Synthesis of enzymatically noncleavable carbocyclic nucleosides for DNA- N-Glycosylase studies

Article abstract of DOI:10.1021/tx9701539

Carbocyclic nucleosides have been of great interest as antiviral agents and in studies in the area of antisense technology. The recent finding that the replacement of a single 2'-deoxynucleoside in DNA by a carba analogue does not alter the Watson-Crick base pairing, yet at the same time provides a chemically and enzymatically stable 'glycosidic' linkage, led us to examine this class of compound as enzyme inhibitors of the DNA-repair enzymes involved in oxidative damage. We now report the synthesis and incorporation into oligomeric DNA via suitable derivatives, the carbanucleosides 8-oxo- 7,8-dihydro-2'-deoxycarbainosine, 8-oxo-7,8- dihydro-2'-deoxycarbaguanosine, and 2'-deoxyaristeromycin. Aristeromycin (1) was deoxy genated at the 2'- position as follows. Treatment of 1 with TPDSCl₂ gave the 3',5'-protected derivative 3 (76%) which on phenylthiocarbonylation at the 2'-position gave 4 in 51% yield. The latter compound on reduction with Bu₃SnH led to the 2'- deoxy derivative 5 (90%). Benzoylation followed by deprotection with TBAF in THF then gave the desired intermediate (6) in 65% yield. N²-Isobutyryl-8- oxo-7,8-dihydro-2'-deoxycarbaguanosine (16) was synthesized from 3-chloro- 2'-deoxycarbainosine (9). Treatment of 9, either with hydrazine followed by catalytic reduction of the 2-hydrazino derivative or with 1-(2- nitrophenyl)ethylamine followed by photolysis of the resulting 2-substituted derivative, in both instances gave the desired 2'- deoxycarbaguanosine (12) in ~50% overall yield in each case. Bromination of 12 gave 13 (90%) which, when treated with BnONa in DMSO at 65 °C, led to the 8-benzyloxy derivative 14 (46%). Isobutyrylation of 14 followed by catalytic reduction then afforded 16. 8-Oxo-7,8- dihydro-2'-deoxycarbainosine (23) was prepared in four steps. Bromination of 2'-deoxyaristero-mycin (19) at the 8-position gave 20 (> 95%) which was converted to the 8-benzyloxy derivative 21 (61%) using BnONa/DMSO at 80 °C. Reductive debenzylation of 21 then led to 8-oxo-7,8-dihydro-2'- deoxyaristeromycin (~100%) which, when treated with adenosine deaminase, provided the desired carbainosine derivative 23 in quantitative yield. Compounds 6, 16, and 23 were converted to their respective 5'-O-DMT, 3-O- [(2-cyanoethoxy)-(N,N-diisopropylamino)-phosphine] derivatives (8, 18, and 25) in excellent overall yields. The latter were then used to synthesize a series of DNA oligomers by automated procedures.

Full text of DOI:10.1021/tx9701539

Chem. Res. Toxicol. 1998, 11, 193-202
193
Syn th esis of En zym a tica lly Non clea va ble Ca r bocyclic
Nu cleosid es for DNA-N-Glycosyla se Stu d ies
Francis J ohnson,* Gyorgy Dorma´n, Robert A. Rieger, Ryuji Marumoto,†
Charles R. Iden, and Radha Bonala
Department of Pharmacological Sciences, Graduate Chemistry Building, State University of
New York at Stony Brook, Stony Brook, New York 11794-3400, and Animal Health Research
Laboratories, Takeda Chemical Industries, Ltd., 17-85 J uso-Honmachi 2-chome, Yodogawa-ku,
Osaka 532, J apan
Received August 25, 1997
Carbocyclic nucleosides have been of great interest as antiviral agents and in studies in the
area of antisense technology. The recent finding that the replacement of a single 2′-
deoxynucleoside in DNA by a carba analogue does not alter the Watson-Crick base pairing,
yet at the same time provides a chemically and enzymatically stable “glycosidic” linkage, led
us to examine this class of compound as enzyme inhibitors of the DNA-repair enzymes involved
in oxidative damage. We now report the synthesis and incorporation into oligomeric DNA via
suitable derivatives, the carbanucleosides 8-oxo-7,8-dihydro-2′-deoxycarbainosine, 8-oxo-7,8-
dihydro-2′-deoxycarbaguanosine, and 2′-deoxyaristeromycin. Aristeromycin (1) was deoxy-
genated at the 2′-position as follows. Treatment of 1 with TPDSCl2 gave the 3′,5′-protected
derivative 3 (76%) which on phenylthiocarbonylation at the 2′-position gave 4 in 51% yield.
The latter compound on reduction with Bu3SnH led to the 2′-deoxy derivative 5 (90%).
Benzoylation followed by deprotection with TBAF in THF then gave the desired intermediate
(6) in 65% yield. N2-Isobutyryl-8-oxo-7,8-dihydro-2′-deoxycarbaguanosine (16) was synthesized
from 3-chloro-2′-deoxycarbainosine (9). Treatment of 9, either with hydrazine followed by
catalytic reduction of the 2-hydrazino derivative or with 1-(2-nitrophenyl)ethylamine followed
by photolysis of the resulting 2-substituted derivative, in both instances gave the desired 2′-
deoxycarbaguanosine (12) in ∼50% overall yield in each case. Bromination of 12 gave 13 (90%)
which, when treated with BnONa in DMSO at 65 °C, led to the 8-benzyloxy derivative 14
(46%). Isobutyrylation of 14 followed by catalytic reduction then afforded 16. 8-Oxo-7,8-
dihydro-2′-deoxycarbainosine (23) was prepared in four steps. Bromination of 2′-deoxyaristero-
mycin (19) at the 8-position gave 20 (>95%) which was converted to the 8-benzyloxy derivative
21 (61%) using BnONa/DMSO at 80 °C. Reductive debenzylation of 21 then led to 8-oxo-7,8-
dihydro-2′-deoxyaristeromycin (∼100%) which, when treated with adenosine deaminase,
provided the desired carbainosine derivative 23 in quantitative yield. Compounds 6, 16, and
23 were converted to their respective 5′-O-DMT, 3′-O-[(2-cyanoethoxy)-(N,N-diisopropylamino)-
phosphine] derivatives (8, 18, and 25) in excellent overall yields. The latter were then used to
synthesize a series of DNA oligomers by automated procedures.
use of a tight-binding product mimic, or more rarely by
the omission of some vital cofactor such as magnesium.
Compounding the problem is the fact that certain DNA
lesions are not hydrolytically stable at the glycosidic
linkage, so their incorporation into oligomeric DNA by
standard solid-phase techniques is not feasible. The
hydrolytic susceptibility of this linkage can be diminished
by introducing an electronegative group at the 2′-position
as is readily evident from the greater stability of ribo-
nucleosides compared with their 2′-deoxy analogues. The
2′-fluoro-substituted derivatives also share enhanced
stability (3).
In tr od u ction
DNA damage, induced by free radicals, ionizing radia-
tion, or environmental toxicants, causes mutation in DNA
and if unrepaired leads frequently to alteration of gene
function which may result in cancer (1). In the DNA-
repair processes there are several glycosylase enzymes
which recognize and cleave primarily the damaged bases
from DNA (2). In the study of these processes, significant
synthetic efforts have been made to incorporate the
modified nucleosides representing the damage into oli-
gomeric DNA in a site-specific manner. The objectives
in any one case are to examine the damage recognition
process, to uncover the mechanistic details of the base
cleavage, and to detail the precise manner in which the
enzyme manipulates the gross DNA structure while
effecting the repair. These aims pose significant prob-
lems because the action of the bound enzyme is difficult
to impede, except by lowering the temperature, by the
Since the completion of the current work, others have
shown that the problem of hydrolytic (and enzymatic)
instability of the deoxynucleosides may be solved either
by inserting a methylene group between the heterocyclic
moiety and the furanose ring (4) or by the use of modified
nucleosides such as 2′-deoxytubercidin or 2′-deoxyformy-
cin (5). However, so far these approaches have been
limited to mimics of 2′-deoxyadenosine.
† Takeda Chemical Industries, Ltd.
S0893-228x(97)00153-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/07/1998

194 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
J ohnson et al.
Ch a r t 1. Ma jor Oxid a tive Lesion Occu r r in g in
Na tive DNA
It appeared to us, nevertheless, that this endemic
problem of the hydrolytic instability of the glycosidic
linkage might be solved simply and generally by replac-
ing the furanose ring oxygen atom by carbon. This
removes the element responsible for the instability of the
O-C-N bond system (hemiamine acetal) and generates
a C-C-N group that is very stable to chemical modifica-
tion yet, at the same time, should have little influence
on the ability of the repair enzyme to bind to DNA.
Several compounds of this class, the carbocyclic nucleo-
sides, were originally developed as antiviral agents (6,
7) either by direct total synthesis or by transformation
from aristeromycin (1, Aris)1 which is available by
fermentation (8). During the past 2 years, new totally
synthetic routes (9-11) have been developed making
these unnatural nucleosides readily available thus en-
hancing the opportunity to incorporate them into DNA.
Exp er im en ta l P r oced u r es
Dry pyridine and acetonitrile were distilled from calcium
hydride. DMT-Cl was recrystallized from 5 volumes of cyclo-
hexane/acetyl chloride (10:1). 1H and 13C NMR spectra were
obtained from a Bruker AC-250 or General Electric QE 300
spectrometer. FAB mass spectra were recorded on a Kratos MS-
890/DS-90 instrument using 50% glycerol/thioglycerol as the
matrix. Electrospray mass spectra were taken on a Micromass
Trio 2000 spectrometer. High-resolution mass spectra were
obtained from the UCR Mass Service Facility, Department of
Chemistry, University of California, Riverside, CA. The purity
of all new compounds was judged to be better than 97% by TLC
analysis in two solvent systems, either chloroform/methanol or
ethyl acetate/ethanol.
The rapidly growing application of carbocyclic nucleo-
sides in DNA research is in the area of antisense
technology (12). This new potentially therapeutic ap-
proach aims to control protein synthesis, which specifi-
cally is involved in diseases, at the DNA or RNA level
by using sugar or base modifications to increase stability
toward nucleases and against simple depurination. Moser
has reported extended stability studies (13) on a complete
carbocyclic sequence annealed to an RNA strand and
separately to a DNA strand. The chemical stability of
the carbaDNA was found to be enhanced. He also
studied the melting behavior of carbaDNA-DNA and
carbaDNA-RNA complexes to determine the stability of
the unnatural double strands. The TM value was slightly
decreased (-0.4 °C/base) in the carbaDNA-DNA case,
whereas the stability of carbaDNA-RNA was slightly
higher than the unmodified form (DNA-RNA). From
these results, it appears that a single carbanucleoside
replacement in DNA does not alter the Watson-Crick
base pairing (14) yet provides a chemically and enzy-
matically stable entity suitable to study those processes
which normally lead to the cleavage of the glycosidic
linkage. This strategy was used first to identify inter-
mediates involved in the proposed mechanism (15) of
DNA cleavage by bleomycin. In this case, a single
carbocyclic adenosine (2′-deoxyaristeromycin, dAris, 1)
residue was site-specifically introduced into DNA (16).
Syn th esis of 2′-Deoxya r ister om ycin . 3′,5′-O-TP DS-a r is-
ter om ycin (3). An amount of 540 mg (2.1 mmol) of aristero-
mycin (1) was partially dissolved in freshly distilled, dry
pyridine (20 mL) under nitrogen by means of sonication. The
resulting white suspension was then treated with TPDS-Cl2
(0.74 mL, 2.3 mmol), and the mixture was stirred for 12 h, when
it became clear. The pyridine was evaporated to dryness giving
a yellow foamy syrup, and water (10 mL) was added followed
by ethyl acetate (100 mL); with vigorous stirring, the two-phase
system was made acidic with 0.2 M NaHSO4, and the phases
were separated. The organic layer was immediately washed
with water (15 mL), saturated NaHCO3 solution (15 mL), then
water (15 mL), and finally brine. The organic extract was dried
over MgSO4 and evaporated in vacuo giving a glassy foam (1.33
g). The crude material was purified on a silica gel column (110
g), elution being accomplished with chloroform/methanol (100:
5). The required product 3 (810 mg) was obtained as a viscous
colorless oil in 76% yield. 1H NMR (250 MHz, δ, CDCl3): 8.25
(s, 1 H, H-8), 7.82 (s, 1 H, H-2), 5.78 (brs, 2 H, NH2); 4.64-4.72
(m, 1H, H-1′); 4.61-4.64 (dd, 1H, H-2′); 4.35-4.39 (m, 1H, H-3′),
4.02 and 3.99 (2 × dd, 2 H, H-5′), 3.15 (brs, 1 H, OH), 2.25 (m,
1H, H-4′), 1.90-2.11 (m, 2 H, carba-CH2), 0.95-1.1 (m, 28 H,
isopropyl). 13C NMR (63 MHz, δ, CDCl3): 157.0, 152.6, 139.9,
75.5, 72.0, 61.8, 60.9, 45.8, 29.1, 17.5, 17.3, 17.1, 13.4, 12.9.
HRMS (DCI, NH3): calcd for C23H42O4Si2N5 (M + H)+ at m/z
508.2775, found 508.2770.
Our research on DNA repair currently is focused
primarily on oxidative damage, specifically the C-8
oxidation product 8-oxo-7,8-dihydro-2′-deoxyguanosine (2;
Chart 1) (17). This lesion is recognized principally by
MutM enzyme (FaPy-DNA glycosylase), which cleaves it
efficiently. On the other hand, during DNA replication,
unrepaired 8-oxo-dG residues, besides being paired with
a dC residue, also form a mismatched base pair with 2′-
deoxyadenosine. This latter lesion is repaired by MutY
enzyme, which excises the opposite base (dA) creating a
gap, which in the subsequent iterative repair process is
filled with either a 2′-dC (correct) or repetitively with a
dA (incorrect) residue (Scheme 1).
2′-O-[(P h en oxyth io)ca r bon yl]-3′,5′-O-TP DS-a r ister om y-
cin (4). 3′,5′-O-TPDS-aristeromycin (3; 375 mg, 0.74 mmol)
was dissolved in dry acetonitrile (8 mL) under nitrogen at 55
°C. To this solution were added N,N-(dimethylamino)pyridine
(253 mg, 2.07 mmol) and phenyl chlorothionoformate (0.105 mL,
0.75 mmol). The yellow solution was stirred at 55 °C. After 10
h, the acetonitrile was removed in vacuo, and the resulting
yellow powder was dissolved in methylene chloride and washed
successively with 0.2 M NaHSO4 (10 mL), saturated NaHCO3
(2 × 10 mL), water (10 mL), and then brine. The organic extract
was dried over MgSO4 and evaporated in vacuo giving a yellow
oil. The crude material (610 mg) was purified by silica gel
column chromatography [silica gel, 45 g; mobile phase, chloroform/
methanol (100:2.5)] to give title compound 4 in 90% yield (432
mg). 1H NMR (250 MHz, δ, CDCl3): 8.23 (s, 1 H, H-8), 7.78 (s,
1 H, H-2), 7.35-7.42 (m, 2 H, phenyl), 7.26 (t, 1 H, phenyl),
7.08 (d, 2 H, phenyl), 6.06 (m, 1 H, H-2′), 5.96 (brs, 2 H, NH2),
5.09 (m, 1 H, H-1′), 4.91 (1 H, m, H-3′), 4.05 and 3.85 (2 × dd,
1 Abbreviations. Compounds: Aris, aristeromycin; dAris, 2′-deoxyaris-
teromycin; 8-oxo-dG, 8-oxo-7,8-dihydro-2′-deoxyguanosine; dA, 2′-
deoxyadenosine; dC, 2′-deoxycytidine; carba-dG, 2′-deoxycarbagua-
nosine; 8-oxo-dI, 8-oxo-2′-deoxyinosine; 8-oxo-carba-dI, 8-oxo-2′-deoxy-
carbainosine. Reagents: AIBN, 2,2′-azobis(isobutyronitrile); DMT-Cl,
4,4′-dimethoxytrityl chloride; DMSO, dimethyl sulfoxide; TBAF/THF,
tetrabutylammonium fluoride in tetrahydrofuran; TPDS-Cl2, 1,3-
dichlorotetraisopropyldisiloxane.

Glycosylase-Resistant Carbocyclic Nucleosides
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 195
Sch em e 1a
a
Grollman, A. P., and Moriya, M. (1993) Trends Genet. 9, 246-249. Reproduced by kind permission of the Editor.
2 H, H-5′), 2.21-2.27 (m, 3 H, H-4′, carba-CH2), 0.93-1.15 (m,
28 H, isopropyl). 13C NMR (63 MHz, δ, CDCl3): 157.1, 153.0,
152.5, 150.2, 140.2, 129.5, 126.6, 121.8, 120.3, 86.3, 70.8, 61.6,
61.4, 45.6, 29.1, 17.5, 17.3, 17.1, 13.4, 12.9. HRMS (FAB,
NBA): calcd for C30H46O5Si2N5S (M + H)+ at m/z 644.2758,
found 644.2751.
The mixture was cooled again in an ice bath, and water (4 mL)
was added followed by concentrated ammonia (8 mL). The
resulting mixture was stirred for 40 min at 24 °C and then
evaporated to dryness. The residue was dissolved in 60 mL of
water, and the solution was extracted with ethyl acetate (3 ×
30 mL), saturated sodium bicarbonate solution (30 mL), water
(30 mL), and then brine (30 mL). The organic phase was dried
over MgSO4 and evaporated, giving an almost pure-yellow
powder (2.95 g). The residue was treated with 1 M TBAF/THF
solution (20 mL) overnight at 24 °C. The volatile materials were
evaporated, and the residue was dissolved in ethyl acetate and
applied to a silica gel column. The column was eluted with an
increasing polarity gradient [ethyl acetate/methanol (10:1 < 5:1
< 3:1)]. The desired product 6 (1.05 g) was isolated as a yellow
powder (72%), mp 165-166 °C. 1H NMR (250 MHz, δ, CDCl3):
8.88 (s, 1 H, H-8), 8.56 (s, 1 H, H-2), 8.19 (d, 2 H, Ph), 7.52 (m,
1 H, Ph), 7.39 (m, 2 H, Ph), 6.23 (brs, 1 H, NH), 5.42 (qt, 1 H,
H-1′), 4.73 (m, 1 H, H-3′), 3.98 (m, 2 H, H-5′), 2.18-2.62 (2 × m
+ 2 × dt, 5 H, H-4′, H-2′, carba-CH2). 13C NMR (63 MHz, δ,
CDCl3): 167.4, 153.8, 151.77, 149.9, 149.04, 142.6, 135.8, 134.9,
132.34, 128.8, 126.5, 123.6, 123.3, 122.84, 72.9, 63.5, 54.2, 50.13,
41.37, 33.9. HRMS (DCI/NH3): calcd for C18H20O3N5 (M + H)+
at m/z 354.1566, found 354.1556.
5′-O-DMT-N6-ben zoyl-2′-d eoxya r ister om ycin (7). To a
solution of N6-benzoyl-2′-deoxyaristeromycin (6; 106 mg, 0.3
mmol; dried three times by coevaporation with dry pyridine) in
dry pyridine (1.5 mL) were added DMT-Cl (recrystallized, 132
mg, 0.39 mmol) and DMAP (3.5 mg). The reaction was
quenched after 4 h by pouring into ice-cold saturated sodium
bicarbonate solution (10 mL). This mixture was extracted with
ethyl acetate (3 × 5 mL), and the combined fractions were dried
over MgSO4. After partial evaporation, the solution was applied
2′-Deoxy-3′,5′-O-TP DS-a r ist er om ycin (5). A solution of
dry 2-O-[(phenoxythio)carbonyl]-3′,5′-O-TPDS-aristeromycin (4;
240 mg, 0.237 mmol) in freshly distilled toluene (7.5 mL) was
degassed by repeated gentle evacuation followed by bubbling
nitrogen through the solution for 20 min. To the solution were
then added tributyltin hydride (0.15 mL, 0.56 mmol) and dry
AIBN (12 mg). The reaction mixture was stirred at 75 °C for 3
h under nitrogen. Thereafter, the colorless liquid was evapo-
rated to dryness, and the residue was dissolved in chloroform
and applied directly onto a silica gel-packed column [mobile
phase, chloroform/methanol (100:2.5)]. Evaporation of the
combined fractions gave the desired compound 5 (0.17 g, yield
91%) as a crystalline solid, mp 110-111 °C. 1H NMR (250 MHz,
δ, CDCl3): 8.30 (s, 1 H, H-8), 7.80 (s, 1 H, H-2), 5.68 (brs, 2 H,
NH2), 5.05 (m, 1 H, H-1′), 4.67 (m, 1 H, H-3′), 4.03 and 3.77 (2
× dd, 2 H, H-5′), 2.35 (m, 2 H, H-4′, H-2a′), 1.70-2.11 (m, 3 H,
H-2b′, carba-CH2), 0.95-1.1 (m, 28 H, isopropyl). 13C NMR (63
MHz, δ, CDCl3): 156.5, 153.6, 139.9, 71.6, 62.2, 51.6, 50.1, 40.0,
27.6, 17.5, 17.3, 17.1, 13.4, 12.9. HRMS (DCI/NH3): calcd for
C
23H42O3Si2N55 (M + H)+ at m/z 492.2826, found 492.2832.
N6-Ben zoyl-2′-d eoxya r ister om ycin (6). A solution of 2′-
deoxy-3′,5′-O-TPDS-aristeromycin (5; 2.025 g, 4.1 mmol; dried
three times by azeotroping with dry pyridine) in dry pyridine
(20 mL) was cooled to 0 °C by means of an ice bath. To this
solution was added benzoyl chloride (2.4 mL, 20.5 mmol), and
the mixture was allowed to stand at room temperature for 3 h.

196 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
J ohnson et al.
to a silica gel column (15 g) and eluted with ethyl acetate/
methanol/TEA (50:1:0.5 then 25:1:0.25). The fractions contain-
ing the pure material were collected and evaporated giving 7
(130 mg, 65% yield) as a pale-yellow solid, mp 95-97 °C. 1H
NMR (250 MHz, δ, CDCl3): 8.70 (s, 1 H, H-8), 8.15 (d, 2 H, Ph),
7.91 (s, 1 H, H-2), 7.18-7.52 (m, 14 H, Ph), 6.79 (d, 2 H, Ph),
5.19 (qt, 1 H, H-1′), 4.42 (m, 1 H, H-3′), 3.76 (s, 6 H, 2 × OCH3),
3.21-3.23 (m, 2 H, H-5′), 1.95-2.49 (2 × m + 2 × dt + m, 5 H,
H-2′, H-4′, carba-CH2). 13C NMR (63 MHz, δ, CDCl3): 169.2,
158.5, 153.5, 153.1, 150.5, 145.1, 142.3, 136.9, 134.2, 133.4,
132.9, 130.0, 128.8, 128.1, 127.9, 126.9, 113.18, 75.5, 65.1, 55.2,
53.7, 47.5, 40.1, 34.2. HRMS (FAB, NBA): calcd for C39H37O5-
Na (M + Na)+ at m/z 678.2692, found 678.2678.
reaction mixture was kept in an oil bath at 100 °C for 10 h in
the dark. The DMSO was evaporated at 70 °C under reduced
pressure, and the residue was triturated with water. After the
liquid phase had been separated from the insoluble material,
the oily residue was stirred with tert-butyl methyl ether followed
by ethyl acetate/2-propanol (3:1). The insoluble residue was
identified spectroscopically as the title compound 11, ∼90%
purity. 1H NMR (300 MHz, δ, methanol-d4): 7.3-7.7 (s + m, 5
H, H-2, aromatic), 5.31 (d, 1 H, NH), 4.68 (m, 1 H, H-1′), 4.10
(m, 1 H, H-3′), 3.86 (q, 1 H, CHCH3), 3.59 (2 × dd, 2 H, H-5′),
2.36 (m, 1 H, H-4′), 1.58, 2.12, 2.36 (m, 4 H, H-2′, carba-CH2),
1.15 (d, 3 H, CH3). HRMS (DCI/NH3): calcd for C19H23N6O5
(M + H)+ at m/z 415.1730, found 415.1746.
5′-O-DMT-N6-ben zoyl-2′-d eoxya r ister om ycin 3′-O-(2-Cy-
a n oeth yl N,N-d iisop r op ylp h osp h or a m id ite) (8). To a solu-
tion of 5′-O-DMT-N6-benzoyl-2′-deoxyaristeromycin (7; 125 mg,
0.19 mmol) in 2 mL of dry methylene chloride and 0.1 mL of
dry triethylamine was added dropwise 0.085 mL of 2-cyanoethyl
N,N-diisopropylchlorophosphoramidite (0.38 mmol). The reac-
tion mixture was stirred for 6 h and then diluted with 15 mL of
dry toluene. The precipitated triethylamine hydrochloride was
filtered under nitrogen, and the toluene was evaporated under
reduced pressure. The residue 8 was used directly in the DNA
synthesizer in dry acetonitrile solution. 1H NMR (250 MHz, δ,
CDCl3): 8.72 (s, 1 H, H-8), 8.00 (d, 2 H, Ph), 7.98 (s, 1 H, H-2),
7.25-7.58 (m, 14 H, Ph), 6.79 (d, 2 H, Ph), 5.17 (qt, 1 H, H-1′),
4.56 (m, 1 H, H-3′), 3.77 (s, 6 H, OCH3), 3.48 (m, 2 H, 2 × CHN),
3.27-3.31 (m, 2 H, H-5′), 2.10-2.49 (2 × m + 2 × dt, 5 H, H-2′,
H-4′, carba-CH2), 1.24 (d, 12 H, isopropyl). 31P NMR (101 MHz,
δ, CDCl3): 149.4, 149.1.
P h otolytic Deben zyla tion of 11. The oily nitrophenethyl
derivative 11 was dissolved in a mixture of methanol (7 mL)
and water (3 mL), and the solution was placed in a quartz
cuvette and then photolyzed for 4.5 h between two clamp-held
lamps at 360 nm. The progress of the reaction was monitored
by UV spectrometry and TLC. After completion of the reaction,
the solution was evaporated to dryness, and the residue was
stirred with methanol. The insoluble product was removed by
filtration and proved to be identical spectroscopically with the
2′-deoxycarbaguanosine (12) obtained by procedure A. The
overall yield of 12, mp 242-244 °C dec (145 mg), from 2-chloro-
2′-deoxycarbainosine was 51%.
8-Br om o-2′-d eoxyca r ba gu a n osin e (13). To a vigorously
stirred suspension of 2′-deoxycarbaguanosine (12; 2.18 g, 8.23
mmol) in 15 mL of water was added portionwise 52 mL of
bromine-saturated water (made by shaking 1.4 mL of Br2 with
120 mL of water at room temperature) over 1.5 h. Stirring was
continued at room temperature after the solution became
homogeneous. At the end of 4 h, a solid began to separate from
the reaction mixture. After 24 h, the thick solid was filtered
and then crystallized from methanol to give 8-bromo-2′-deoxy-
carbaguanosine (13; 2.0 g). Evaporation to dryness of the
mother liquor and crystallization of the residue twice from
methanol yielded 0.55 g of a second crop, total yield 90%, mp
238-240 °C. 1H NMR (300 MHz, δ, DMSO-d6): 4.85 (dt, 1 H,
H-1′), 4.01 (m, 1 H, H-3′), 3.27 and 3.46 (2 × dd, 2 H, H-5′),
2.40 (m, 1 H, H-4′), 1.66-1.95 (m, 4 H, H-2′, carba-CH2). 13C
NMR (63 MHz, δ, D2O): 155.7, 153.5, 152.4, 121.2, 117.3, 71.8,
63.4, 55.3, 49.7, 43.8, 38.1, 32.00. HRMS (EI): calcd for
C11H14N5O3Br (M+) at m/z 343.0280, found 343.0294.
Syn th esis of Der iva tives of 8-Oxo-7,8-d ih yd r o-2′-d eoxy-
ca r ba gu a n osin e. P r oced u r e A: Hyd r a zin e Meth od . 2′-
Deoxyca r ba gu a n osin e (12). A mixture of methanol (160 mL)
and deionized water (80 mL) in a 500-mL flask was degassed
with nitrogen for 30 min. To this solution was added 2-chloro-
2′-deoxycarbainosine (9; 5.0 g) under nitrogen followed by
anhydrous hydrazine (31 mL). The mixture was refluxed gently
for 24 h, then solvent was removed in vacuo, and the residue
was triturated with degassed 95% ethanol (200 mL). The
resulting white precipitate was filtered under nitrogen to give
2-hydrazino-2′-deoxycarbainosine (10; 3.6 g). This compound
readily decomposes upon standing and more rapidly on exposure
to light. The subsequent reductive deamination was therefore
carried out without further purification. In a hydrogenation
flask was placed a 1:1 mixture of methanol and water (120 mL),
and the mixture was degassed with nitrogen for 30 min. To
this mixture was added wet Raney Ni (15 g), and the catalyst
was saturated with H2 for 1 h at 50 psi. The 2-hydrazino-2′-
deoxycarbainosine (10; 3.6 g) dissolved in the same solvent
mixture (120 mL, degassed) was added to the catalyst, and the
mixture was hydrogenated for 24 h at 55 °C in a Parr apparatus
at 50 psi. The reduction was monitored by HPLC analysis
performed on a C-18 column using triethylammonium acetate
buffer in acetonitrile in a linear gradient (0-10% acetonitrile
over 50 min). After completion of the hydrogenation, the
catalyst was removed using a filter aid and washed with a
methanol-water (hot) mixture. The filtrate was concentrated
to give the desired 2′-deoxycarbaguanosine (12; 2.8 g) as a white
powder which was recrystallized from water, mp 243-245 °C
dec (solubility: 7.1 g/100 g of water at 90 °C). 1H NMR (300
MHz, δ, D2O): 7.74 (s, 1 H, H-2), 4.72 (m, 1 H, H-1′), 4.15 (m,
1 H, H-3′), 3.51 and 3.60 (2 × dd, 2 H, H-5′), 2.36 (m, 1 H, H-4′),
1.58, 2.11, 2.36 (m, 4 H, H-2′, carba-CH2). 13C NMR (63 MHz,
δ, D2O): 157.2, 153.7, 151.4, 135.7, 71.9, 63.1, 52.5, 49.7, 41.0,
34.4. HRMS (DCI, NH3): calcd for C11H16N5O3 (M + H)+ at
m/z 266.1253, found 266.1256.
8-(Ben zyloxy)-2′-d eoxyca r ba gu a n osin e (14). A solution
of 8-bromo-2′-deoxycarbaguanosine (13; 2.5 g, 7.29 mmol) in dry
dimethyl sulfoxide (11 mL) was added to a solution of sodium
benzyl oxide (prepared from 19.4 mL of benzyl alcohol and 610
mg of sodium heated at 80 °C for 2 h) under nitrogen. The heavy
syrup was stirred for 24 h at 65 °C and then allowed to cool to
room temperature. After neutralization with glacial acetic acid,
the bulk of the dimethyl sulfoxide was removed by vacuum
distillation at 100 °C (bath temperature). The remaining liquid
was then poured into tert-butyl methyl ether (500 mL) with
stirring. The ether layer was decanted thereby removing the
excess benzyl alcohol. TLC analysis of the residue showed two
new substances in addition to unreacted starting material. The
new materials were purified by column chromatography using
methylene chloride/methanol (4:1) as the eluant. The desired
compound 14 was isolated as an oil in 46% yield (1.24 g). 1H
NMR (300 MHz, δ, MeOH-d4): 7.30-7.50 (d, 5 H, aromatic),
5.47 (s, 1 H, benzylic CH2), 4.83 (m, 1 H, H-1′), 3.97 (m, 1 H,
H-3′), 3.35 and 3.45 (2 × dd, 2 H, H-5′), 2.23 (m, 1 H, H-4′),
1.77-2.05 (m, 4 H, H-2′, carba-CH2). 13C NMR (63 MHz, δ,
MeOH-d4): 160.1, 155.1, 154.1, 153.1, 137.1, 130.0, 129.9, 112.3,
74.3, 73.3, 65.0, 52.7, 51.1, 40.8, 39.5, 33.3. HRMS (EI): calcd
for C18H21N5O4 (M+) at m/z 371.1594, found 371.1597.
P r oced u r e B: P h otolytic Meth od . 2-[[1-(2-Nitr op h en -
yl)]eth yl]a m in o]-2′-d eoxyca r ba in osin e (11). To a solution
of 2-chloro-2′-deoxycarbainosine (9; 300 mg, 1.06 mmol) in dry
DMSO (10 mL) was added 2-(nitrophenyl)ethylamine (350 mg,
2.12 mmol; prepared from 2-nitroacetophenone by reductive
amination), and the solution was placed in a sealed ampule. The
The second product was identified as 8-oxo-2′-deoxycarba-
guanosine (0.6 g, 29%). 1H NMR (300 MHz, δ, MeOH-d4): 4.78
(m, 1 H, H-1′), 4.05 (m, 1 H, H-3′), 3.45 (t, 2 H, H-5′), 2.25 (m,
1 H, H-4′), 1.82-2.1 (m, 4 H, H-2′, carba-CH2). FAB/MS (+ve
ion, thioglycerol matrix): m/z 282 (M + H)+.

Glycosylase-Resistant Carbocyclic Nucleosides
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 197
N2-Isobu tyr yl-8-(ben zyloxy)-2′-deoxycar bagu an osin e (15).
To a suspension of 8-(benzyloxy)-2′-deoxycarbaguanosine (14;
280 mg, 0.75 mmol; dried by evaporating three times with
pyridine) in 5 mL of dry pyridine was added trimethylchlorosi-
lane (0.63 mL, 3.90 mmol), and the reaction mixture was stirred
for 30 min. To the clear solution was added isobutyric anhy-
dride (0.81 mL, 3.75 mmol), and the mixture was then stirred
overnight at room temperature. The solution was cooled to 0
°C in an ice bath, and water (2 mL) was added followed by the
addition of concentrated ammonia (1 mL) at the same temper-
ature. The solution was stirred for 30 min while the temper-
ature warmed to ambient and was then evaporated to dryness.
The residue was dissolved in methanol and mixed with a small
portion of silica gel. The solvent was removed by evaporation,
and the dried residue was applied to a silica gel column and
eluted with a linear gradient of a methylene chloride/methanol
mixture (20:1 up to 15:2). The main fractions yielded 15 as a
white powder in 60% yield (200 mg). 1H NMR (300 MHz, δ,
MeOH-d4): 7.37-7.52 (ddm, 5 H, aromatic), 5.47 (s, 1 H,
benzylic CH2), 5.05 (m, 1 H, H-1′), 4.19 (m, 1 H, H-3′), 3.52 and
3.61 (2 × dd, 2 H, H-5′), 2.71 (2 × t, 1 H, isobutyryl CH), 2.55
(m, 1 H, H-4′), 1.92-2.20 (m, 4 H, H-2′, carba-CH2), 1.24 (2 × s,
2 CH3). 13C NMR (63 MHz, δ, MeOH-d4): 168.5, 157.0, 155.5,
150.5, 147.6, 129.8, 129.7, 117.5, 118.7, 73.6, 73.3, 64.4, 52.9,
50.7, 39.3, 33.0, 19.4, 19.3. HRMS (DCI, NH3): calcd for
C22H28N5O5 (M + H)+ at m/z 442.2090, found 442.2093.
N2-Isob u t yr yl-8-oxo-7,8-d ih yd r o-2′-d eoxyca r b a gu a n o-
sin e (16). A solution of N2-isobutyryl-8-(benzyloxy)-2′-deoxy-
carbaguanosine (15; 200 mg, 0.45 mmol) in methanol (8 mL)
was hydrogenated at room temperature in the presence of 10%
Pd/C catalyst (80 mg) for 24 h. The reaction mixture was then
filtered through a Celite bed, and the filtrate was concentrated
in vacuo. The product (148 mg) was used in the next step
without purification. 1H NMR (300 MHz, δ, MeOH-d4): 4.90
(m, 1 H, H-1′), 4.21 (m, 1 H, H-3′), 3.66 and 3.56 (2 × m, 2 H,
H-5′), 2.63 (2 × t, 1 H, isobutyryl CH), 2.50 (m, 1 H, H-4′), 1.82-
2.05 (m, H-2′, 4 H, carba-CH2), 1.12 (2 × s, 2 CH3). 13C NMR
(63 MHz, δ, MeOH-d4): 181.3, 168.0, 154.0, 148.4, 147.6, 104.3,
73.4, 64.1, 51.2, 50.5, 38.2, 36.9, 31.8, 19.0. HRMS (EI): calcd
for C15H21N5O5 (M+) at m/z 351.1543, found 351.1546.
which time a white precipitate of Et3N‚HCl separated. A second
portion (0.035 mL) of the phosphoramidite reagent was added
to complete the conversion. After a further 1 h, the mixture
was filtered and the collected solid was washed with toluene (2
mL). The filtrate was concentrated. The residual glassy solid
18 (0.175 g) showed two single peaks in the 31P NMR spectrum
as noted below for the expected diastereoisomers with virtually
no other absorption in the spectrum. The material was judged
pure enough to be used directly in DNA synthesis. 1H NMR
(300 MHz, δ, MeOH-d4): 7.03-7.74 (m, 11 H, Ph), 6.72 (d, 2 H,
Ph), 5.03 (m, 1 H, H-1′), 4.18 (m, 1 H, H-3′), 3.76 (s, 6 H, 2 ×
OCH3), 3.44, 3.54, 3.56 (2 × m, 4 H, H-5′, 2 × CHN), 2.64-2.76
(m, 2 H, isobutyryl CH, H-4′), 1.84-2.18 (m, 4 H, H-2′, carba-
CH2), 1.26 (d, 12 H, isopropyl), 1.05 (2 × s, 2 CH3). 31P NMR
(101 MHz, δ, CDCl3): 147.6, 147.9.
8-Br om o-2′-d eoxya r ister om ycin (20). 2′-Deoxyaristero-
mycin (19; 0.5 g) was dissolved in sodium acetate buffer (60 mL,
1 M, pH 4) which was then treated with bromine (150 mL) in
the same buffer (30 mL). The reaction mixture was stirred for
5 h in the dark, and the solution was decolorized with a small
amount of NaHSO3 solution and then saturated with sodium
chloride. The solution was then extracted with chloroform (4
× 5 mL). The organic extracts were dried over MgSO4, filtered,
and evaporated to give 20 (665 mg) of better than 95% purity.
This was used directly in the next step. A small portion was
recrystallized from ethanol/hexane to give the analytically pure
material. 1H NMR (250 MHz, δ, DMSO-d6): 7.80 (s, 1 H, H-2),
7.35 (brs, 2 H, NH2), 5.08 (m, 1 H, H-1′), 4.63 (m, 1 H, H-3′),
4.17 and 3.65 (2 × dd, 2 H, H-5′), 2.75 (m, 1 H, H-4′), 1.95-2.24
(m, 4 H, H-2′, carba-CH2). 13C NMR (63 MHz, δ, DMSO-d6):
175.1, 155.7, 152.8, 151.42, 126.6, 120.9, 73.6, 64.2, 50.6, 48.3,
38.7, 31.8, 20.5. HRMS (EI): calcd for C11H14O2N5Br (M+) at
m/z 327.0331, found 327.0329.
8-(Ben zyloxy)-2′-d eoxya r ister om ycin (21). To dry benzyl
alcohol (15 mL) at 65 °C in a 100-mL flask was added sodium
(408 mg, 17.7 mmol) in small pieces (weighed and cut under
dry pentane). The temperature was raised to 80 °C, and the
metal dissolved within 2 h. 8-Bromo-2′-deoxyaristeromycin (20;
1.02 g, 3.2 mmol) was added, and the reaction mixture was
heated for
a further 10 h at the same temperature and
5′-O-DMT-N2-isobu tyr yl-8-oxo-7,8-d ih yd r o-2′-d eoxyca r -
ba gu a n osin e (17). N2-Isobutyryl-8-oxo-7,8-dihydro-2′-deoxy-
carbaguanosine (16; 120 mg, 0.34 mmol) was dried with pyridine
(3 × 5 mL). The dry material was suspended in pyridine (2
mL) containing DMT-Cl (0.127 mg, 0.37 mmol). After stirring
for 1 h at room temperature, the reaction was quenched by
pouring into an ice-cold saturated solution of sodium bicarbonate
(5 mL). The solution was extracted with methylene chloride (3
× 5 mL), and the combined fractions were dried over MgSO4.
After partial evaporation the solution was applied to a silica
gel column which was then eluted with methylene chloride/
methanol/triethylamine (20:1:0.5). The main fractions yielded
pure 17 (200 mg), mp 164-165 °C. 1H NMR (300 MHz, δ,
MeOH-d4): 7.03-7.74 (m, 11 H, Ph), 6.72 (d, 2 H, Ph), 4.95 (m,
1 H, H-1′), 4.17 (m, 1 H, H-3′), 3.66 (2 × s, 6 H, 2 × OCH3), 3.23
and 3.05 (2 × m, 2 H, H-5′), 2.55 (m, 2 H, isobutyryl CH, H-4′),
1.84-2.18 (m, 4 H, H-2′, carba-CH2), 1.08 (2 × s, 6 H, 2 CH3).
13C NMR (63 MHz, δ, MeOH-d4): 182.2, 160.4, 154.7, 152.0,
149.2, 148.3, 147.1, 138.0, 131.6, 129.7, 129.0, 128.0, 114.3,
105.0, 97.9, 87.5, 74.6, 66.6, 56.1, 38.7, 37.1, 34.0, 19.7. HRMS
(FAB/NBA): calcd for C36H39N5O7Na (M + Na)+ at m/z 676.2744,
found 676.2747.
5′-O-DMT-N 2-isob u t yr yl-8-oxo-2′-d e oxyca r b a gu a n o-
sin e 3′-O-(2-Cya n oeth yl N,N-d iisop r op ylp h osp h or a m id ite)
(18). A modification of the procedure used by Bodepudi et al.
(23) was used. A thoroughly dried sample of 5′-O-DMT-N2-
isobutyryl-8-oxo-7,8-dihydro-2′-deoxycarbaguanosine (0.13 g, 0.2
mmol) was dissolved in a mixture of dry tetrahydrofuran (1 mL)
and dry methylene chloride (1 mL) under N2 at room temper-
ature. To this solution were added dry triethylamine (0.085 mL)
and 2-cyanoethyl N,N-diisopropylchlorophosphoramidite (0.105
mL, 0.22 mmol). The mixture was stirred at 24 °C for 2 h during
monitored by TLC [methylene chloride/methanol (5:1), long
plate]. The benzyl alcohol was then removed under reduced
pressure, and the residue was triturated with tert-butyl methyl
ether giving crude 21 (630 mg) as a brown precipitate. This
was purified through a short silica gel column using ethyl
acetate/methanol (7.5:1) as the mobile phase. This yielded pure
21 (510 mg) as a faintly yellow powder. 1H NMR (300 MHz, δ,
MeOH-d4): 8.05 (s, 1 H, H-2), 7.50 (d, 2 H, aromatic), 7.37 (m,
2 H, aromatic), 7.30 (dd, 1 H, aromatic), 5.53 (s, 1 H, benzylic
CH2), 5.12 (m, 1 H, H-1′), 4.17 (m, 1 H, H-3′), 3.63 and 3.49 (2
× dd, 2 H, H-5′), 2.53 (m, 1 H, H-4′), 1.96-2.24 (m, 4 H, H-2′,
carba-CH2). 13C NMR (63 MHz, δ, MeOH-d4): 156.2, 155.2,
151.7, 150.7, 136.8, 130.03, 129.97, 129.7, 116.42, 74.2, 73.45,
67.0, 52.9, 50.9, 39.48, 33.25. HRMS (FAB): calcd for C18H22O3N5
(M + H)+ at m/z 356.1723, found 356.1727.
8-Oxo-7,8-d ih yd r o-2′-d eoxya r ister om ycin (22). A sample
of purified 8-(benzyloxy)-2′-deoxyaristeromycin (21; 160 mg) was
hydrogenated at 50 °C and 50 psi in methanol (7 mL) using a
10% Pd/C catalyst (35 mg). After 3 h, TLC analysis [solvents,
ethyl acetate/methanol (3:1)] showed complete conversion to
product. The reaction mixture was filtered through a Celite bed
and evaporated to dryness. This gave an almost quantitative
1
yield (117 mg) of 22 which was used directly to prepare 23. H
NMR (300 MHz, δ, MeOH-d4/D2O): 7.95 (s, 1 H, H-2), 5.07 (m,
1 H, H-1′), 4.19 (m, 1 H, H-3′), 3.66 and 3.50 (2 × dd, 2 H, H-5′),
2.54 (m, 1 H, H-4′), 184-2.19 (m, 4 H, H-2′, carba-CH2). 13C
NMR (63 MHz, δ, MeOH-d4/D2O): 155.1, 152.4, 149.3, 149.0,
105.8, 74.7, 65.4, 52.0, 51.2, 38.8, 32.7. HRMS (DCI/NH3): calcd
for C11H16N5O3 (M + H)+ at m/z 266.1253, found 266.1256.
8-Oxo-2′-d eoxy-7,8-d ih yd r oca r ba in osin e (23). A dilute
solution of 8-oxo-2′-deoxyaristeromycin (22) was prepared (110
mg in 750 mL of deionized water) and treated with adenosine

198 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
J ohnson et al.
deaminase (30 mg, type II, purified, calf intestine; Sigma). After
8 days the enzymatic deamination was complete according to
TLC analysis [methylene chloride/methanol/ammonium hydrox-
ide (30:10:2)]. The solution was evaporated to one-third its
volume, and the title compound precipitated out. The solid was
isolated by centrifugation, washed with cold water, and dried,
yielding 95 mg (86%) of essentially pure 23. Because this
compound has a very low solubility in the generally employed
solvents, the crude product was again used in the next step
without further purification. 1H NMR (300 MHz, δ, pyridine-
d5/D2O): 7.66 (s, 1 H, H-2), 4.56 (m, 1 H, H-1′), 4.07 (m, 1 H,
H-3′), 3.55 and 3.42 (2 × dd, 2 H, H-5′), 2.22 (m, 1 H, H-4′),
1.84-1.94 (m, 4 H, H-2′, carba-CH2). 13C NMR (63 MHz, δ,
pyridine-d5/D2O): 152.9, 148.1, 145.5, 145.3, 108.1, 72.9, 63.4,
50.3, 48.8, 37.1, 31.3. HRMS (EI): calcd for C11H14N4O4 (M+)
at m/z 266.1015, found 266.1023.
Instruments, Inc., Farmingdale, NY), and the crude product was
resuspended in water. Purification was accomplished on a
Waters Chromatography (Milford, MA) system equipped with
a 990 photodiode array detector. For DMT-protected DNA, a
µBondapak C18 column was employed using a gradient of 16-
35% acetonitrile and triethylammonium acetate buffer (0.1 M,
pH 7.2) in 30 min. Collections were pooled and evaporated to
dryness (Savant Instruments, Inc., Farmingdale, NY), and the
crude product was resuspended in water.
A. 5′-TT(d Ar is)TT-3′. This was prepared as described above
using 8. The DMT oligomer was found to have a retention time
of 20.4 min using the conditions outlined above. The collected
material was dried and treated with 80% acetic acid (0.2 mL)
for 30 min. The resulting solution was dried and rechromato-
graphed by HPLC. A single peak was obtained with a retention
time of 16.56 min using 0-15% acetonitrile in 0.1 M triethyl-
ammonium acetate over 30 min. The collected portion was
submitted for mass analysis. MS (FAB, Gly): calcd 1465.3 Da
(M - H-), found m/z 1464.
B. 5′-CTCTCCCTTC(d Ar is)CTCCTTTCCTCT-3′. The
method of synthesis again followed the procedure described
above. The DMT oligomer was purified and found to have a
retention time of 21.5 min on the C18 column. The DMT group
was removed with 80% acetic acid (room temperature, 17.5 min)
and resulted in a single substance when rechromatographed.
This was collected and submitted for mass analysis. MS (ESI):
calcd 6761.4 Da, found 6760.8 Da.
C. 5′-CTCTCCCTTC(8-oxo-ca r ba -d G)G-CTCCTTTCCT-
CT-3′. This was prepared according to the procedure described
above using 18. â-Mercaptoethanol (7 µL, 0.1 N) was added to
the ammonia solution to avoid any oxidation of the modified
base. The DMT oligomer was found to have a retention time of
18.4 min using the conditions outlined above. The collected
material was dried and treated with 80% acetic acid (0.2 mL)
for 30 min. The resulting solution was then dried and rechro-
matographed by HPLC. A single peak was obtained with a
retention time of 23.5 min using a linear gradient of 0-15%
acetonitrile in 0.1 M triethylammonium acetate over 30 min.
The UV spectrum was consistent with an oligomer containing
an 8-oxo-dG moiety with a peak at 266 nm and a broad shoulder
from 292 to 310 nm. The collected portion was submitted for
mass analysis. MS (ESI): calcd 6793.3 Da, found 6793.7 Da.
D. 5′-TTCAGTCA(8-oxo-ca r ba -d I)TCAGTCGTA-3′. This
was prepared as described above using 25. â-Mercaptoethanol
(7 µL, 0.1 N) was added to the ammonia solution to avoid further
oxidation of the modified base. The DMT oligomer was found
to have a retention time of 27.5 min using the conditions
outlined above. The collected material was dried and treated
with 80% acetic acid (0.2 mL) for 30 min. The resulting solution
was evaporated to dryness and rechromatographed by HPLC.
A single peak was obtained with a retention time of 36.3 min
using a linear gradient of 0-15% acetonitrile in 0.1 M triethyl-
ammonium acetate over 30 min. The UV spectrum showed a
maximum at 260 nm and was consistent with an oligomer
previously synthesized that contained the analogous nucleoside
residue 8-oxo-dI. The collected portion was submitted for mass
analysis. MS (ESI): calcd 5488.6 Da, found 5488.1 Da.
5′-O-DMT-8-oxo-7,8-d ih yd r o-2′-d eoxyca r ba in osin e (24).
To a suspension of 8-oxo-7,8-dihydro-2′-deoxycarbainosine (23;
95 mg, 0.36 mmol) in dry pyridine (1.5 mL) was added 4,4′-
DMT-Cl (134 mg, 0.39 mmol) in one portion. The reddish
suspension was stirred overnight which resulted in a clear
solution. The reaction was monitored by TLC using methylene
chloride/methanol (2:1) as the eluant. The mixture was then
cooled in an ice bath, quenched with water (5 mL), and extracted
with methylene chloride (3 × 5 mL). The combined organic
extracts were washed with water (2 × 4 mL) and brine (4 mL)
and then dried over MgSO4. After evaporation of the solvent,
the crude product was purified on a silica gel column using
methylene chloride/methanol (20:1) containing 2% triethylamine
as the eluting solvent mixture. The desired compound 23 was
obtained in 67% yield (140 mg). 1H NMR (300 MHz, δ, pyridine-
d5): 8.02 (s, 1 H, H-2), 7.23-7.64 (m, 11 H, Ph), 6.77 (d, 2 H,
Ph), 5.52 (m, 1 H, H-1′), 4.69 (m, 1 H, H-3′), 3.47 (s, 6 H, 2 ×
OCH3), 3.38 and 3.65 (2 × dd, 2 H, H-5′), 2.90 (m, 1 H, H-4′),
2.22-2.55 (2 × m + 2 × dt, 4 H, H-2′, carba-CH2). 13C NMR
(63 MHz, δ, pyridine-d5): 158.5, 153.1, 151.8, 146.1, 145.9, 143.8,
136.7, 130.4, 128.4, 127.8, 126.1, 113.2, 109.4, 99.5, 85.8, 73.9,
65.5, 54.7, 50.2, 36.5, 32.6. HRMS (FAB/NBA): calcd for
C32H32N4O6Na (M + Na+) at m/z 591.2220, found 591.2212.
5′-O-DMT-8-oxo-7,8-d ih yd r o-2′-d eoxyca r ba in osin e 3′-O-
(2-Cya n oeth yl N,N-d iisop r op ylp h osp h or a m id ite) (25). A
thoroughly dried portion of 5′-O-DMT-8-oxo-7,8-dihydro-2′-
deoxycarbainosine (24; 85 mg, 0.15 mmol) was dissolved in dry
tetrahydrofuran (0.8 mL) and dry methylene chloride (0.8 mL)
under nitrogen. To this solution were added dry triethylamine
(0.064 mL) and 2-cyanoethyl N,N-diisopropylchlorophosphor-
amidite (0.079 mL, 0.17 mmol). The reaction mixture was
strirred for 2 h at room temperature, and a white precipitate
separated (Et3N‚HCl). After 2 h, a second portion of phosphor-
amidite reagent (0.026 mL) was added to the reaction mixture
to achieve total conversion. After the mixture had been stirred
for a further 1 h, dry toluene (10 mL) was added to complete
the precipitation of triethylamine hydrochloride. The precipitate
was filtered under nitrogen and washed with dry toluene (2 mL),
and the filtrate was concentrated. The title compound 25,
obtained as a glassy solid (115 mg), was pure enough to use
directly for DNA synthesis as confirmed by 31P NMR analysis.
1H NMR (250 MHz, δ, CDCl3): 7.94 (s, 1 H, H-2), 7.25-7.49
(m, 14 H, Ph), 6.80 (d, 2 H, Ph), 5.11 (m, 1 H, H-1′), 4.33 (m, 1
H, H-3′), 3.76 (s, 6 H, 2 × OCH3), 3.51 (m, 2 H, 2 × CHN), 3.32-
3.42 (m, 2 H, H-5′), 2.72 (m, 1 H, H-4′), 2.05-2.49 (2 × m + 2
× dt, 4 H, H-2′, carba-CH2), 1.24 (d, 12 H, isopropyl). 31P NMR
(101 MHz, δ, CDCl3): 149.3, 148.9.
Resu lts a n d Discu ssion
In our approach to the study of these two enzymes
(MutM and MutY), we now report the synthesis and
incorporation into oligomeric DNA of the carbocyclic
analogues of the 2-deoxynucleosides discussed above,
namely, 2′-deoxyaristeromycin, 8-oxo-7,8-dihydro-2′-
deoxycarbaguanosine, and 8-oxo-7,8-dihydro-2′-deoxy-
carbainosine. The oligomers are useful (1) for the
preparation of affinity chromatography substrates for
enzyme isolation and (2) for the characterization of, and
mechanism studies on, the DNA-repair processes.
Syn th esis a n d In cor p or a tion of d Ar is in to DNA.
Our initial synthesis started from aristeromycin (1) and
Syn th esis of DNA Oligom er s. The examples described
below illustrate the general methods employed for the synthesis
of all of the oligomers described in this and a related paper (17).
Oligomers were synthesized on an ABI (Foster City, CA) 394
DNA synthesizer using standard phosphoramidite protocols.
The selected modified phosphoramidite was introduced into
several oligomers. After each synthesis, the product was
deprotected by incubation in 1 mL of 28% ammonium hydroxide,
16 h at 55 °C. The solution was evaporated to dryness (Savant

Glycosylase-Resistant Carbocyclic Nucleosides
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 199
Sch em e 2
followed the Markiewicz-Robins deoxygenation (18)
route described for 2-deoxyadenosine, rather than the
original procedures used by Marumoto (19). General
phosphoramidite chemistry was used to incorporate dAris
into DNA oliogomers using the standard protecting
groups, namely, 5′-O-DMT and N6-benzoyl (Scheme 2).
The coupling efficiency was comparable to that of the
natural base during the synthesis of a 23-mer [5′-CTC
TCC CTT C(dAris)C TCC TTT CCT CT-3′]. The oligo-
mers were then characterized by electrospray mass
spectrometry. In simple binding experiments (20), the
duplex containing dAris opposite to dG was recognized
but not cleaved by the MutY enzyme as might be
anticipated from the postulated mechanism of cleavage
(SN2 displacement).
Syn th esis of 8-Oxo-7,8-d ih yd r o-2′-d eoxyca r ba gu a -
n osin e a n d In cor p or a tion in to DNA. The synthesis
of this compound began with 2-chloro-2′-deoxycarbaino-
sine (9; Scheme 3), which can be obtained from aristero-
mycin (1) in several steps (21). In our first approach to
convert this compound to carba-dG (12), we attempted a
direct amination by using ammonia under pressure, but
after several hours mainly starting material was recov-
ered. In a second attempt, the N2-benzyl derivative of
carba-dG was prepared from 9 in 51% yield, but the
subsequent attempt to effect debenzylation by hydroge-
nation at 50 psi and 60 °C was not successful. Photo-
chemical benzylic cleavage, however, readily yielded
carba-dG (12) when the 2-nitrobenzyl analogue 11 was
employed. The latter was easily prepared by treating 9
with an excess of 2-(nitrophenyl)ethylamine (22) in
dimethyl sulfoxide at 100 °C for 12 h in the dark. On a
small scale the photolytic removal of the 2-nitrophenethyl
group was conducted at room temperature with lamps
radiating at ∼360-nm wavelength, and this readily
provided the desired carbocyclic nucleoside (23). How-
ever, on a larger scale, we simply followed the Long-
Robins-Townsend methodology (24) for the ribonucleo-
side series and converted 2-chloro-2′-deoxycarbainosine
(9) into the corresponding 2-hydrazino derivative 10 in
73% yield. Subsequent hydrogenation of this oxygen-
sensitive compound at 50 °C and 50 psi for 20 h followed
by crystallization of the product from water gave very
pure carba-dG (12) in 72% yield.
The synthesis of the protected 8-oxo-carba-dG (17) was
carried out in five steps using a modification of Bodepu-
di’s procedure (25) for 8-oxo-dG (2). However, the initial
bromination to give 13, using aqueous bromine at room
temperature for 5 h, was considerably slower than in the
case of the 2′-deoxyribose analogue. This suggests either
that in the carbocyclic analogues, the 7,8-double bond of
the purine ring has a considerably reduced electron
density or that in the natural nucleoside, the ring oxygen
can assist the process, perhaps through bromine com-
plexation. The replacement of the 8-bromo group by
benzyloxy was then accomplished in 46% yield using
sodium benzyl oxide and DMSO at 70 °C for 24 h. This
reaction was not clean and afforded, besides 14, some
recovered starting material and a 29% yield of 8-oxo-
carba-dG. Nevertheless, 14 was readily separated by
chromatography. Surprisingly, it did not crystallize as
readily as the analogous compound derived from the
natural nucleoside, and the benzyl group showed an
increased tendency to be cleaved under acidic conditions.

200 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
J ohnson et al.
Sch em e 3
Protection of the N2-amino group of 14 by the ‘transient
protection’ method (26) to give 15 was best carried out
at this stage. If the benzyl group was removed first, the
subsequent acylation led to undesired side products. The
N2-protected 8-benzyl derivative 15 was then hydroge-
nated over 10% Pd/C at room temperature and gave N2-
isobutyryl-8-oxo-carba-dG (16). The DMT protection of
the 5-hydroxyl to give 17 was carried out by the standard
procedure (DMT-Cl/Py), but the reaction was consider-
ably slower (12 h) than that in the deoxyribose series (15
min). Attempts to catalyze the process by DMAP led only
to increased formation of the 3′,5′-bis-O-DMT derivative.
The preparation of the phosphoramidite 18 was carried
out by the ‘nonaqueous workup technique’ developed by
Bodepudi (27) in our laboratory, and again the reaction
proved to be slower than that of the deoxyribose analogue
requiring both a much longer reaction period and a larger
excess of reagent.
The use of 18 to incorporate 8-oxo-carba-dG residues
into DNA [pentamer, TT(8-oxo-carba-dG)TT; 23-mer, 5′-
CTC TCC CTT C(8-oxo-carba-dG)C TCC TTT CCT CT-
3′] was carried out by automated synthesis and proceeded
with excellent coupling efficiency (>95%). Deprotection
and isolation of these oligomers followed the normal
procedures used for oligomers containing 8-oxo-dG resi-
dues (25). In preliminary experiments it was found that
MutM binds to a 23-mer duplex containing 8-oxo-carba-
dG opposite to dC in the sequence indicated above but
that cleavage of the glycosidic bond is totally inhibited
(Kd ) 15-25 nM). The detailed study has been published
elsewhere (20, 28).
In previous studies with MutM enzyme (29) (FaPy
protein), it was found that 8-oxo-2′-deoxyinosine had an
enzyme binding affinity similar to that of 8-oxo-dG.
Therefore, it appeared that the carbocylic analogue might
serve as a good nonenzymatically cleavable substrate for
biochemical studies and for enzyme purification. The
synthesis from 2′-deoxyaristeromycin (19) is reasonably
short (Scheme 4). Bromination of 19, obtained by de-
benzoylation of 6, was carried out according to Ettner
(30), but an extractive workup, rather than direct crys-
tallization, was used to obtain 8-bromo-dAris (20) of
acceptable purity in 65% yield. This was used directly
in the benzylation step, which afforded 21 in 61% yield.
Hydrogenation of 21 then removed the benzyl group in
excellent yield. Chemical deamination of 8-oxo-dAris (22,
8-oxo-7,8-dihydro-2′-deoxycarbaadenosine) did not pro-
ceed well, whereas deamination prior to reductive de-
benzylation (i.e., of 21) led to partial loss of the benzyl
group, thus making purification of the product difficult.
Enzymatic deamination proved more successful. Ad-
enosine deaminase is known to have a low substrate
specificity (31, 32) for purine nucleosides, having sub-
stituents at the 8-position. Thus, when 22 was treated
with this enzyme for 7 days, the desired 8-oxo-7-dehydro-
2′-deoxycarbainosine (23) was obtained in quantitative
yield. The formation of the DMT-phosphoramidite (25)
and subsequent incorporation into oligomeric DNA were
accomplished exactly as described above for 8-oxo-carba-
dG.
Sta bility of 8-Oxo-ca r ba -d G tow a r d Aer obic Oxi-
d a tion . The deprotection step during the synthesis of
oligomeric DNA requires ammonia treatment at 55 °C
for 5-15 h. When 8-oxo-dG residues are present, exten-
Syn th esis of 8-Oxo-7,8-d ih yd r o-2′-d eoxya r ister o-
m ycin a n d 8-Oxo-7,8-d ih yd r o-2′-d eoxyca r ba in osin e.

Glycosylase-Resistant Carbocyclic Nucleosides
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 201
Sch em e 4
(8) J enkins, G. N., and Turner, N. (1995) The Biosynthesis of
sive degradation (33, 34) and cleavage of the DNA occur
if an antioxidant such as mercaptoethanol or ascorbic
acid is not added (35, 36). Although no cleavage of DNA
containing 8-oxo-carba-dG residues occurs during this
step, there is substantial degradation of the heterocyclic
base, and an antioxidant must also be included in this
case. Further work on the nature of this oxidation will
be the subject of a future publication.
Carbocyclic Nucleosides. Chem. Soc. Rev. 169-176.
(9) Be´res, J ., Sa´gi, Gy., Baitz-Ga´cs, E., To¨mo¨sko¨zi, I., and O¨ tvo¨s, L.
(1988) Stereospecific Synthesis of (+)-Carbocyclic 2′-Deoxyadeno-
sine. An Improved Procedure for the Preparation of (+)-(1R,2S,-
4R)-4-Amino-2-hydroxy-1-hydroxymethylcyclopentane. Tetrahe-
dron 44, 6207-6216.
(10) Agrofolio, L., Suhas, E., Farese, A., Condom, R., Challand, S. R.,
Earl, R. A., and Guedj, R. (1994) Synthesis of Carbocyclic
Nucleosides. Tetrahedron 50, 10611-10670.
(11) Taniyama, Y., Fukuda, T., and Marumoto, R. (1992) Carbocyclic
Purine Nucleosides Derived From Aristeromycin Through Two
Key Intermediates, Carbocyclic 5-Amino-4-imidiazolecarboxamide
Riboside and 2′-Deoxyriboside. Nucleosides Nucleotides 11 (2-
4), 529-549.
Ack n ow led gm en t. We are very grateful to Drs.
Arthur P. Grollman, Kenneth Breslauer, Craig Gelfand,
and Carlos de los Santos for their interest in these
studies. This research was supported by a grant (CA
4799506) from the National Cancer Institute of the
National Institutes of Health.
(12) De Desmaeker, A., Ha¨ner, R., Martin, P., and Moser, H. E. (1995)
Antisense Oligonucleotides. Acc. Chem. Res. 28, 366-374.
(13) Moser, H. E. (1993) Strategies and Chemical Approaches towards
Oligonucleotide Therapeutics. In Perspectives in Medicinal Chem-
istry (Testa, B., Fuhrer, W., Kyburz, E., and Giger, R., Eds.) pp
275-297, Verlag Helvetica Chimica Acta, Basel.
Refer en ces
(14) Gelfand, C., Breslauer, K., and de los Santos, C. Unpublished
results.
(15) Giloni, L., Takeshita, M., J ohnson, F., and Grollman, A. P. (1981)
Bleomycin Induced Strand-Scission of DNA: Mechanism of
Deoxyribose Cleavage. J . Biol. Chem. 256, 8609-8615.
(1) Breimer, L. H. (1990) Molecular Mechanisms of Oxygen Radical
Carcinogenesis and Mutagenesis: The Role of DNA Base Damage.
Mol. Carcinog. 3, 188-197.
(2) Myers, L. C., and Verdine, G. L. (1994) DNA Repair Proteins.
Curr. Op. Struct. Biol. 4, 51-59.
(3) Scha¨fer, D. O., and Verdine, G. (1995) A Designed Inhibitor of
Base-Excision DNA Repair. J . Am. Chem. Soc. 117, 10781-10782.
(4) Deng, L., Scharer, O. D., and Verdine, G. L. (1997) Unusually
Strong Binding of a Designed Transition-State Analogue to a
Base-Excision DNA Repair Protein. J . Am. Chem. Soc. 119, 7865-
7866.
(16) Sugiyama, H., Takashi, S., Dannoue, Y., Marumoto, R., and Saito,
I. (1991) Bleomycin-Mediated Degradation of Aristeromycin-
Containing DNA. Novel Dehydrogenation Activity of IronII
-
Bleomycin. J . Am. Chem. Soc. 113, 2290-2295.
(17) Grollman, A. P., and Moriya, M. (1993) Mutagenesis by 8-oxo-
guanine: An Enemy Within. Trends Genet. 9, 246-249.
(5) Porello, S. L., Williams, S. D., Kuhn, H., Michales, M. L., and
David, S. S. (1996) Specific Recognition of Substrate Analogues
by the DNA Mismatch Repair Enzyme Mut Y. J . Am. Chem. Soc.
118, 10684-10692.
(6) Coe, D. M., Roberts, S. M., and Storer, R. (1992) The Potential of
Carbocyclic Nucleosides for the Treatment of AIDS: Synthesis
of some Diphosphorylphosphonates Possessing Potent Activity
Against HIV-Coded Reserve Transcriptase. J . Chem. Soc., Perkin
Trans. I 2695-2701.
(18) Markiewicz, W. T. (1979) Tetraisopropyldisiloxane-1,3-diyl, a
Group for Simultaneous Protection of 3′- and 5′-Nucleosides and
Nucleic Related Compounds. J . Chem. Res. (Synop) 24-25.
Robins, M. J ., Wilson, J . S., and Hansske, F. (1983) A General
Procedure for the Efficient Deoxygenation of Secondary Alcohols,
Regiospecific and Stereoselective Conversion of Ribonucleosides
to 2′-Deoxynucleosides. J . Am. Chem. Soc. 105, 4059-4065.
(19) Miyashita, O., Taniyama, Y., Fujisawa, Y., and Marumoto, R.
(1985) Protected 2′-Deoxyaristeromycin on Polymer Support: A
Substitute for the Acid-labile 2′-Deoxyadenosine Moiety at the
3′-Terminus of Oligodeoxynucleotides. Nucleic Acid Symp. Ser.
No. 16 141-144.
(7) Patil, S. D., Koga, M., and Schneller, S. W. (1992A) (()-Carbocyclic
5′-nor-2′-deoxyguanosine and Related Purine Derivatives: Syn-
thesis and Antiviral Properties. J . Med. Chem. 35, 2191-2195.

202 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
J ohnson et al.
(20) Bulychev, N. V., Chamakura, V. V., Dorma´n, G., Miller, J . H.,
Eisenberg, M., Grollman, A. P., and J ohnson, F. (1996) Substrate
Specificity of Escherichia coli MutY Protein. Biochemistry 35,
13147-13156.
(21) Marumoto, R., Hamana, T., and Kashiwabara, Y. (1989) Prepara-
tion, Testing, and Formulation of Carbocyclic Purine Nucleosides
as Virucides. J pn. Kokai Tokyo Koho J P01 75499[8975499]; Chem.
Abstr. 111, 214892b.
(22) Walker, J . W., McCray, J . A., and Hess, G. P. (1986) Photolabile
Protecting Groups for an Acetylcholine Receptor Ligand. Synthe-
sis and Photochemistry of a New Class of O-Nitrobenzyl Deriva-
tives and Their Effects on Receptor Function. Biochemistry 25,
1799-1805.
cosylase/AP-lyase Reaction Catalyzed by Escherichia coli Forma-
midopyrimidine-DNA Glycosylase (Fpg) Protein. J . Biol. Chem.
272, 5335-5341.
(29) Tchou, J ., Bodepudi, V., Shibutani, S., Antosheckin, I., Miller, J .,
Grollman, A. P., and J ohnson, F. (1994) Substrate Specificity of
Fpg Protein. Recognition and Cleavage of Oxidatively Damaged
DNA. J . Biol. Chem. 21, 15318-15324.
(30) Ettner, N., Haak, U., Niederweis, M., and Hillen, W. (1993)
Synthesis of 8-Bromo- and 8-Azido-2′-deoxyadenosine-5′-O-(1-
thiotriphosphate). Nucleosides Nucleotides 12 (7), 757-771.
(31) Gala, D., and Schumacher, D. P. (1992) Deamination of Carbocy-
clic Adenosine Analogues with Adenosine Deaminase and Its
Synthetic Utility. Synlett 61-63.
(23) For reviews of recent applications of 2-nitrobenzyl photochemistry,
see: (a) Pillai, V. N. R. (1987) Photolytic Deprotection and
Activation of Functional Groups. Org. Photochem. 9, 225-341.
(b) Dorma´n, G., Ceruso, M., and Prestwich, G. D. (1995)
ChemTracts 137-141.
(24) Long, R. A., Robins, R. K., and Townsend, L. B. (1967) Purine
Nucleosides. XV. The Synthesis of 8-Amino- and 8-Substituted
Aminopurine Nucleosides. J . Org. Chem. 32, 2751-2756.
(25) Bodepudi, V., Shibutani, S., and J ohnson, F. (1992) Synthesis of
2′-Deoxy-7,8-dihydro-8-oxo-2′-deoxyguanosine and 7,8-Dihydro-
8-oxo-2′-deoxyadenosine and Their Incorporation into Oligomeric
DNA. Chem. Res. Toxicol. 5, 608-617.
(26) Ti, G. S., Gaffney, B. L., and J ones, R. A. (1982) Transient
Protection. Efficient One-Flask Syntheses of Protected Deoxy-
nucleosides. J . Am. Chem. Soc. 104, 1316-1319.
(27) Bodepudi, V., Iden, C. R., and J ohnson, F. (1991) An Improved
Method for the Preparation of the Phosphoramidites of Modified
2′-Deoxynucleotides: Incorporation of 8-Oxo-2′-deoxy-7H-guano-
sine into Synthetic Oligomers. Nucleosides Nucleotides 10 (4),
755-761.
(32) Buenger, G. S., and Nair, V. (1990) Dideoxygenated Purine
Nucleosides Substituted at the 8-Position: Chemical Synthesis
and Stability. Synthesis 962-966.
(33) J ohnson, F. (1991) Synthesis and Chemistry of Carcinogenic
Adducts of DNA. In Studies in Natural Products Chemistry
(Rahman, A.-U., Ed.) Vol. 8, pp 373-394, Elsevier Science
Publishers B.V., Amsterdam.
(34) Cullis, P. M., Malone, M. E., and Merson-Davies, L. A. (1996)
Guanine Radical Cations Are Precursors of 7,8-Dihydro-8-oxo-2′-
deoxyguanosine But Are Not Precursors of Immediate Strand
Breaks in DNA. J . Am. Chem. Soc. 118, 2775-2781.
(35) Shibutani, S., Gentles, R. G., Iden, C. R., and J ohnson, F. (1990)
Facile Aerial Oxidation of the DNA-Base Adduct N-(2′-Deoxygua-
nosin-8-yl)-2-aminofluorene [dG(C8)AF]. J . Am. Chem. Soc. 112,
5667-5668.
(36) Shibutani, S., Gentles, R., and J ohnson, F. (1991) Aerial Oxidation
of Acetylaminofluorene-derived DNA Adducts. In Nitroarenes:
Occurrence, Metabolism and Biological Impact (Howard, P. C.,
Beland, F. A., and Hecht, S. S., Eds.) pp 135-148, Plenum Press,
New York.
(28) Zharkov, D. O., Rieger, R. A., Iden, C. R., and Grollman, A. P.
(1997) NH2-terminal Proline Acts as a Nucleophile in the Gly-
TX9701539