Novel stereoselective control over cis vs trans opening of benzo[c]phenanthrene 3,4-diol 1,2-epoxides by the exocyclic N2-amino group of deoxyguanosine in the presence of hexafluoropropan-2-ol

Article abstract of DOI:10.1021/jo020418a

We describe a novel and efficient synthesis (62-84% yields) of the eight possible, diastereomerically pure, cis and trans, R and S O⁶-allyl-protected N²-dGuo phosphoramidite building blocks derived through cis and trans opening of (±)-3α,4β-dihydroxy-1β,2β-epoxy-1,2,3, 4-tetrahydrobenzo[c]-phenanthrene [BcPh DE-1 (1)] and (±)-3α,4β-dihydroxy-1α,2α-epoxy-1,2,3, 4-tetrahydrobenzo[c]-phenanthrene [BcPh DE-2 (2)] by hexafluoropropan-2-ol (HFP)-mediated addition of O⁶-allyl-3′,5′- di-O-(tert-butyldimethylsilyl)-2′-deoxyguanosine (3) at C-1 of the epoxides. Simply changing the relative amount of HFP used in the reaction mixture can achieve a wide ratio of cis/trans addition products. Thus, the observed cis/trans adduct ratio for the reaction of DE-1 (1) in the presence of 5 equiv of 3 varied from 17/83 to 91/9 over the range of 5-532 equiv of HFP. The corresponding ratios for DE-2 (2) varied from 2/98 to 61/39 under the same set of conditions. When 1 or 2 was fused with a 20-fold excess of 3 at 140°C in the absence of solvent HFP, almost exclusive trans addition (>95%) was observed for the both DEs. Through the use of varying amounts of HFP in the reaction mixture as described above, each of the eight possible phosphoramidite oligonucleotide building blocks (DE-1/DE-2, cis/trans, R/S) of the BcPh DE N²-dGuo adducts can be prepared in an efficient fashion. To rationalize the varying cis-to-trans ratio, we propose that the addition of 3 to 1 or 2 in the absence of solvent or in the presence of small amounts of HFP proceeds primarily via an S_N2 mechanism to produce mainly trans-opened adducts. In contrast, increasing amounts of HFP promote increased participation of an S_N1 mechanism involving a relatively stable carbocation with two possible conformations. One of these conformations reacts with 3 to give mostly trans adduct, while the other conformation reacts with 3 to give mostly cis adduct.

Full text of DOI:10.1021/jo020418a

Novel Ster eoselective Con tr ol over Cis vs Tr a n s Op en in g of
Ben zo[c]p h en a n th r en e 3,4-Diol 1,2-Ep oxid es by th e Exocyclic
N2-Am in o Gr ou p of Deoxygu a n osin e in th e P r esen ce of
Hexa flu or op r op a n -2-ol
Haruhiko Yagi,*,† Andagar R. Ramesha,† Govind Kalena,† J ane M. Sayer,†
Subodh Kumar,‡ and Donald M. J erina*,†
Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases,
The National Institutes of Health, Bethesda, Maryland 20892, and Environmental Toxicology and
Chemistry Laboratory, Great Lakes Center, State University of New York College at Buffalo,
1300 Elmwood Avenue, Buffalo, New York 14222
dmjerina@nih.gov
Received J une 19, 2002
We describe a novel and efficient synthesis (62-84% yields) of the eight possible, diastereomerically
pure, cis and trans, R and S O6-allyl-protected N2-dGuo phosphoramidite building blocks derived
through cis and trans opening of (()-3R,4â-dihydroxy-1â,2â-epoxy-1,2,3,4-tetrahydrobenzo[c]-
phenanthrene [BcPh DE-1 (1)] and (()-3R,4â-dihydroxy-1R,2R-epoxy-1,2,3,4-tetrahydrobenzo[c]-
phenanthrene [BcPh DE-2 (2)] by hexafluoropropan-2-ol (HFP)-mediated addition of O6-allyl-3′,5′-
di-O-(tert-butyldimethylsilyl)-2′-deoxyguanosine (3) at C-1 of the epoxides. Simply changing the
relative amount of HFP used in the reaction mixture can achieve a wide ratio of cis/trans addition
products. Thus, the observed cis/trans adduct ratio for the reaction of DE-1 (1) in the presence of
5 equiv of 3 varied from 17/83 to 91/9 over the range of 5-532 equiv of HFP. The corresponding
ratios for DE-2 (2) varied from 2/98 to 61/39 under the same set of conditions. When 1 or 2 was
fused with a 20-fold excess of 3 at 140 °C in the absence of solvent HFP, almost exclusive trans
addition (>95%) was observed for the both DEs. Through the use of varying amounts of HFP in
the reaction mixture as described above, each of the eight possible phosphoramidite oligonucleotide
building blocks (DE-1/DE-2, cis/trans, R/S) of the BcPh DE N2-dGuo adducts can be prepared in an
efficient fashion. To rationalize the varying cis-to-trans ratio, we propose that the addition of 3 to
1 or 2 in the absence of solvent or in the presence of small amounts of HFP proceeds primarily via
an SN2 mechanism to produce mainly trans-opened adducts. In contrast, increasing amounts of
HFP promote increased participation of an SN1 mechanism involving a relatively stable carbocation
with two possible conformations. One of these conformations reacts with 3 to give mostly trans
adduct, while the other conformation reacts with 3 to give mostly cis adduct.
In tr od u ction
region (benzo[c]phenanthrene, BcPh) as well as the DE-1
and DE-2 diastereomers are shown in Figure 1. Interest-
ingly, the highly hindered fjord-region DEs are much
more tumorigenic than their bay-region counterparts.1,3
Each diastereomer exists as a pair of enantiomers. These
DEs are known to form stable adducts by cis and trans
addition of the exocyclic amino groups of dAdo and dGuo
in DNA to the benzylic position of the epoxide.4 Damage
to genes such as the ras family of proto-oncogenes5 and
Metabolically formed bay-region and fjord-region diol
epoxides (DEs) are known ultimate carcinogens of the
environmentally prevalent polycyclic aromatic hydrocar-
bons.1 These diol epoxides are formed from trans-dihy-
drodiols and exist as pairs of diastereomers in which the
epoxide oxygen is either cis (DE-1) or trans (DE-2) to the
benzylic hydroxyl group.2 Typical examples of hydrocar-
bons with a bay region (benzo[a]pyrene, BaP) and a fjord
* Please direct correspondence to: Section on Oxidation Mecha-
nisms, Laboratory of Bioorganic Chemistry, The National Institutes
of Health, NIDDK, Building 8A, Room 1A01, Bethesda, MD 20892.
Phone: 301 496-4560. Fax: 301 402-0008.
(3) Levin, W.; Wood, A. W.; Chang, R. L.; Ittah, Y.; Croisy-Delcey,
M.; Yagi, H.; J erina, D. M.; Conney, A. H. Cancer Res. 1980, 40, 3910-
3914.
(4) (a) J erina. D. M.; Chadha, A.; Cheh, A. M.; Schurdak, M. E.;
Wood, A. W.; Sayer, J . M. In Biological Reactive Intermediates IV.
Molecular and Cellular Effects and Their Impact on Human Health;
Witmer, C. M., Snyder, R., J ollow, D. J ., Kalf, G. F., Kocsis, J . J ., Sipes,
I. G., Eds.; Plenum Press: New York, 1991; pp 533-553. (b) Dipple,
A. In DNA Adducts: Identification and Biological Significance; Hem-
minki, K., Dipple, A., Shuker, D. E. G., Kadlubar, F. F., Segerbaeck,
D., Bartsch, H., Eds.; IARC Publications; Lyon, 1994; pp 107-129.
(5) (a) Marshall, C. J .; Vousden, K. H.; Phillips, D. H. Nature 1984,
310, 586. (b) Vousden, K. H.; Bos, J . L.; Marshall, C. J .; Phillips, D. H.
Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 1222.
† The National Institutes of Health.
‡ State University of New York College at Buffalo.
(1) J erina, D. M.; Sayer, J . M.; Agarwal, S. K.; Yagi, H.; Levin, W.;
Wood, A. W.; Conney, A. H.; Pruess-Schwartz, D.; Baird, W. M.; Pigott,
M. A.; Dipple, A. In Biological Reactive Intermediates III; Kocsis, J .
J ., J ollow, D. J ., Witmer, C. M., Nelson, J . O., Snyder, R., Eds.; Plenum
Press: New York, 1986; pp 11-30.
(2) Thakker, D. R.; Yagi, H.; Levin, W.; Wood, A. W.; Conney, A.
H.; J erina, D. M. In Bioactivation of Foreign Compounds; Anders, M.
W., Ed.; Academic Press: New York, 1985; pp 177-242.
10.1021/jo020418a CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/29/2002
6678
J . Org. Chem. 2002, 67, 6678-6689

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
shown that fidelity versus misincorporation is enzyme,
sequence, and adduct dependent.7e-g Studies of human
topoisomerase I have shown that, depending on the
adduct, cleavage of the DNA can be blocked or subse-
quent religation of the cleaved DNA can be inhibited.7h,i
Rates of DNA repair are highly dependent on the
hydrocarbon and the structure of the DE adduct.7j,k Thus,
synthetic access to oligonucleotides containing a variety
of different DE adducts in different sequence contexts is
of high significance.
To date, three strategies have been successful in the
preparation of oligonucleotides containing N6-dAdo or N2-
dGuo adducts of the polycyclic aromatic hydrocarbon
DEs: direct reaction of DEs with short oligonucleotides,
postsynthetic modification in which an oligonucleotide
containing an activated nucleotide unit is allowed to react
with an amino triol derived from a DE, and synthesis of
a DE-adducted phosphoramidite suitably protected for
use in standard, solid-phase DNA synthesis. The advan-
tages and disadvantages of each of these approaches have
been detailed elsewhere.8 We have preferred the third
approach in that it allows the greatest flexibility in terms
of the choice of specific adduct in any sequence context.
In addition, the desired oligonucleotides are generally
much more easily purified from the crude product mix-
tures compared to the other two approaches. Initially,
the stepwise synthesis of phosphoramidites involved
nucleophilic displacement by aminotriols derived from
DEs on appropriately protected purine nucleosides with
either 6-fluoro-9a,10a or 6-sulfonate-9b,c leaving groups
(dAdo precursors) and 2-fluoro-7j,10b or 2-triflate-10c de-
oxyinosine derivatives (dGuo precursors). Although trans-
opened aminotriols are readily available by direct ami-
nolysis of DEs, synthesis of the corresponding cis-
aminotriols requires multiple steps.11,12a This difficulty
was partially resolved by our discovery that cis-opened
N6-dAdo adducts of BaP DE-2 and BcPh DE-2 (2) could
be prepared in a single12 step from the corresponding
dihydrodiols and 3′,5′-di-O-(tert-butyldimethylsilyl)-2′-
deoxyadenosine in a highly stereoselective fashion under
Sharpless aminohydroxylation conditions. Notably, how-
ever, we have been unable to find reaction conditions for
preparation of dGuo adducts by this procedure.
F IGURE 1. Structures of benzo[a]pyrene (BaP) and benzo-
[c]phenanthrene (BcPh) with their respective bay region and
fjord regions as well as key benzo-rings (bold) indicated. Partial
structures of the DE-1 (1) and DE-2 (2) diol epoxide diaster-
eomers and the dGuo reactant O6-allyl-3′,5′-di-O-(tert-bu-
tyldimethylsilyl)-2′-deoxyguanosine (3) are also shown.
the tumor suppressor genes such as p536 and p216c has
been implicated in the causation of cancer by DEs.
There exists a great deal of interest in small oligo-
nucleotides containing DE adducts. Determination of the
solution conformation of DNA adducts by two-dimen-
sional NMR has identified minor-groove-bound as well
as intercalated adducts.7a,b Site specific mutagenesis
studies have indicated a strong dependence on sequence
context and the nature of the adduct.7c,d Translesional
bypass synthesis with a variety of DNA polymerases has
(6) (a) Bjelogrlic, N. M.; Makinen, M.; Stenback, F.; Vahakangas,
K. Carcinogensis 1994, 15, 771. (b) Denissenko, M. F.; Pao, A.; Tang,
M. S.; Pfeifer, G. P. Science 1996, 274, 430. (c) Kaspin, L. C.; Baird,
W. M. Polycyclic Aromat. Compd. 1996, 10, 299.
Direct reaction of the exocyclic amino groups of the
purines with the DEs constitutes an attractive route to
adducts. Only recently have we been able to establish
(7) For leading references, see: (a) Lin, C. H.; Huang, X.; Kolban-
ovskii, A.; Hingerty, B. E.; Amin, S.; Broyde, S.; Geacintov, N. E.; Patel,
D. J . J . Mol. Biol. 2001, 306, 1059-1080. (b) Pradhan, P.; Tirumala,
S.; Liu, X.; Sayer, J . M.; J erina, D. M.; Yeh, H. J . C. Biochemistry 2001,
40, 5870-5881. (c) Ponte´n, I.; Sayer, J . M.; Pilcher, A. S.; Yagi, H.;
Kumar, S.; J erina, D. M.; Dipple, A. Biochemistry 2000, 39, 4136-
4144. (d) Fernandes, A.; Liu, T.; Amin, S.; Geacintov, N. E.; Grollman,
A. P.; Moriya, M. Biochemistry 1998, 37, 10164-10172. (e) Chiap-
perino, C.; Kroth, H.; Kramarczuk, I. H.; Sayer, J . M.; Masutani, C.;
Hanaoka, F.; J erina, D. M.; Cheh, A. M. J . Biol. Chem. 2002, 277,
11765-11771. (f) Suzuki, N.; Ohashi, E.; Kolbanovskiy, A.; Geacintov,
N. E.; Grollman, A. P.; Ohmori, H.; Shibutani, S. Biochemistry 2002,
41, 6100-6106. (g) Shen, X.; Sayer, J . M.; Kroth, H.; Ponte´n, I.;
O’Donnell, M.; Woodgate, R.; J erina, D. M.; Goodman, M. F. J . Biol.
Chem. 2002, 277, 5265-5274. (h) Pommier, Y.; Kohlhagen, G.; Pour-
quier, P.; Sayer, J . M.; Kroth, H.; J erina, D. M. Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 2040. (i) Pommier, Y.; Kohlhagen, G.; Laco, G. S.;
Kroth, H.; Sayer, J . M.; J erina, D. M. J . Biol. Chem. 2002, 277, 13666-
13672. (j) Custer, L.; Zajc, B.; Sayer, J . M.; Cullinane, C.; Phillips, D.
R.; Cheh, A. M.; J erina, D. M.; Bohr, V. A.; Mazur, S. J . Biochemistry
1999, 38, 569-581. (k) Buterin, T.; Hess, M. T.; Luneva, N.; Geacintov,
N. E.; Amin, S.; Kroth, H.; Seidel, A.; Naegeli, H. Cancer Res. 2000,
60, 1849-1856.
(8) (a) Kroth, H.; Yagi, H.; Seidel, A.; J erina, D. M. J . Org. Chem.
2000, 65, 5558-5564. (b) Kroth, H.; Yagi, H.; Sayer, J . M.; Kumar, S.;
J erina, D. M. Chem. Res. Toxicol. 2001, 14, 708-719 and references
therein.
(9) (a) Lakshman, M. K.; Sayer, J . M.; J erina, D. M. J . Am. Chem.
Soc. 1991, 113, 6589-6594. (b) Steinbrecher, T.; Becker, A.; Stezowski,
J . J .; Oesch, F.; Seidel, A. Tetrahedron Lett. 1993, 34, 1773-1774. (c)
Kroth, H.; Hertkorn, N.; Oesch, F.; Seidel, A. Polycyclic Aromat.
Compd. 1996, 11, 349.
(10) (a) Lakshman, M. K.; Sayer, J . M.; J erina, D. M. J . Org. Chem.
1992, 57, 3438-3443. (b) Zajc, B.; Lakshman, M. K.; Sayer, J . M.;
J erina, D. M. Tetrahedron Lett. 1992, 33, 3409-3412. (c) Cooper, M.
D.; Hodge, R. P.; Tamura, P. J .; Wilkinson, A. S.; Harris, C. M.; Harris,
T. M. Tetrahedron Lett. 2000, 41, 3555-3558.
(11) J hingan, A. K.; Meehan, T. J . Chem. Res., Synop. 1991, 122-
123.
(12) (a) Pilcher, A. S.; Yagi, H.; J erina, D. M. J . Am. Chem. Soc.
1998, 120, 3520-3521. (b) Ilankumaran, P.; Pannell, L. K.; Gebrese-
lassie, P.; Pilcher, A. S.; Yagi, H.; Sayer, J . M.; J erina, D. M. Chem.
Res. Toxicol. 2001, 14, 1330-1338.
J . Org. Chem, Vol. 67, No. 19, 2002 6679

Yagi et al.
SCHEME 1a
a
Key: (i) (CF3)2CHOH; (ii) HPLC or column; (iii) Ac2O, DMAP, pyridine; (iv) HPLC.
satisfactory conditions for this reaction.8 Blocking the
sugar hydroxyl groups of dGuo with the very nonpolar
tert-butyldimethylsilyl ether group (TBDMS) and O6-allyl
protection (O6-allyl-3′,5′-di-O-(tert-butyldimethylsilyl)-2′-
deoxyguanosine (3), Figure 1) provides an organic soluble
form of the purine with enhanced nucleophilicity of the
2-amino group that efficiently reacts with the DEs in
dimethylacetamide (DMA) at elevated temperature. For
BaP DE-1 and DE-2 (100 °C, 2 h), a mixture of cis and
trans adducts was formed (∼50% yield) with a cis/trans
ratio of ∼1/1.8a The more hindered and less reactive BcPh
DE-2 (2) (100 °C, 5 h) gave a cis/trans ratio of 27/73 in
much lower yield (26%).8b Surprisingly, we have been
unable to find similar conditions where 3′,5′-di-O-(tert-
butyldimethylsilyl)-2′-deoxyadenosine reacts with DEs to
form adducts.
to give a mixture of cis and trans adducts in 65 and 43%
overall yields, respectively, with cis/trans ratios of 85/15
(DE-1) and 40/60 (DE-2).13 Unlike the reaction in DMA,
reaction of 3′,5′-di-O-(tert-butyldimethylsilyl)-2′-deoxy-
adenosine in TFE with BaP DE-1 and DE-2 afforded
exclusively cis-opened adducts, albeit in low yield (22-
33%).13 Encouraged by such facile reaction at room
temperature, we explored the possibility of solvent-free
reaction.14 Simply grinding BaP DE-1 or DE-2 with 3
with a pestle in a porcelain mortar afforded a mixture of
the corresponding cis and trans adducts (cis/trans ratio
for DE-1: 50/50, 45% yield; for DE-2: 15/85, 54% yield).14
As with reaction in DMA, very little dAdo adduct forms
in the solvent-free reaction.
(13) Ramesha, A. R.; Kroth, H.; J erina, D. M. Tetrahedron Lett.
2001, 42, 1003-1005. In this study, we observed cis addition of solvent
TFE to the BaP DEs. On page 1005, line 6, the products were stated
to derive from trans addition due to a typographical error.
(14) Ramesha, A. R.; Kroth, H.; J erina, D. M. Org. Lett. 2001, 3,
531-533.
Fluorinated alcohols have a dramatic catalytic effect
on the addition reactions between blocked purines and
DEs. Thus, reaction of BaP DE-1 and DE-2 with 3 in
2,2,2-trifluoroethanol (TFE) proceeds at room temperature
6680 J . Org. Chem., Vol. 67, No. 19, 2002

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
TABLE 1. HF P -Med ia ted Syn th esis of Cis- a n d
Although there have been a number of studies using
site-specifically modified oligonucleotides containing dGuo
adducts of BaP DEs, corresponding oligonucleotides
containing BcPh DE-dGuo adducts have been less acces-
sible due to the greater difficulty in synthesizing such
adducts from the more hindered and less reactive inter-
mediates in the BcPh series. Thus there is a pressing
need for new synthetic routes to oligonucleotides contain-
ing such BcPh DE-dGuo adducts for physical and bio-
chemical studies. Since reaction of BcPh DE-2 with 3 in
DMA proceeded in modest yield,8b we have investigated
in the present study the potential of fluorinated alcohols
as catalysts in the formation of the desired adducts. We
have found that reaction in 1,1,1,3,3,3-hexafluoropropan-
2-ol (HFP) results in much higher yields (62-84%) and
that the cis/trans ratio of adducts is highly dependent
on the amount of fluorinated alcohol used. The use of
HFP provides ready access to the eight possible diaste-
reomerically pure, fully protected N2-dGuo adducts of
BcPh DE-1 (1) and BcPh DE-2 (2) for use in oligonucleo-
tide synthesis (Scheme 1).
Tr a n s-Op en ed N2-d Gu o Ad d u cts fr om BcP h DE-1 (1) a n d
DE-2 (2)a
diol
epoxide
HFP
(molar equiv)
temp
(°C)
cis:trans
yield
(%)c
adduct ratiob
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
532
532
266
266
100
100
75
75
30
30
24
24
5
rt
rt
rt
rt
rt
rt
45
60
50
70
75
75
110
110
140
140
91:9
66
78
67
80
63
84
62
83
69
81
68
72
62
74
74
60
61:39
90:10
53:47
84:16
40:60
83:17
36:64
52:48
18:82
42:58
14:86
17:83
2:98
5
0
0
5:95
2:98
a
All reactions were run for 3 h with a molar ratio of the DE to
dGuo nucleoside (3) of 1 to 5, except for the fusion reaction in the
b
absence of solvent, where the ratio was 1 to 20. Ratio was
determined by HPLC (300 nm) as disilyl triacetates for DE-1 and
directly as disilyl trihydroxy derivatives for DE-2. c Actual isolated
yield of combined cis and trans adducts (as triacetates for DE-1,
50-100 mg scale).
Resu lts a n d Discu ssion
Direct opening of the highly hindered fjord-region BcPh
DEs (1 and 2) with 3 was investigated using TFE (pKa )
12.4),15 HFP (pKa ) 9.3),15 and perfluoro-tert-butyl alcohol
(PFTB; pKa ) 5.2)15b as solvents. HFP was found to be
the solvent of choice. Although the addition reactions
occurred in PFTB, the yield was lower than in HFP, and
several unidentified products were formed. Increased
ionizing power and acidity of PFTB might have caused
degradation through N7-adduct formation and deglyco-
sylation. Unlike BaP DEs,13 the corresponding BcPh DEs
did not undergo this addition reaction in the less acidic
TFE at room temperature. The C-1 carbocation corre-
sponding to the BcPh DEs (∆Edeloc ) 0.600â) is higher in
energy and less easily formed than the C-10 carbocation
corresponding to the BaP DEs (∆Edeloc ) 0.794â).16 Thus,
the BcPh DEs are much less reactive17 than the BaP DEs.
Increased steric hindrance of the fjord region of BcPh DEs
compared with the bay region of BaP DEs may also
retard their reaction rates.
Reaction of a 5-fold molar excess of 3 with BcPh DEs
(1 or 2) was complete within 3 h at room temperature
when a 100-fold molar equivalent or more of HFP was
present (Table 1). When a lesser amount of HFP was used
(75 molar equiv or less), the reaction required heating
(45-140 °C, 3 h) to reach homogeniety, at which time
reaction was complete. Yields of the adducts as their
triacetates for the DE-1 reaction (62-67%) and the DE-2
reaction (72-84%) were much higher than that for DE-2
in DMA (26%).8b Although decomposition products of the
DEs were not detected, minor amounts of solvent addition
of HFP (∼10%) were observed (Scheme 1). In reactions
of DE-1 (1), only the cis-HFP adduct (8) was formed. In
reactions of DE-2, the cis-HFP adduct (9) was ac-
companied by a lesser amount (1/5) of the trans adduct
(10). For reactions of DE-1 (1), residual 3 was removed
by column chromatography on silica gel, and the result-
ant adduct fraction was acetylated. The triacetate of the
cis-HFP adduct (8) and the diastereomeric triacetates of
the O-allyl disilyl dGuo adducts cis-(1R)-4a , cis-(1S)-4b,
trans-(1S)-5a , and trans-(1R)-5b (Scheme 1) were readily
separated by HPLC. For the reactions of DE-2 (2), direct
HPLC provided four fractions: the cis-HFP adduct (9),
the trans-HFP adduct (10), the cis-opened diastrereo-
meric adduct mixture, and the trans-opened diastereo-
meric adduct mixture. After acetylation, the diastereo-
isomeric pair of cis isomers (cis-(1S)-6a and cis-(1R)-6b)
and the diastereoisomeric pair of trans isomers (trans-
(1S)-7a and trans-(1R)-7b) were separated by HPLC.
These DE-2 dGuo adducts were identified (1H NMR, CD,
and mass spectra) by comparison with authentic samples.8b
The cis/trans ratio of adducts from both DE-1 (1) and
DE-2 (2) in HFP changes with the relative amount of
HFP solvent used. As shown in Table 1 and in Figure 2,
a maximum cis/trans ratio (91/9) for DE-1 was obtained
when 532 molar equiv of HFP was present. A significant
increase in the formation of the trans adduct began below
∼150 molar equiv of HFP. At 5 molar equiv of HFP, the
ratio reached 17/83. A similar, but less dramatic change
was observed for DE-2. The observed cis/trans ratios
ranged from 61/39 to 2/98 for this DE.
Unlike the BaP DEs,14 the less reactive BcPh DEs (1
or 2) did not undergo solvent-free reaction with 3 at room
temperature. However, when 1 or 2 and a 20-fold molar
excess of 3 were fused at 140 °C (Table 1), reaction was
complete within 3 h (74 and 60% yields, respectively).
Almost exclusive formation of the trans isomers was
observed for both DE-1 (1) (cis/trans ) 5/95) and DE-2
(2) (cis/trans ) 2/98). This high trans selectivity provides
an excellent route to these trans isomers. Much less
selectivity was observed with the BaP DEs in the absence
of solvent.14
(15) (a) Matesich, M. A.; Knoefel, J .; Feldman, H.; Evans, D. F. J .
Phys. Chem. 1973, 77, 366-369. (b) Filer, R.; Schure, R. M. J . Org.
Chem. 1967, 32, 1217-1219.
(16) J erina, D. M.; Lehr, R. E. In Microsomes and Drug Oxidation;
Ullrich, V., Roots, I., Hildebrandt, A. G., Estabrook, R. W., Conney, A.
H., Eds.; Pergamon Press: Oxford, 1977; pp 709-720.
(17) Sayer, J . M.; Yagi, H.; Croisy-Delcey, M.; J erina, D. M. J . Am.
Chem. Soc. 1981, 103, 4970-4972.
J . Org. Chem, Vol. 67, No. 19, 2002 6681

Yagi et al.
TABLE 2. Com p a r ison of th e BcP h DE-1 Ben zo-Rin g 1H NMR Da ta (500 MHz, Aceton e-d 6) for th e O6-Allyl-P r otected
Cis- a n d Tr a n s-Op en ed N2-d Gu o Ad d u cts a s Th eir Disilyl Tr ia ceta tesa
compd H1 H2
4a (1R) 6.92 4.90
H3
5.97
H4
6.17
cis-O-allyl-dGuo-DE-1
4b (1S)
J 1,2 ) 3.3
6.95
J 2,3 ) 9.3
4.92
J 3,4 ) 3.3
5.98
6.18
J 1,2 ) 3.3
6.19
J 2,3 ) 9.3
5.84
J 3,4 ) 3.4
5.12
5a (1S)
6.88
trans-O-allyl-dGuo-DE-1
5b (1R)
J 1,2 ) 3.3
6.19
J 2,3 ) 2.4
5.84
J 1,3 ) 0.9
5.13
J 3,4 ) 7.8
6.85
J 1,2 ) 3.3
J 2,3 ) 2.4
J 1,3 ) 0.9
J 3,4 ) 8.2
a
Assignments and coupling constants were confirmed by decoupling experiments.
only the cis-HFP adduct (8) was formed as a side product
in the reaction of 1.
In contrast to 1, the fjord region has less of an effect
on the preferred conformation of 2 since relief of steric
strain between H1 and H12 in 2a is offset by an adverse
eclipsing interaction between the epoxide C2-O bond and
the nonbenzylic hydroxyl group (C3-OH) in 2b (Scheme
2). Consequently, fjord-region DE-2 (2) prefers the same
conformation17 (2a ) as DE-2 isomers with a bay re-
gion.18,19 The corresponding carbocation (2c) would be
expected to undergo preferential axial attack, as above,
to give trans opened products, and indeed only trans-
tetrol is observed in the acid-catalyzed hydrolysis of 2.17
In the present case, a marked increase in cis addition of
3 [∼60% cis-2 adduct] is observed. Similarly, cis-(9) and
trans-(10) HFP adducts were formed from 2 in a ratio of
80 to 20. We speculate that this preference for cis addition
in HFP may result from an increased lifetime of the
carbocation, which allows the initially formed 2c to
undergo conformational equilibration with 2d at a rate
faster than its capture by solvent or by 3. Thus, assuming
that capture of 2c and 2d by axial attack of 3 occurs at
similar rates, the product ratio under SN1 conditions will
be largely determined by the position of the equilibrium
between these conformers. Although previous studies20,22
indicate that SN1 solvolyses of DE-2 isomers in water
proceed predominantly via the conformation correspond-
ing to 2c to give trans-tetrols, the present results suggest
that for BcPh DE-2 in HFP conformations 2c and 2d may
be of comparable stability and have sufficiently long
lifetimes to equilibrate before they are captured by a
nucleophile.
F IGURE 2. Stereoselectivity of the HFP-mediated cis and
trans opening of BcPh DE-1 (1) and DE-2 (2) by the dGuo
reactant O6-allyl-3′,5′-di-O-(tert-butyldimethylsilyl)-2′-deoxy-
guanosine (3) in a molar ratio of 1:5. See Table 1 for details.
In the presence of very low amounts of HFP, reaction
of the DEs (1 and 2) with 3 is presumed to proceed largely
by an SN2 mechanism. As the amount of HFP is increased
relative to the other reactants, the medium becomes more
polar, and an SN1 reaction mechanism becomes more
favorable. This would account for the increase in the
amount of cis-opened products. Conformations of the
starting DEs and their intermediate carbocations (Scheme
2) provide a plausible explanation for the cis/trans
product ratios observed in the presence of high amounts
of HFP (Figure 2). On the basis of observed NMR spectra
in Me2SO-d6,17 1 is expected to prefer conformation 1b
(Scheme 2) in polar media such as HFP. This conforma-
tion differs from that observed for DE-1 isomers with a
bay region18,19 as a result of steric crowding between H1
and H12 in the fjord region. The conformation of the
carbocation formed upon initial epoxide ring opening of
1b is 1d . This conformation is also expected to be the
preferred one at equilibrium.20 Energetically favorable
axial attack20,21 by nucleophiles on this conformation will
lead to cis products. Thus, under SN1 conditions, acid-
catalyzed hydrolysis of 1 and its reaction with 3 in the
presence of solvent HFP give predominantly cis-tetrol
(∼85%)17 and cis-1 adduct (∼90%), respectively. Similarly,
1H NMR spectral data for the saturated benzo-ring
protons of the DE-1 O6-allyl disilyl dGuo triacetates (4a ,b
and 5a ,b) are given in Table 2. Practically identical
chemical shifts and coupling constants were observed for
members of the cis pair and of the trans pair of diaster-
eomers. Assignment of relative stereochemistry for these
adducts was readily achieved by comparison of the
present NMR data with those obtained for the BcPh
tetrol tetraacetates23 and the peracetates of the dGuo
adducts23 derived from DE-1. Notably, only products
derived from cis opening of DE-1 (i.e., 4a ,b) have a large
(21) Goering, H. L.; J osephson, R. R. J . Am. Chem. Soc. 1962, 84,
2779-2785. Goering, H. L.; Vlazny, J . C. J . Am. Chem. Soc. 1979, 101,
1801-1805.
(18) Yagi, H.; Thakker, D. R.; Hernandez, O.; Koreeda, M.; J erina,
D. M. J . Am. Chem. Soc. 1977, 99, 1604-1611.
(22) Sayer, J . M.; Lehr, R. E.; Kumar, S.; Yagi, H.; Yeh, H. J . C.;
Holder, G. M.; Duke, C. C.; Silverton, J . V.; Gibson, C.; J erina, D. M.
J . Am. Chem. Soc. 1990, 112, 1177-1185.
(19) Lehr, R. E.; Schaefer-Ridder, M.; J erina, D. M. Tetrahedron Lett.
1977, 539-542.
(23) Agarwal, S. K.; Sayer, J . M.; Yeh, H. J . C.; Pannell, L. K.; Hilton,
B. D.; Pigott, M. A.; Dipple, A.; Yagi, H.; J erina, D. M. J . Am. Chem.
Soc. 1987, 109, 2497-2504.
(20) Sayer, J . M.; Yagi, H.; Silverton, J . V.; Friedman, S. L.; Whalen,
D. L.; J erina, D. M. J . Am. Chem. Soc. 1982, 104, 1972-1978.
6682 J . Org. Chem., Vol. 67, No. 19, 2002

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
SCHEME 2
value for J 2,3 (9.3 Hz) and a small value for J 3,4 (3.3 Hz)
compared to a small value of J 2,3 (2.4 Hz) and a large
value of J 3,4 (7.8-8.2 Hz) for the trans-opened DE-1
products (i.e., 5a ,b). In the case of the cis-opened adducts
of DE-1 (4a ,b), the substituent at C-1 prefers a quasiaxial
conformation (constrained by steric hindrance in the fjord
region), which requires the acetoxy groups at C-2 and
C-3 to be quasidiequatorial in a twisted half chair
conformation; thus, J 2,3 is large (cis-1 in Figure 3). For
the trans adducts (5a ,b), a twisted half boat conformation
is preferred in which the substituent at C-1 is again
quasiaxial (trans-1 in Figure 3). Consequently, the C-3
acetoxy group adopts a quasiaxial conformation (trans-1
in Figure 3). The dihedral angle between C2-H and
C3-H is ∼110° (J 2,3 ) 2.4 Hz), and the angle between
C3-H and C4-H is large (J ≈ 8 Hz). A characteristic
W-type long-range coupling (J 1,3 ) 0.9 Hz) was observed
only for the trans adducts (5a ,b).
given in Table 3. In all cases (8 from DE-1, 9 and 10 from
DE-2), the most downfield doublets were assigned as H-1
on the carbon bearing the HFP substituent, as they
underwent little downfield shift on acetylation in contrast
to marked downfield shifts of the other signals. The large
values of J 2.3 and the intermediate values of J 3,4 for the
HFP adduct 8 and its acetate indicate cis addition and
are in good agreement with those of the cis addition
products of BcPh DE-1 (1) discussed above. HFP adduct
9 also results from cis addition but has somewhat
unusual intermediate values for J 2,3 and J 3,4, which did
not agree with the previously observed values for either
cis or trans adducts derived from BcPh DE-2.23 The cis-
2A (half chair) and preferred cis-2B (half boat) conforma-
tions consistent with these coupling constants are shown
in Figure 3. Notably, in the presence of 3′,5′-di-O-(tert-
butyldimethylsilyl)-2′-deoxyadenosine, reaction of solvent
TFE with BaP DEs occurred by exclusive cis addition.13
A minor HFP adduct 10 derived from DE-2 had a large
1H NMR spectral data for the saturated benzo-ring
protons of the HFP adducts and their triacetates are
J . Org. Chem, Vol. 67, No. 19, 2002 6683

Yagi et al.
F IGURE 3. Preferred conformations of adducts derived from BcPh DEs.
TABLE 3. Com p a r ison of th e BcP h DE-1 a n d DE-2
ration of the BcPh DE-1 O6-allyl disilyl triacetyl N2-dGuo
adducts (4a ,b and 5a ,b) by removal of the O6-allyl
protecting group and determination of the CD spectra
(Figure 4) of the resultant disilyl triacetates of the N2-
dGuo adducts (11a ,b and 12a ,b). Thus, the positive CD
band at 258 nm for (1S)-12a , which was obtained on O6-
allyl deprotection of (1S)-5a , establishes that these
configurationally related compounds have a 1S absolute
configuration and were formed from (-)-[1R,2S,3R,4S]-
BcPh DE-1 by inversion of the configuration at C-1
(Scheme 1).
Ben zo-Rin g 1H NMR Da ta for Disilyl HF P Ad d u cts befor e
a n d a fter Acetyla tion a
compd
H1
6.65
H2
4.25
H3
4.27
H4
8b
4.79
J 1,2 ) 1.8
J 2,3 ) 10.3
J 3,4 ) 7.3
triacetate of 8c
9b
6.82
5.55
5.95
6.47
5.16
6.76
5.10
6.63
J 1,2 ) 1.9
6.52
J 1,2 ) 3.0
6.63
J 1,2 ) 1.7
6.82
J 1,2 ) 4.6
6.72
J 2,3 ) 9.6
4.44
J 2,3 ) 8.2
5.54
J 2,3 ) 7.4
4.64
J 2,3 ) 2.7
6.03
J 3,4 ) 4.1
3.95
J 3,4 ) 7.1
5.55
J 3,4 ) 5.1
4.20
J 3,4 ) 8.4
5.74
triacetate of 9d
10e
In the preceding paper,8b we have reported the syn-
thesis of 5′-dimethoxytrityl (DMT)-protected 3′-phos-
phoramidite building blocks of the N2-dGuo adducts
derived from cis and trans opening of BcPh DE-2 (6a ,b
and 7a ,b). Those phosphoramidites are 1R/1S diastereo-
meric mixtures at C-1. Their use in oligonucleotide
synthesis results in a mixture of 1R/1S constructs.
Although separation of the two adducted oligonucleotides
has generally been possible, we have recently encoun-
tered a case in which the R/S pair of oligonucleotides
could not be separated by HPLC under a variety of
conditions.24 In the present study, each of the pure
diastereomers at C-1, 6a ,b and 7a ,b derived from cis and
trans opening of BcPh DE-2, as well as newly synthesized
N2-dGuo adducts (4a ,b and 5a ,b) derived from DE-1 were
separately used as the starting materials for the synthe-
sis of the corresponding phosphoramidites (Scheme 3).
Although more work synthetically, use of single phos-
phoramidite diastereomers with a known absolute con-
figuration greatly simplifies the HPLC purification of the
resultant oligonucleotides since only one of two closely
triacetate of 10c
J 1,2 ) 4.7
J 2,3 ) 2.9
J 3,4 ) 8.6
a
Assignments and coupling constants were confirmed by de-
b
coupling experiments. Measured at 500 MHz (CDCl3-CD3OD).
c Measured at 300 MHz (CDCl3). Measured at 500 MHz (CDCl3).
d
e Measured at 300 MHz (CDCl3-CD3OD).
value for J 3,4 and a small value for J 2,3 and J 1,2 compat-
ible23 with the trans-2 (half chair) form shown in Figure
3.
Previously, we had established the absolute configu-
ration of the O6-allyl disilyl triacetyl N2-dGuo adducts
(6a ,b and 7a ,b) derived from BcPh DE-2 by removal of
the O6-allyl protecting group and comparison of the CD
spectra of the resulting disilyl triacetates8b with those of
the corresponding unprotected nucleoside adducts of
known absolute configuration.23 For BcPh DE N2-dGuo
adducts, a negative CD band at ∼258 nm is typical of an
R absolute configuration at C-1 of the hydrocarbon. The
nonchromophoric TBDMS and acetyl groups have little
effect on the CD spectra of unprotected nucleoside
adducts.12b In contrast, O6-allyl protection significantly
alters the guanine chromophore and causes marked
changes in the exciton CD spectra.8 The same strategy
has been used here to determine the absolute configu-
(24) We have been unable to separate a pair of 16-mer oligonucleo-
tides containing the trans-DE-2 adducts 7a and 7b. See: Ramos, L.
A.; Ponte´n, I.; Yagi, H.; Sayer, J . M.; Kroth, H.; Kalena, G.; Kumar,
S.; Dipple, A.; J erina, D. M. Chem. Res. Toxicol., submitted for
publication.
6684 J . Org. Chem., Vol. 67, No. 19, 2002

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
F IGURE 4. Circular dichroism spectra (normalized to 1.0 absorbance unit at λmax in methanol) of the cis- and trans-opened
O6-allyl-protected N2-dGuo adducts 4a ,b and 5a ,b and of the O6-allyl-deprotected N2-dGuo adducts 11a ,b and 12a ,b, all as disilyl
triacetetes. The strong CD bands at 261 nm for the O6-allyl-deprotected cis-opened N2-dGuo adducts 11a ,b (solid line, 1S) and at
258 nm for the O6-allyl-deprotected trans-opened N2-dGuo adducts 12a ,b (solid lines, 1S) are diagnostic of their absolute
configuration. See text and Scheme 1.
eluting adducted oligonucleotide peaks is present and the
adduct configuration is known in advance. In addition,
there is twice as much of this adduct-containing peak
relative to failure sequences. Also, separation of R/S pairs
of adducted oligonucleotides is likely to become increas-
ingly more difficult as the sequence length increases.
Previously, desilylation of 6a ,b and 7a ,b was achieved
with 2% HF in a mixture of acetonitrile and pyridine (4:
1).8b Although the yields for desilylation under these mild
conditions were reasonably high (70-75%), conversion
of starting material into desilylated products often
required recycling recovered starting material (24 h) two
to three times. This time-consuming recycling procedure
could be avoided by simply using 7% HF in pyridine for
12 h at room temperature. Thus, 4a ,b, 5a ,b, 6a ,b, and
7a ,b were converted to the sugar desilylated derivatives
13a ,b, 14a ,b, 15a ,b, and 16a ,b, respectively, in >95%
yield (Scheme 3).
To date, two methods have been utilized for the
selective introduction of the 5′-O-dimethoxytrityl (DMT)
protecting group into polycyclic aromatic hydrocarbon-
adducted nucleotide building blocks. The first method
uses DMT chloride in pyridine under strictly anhydrous
conditions.25 Although this method has seen considerable
use,8b,9b,c yields can be variable and sometimes low. Some
improvements have been observed when a 2.25-fold
excess of DMT chloride and DMAP are present.8b The
second method uses DMT fluoroborate26 with lutidine and
Li2CO3 as bases in THF.10,12,27 We have had generally
good results with the second method when the reaction
proceeds normally (>90% conversion within 2 h). How-
ever, when extended reaction times are required (> 20
h), substantial byproducts form. For example on trityla-
tion of 14a in THF (Scheme 3), a major secondary
reaction occurs in which the desired product (19a ) reacts
with THF at the 3′-hydroxyl group to produce a diaster-
eomeric pair of acetals (17a ,b) on consumption of the
starting material. Presumably, hydride abstraction from
the C-2 position of THF by the DMT carbocation oc-
curred,28 and the resultant carbocation alkylated the 3′-
hydroxyl group of 19a . In fact, when DMT fluoroborate
was stored in THF, the characteristic reddish color of the
DMT carbocation gradually faded and the hydride-
transfer product, bis-(4-methoxyphenyl)phenylmethane,
was produced. Simply replacing THF with CH2Cl2 as
solvent solved the problem. The tritylation reactions were
completed within 2-3 h without formation of any sig-
nificant side products. Thus, the required 5′-DMT deriva-
tives 18a ,b, 19a ,b, 20a ,b, and 21a ,b were obtained from
13a ,b, 14a ,b, 15a ,b, and 16a ,b, respectively, in nearly
quantitative yield (>95%).
Finally, introduction of the phosphoramidite function
at the 3′-deoxyribose hydroxyl group in the 5′-DMT
derivatives (18a,b, 19a,b, 20a,b, and 21a,b) was achieved
with 2-cyanoethyl tetraisopropylphosphorodiamidite and
N,N-diisopropylammonium tetrazolide29 in CH2Cl2 to give
22a ,b, 23a ,b, 24a ,b, and 25a ,b, respectively, in 80-90%
(27) Lakshman, M. K.; Zajc, B. Nucleosides Nucleotides 1996, 15,
1029-1039.
(28) Barton, D. H. R.; Magnus, P. D.; Smith, G.; Streckert, G.; Zurr,
D. J . Chem. Soc., Perkin Trans. 1 1972, 542-552.
(29) Beaucage, S. L. In Methods in Molecular Biology: Protocols for
Oligonucleotides and Analogues; Agrawal, S., Ed.; Humana Press:
Totowa, NJ , 1993; Vol. 20, pp 33-61.
(25) J ones, R. In Oligonucleotide Synthesis: A Practical Approach;
Gait, M. J ., Ed.; IRL Press: Oxford, 1984; pp 22-34.
(26) Bleasdale, C.; Ellwood, S. B.; Golding, B T. J . Chem. Soc., Perkin
Trans. 1 1990, 803-805.
J . Org. Chem, Vol. 67, No. 19, 2002 6685

Yagi et al.
SCHEME 3a
Key. (i) 7%/HF/pyridine; (ii) DMT+BF4--THF; (iii) DMT+BF4--CH2Cl2; (iv) 2-cyanoethyltetraisopropylphosphoramidite.
a
yield. Each of the diastereomeric phosphoramidites thus
obtained consisted of a mixture of two diastereomers due
to the asymmetry of the phosphorus. In general the
mixed phosphorus diastereomers exhibit two peaks with
equal intensities on 31P NMR spectroscopy. Although
these phosphorus diastereomers could be separated by
HPLC (see Experimental Section), such separation was
not routinely done prior to their use in oligonucleotide
synthesis.
dGuo (3) in the absence of solvent provided almost
exclusively the trans adducts in 60-74% yield. Through
the use of appropriate amounts of HFP, the cis/trans ratio
can be varied from 1/1 to as much as 91/9 for DE-1 (66%
yield). For DE-2, the highest cis/trans ratio was 61/39 in
78% yield. The high polarity and higher acidity of HFP
relative to that of TFE are thought to account for the
marked catalytic activity of HFP through hydrogen
bonding and/or protonation of the epoxide oxygen. HFP
may also enhance the nucleophilicity of the exocyclic
amino group of guanine.30
In summary, novel and efficient syntheses of O6-allyl-
protected N2-dGuo phosphoramidite building blocks 22a,b,
23a ,b, 24a ,b, and 25a ,b (cf. Scheme 3) in diastereomeri-
cally pure forms are described. Simply heating the
individual DEs with O6-allyl- and di-O-TBDMS-protected
(30) Moon, K.-Y.; Moschel, R. C. Chem. Res. Toxicol. 1998, 11, 696-
702.
6686 J . Org. Chem., Vol. 67, No. 19, 2002

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
glass (4.5 mg). The ratio of the cis to trans adducts was
Exp er im en ta l Section
estimated to be ∼1/1 by 1H NMR (300 MHz, acetone-d6) on
the basis of the H12 aromatic proton signals of the cis
diastereomers (δ 8.87 ppm) and trans diastereomers (δ 8.70
ppm). The remainder of the crude product was acetylated with
Ac2O (1 mL) in pyridine (5 mL) in the presence of DMAP (20
mg) at room temperature for 18 h. After removal of volatiles
in vacuo, the residue was taken up in EtOAc (100 mL) and
washed with 5% NaHCO3 (2 × 30 mL) and water (50 mL).
Evaporation of the solvent gave a gum, which was separated
by HPLC on a Vertex column (LiChrosorb Si-60, 2 × 25 cm,
Sonntek, Inc., Upper Saddle River, NJ ) eluted with 20% EtOAc
in n-hexane at a flow rate of 25 mL/min (detected at 300 nm).
The triacetate of the cis-HFP adduct 8 had tR ) 8.4 min (10%
as detected at 300 nm; 45 mg, 8% isolated yield). The triacetate
of the cis-N2-dGuo adducts had tR(early) ) 19.1 min [(1R)-4a ,
25% (300 nm); 158 mg, 17.3% isolated yield] and tR(late) ) 25.6
min [(1S)-4b, 22% (300 nm); 114 mg, 13% isolated yield]. The
triacetate of the trans-N2-dGuo adducts had tR(early) ) 31.3 min
[(1S)-5a , 18% (300 nm); 145 mg, 16% isolated yield] and
tR(late) ) 34.1 min [(1R)-5b, 26% (300 nm); 205 mg, 23% isolated
yield]. (()-1â-Hexaflu or o-2-pr opyloxy- 2â,3r,4â-tr iacetoxy-
1,2,3,4-tetr a h yd r oben zo[c]p h en a n th r en e (Tr ia ceta te of
Ca u tion : Benzo[c]phenanthrene 3,4-dihydrodiol and diol
epoxides DE-1 and DE-2 are mutagenic and carcinogenic and
must be handled carefully in accordance with NIH guidelines.31
1
Gen er a l Meth od s. H NMR spectra were recorded at 300 or
500 MHz as indicated. Chemical shifts are reported in parts
per million (δ) downfield from TMS. Coupling constants (J )
are given in hertz, and spin multiplicities are indicated by the
following symbols:
s (singlet), d (doublet), t (triplet), q
(quartet), m (multiplet), br (broad). Assignments are based on
decoupling experiments in all cases. For adducts and related
compounds, singly primed numbers are used for the protons
on the ribose moiety (1′-5′) and the purine protons are doubly
primed (8′′). For the vinyl protons of the allyl-protecting group,
Hv designates the vinyl hydrogen adjacent to the methylene,
and Hc and Ht are the terminal vinyl protons cis and trans to
Hv. 31P NMR spectra were recorded at 300 MHz in CD3CN
with 85% H3PO4 as an external standard. High-resolution
mass spectroscopy (HRMS) was performed using standard
methods. Methylene chloride (CH2Cl2), lutidine, triethylamine
(TEA), and pyridine were dried over 4 Å molecular sieves after
distillation. Tetrahydrofuran (THF) was distilled over LiAlH4
prior to use. O6-Allyl-3′,5′-di-O-(tert-butyldimethylsilyl)-2′-
deoxyguanosine (3) was prepared from 2′-deoxyguanosine as
described.32 The diol epoxides, DE-1 and DE-2 of BcPh, were
synthesized from the BcPh 3,4-dihydrodiol as described.17
Flash column chromatography was performed using thick-
walled glass columns with 230-400 mesh silica gel 60. Melting
points were not corrected.
1
8). Colorless prisms; mp 141 °C. H NMR (300 MHz, CDCl3)
δ: 2.16, 2.17 and 2.23 (each s, each 3H, 3 × CH3CO), 3.33 (m,
1H, (CF3)2CHO), 5.55 (dd, 1H, H2, J ) 9.6, 1.9), 5.95 (dd, 1H,
H3, J ) 9.6, 4.1), 6.47 (d, 1H, H4, J ) 4.1), 6.82 (d, 1H, H1,
J ) 1.9), 7.50-8.00 (m, 7H, aromatic protons), 8.45 (m,1H,
H
12). HRMS (FAB+) calcd for C27H22O7F6 (M+): 572.1270.
Found: 572.1258. N2-[1R-(2R,3R,4S-Tr ia cetoxy-1,2,3,4-tet-
r a h yd r oben zo[c]p h en a n th r en yl)]-O6-a llyl-3′,5′-d i-O-(ter t-
bu tyld im eth ylsilyl)-2′-d eoxygu a n osin e (4a ). 1H NMR (500
MHz, acetone-d6) δ: 0.11(s, 6H, Si bonded Me), 0.14 (s, 6H, Si
bonded Me), 0.88 (br s, 9H, tert-butyl Me), 0.94 (s, 9H, tert-
butyl Me), 1.80, 2.06 and 2.23 (each s, each 3H, 3 x CH3CO),
2.50 (m, 1H, H2′), 2.80 (m, 1H, H2′), 3.86 (m, 2H, H5′,5′), 3.98
(m, 1H, H4′), 4.72 (br s, 1H, H3′), 4.90 (dd, 1H, H2, J ) 9.3,
3.3), 5.03 (m, 2H, CH2(allyl)), 5.26 (br d, 1H, Hc, J ) 10.5), 5.47
(br d, 1H, Ht, J ) 16.8), 5.97 (dd, 1H, H3, J ) 9.3, 3.3), 6.16
(m, 1H, Hv), 6.17 (dd, 1H, H4, J ) 3.3), 6.50 (m, 1H, H1′), 6.74
(1H, d, NH, J ) 10.0), 6.92 (d, 1H, H1, J ) 10.0, 3.3), 7.32 (m,
1H, H11), 7.60 (br t, 1H, H10), 7.71 (d, 1H, H5, J ) 8.2), 7.86
(AB q, 2H, H7/H8, J ) 7.8), 8.00 (dd, 1H, H9, J ) 8.8, 0.9), 8.07
(d, 1H, H6, J ) 8.2), 8.12 (s, 1H, H8′′), 8.70 (br s, 1H, H12).
HRMS (FAB+) calcd. for C49H66O10N5Si2 (M+): 940.4348.
Found: 940.4342. CD spectrum (see Figure 4). N2-[1S-(2S,-
3S,4R-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo[c]p h en a n th r e-
n yl)]-O6-a llyl-3′,5′-d i-O-(ter t-bu tyld im eth ylsilyl)-2′-d eoxy-
gu a n osin e (4b). 1H NMR (500 MHz, acetone-d6) δ: 0.00-
0.20 (m, 12H, Si bonded Me), 0.90 (br s, 18H, tert-butyl Me),
1.76, 2.06 and 2.24 (each s, each 3H, 3 × CH3CO), 2.50 (m,
1H, H2′), 2.80 (m, 1H, H2′), 3.88 (m, 2H, H5′,5′), 3.97 (m, 1H,
H4′), 4.72 (br s, 1H, H3′), 4.92 (dd, 1H, H2, J ) 9.3, 3.3), 5.03
(m, 2H, CH2(allyl)), 5.28 (br d, 1H, Hc, J ) 10.5), 5.48 (br d, 1H,
Ht, J ) 16.8), 5.98 (dd, 1H, H3, J ) 9.3, 3.4), 6.20 (m, 1H, Hv),
6.18 (d, 1H, H4, J ) 3.4), 6.48 (m, 1H, H1′), 6.75 (1H, d, NH,
J ) 10.0), 6.95 (dd, 1H, H1, J ) 10.0, 3.3), 7.46 (m, 1H, H11),
7.62 (m, 1H, H10), 7.71 (d, 1H, H5, J ) 8.2), 7.88 (AB q, 2H,
H7/H8, J ) 7.8), 8.00 (dd, 1H, H9, J ) 7.6, 0.9), 8.08 (d, 1H,
H6, J ) 8.2), 8.11 (s, 1H, H8′′), 8.82 (br d, 1H, H12, J ) 8.5).
HRMS (FAB+) calcd for C49H66O10N5Si2 (M+): 940.4348.
Found: 940.4358. CD spectrum (see Figure 4). N2-[1S-(2R,-
3R,4S-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo[c]p h en a n th r e-
n yl)]-O6-a llyl-3′,5′-d i-O-(ter t-bu tyld im eth ylsilyl)-2′-d eoxy-
gu a n osin e (5a ). 1H NMR (500 MHz, acetone-d6) δ: 0.00-0.20
(m, 12H, Si bonded Me), 0.85-0.92 (m, 9H, tert-butyl Me), 0.94
(s, 9H, tert-butyl Me), 1.78, 1.94 and 2.18 (each s, each 3H,
3 × CH3CO), 2.34 (m, 1H, H2′), 2.92 (m, 1H, H2′), 3.78 (m, 2H,
H5′,5′), 3.92 (m, 1H, H4′), 4.62 (br s, 1H, H3′), 5.00 (m, 2H,
CH2(allyl)), 5.12 (ddd, 1H, H3, J ) 7.8, 2.4, 0.9), 5.20 (br d, 1H,
Hc, J ) 10.5), 5.41 (dd, 1H, Ht, J ) 17.2, 2.6), 5.84 (dd, 1H,
H2, J ) 3.3, 2.4), 6.14 (m, 1H, Hv), 6.19 (dd, 1H, H1, J ) 3.3,
Rea ction of BcP h DE-1 (1) w ith O6-Allyl-3′,5′-d i-O-(ter t-
bu tyld im eth ylsilyl)-2′-d eoxygu a n osin e (3) in HF P . To a
solution of O6-allyl dGuo di-TBDMS ether (3) (2.6 g, 4.85 mmol)
in HFP (4.9 g, 29.2 mmol, 30 molar equiv) was added BcPh
DE-1 (1) (270 mg, 0.97 mmol), and the mixture was stirred at
50 °C in a sealed vial filled with Ar gas. The crystals of 1
gradually dissolved, and the mixture was stirred at room
temperature for 3 h or until no starting material was detected
by HPLC. If 100 or more molar equiv of HFP was used, the
sealed reaction mixture was stirred for 3 h at room temper-
ature. If less than 100 molar equiv was used, the sealed
reaction mixture was heated with stirring at various temper-
atures (45-140 °C) for 3 h, at which time a uniform solution
was obtained (see Table 1). When no solvent HFP was used, a
mixture of 1 and 20 molar equiv of 3 was melted in a sealed
vial under Ar gas with stirring at 140 °C for 3 h.
The solvent HFP was evaporated in vacuo, and the residue
was subjected to column chromatography on silica gel (2.5 ×
30 cm) eluted with 40% n-hexane in EtOAc to give unreacted
3 (1.8 g, 61% recovery). The column was then eluted with 5%
MeOH in CH2Cl2, and the solvent was evaporated to give a
colorless glass (550 mg). A portion (∼5 mg) of this crude
product was subjected to HPLC on two coupled Axxiom silica
gel columns (9.5 × 250 mm, 5 µm) using 25% n-hexane in
EtOAc as a solvent at a flow rate of 6 mL/min (detected at
260 nm) and separated into two fractions (tR ) 9.2 and 10.4
min). Evaporation of the early fraction gave (()-1â-h exaflu or o-
2-p r op yloxy-2â,3r,4â-tr ih yd r oxy-1,2,3,4-tetr a h yd r oben -
zo[c]p h en a n th r en e (8) as a colorless solid (0.5 mg). 1H NMR
(500 MHz, CDCl3-CD3OD) δ: 3.81 (m, 1H, (CF3)2 CHO), 4.22
(dd, 1H, H2, J ) 10.3, 1.8), 4.27 (dd, 1H, H3, J ) 10.3, 7.3),
4.79 (d, 1H, H4, J ) 7.3), 6.65 (d, 1H, H1, J ) 1.8), 7.48-7.90
(m, 7H, aromatic protons), 8.27 (br d, 1H, H12, J ) 8.0). HRMS
(FAB+) calcd for C21H16O7F6 (M+): 446.0953. Found: 446.0941.
Evaporation of the late fraction gave a mixture of two cis-
and two trans-N2-dGuo adduct diastereomers as a colorless
(31) NIH Guidelines for the Laboratory Use of Chemical Carcinogens;
NIH Publication No. 81-2385; U.S. Government Printing Office:
Washington, DC, 1981.
(32) Steinbrecher, T.; Wameling, C.; Oesch, F.; Seidel, A. In Poly-
cyclic Aromatic Compounds: Synthesis, Properties, Analytical Measure-
ments, Occurrence and Biological Effects; Garrigues, P., Lamotte, M.,
Eds; Gordon and Breach: Amsterdam, 1993; pp 223-230.
J . Org. Chem, Vol. 67, No. 19, 2002 6687

Yagi et al.
0.9), 6.37 (dd, 1H, H1′, J ) 7.7, 5.7), 6.88 (d, 1H, H4, J ) 7.8),
7.30 (m, 2H, NH and H11), 7.60 (dt, 1H, H10, J ) 7.8, 1.1), 7.72
(d, 1H, H5, J ) 7.9), 7.86 (app. s, 2H, H7/H8), 8.00 (dd, 1H, H9,
J ) 8.8, 1.2), 8.08 (d, 1H, H6, J ) 8.1), 8.12 (s, 1H, H8′′), 8.44
(br d, 1H, H12, J ) 8.2). HRMS (FAB+) calcd for C49H66O10N5-
Si2 (M+): 940.4348. Found: 940.4358. CD spectrum (see Figure
4). N2-[1R-(2S,3S,4R-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo-
[c]p h en a n th r en yl)]-O6-a llyl-3′,5′-d i-O-(ter t-bu tyld im eth -
ylsilyl)-2′-deoxygu an osin e (5b). 1H NMR (500 MHz, acetone-
d6) δ: (-0.38)-0.12 (m, 12H, Si bonded Me), 0.86 (m, 9H, tert-
butyl Me), 0.89 (m, 9H, tert-butyl Me), 1.72, 1.96 and 2.27 (each
s, each 3H, 3 × CH3CO), 2.35 (m, 1H, H2′), 2.95 (m, 1H, H2′),
3.75 (m, 2H, H5′,5′), 3.90 (m, 1H, H4′), 4.60 (br s, 1H, H3′), 5.00
(m, 2H, CH2(allyl)), 5.13 (ddd, 1H, H3, J ) 8.2, 2.4, 0.9), 5.20 (br
d, 1H, Hc, J ) 10.5), 5.41 (dd, 1H, Ht, J ) 17.2, 2.6), 5.84 (dd,
1H, H2, J ) 3.3, 2.4), 6.14 (m, 1H, Hv), 6.19 (dd, 1H, H1, J )
3.3, 0.9), 6.37 (dd, 1H, H1′, J ) 7.7, 5.7), 6.85 (d, 1H, H4, J )
tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo[c]p h en a n th r en e (tr i-
a ceta te of 10) as colorless solid (95% yield). 1H NMR (300
MHz, CDCl3) δ: 2.10, 2.13 and 2.27 (each s, each 3H, 3 × CH3-
CO), 4.42 (m, 1H, (CF3)2CHO), 5.74 (dd, 1H, H3, J ) 8.6, 2.9),
6.03 (dd, 1H, H2, J ) 4.7, 2.9), 6.63 (d, 1H, H4, J ) 8.6), 6.72
(d, 1H, H1, J ) 4.7), 7.55-8.10 (m, 7H, aromatic protons), 8.30
(br d, 1H, H12, J ) 8.2 ). HRMS (FAB+) calcd for C27H22O7F6
(M+): 572.1270. Found: 572.1265.
Evaporation of the combined fraction at tR ) 4.4 min
afforded a mixture of the two trans-N2-dGuo diastereomers as
a colorless solid (128 mg, 54%). Evaporation of the combined
fraction at tR ) 5.1 min afforded a mixture of the two cis-N2-
dGuo diastereomers as a colorless solid (69 mg, 30%). The cis/
trans adduct ratio obtained from HPLC (300 nm) was 36/64.
These adduct mixtures were separately acetylated and sepa-
rated by HPLC into the diastereomerically pure triacetates
(1S)-6a and (1R)-6b and (1S)-7a and (1R)-7b, respectively,
according to the conditions previously reported,8b and their
structures were confirmed by comparison of their 1H NMR,
HRMS, and CD spectra with those of the authentic samples.8b
N2-[1R-(2R,3R,4S-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo-
[c]p h en a n th r en yl)]-3′,5′-d i-O-(ter t-bu tyld im eth ylsilyl)-2′-
d eoxygu a n osin e (11a ) a n d N2-[1S-(2S,3S,4R-Tr ia cetoxy-
1,2,3,4-tetr ah ydr oben zo[c]ph en an th r en yl)]-3′,5′-di-O-(ter t-
bu tyldim eth ylsilyl)-2′-deoxygu an osin e (11b). Palladium(0)
catalysts are effective reagents for the removal of O-allyl
protecting groups.33 To a solution of either 4a (7 OD at 260
nm) or 4b (8 OD at 260 nm) in CH2Cl2 (3 mL) was added a
solution of Pd(Ph3)4 (10.2 mg) and morpholine (620 µL) in CH2-
Cl2 (1 mL), and the mixture was stirred at room temperature
under dry Ar gas for 1 h. The reaction mixture was washed
with brine (2 mL), saturated NaHCO3 (2 mL), and water (2
mL). The extract was dried (MgSO4) and evaporated to give a
solid, which was purified by HPLC on an Axxiom silica gel
column (9.5 × 250 mm, 5 µm) eluted with 10% n-hexane in
EtOAc at a flow rate of 5 mL/min (detected at 260 nm) to give
11a (4.2 OD at 260 nm; tR ) 9.1 min) or 11b (5.0 OD at 260
nm; tR ) 10.7 min). CD spectrum (see Figure 4). HRMS
(FAB+) calcd for C46H61N5O10Si2Cs (M+ + Cs): 1032.3011.
Found: 1032.3029 (11a ); 1032.3007 (11b).
N2-[1S-(2R,3R,4S-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo-
[c]p h en a n th r en yl)]-3′,5′-d i-O-(ter t-bu tyld im eth ylsilyl)-2′-
d eoxygu a n osin e (12a ) a n d N2-[1R-(2S,3S,4R-Tr ia cetoxy-
1,2,3,4-tetr ah ydr oben zo[c]ph en an th r en yl)]-3′,5′-di-O-(ter t-
bu tyld im eth ylsilyl)-2′-d eoxygu a n osin e (12b). To a solution
of either 5a (3 OD at 260 nm) or 5b (3 OD at 260 nm) in CH2-
Cl2 (3 mL) was added a solution of Pd(Ph3)4 (10.2 mg) and
morpholine (620 µL) in CH2Cl2 (1 mL), and the mixture was
stirred at room temperature under dry Ar gas for 1 h. The
reaction mixture was washed with brine (2 mL), saturated
NaHCO3 (2 mL), and water (2 mL). The extract was dried
(MgSO4) and evaporated to give a solid, which was purified
by HPLC on an Axxiom silica gel column (2.0 × 250 mm, 5
µm) eluted with 35% n-hexane in EtOAc at a flow rate of 0.95
mL/min (detected at 260 nm) to give 12a (1.5 OD at 260 nm;
tR ) 9.4 min) or 12b (0.42 OD at 260 nm; tR ) 9.1 min). CD
8.2), 7.30 (dt, 1H, H11), 7.43 (br s, 1H, NH), 7.58 (br t, 1H, H10
,
J ) 7.8, 1.1), 7.72 (d, 1H, H5, J ) 8.2), 7.86 (app. s, 2H, H7/
H8,), 8.00 (dd, 1H, H9, J ) 8.8, 1.1), 8.07 (d, 1H, H6, J ) 8.2),
8.09 (s, 1H, H8′′), 8.49 (br d, 1H, H12, J ) 8.6). HRMS (FAB+)
calcd for C49H66O10N5Si2 (M+): 940.4348. Found: 940.4359. CD
spectrum (see Figure 4).
Rea ction of BcP h DE-2 (2) w ith O6-Allyl-3′,5′-d i-O-(ter t-
bu tyld im eth ylsilyl)-2′-d eoxygu a n osin e (3) in HF P . To a
solution of O6-allyl dGuo di-TBDMS ether (3) (783 mg, 1.46
mmol) in HFP (3.69 g, 21.93 mmol, 75 molar equiv) was added
BcPh DE-2 (2) (81.4 mg, 0.29 mmol) with stirring at 60 °C in
a sealed vial filled with Ar gas. The crystals of 2 gradually
dissolved, and the reaction mixture was stirred at room
temperature for 3 h or until no starting material was detected
by HPLC. If more than 100 molar equiv was used, the sealed
reaction mixture was heated with stirring at various temper-
atures (60-140 °C) for 3 h, at which time uniform solutions
were obtained (see Table 1). When no HFP was used, a mixture
of 1 and 20 molar equiv of 3 was melted and stirred in a sealed
vial under Ar gas at 140 °C for 3 h.
The solvent HFP was evaporated in vacuo, and the residue
was subjected to HPLC separation on an Axxiom silica gel
column (9.5 × 250 mm) eluted with a mixture of TEA (1%),
n-hexane (29%), and EtOAc (70%) as a solvent at a flow rate
of 10 mL/min (detected at 300 nm). Evaporation of the
combined fraction at tR ) 1.8 min gave recovered 3 (600 mg,
77% recovery). Evaporation of the combined fraction at tR
)
2.57 min gave (()-1r-h exa flu or o-2-p r op yloxy-2r,3r,4â-
tr ih yd r oxy-1,2,3,4-tetr a h yd r oben zo[c]p h en a n th r en e (9)
as a colorless solid (8 mg, 6%). 1H NMR (500 MHz, CDCl3-
CD3OD) δ: 3.93 (m, 1H, (CF3)2CHO), 3.95 (dd, 1H, H3, J )
8.2, 7.1), 4.44 (dd, 1H, H2, J ) 8.2, 3.0), 5.16 (d, 1H, H4, J )
7.1), 6.52 (d, 1H, H1, J ) 3.0), 7.48-7.90 (m, 7H, aromatic
protons), 8.10 (br d, 1H H12, J ) 8.0). HRMS (FAB+) calcd for
C
21H16O7F6 (M+): 446.0953. Found: 446.0948. Acetylation of
9 by standard methods afforded (()-1r-h exa flu or o-2-p r o-
p yloxy-2r,3r,4â-t r ia cet oxy-1,2,3,4-t et r a h yd r ob en zo[c]-
p h en a n th r en e (tr ia ceta te of 9) as colorless prisms from
1
Et2O-n-hexane (mp 163 °C, 95% yield). H NMR (500 MHz,
spectrum (see Figure 4). HRMS (FAB+) calcd for C46H61N5O10
-
CDCl3) δ: 2.15, 2.17 and 2.27 (each s, each 3H, 3 × CH3CO),
4.08 (m, 1H, (CF3)2CHO), 5.54 (dd, 1H, H2, J ) 7.4, 1.7), 5.55
(dd, 1H, H3, J ) 7.4, 5.1), 6.63 (d, 1H, H1, J ) 1.7), 6.76 (d,
1H, H1, J ) 5.1), 7.50-7.93 (m, 7H, aromatic protons), 8.42
(m, 1H, H12). HRMS (FAB+) calcd for C27H22O7F6 (M+):
572.1270. Found: 572.1255.
Si2Cs (M+ + Cs): 1032.3011. Found: 1032.3005 (12a );
1032.3029 (12b).
Syn th esis of O6-Allyl-d Gu o P h osp h or a m id ite Bu ild in g
Block s. (1) Desilyla tion of th e Su ga r Hyd r oxyl Gr ou p s.
The diastereomerically pure O6-allyl-protected cis-(1R)-4a or
cis-(1S)-4b and trans-(1S)-5a or trans-(1R)-5b, derived from
BcPh DE-1 (1), as well as cis-(1S)-6a or cis-(1R)-6b and trans-
(1S)-7a and trans-(1R)-7b, derived from BcPh DE-2 (2) (each
on a 120 mg scale), were separately dissolved in 7% HF-
pyridine (14 mL) under cooling at 5 °C in a polyethylene vial.
The reaction mixture was stirred for 12 h at room temperature
Evaporation of the combined fraction of tR ) 2.63 min gave
(()-1â-h exaflu or o-2-pr opyloxy-2r,3r,4â-tr ih ydr oxy-1,2,3,4-
tetr a h yd r oben zo[c]p h en a n th r en e (10) as a colorless solid
(2 mg, 1.5%). 1H NMR (300 MHz, CDCl3-CD3OD) δ: 3.31 (m,
1H, (CF3)2CHO), 4.20 (dd, 1H, H3, J ) 8.4, 2.7), 4.64 (dd, 1H,
H2, J ) 4.6, 2.7), 5.10 (d, 1H, H4, J ) 8.4), 6.82 (d, 1H, H1,
J ) 4.6), 7.50-8.01 (m, 7H, aromatic protons), 8.51 (br d, 1H
H
12, J ) 8.1). HRMS (FAB+) calcd for C21H16O7F6 (M+):
(33) (a) Kunz, H.; Waldmann, H. Angew. Chem., Int. Ed. Engl. 1984,
23, 71-73. (b) Schmittberger, T.; Cotte`, A.; Waldmann, H. J . Chem.
Soc., Chem. Commun. 1998, 937-938.
446.0953. Found: 446.0956. Acetylation of 10 by standard
methods afforded (()-1â-h exa flu or o-2-p r op yloxy-2r,3r,4â-
6688 J . Org. Chem., Vol. 67, No. 19, 2002

Cis vs Trans Opening of Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
and diluted with EtOAc (300 mL), which was washed with
saturated NaHCO3 (50 mL) and water (3 × 50 mL). The extract
was dried (Na2SO4) and evaporated in vacuo to give a colorless
solid (>95% yield). HRMS (FAB+) calcd for C37H37N5O10
(M+): 712.2619. Found: 712.2634 [(1R)-13a ]; 712.2603 [(1S)-
13b]; 712.2612 [(1S)-14a ]; 712.2593 [(1R)-14b]; 712.2632 [(1S)-
15a ]; 712.2613 [(1R)-15b]; 712.2622 [(1S)-16a ]; 712.2599 [(1R)-
16b].
(2) Syn th esis of 5′-O-Dim eth oxytr ityl Der iva tives. (i)
The desilylated triacetates 13a ,b, 14a ,b, 15a ,b, and 16a ,b
were transformed into the corresponding 5′-O-DMT-triacetates
by the treatment with a 2.6-fold excess of DMT+BF4-, a 5-fold
excess of lutidine, and a 4-fold excess of Li2CO3 in dry CH2Cl2
(2 mL). Reaction mixtures were stirred under Ar gas at room
temperature for 2-4 h until no starting material was detected
by TLC or HPLC. The reaction mixtures were purified by silica
gel column chromatography using 0-40% n-hexane in EtOAc
(with a trace of TEA) as the mobile phase to give colorless
H3′), 5.15 (m, 2H, CH2(allyl)), 5.22 (dd, 1H, H3, J ) 8.1, 2.5),
5.28-5.36 (m, 2H, H2 and C-2 methine proton of THF acetal),
5.54 (d, 1H, Ht, J ) 17.5), 5.98 (m, 1H, H2), 6.26 (m, 1H, Hv),
6.30 (dd, 1H, H1, J ) 3.3, 0.9), 6.46 (t, 1H, H1′, J ) 6.5), 6.80-
7.72 (m, 15H, aromatic protons), 6.98 (d, 1H, H4, J ) 8.1), 7.84
(d, 1H, H5, J ) 8.2), 7.98 (app. s, 2H, H7/H8), 8.12 (br d, 1H,
H9, J ) 8.0), 8.14 (s, 1H, H8′′), 8.20 (d, 1H, H6, J ) 8.3), 8.65
(br d, 1H. H12, J ) 8.4). HRMS (FAB+) calcd for C62H62N5O13
(M+ + 1): 1084.4349. Found: 1084.4353.
(iii) Rea ction of DMT+BF 4 in THF . A solution of
-
DMT+BF4- (390 mg, 1.0 mmol) in THF (10 mL) was stirred at
room temperature for 48 h. The bright red color of the reaction
mixture gradually faded to give a yellowish solution. The
reaction mixture was evaporated in vacuo, and the residue was
loaded onto a silica gel column and eluted with 20% EtOAc in
n-hexane (with trace of TEA) to give bis-(4-m eth oxyp h en yl)-
p h en ylm eth a n e as a slightly yellowish oil (216 mg, 71%). 1H
NMR (300 MHz, CDCl3) δ: 3.76 (s, 5H, 2 × CH3O), 5.18 (s,
1H, C-1 methine proton), 6.80-7.30 (m, 9H, aromatic protons).
solids (>95% yield). HRMS (FAB+) calcd for C58H55N5O12
-
Cs (M+ + Cs): 1146.2902. Found: 1146.2904 [(1R)-18a ];
1146.2931 [(1S)-18b]; 1146.2899 [(1S)-19a ]; 1146.2898 [(1R)-
19b]; 1146.2870 [(1S)-20a ]; 1146.2921 [(1R)-20b]; 1146.2908
[(1S)-21a ]; 1146.2902 [(1R)-21b].
HRMS (EI) calcd for
304.1461.
C
21H20O2 (M+): 304. 1463. Found:
(3) 3′-(2-Cya n oeth yl-N,N-d iisop r op yl)p h osp h or a m id -
ites. Each of the 5′-O-DMT-derivatives 18a ,b, 19a ,b, 20a ,b,
and 21a ,b was combined with a 10-fold excess of N,N-diiso-
propylammonium tetrazolide and dried for 6 h under high
vacuum (0.3 mmHg). To this mixture was added a 10-fold
excess of 2-cyanoethyl tetraisopropylphosphorodiamidite in dry
CH2Cl2 (10 mL). The mixture was stirred under Ar gas at room
temperature for 3-4 h until no starting material was detected
by HPLC. The reaction mixture was evaporated in vacuo, and
the residue was purified by silica gel column chromatography
under Ar gas using 40% EtOAc in n-hexane (with 0.1% TEA)
as a solvent. Fractions containing the desired products were
pooled and evaporated to give a colorless glass to which
n-hexane (∼2 mL) was added. The mixture was sonicated until
the products formed colorless powders. The solvent was
decanted, and the resulting powders were washed several
times with n-hexane and dried in vacuo. Yields were 80-90%.
HPLC on an Axxiom silica gel column (9.5 × 250 mm) eluted
at 8 mL/min with the proportions of EtOAc in n-hexane
designated in parentheses showed the following two peaks of
equal areas (260 nm) corresponding to diastereomers at
phosphorus: (1R)-22a , tR ) 7.7 and 8.6 min (40%); (1S)-22b,
tR ) 12.4 and 20.4 min (40%); (1S)-23a , tR ) 5.4 and 7.4 min
(ii) N2-[1S-(2R,3R,4S-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben -
zo[c]ph en an th r en yl)]-O6-allyl-3′-O-(2R-tetr ah ydr ofu r an yl)-
5′-O-(d im eth oxytr ityl)-2′-d eoxygu a n osin e (17a ) a n d /or
N2-[1S-(2R,3R,4S-Tr ia cetoxy-1,2,3,4-tetr a h yd r oben zo[c]-
p h en a n th r en yl)]-O6-a llyl-3′-O-(2S-tetr a h yd r ofu r a n yl)-5′-
O-(d im eth oxytr ityl)-2′-d eoxygu a n osin e (17b). Tritylation
of 14a (147 mg, 0.207 mmol) was carried out the same way as
above except that THF (2 mL) was used instead of CH2Cl2.
The reaction mixture was stirred for 20 h until ∼90% of the
starting material was consumed as judged by HPLC. The crude
product was subjected to chromatography on a silica gel
column eluted with 50% n-hexane in EtOAc (with a trace of
TEA) to give a glass (240 mg), which was separated by HPLC
on an Axxiom silica gel column (9.5 × 250 mm) eluted with
50% n-hexane in EtOAc (with a trace of TEA) at a flow rate of
10 mL/min (detected at 300 nm). The combined fraction of
tR ) 8.8 min afforded a glass (17a or 17b; 88 mg, 39%). 1H
NMR (300 MHz, acetone-d6) δ: 1.73-2.01 (m, 4H, 2 × C-3 and
C-4 methine protons of the THF acetal), 1.71, 2.02 and 2.14
(each s, each 3H, 3 × COCH3), 2.70 (m, 1H, H2′), 3.00 (m, 1H,
H2′), 3.25-3.50 (m, 1H, H5′, 5′), 3.85 and 3.84 (each s, each 3H,
2 × OCH3), 3.95 (m, 2H, C-5 methylene protons of the THF
acetal), 4.02 (m, 1H, H4′), 4.70 (m, 1H, H3′), 5.10 (m, 2H,
CH2(allyl)), 5.24 (dd, 1H, H3, J ) 8.1, 2.5), 5.28-5.35 (m, 2H, H2
and C-2 methine proton of the THF acetal), 5.54 (d, 1H, Ht,
J ) 17.5), 5.96 (m, 1H, H2), 6.26 (m, 1H, Hv), 6.34 (dd, 1H, H1,
J ) 3.3, 0.9), 6.46 (t, 1H, H1′, J ) 6.5), 6.90-7.70 (m, 15H,
aromatic protons), 7.18 (d, 1H, H4, J ) 8.1), 7.84 (d, 1H, H5,
J ) 8.3), 7.98 (app. s, 2H, H7/H8), 8.12 (br d, 1H, H9, J ) 8.0),
(50%); (1R)-23b, tR ) 5.5 and 6.4 min (50%); (1S)-24a , tR
)
7.2 and 10.7 min (45%); (1R)-24b, tR ) 9.6 and 25.8 min (45%).
HPLC on a Vertex LiChrosorb Si-60 column (20 × 250 mm)
eluted at 9 mL/min with a linear gradient of 30% -70% EtOAc
in n-hexane over 60 min gave two peaks with equal areas:
(1S)-25a , tR ) 33.8 and 43.3 min; (1R)-25b, tR ) 30.5 and 36.2
min. 31P NMR (300 MHz, CD3CN) δ: 148.810 (s) [(1R)-22a ];
149.212 (s) [(1S)-22b]; 148.885 (s) and 149.288 (s) [(1S)-23a ];
149.086 (s) [(1R)-23b ]; 149.272 (s) [(1S)-24a ]; 149.020 (s)
[(1R)-24b]; 148.793 (s) and 149.185 (s) [(1S)-25a ]; 148.974 (s)
8.14 (s, 1H, H8′′), 8.20 (d, 1H, H6, J ) 8.3), 8.65 (br d, 1H. H12
,
J ) 8.4). HRMS (FAB+) calcd for C62H62N5O13 (M+ + 1):
1084.4349. Found: 1084.4384.
The combined fraction at tR ) 10.5 min afforded a glass (17a
or 17b; 80 mg, 36%). 1H NMR (300 MHz, acetone-d6) δ: 1.73-
2.02 (m, 4H, 2 × C-3 and C-4 methine protons of the THF
acetal), 1.80, 2.02 and 2.10 (each s, each 3H, 3 × COCH3), 2.67
(m, 1H, H2′), 2.98 (m, 1H, H2′), 3.40 (m, 2H, H5′, 5′), 3.77 (m,
2H, C-5 methylene protons of the THF acetal), 3.85 and 3.87
(each s, each 3H, 2 × OCH3), 4.26 (m, 1H, H4′), 4.74 (m, 1H,
[(1R)-25b]. HRMS (FAB+) calcd for C67H72N7O13Cs (M+
+
Cs): 1346.3980. Found: 1346.3962 [(1R)-22a ]; 1346.3970 [(1S)-
22b]; 1346.3991 [(1S)-23a ]; 1346.3978 [(1R)-23a ]; 1346.3942
[(1S)-24a ]; 1346.3998 [(1R)-24b ]; 1346.3993 [(1S)-25a ];
1346.3988 [(1R)-25b].
J O020418A
J . Org. Chem, Vol. 67, No. 19, 2002 6689