impure product. This off-white solid was recrystallized from
MeOH to yield 49 mg (73%) of 15 as an off-white powder.
Acknowledgment. We thank Dr. Nick Greco for critical
discussions. This work was supported by an award from the
NSF to L.W.M. (MCB 0451488).
Supporting Information Available: 1H NMR and 13C NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
(26) Kazimierczuk, Z.; Mertens, R.; Kawczhnski, W.; Seela, F. Helv.
Chim. Acta 1991, 74, 1742–1752.
J. Org. Chem. Vol. 75, No. 5, 2010 1365