Convergent synthesis of polyhalogenated quinoline C-nucleosides as potential antiviral agents

Article abstract of DOI:10.1021/jo020643s

2,5,6-Trichloro-1-(β-D-ribofuranosyl)benzimidazole (TCRB) and 2-bromo-5,6-dichloro-1-(β-D-ribofuranosyl)benzimidazole (BDCRB) are benzimidazole nucleosides that exhibit strong and selective anti-HCMV activity. We proposed to synthesize 2-halo-6,7-dichloro-4-(β-D-ribofuranosyl)quinolines as 6 + 6 bicyclic analogues of TCRB. The synthesis used Wittig reactions in two key steps. The first Wittig reaction coupled a fully functionalized benzene with a ribofuranose derivative to provide (Z)-6-O-(tert-butyldimethylsilyl)-1-(4,5-dichloro-2-nitrophenyl)- 1,2-dideoxy-3,4-O-isopropylideneD-allo-1-enitol (5) as the basic skeleton for the target compounds. The following electrophile-mediated intramolecular cyclization of the cis-alkene (5) was found to afford (1S,2S)-2,5-anhydro-1-bromo6-O-(tert-butyldimethylsilyl)- 1-deoxy-1-(4,5-dichloro-2-nitrophenyl)-3,4-O-isopropylidene-D-allitol (8) as the major product. This α-stereoselectivity was contrary to the literature precedence. A doublebond isomerization was established to be the cause of the unexpected stereochemistry. The bromo group of 8 was displaced by a hydroxyl group. Oxidation of the hydroxy group and the reduction of a phenylnitro group provided (2S)-1-(2-amino-4,5-dichlorophenyl)-2,5-anhydro-6-O- (tert-butyldimethylsilyl)-3,4-O-isopropylidene-D-allose (11), which was subjected to the second Wittig reaction with a phosphacumulene to construct 4-[5-O -(tert-butyldimethylsilyl)-2,3-O- isopropylidene-α-Dribofuranosyl]-6,7-dichloroquinolin-2-one (13). Halogenation followed by deprotection of 13 and led to the synthesis of 4-(α -D-ribofuranosyl)-2,6,7-trichloroquinoline (17) as the major product. The 2-aminophenone α-nucleoside (11) was successfully anomerized to the β-anomer (19), which led to the synthesis of the targeted 2-chloro- and 2-bromo-6,7-dichloro-4-(β-D-ribofuranosyl)quinolines (18 and 21, respectively).

Full text of DOI:10.1021/jo020643s

Con ver gen t Syn th esis of P olyh a logen a ted Qu in olin e C-Nu cleosid es
a s P oten tia l An tivir a l Agen ts
J iong J . Chen,† J ohn C. Drach, and Leroy B. Townsend*
Department of Chemistry, College of Literature, Sciences and Arts, Department of Medicinal Chemistry,
College of Pharmacy, Biologic and Materials Sciences, School of Dentistry, University of Michigan,
Ann Arbor, Michigan 48109-1065
ltownsen@umich.edu
Received October 9, 2002
2,5,6-Trichloro-1-(â-D-ribofuranosyl)benzimidazole (TCRB) and 2-bromo-5,6-dichloro-1-(â-D-ribo-
furanosyl)benzimidazole (BDCRB) are benzimidazole nucleosides that exhibit strong and selective
anti-HCMV activity. We proposed to synthesize 2-halo-6,7-dichloro-4-(â-D-ribofuranosyl)quinolines
as 6 + 6 bicyclic analogues of TCRB. The synthesis used Wittig reactions in two key steps. The
first Wittig reaction coupled a fully functionalized benzene with a ribofuranose derivative to provide
(Z)-6-O-(tert-butyldimethylsilyl)-1-(4,5-dichloro-2-nitrophenyl)-1,2-dideoxy-3,4-O-isopropylidene-
D-allo-1-enitol (5) as the basic skeleton for the target compounds. The following electrophile-mediated
intramolecular cyclization of the cis-alkene (5) was found to afford (1S,2S)-2,5-anhydro-1-bromo-
6-O-(tert-butyldimethylsilyl)-1-deoxy-1-(4,5-dichloro-2-nitrophenyl)-3,4-O-isopropylidene-D-allitol (8)
as the major product. This R-stereoselectivity was contrary to the literature precedence. A double-
bond isomerization was established to be the cause of the unexpected stereochemistry. The bromo
group of 8 was displaced by a hydroxyl group. Oxidation of the hydroxy group and the reduction of
a phenylnitro group provided (2S)-1-(2-amino-4,5-dichlorophenyl)-2,5-anhydro-6-O-(tert-butyldi-
methylsilyl)-3,4-O-isopropylidene-D-allose (11), which was subjected to the second Wittig reaction
with a phosphacumulene to construct 4-[5-O-(tert-butyldimethylsilyl)-2,3-O-isopropylidene-R-D-
ribofuranosyl]-6,7-dichloroquinolin-2-one (13). Halogenation followed by deprotection of 13 and led
to the synthesis of 4-(R-D-ribofuranosyl)-2,6,7-trichloroquinoline (17) as the major product. The
2-aminophenone R-nucleoside (11) was successfully anomerized to the â-anomer (19), which led to
the synthesis of the targeted 2-chloro- and 2-bromo-6,7-dichloro-4-(â-D-ribofuranosyl)quinolines (18
and 21, respectively).
In tr od u ction
A series of 2-substituted benzimidazole nucleosides
have been synthesized in our laboratory and displayed
strong antiviral activities.1 The lead compounds, includ-
ing 2,5,6-trichloro-1-(â-D-ribofuranosyl)benzimidazole
(TCRB) and 2-bromo-5,6-dichloro-1-(â-D-ribofuranosyl)-
benzimidazole (BDCRB) (Figure 1), have shown potent
activity against HCMV with low cellular toxicity at
concentrations inhibiting viral growth (TCRB: IC50 ) 2.8
µM, CC50 ) 238 µM, BDCRB: IC50 ) 0.7 µM, CC50 ) 118
µM in a plaque assay in human foreskin fibroblasts).
TCRB is very selective against HCMV. It was found that
TCRB displays only weak antiviral activity against HSV-
1, HSV-2, VZV, and even murine cytomegalovirus. TCRB
was also found to be inactive against a selection of RNA
viruses including the A and B strains of influenza,
respiratory syncytia, measles virus, and HIV-1. Moreover,
TCRB was found to be active against three distinct
clinical isolates of HCMV that had become resistant to
ganciclovir. Overall, TCRB and BDCRB were determined
F IGURE 1.
to be promising anti-HCMV agents with high potency,
high selectivity, and low cytotoxicity.2 However, subse-
quent studies in rats and monkeys revealed the instabil-
ity of the glycosidic bond in vivo.3
(2) (a) Gudmundsson, K. S.; Tidwell, J .; Mezzano, N.; Koszalka, G.
W.; Van Draanen, N. A.; Ptak, R. G.; Drach, J . C.; Townsend, L. B. J .
Med. Chem. 2000, 43, 2464-2472. (b) Zou, R.; Kawashima, E.;
Freeman, G. A.; Koszalka, G. W.; Drach, J . C.; Townsend, L. B.
Nucleosides Nucleotides Nucleic Acids 2000, 19, 125-154. (c) Chulay,
J .; Biron, K. K.; Wang, L.; Underwood, M.; Chamberlain, S.; Frick, L.;
Good, S.; Davis, M.; Harvey, R.; Townsend, L. B.; Drach, J . C.;
Koszalka, G. W. Adv. Exp. Med. Biol. 1999, 458, 129-134. (d)
Gudmundsson, K. S.; Freeman, G. A.; Drach, J . C.; Townsend, L. B.
J . Med. Chem. 2000, 43, 2473-2478. (e) Gudmundsson, K. S.; Drach,
J . C.; Wotring, L. L.; Townsend, L. B. J . Med. Chem. 1997, 40, 785-
793.
† Present address: Pharmacia Corp., 7000 Portage Road, Kalama-
zoo, MI 49001.
(1) Townsend, L. B.; Devivar, R. V.; Turk, S. R.; Nassiri, M. R.;
Drach, J . C. J . Med. Chem. 1995, 38, 4098-4105.
10.1021/jo020643s CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/02/2003
4170
J . Org. Chem. 2003, 68, 4170-4178

Synthesis of Polyhalogenated Quinoline C-Nucleosides
SCHEME 1
This prompted a broadened synthetic effort to prepare
analogues with increased glycosidic bond stability. The
efforts produced several other promising compounds4
including 1263W94 (maribavir).5 1263W94 is an L-ribosyl
benzimidazole that is a potent and selective inhibitor of
HCMV replication.5 Its mechanism of action, however,
is different from TCRB and BDCRB and involves the
inhibition of viral DNA synthesis.5 A mutation in the
HCMV UL97 protein kinase produced resistance imply-
ing UL97 is the target for this compound.5 Development
of 1263W94 has progressed to phase I/II clinical trials
for the treatment of HCMV infections.5
Another approach to analogues with increased glyco-
sidic bond stability involves indole and imidazo[1,2-a]-
pyridine bicyclic heterocycles structurally related to the
benzimidazoles. Several polyhalogenated indole N-nu-
cleosides and imidazo[1,2-a]pyridine C-nucleosides have
been prepared as TCRB analogues.4 Some of these
nucleosides have also displayed some interesting antivi-
ral activities. Indole and imidazo[1,2-a]pyridine are 6 +
5 bicyclic ring systems, which may bear the same
exocyclic substitution pattern as that of TCRB. Several
6 + 6 bicyclic ring systems, such as a quinoline, have
the potential for a substitution pattern similar to that of
TCRB. Therefore, we initiated a study designed to
synthesize some quinoline C-nucleosides (Figure 1) as
part of our substantial SAR studies on TCRB. The
2-chloro- and 2-bromo-6,7-dichloro-4-(â-D-ribofuranosyl)-
quinolines should be sufficient to evaluate the potential
antiviral activity of this type of nucleoside as analogues
of TCRB.
could be used to install a functional group at the benzylic
position of the resulting homo-C-nucleoside. This homo-
C-nucleoside could then be elaborated to a 2-aminophe-
none nucleoside intermediate. From this intermediate,
several methods are available for the construction of a
quinoline ring.12-14
Resu lts a n d Discu ssion
4,5-Dichloro-2-nitrotoluene (1)15 under radical condi-
tions was brominated to yield 4,5-dichloro-2-nitrobenzyl
bromide (2) (Scheme 2). (4,5-Dichloro-2-nitrobenzyl)-
triphenylphosphonium bromide (3) was then obtained by
simply stirring compound 2 with triphenylphosphine at
room temperature for 4 days. The Wittig reaction has
been widely used in C-nucleoside syntheses.16 In most
reports, the alkene products were not isolated; instead,
the alkenes were cyclized to afford homo-C-nucleosides
under basic conditions as a result of an intramolecular
Michael addition.9,17-19 If compound 3 was directly coupled
with a ribofuranose derivative in the presence of triethyl-
amine as a base, the desired alkene products, i.e., (Z)-6-
O-(tert-butyldimethylsilyl)-1-(4,5-dichloro-2-nitrophenyl)-
1,2-dideoxy-3,4-O-isopropylidene-D-allo-1-enitol (5) and
(E)-6-O-(tert-butyldimethylsilyl)-1-(4,5-dichloro-2-nitro-
phenyl)-1,2-dideoxy-3,4-O-isopropylidene-D-allo-1-enitol
(6), were formed. However, compounds 5 and 6 under-
went an intramolecular Michael addition under the basic
conditions, and undesired (2S)-2,5-anhydro-6-O-(tert-
butyldimethylsilyl)-1-deoxy-1-(4,5-dichloro-2-nitrophenyl)-
3,4-O-isopropylidene-D-allitol (7) and its R-anomer were
obtained in a substantial amount along with 5 and 6. To
prevent this base-induced cyclization, compound 3 was
first converted to (4,5-dichloro-2-nitrobenzyl)triphen-
Quinoline C-nucleosides have been reported in the
literature.6,7 However, most of these compounds were
nucleosides with a ribofuranosyl group substituted at the
2-position of a simple quinoline.6 The only reported 4-(â-
D-ribofuranosyl)quinoline was prepared via a unique
enaminone glycoside procedure.7 However, this procedure
was not suitable for the synthesis of our target com-
pounds. To prepare the polyhalogenated quinoline C-
nucleosides regioselectively, we elected to use a homo-
C-nucleoside approach (Scheme 1). In this approach, a
ribofuranose derivative could be coupled with a fully
functionalized benzene through a Wittig reaction to give
a key alkene intermediate.8,9 Then an electrophile-
10,11
mediated intramolecular cyclization (EMIC) reaction
(3) Good, S. S.; Owens, B. S.; Townsend, L. B.; Drach, J . C. Antiviral
Res. 1994, 23, Suppl. I, 103.
(4) (a) Townsend, L. B.; Gudmundsson, K. S.; Daluge, S. M.; Chen,
J . J .; Zhu, Z.; Koszalka, G. W.; Boyd, L.; Chamberlain, S. D.; Freeman,
G. A.; Biron, K. K.; Drach, J . C. Nucleosides Nucleotides 1999, 18, 509-
519. (b) Chen, J . J .; Wei, Y.; Drach, J . C.; Townsend, L. B. J . Med.
Chem. 2000, 43, 2449-2456.
(5) (a) Biron, K. K.; Harvey, R. J .; Chamberlain, S. C.; Good, S. S.;
Smith, A. A., III; Davis, M. G.; Talarico, C. L.; Miller, W. H.; Ferris,
R.; Dornsife, R. E.; Stanat, S. C.; Drach, J . C.; Townsend, L. B.;
Koszalka, G. W. Antimicrob. Agents Chemother. 2002, 46, 2365-2372.
(b) Lalezari, J . P.; Aberg, J . A.; Wang, L. H.; Wire, M. B.; Miner, R.;
Snowden, W.; Talarico, C. L.; Shaw, S.; J acobson, M. A.; Drew, W. L.
Antimicrob. Agents Chemother. 2002, 46, 2969-2976.
(6) Yokoyama, M.; Tanabe, T.; Toyoshima, A.; Togo, H. Synthesis
1993, 517-520. (b) Yokoyama, M.; Toyoshima, A.; Akiba, T.; Togo, H.
Chem. Lett. 1994, 265-268. (c) Yokoyama, M.; Toyoshima, H.; Shimizu,
M.; Togo, H. J . Chem. Soc., Perkin Trans. 1 1997, 29-33.
(7) Maeba, I.; Ito, Y.; Wakimura, M.; Ito, C. Heterocycles 1993, 36,
2805-2810.
(10) Freeman, F.; Robarge, K. D. J . Org. Chem. 1989, 54, 346-359.
(11) Kane, P. D.; Mann, J . J . Chem. Soc., Perkin Trans. 1 1984, 657-
660.
(12) Chorbadjiev, S. Synth. Commun. 1990, 20, 3497-3505.
(13) Hino, K.; Kawashima, K.; Oka, M.; Nagai, Y.; Uno, H.; Mat-
sumoto, J . A Chem. Pharm. Bull. 1989, 37, 110-115.
(14) Lo¨ffler, J .; Schobert, R. D. J . Chem. Soc., Perkin Trans. 1 1996,
2799-2802.
(15) Cohen, J . B.; Dakin, H. D. J . Chem. Soc. 1902, 1344-1349.
(16) Watanabe, K. A. In The Chemistry of Nucleosides and Nucle-
otides; Townsend, L. B., Ed.; Plenum Press: New York, 1994; Vol. 3,
pp 421-535.
(17) Dondoni, A.; Marra, A. Tetrahedron Lett. 1993, 34, 7327-7330.
(18) Katagiri, N.; Takashima, K.; Kato, T. J . Chem. Soc., Perkin
Trans. 1 1983, 201-209.
(19) Barrett, A. G. M.; Broughton, H. B.; Attwood, S. V.; Gunatilaka,
A. A. L. J . Org. Chem. 1986, 51, 495-503.
(8) Chu, C. K.; Wempen, I.; Watanabe, K. A.; Fox, J . J . J . Org. Chem.
1976, 41, 2793-2997.
(9) Herrera, F. J . L.; Gonza´lez, M. S. P.; Aguas, R. P. J . Chem. Soc.,
Perkin Trans. 1 1989, 2401-2406.
J . Org. Chem, Vol. 68, No. 11, 2003 4171

Chen et al.
SCHEME 2
SCHEME 3
To circumvent the epoxidation step, an electrophile-
mediated intramolecular cyclization (EMIC) reaction was
applied. The EMIC reaction has been used previously in
the synthesis of C-ribofuranosides.11,21-23 The stereose-
lectivity of this intramolecular cyclization of ribose-
derived alkenes was systematically studied by Freeman
and co-workers.10 They found that a cis-alkene predomi-
nantly yielded a â-homonucleoside, while a trans-alkene
predominantly gave an R-anomer. Since the cis-alkene 5
was the major Wittig product in our case, the EMIC
reaction should fulfill our requirement of a â-homo-
nucleoside. Thus, the cis-alkene 5 was subjected to an
EMIC reaction using NBS as the electrophile. The
reaction was conducted in acetonitrile at 50 °C and was
completed in 5 h (Scheme 3). This reaction furnished one
major product, (1S,2S)-2,5-anhydro-1-bromo-6-O-(tert-
butyldimethylsilyl)-1-deoxy-1-(4,5-dichloro-2-nitrophenyl)-
3,4-O-isopropylidene-D-allitol (8), along with several in-
separable impurities. According to Freeman’s results,10
we expected to obtain a cyclized â-anomer as the major
product. However, the ∆δ of the isopropylidene of com-
pound 8 was quite small, only equal to 0.06 ppm, while
a â-anomer normally has ∆δ > 0.15 ppm according to
the Imbach rules and Freeman’s experimental results.10
When the trans-alkene 6 was subjected to the same
reaction conditions as those described for 5, compound 6
was converted to compound 8 in 1 h without any of the
impurities that were observed in the EMIC reaction of
the cis-alkene 5. This experimental result and the ∆δ
value of compound 8 suggested that compound 8 was an
ylphosphorane (4) with sodium hydroxide. The deep
purple ylide 4 was then used directly in the Wittig
reaction without any base and provided the cis-alkene 5
and trans-alkene 6 in a combined 85% yield and a 12:1
ratio, respectively. The configuration of the alkenes was
based on the coupling constants for the alkene protons,
i.e., 11.7 Hz for the cis isomer 5 and 15.2 Hz for the trans
isomer 6. Compound 7 was obtained as the major product
when the cis-alkene (5) was treated with NaOH in
MeOH. If the trans-alkene (6) was treated with NaOH
in MeOH, compound 7 and its R-anomer were obtained
as a mixture in almost equal amounts. Compound 7 was
most likely the â-anomer because the chemical shift for
the H-1′ of 7 was upfield from the peak observed for the
H-1′ of the other anomer.
To prepare a 2-aminophenone intermediate, we pro-
posed to oxidize the alkene 5 to an epoxide. An intramo-
lecular cyclization would then give a homo-C-nucleoside
with a benzylic hydroxyl group. Oxidation of this benzylic
hydroxy group and a reduction of the nitro group should
provide the desired 2-aminophenone intermediate. How-
ever, the π-electron-deficient nitrophenyl alkene 5 re-
sisted a number of strong epoxidation conditions (Scheme
3). For example, it did not give any significant amount
of an epoxide even under strong m-CPBA conditions20 or
trifluoroperacetic acid (TFPAA).
(21) Nicotra, F.; Panza, L.; Ronchetti, F.; Toma, L. Tetrahedron Lett.
1984, 25, 5937-5838.
(22) Nicotra, F.; Perego, R.; Ronchetti, F.; Russo, G.; Toma, L. Gazz.
Chim. Ital. 1984, 193-195.
(20) Kishi, Y.; Aratani, M.; Tanino, H.; Fukuyama, T.; Goto, T. J .
Chem. Soc., Chem. Commun. 1972, 64-65.
(23) Freeman, F.; Robarge, K. D. Carbohydr. Res. 1985, 89-97.
4172 J . Org. Chem., Vol. 68, No. 11, 2003

Synthesis of Polyhalogenated Quinoline C-Nucleosides
SCHEME 4
R-homonucleoside. The stereochemical outcome of this
EMIC reaction was quite surprising, since Freeman’s
results were supported by others.21,22 The close resem-
blance of our alkenes and Freeman’s substrates sug-
gested that Freeman’s reactive conformer models were
still applicable. The alkenes may exist in two low energy
reactive conformers A and B. In conformer B, the lone
pair of the oxygen is delocalized into the π-bond. Thus,
this conformer is more abundant than conformer A in
the trans-alkene due to a low energy level and gave the
R-nucleoside as the major product in an EMIC reaction.
In the cis-alkene, the substituent attached to the double
bond interferes sterically with the methyl groups of the
isopropylidene group. Thus, conformer A should be more
abundant than conformer B and result in the â-anomer
as the major product. However, in our case, the nitro-
phenyl substituent of alkene 5 was much bulkier than
the simple methyl group in Freeman’s substrate. Thus,
the predicted conformer A of the cis-alkene may not exist
due to the strong steric hindrance. It was possible that
the cis-alkene 5 isomerized first to the trans-alkene 6.
This isomerization could be catalyzed by the bromonium
ion in the EMIC reaction and provided the conformer B
of the trans-alkene. This conformer would not have the
steric interactions that were associated with the cis-
alkene (A conformer). The trans-alkene would then
proceed to give the R-homo-C-nucleoside 8. This hypoth-
esis was supported by the following facts: (1) both
alkenes 5 and 6 afforded the same product; (2) there was
a dramatic difference between the reaction rates of 5 and
6. If our hypothesis is correct, the trans-alkene should
be detected in the EMIC reaction of the cis-alkene due
to the proposed isomerization. Thus, a small-scale EMIC
reaction of 5 was conducted in deuterated acetonitrile and
the benzyl bromide with an acetate, a soft oxygen
nucleophile, and convert the acetate product to a phe-
none. However, compound 8 also resisted the acetate
displacement under a number of conditions, including
KOAc/18-crown-6, AgOAc, and Hg(OAc)2.26 Therefore, we
elected to displace the bromide of 8 with a hydroxyl group
to give 2,5-anhydro-6-O-(tert-butyldimethylsilyl)-1-(4,5-
dichloro-2-nitrophenyl)-3,4-O-isopropylidene-D-allitol (9)
under strong basic conditions (Scheme 4). A number of
solvents were used, e.g., acetone, THF, MeOH, MeCN,
and 2-propanol. Acetone gave the best results, but the
yield over two steps (EMIC and displacement) was only
30-40%. The low yield was probably due to a possible
elimination on the ribofuranosyl moiety and/or a nucleo-
philic substitution of the dichloronitrobenzene moiety.
Compound 9 consisted of two equal diastereomers be-
cause the nucleophilic substitution of a secondary halide
is not stereospecific. A Swern oxidation of compound 9
afforded the 2-nitrophenone (2S)-2,5-anhydro-6-O-(tert-
butyldimethylsilyl)-1-(4,5-dichloro-2-nitrophenyl)-3,4-O-
isopropylidene-D-allose (10) in a good yield. The nitro
group of compound 10 was then reduced to give (2S)-1-
(2-amino-4,5-dichlorophenyl)-2,5-anhydro-6-O-(tert-bu-
tyldimethylsilyl)-3,4-O-isopropylidene-D-allose (11).
1
followed closely by H NMR. Indeed, the characteristic
1
peaks for the trans-alkene 6 appeared in the H NMR
spectra taken during the reaction process. The amount
of this isomerization product was consistently low, since
it underwent an EMIC reaction much faster than 5.
1
Thus, this H NMR study further supported our isomer-
ization hypothesis.
We also tried other conditions to test the EMIC
reactions of alkene 5. Solvents other than acetonitrile,
e.g., DMF, carbon tetrachloride, and THF, resulted in
either the decomposition of starting material or no
reaction. Electrophiles other than NBS, e.g., bromine,
iodine, and mercury acetate, resulted in either the
decomposition of starting material or no reaction. Al-
though compound 8 possessed the wrong anomeric con-
figuration, its benzylic bromide would be transformed to
a phenone attached to a ribofuranosyl moiety. The
anomeric proton of this phenone R-nucleoside could allow
some anomerization in subsequent reactions. Therefore,
we used compound 8 as the starting material for the
subsequent reactions in our proposed retrosynthetic
route.
Simple quinolin-2-ones have been constructed from
2-aminophenones or 2-acetylaminophenones via an in-
tramolecular aldol condensation.12,13,27 Following the
procedure of Hino et al.,13 compound 11 was first pro-
tected with an acetyl group and then treated with
NaOMe in EtOH. The starting material was consumed
in less than 5 min on the basis of TLC. However, the
major product isolated from the reaction was not a
1
quinolin-2-one nucleoside. The H NMR spectrum of this
compound showed a D2O exchangeable doublet in the
midfield (indicating a secondary hydroxy), a total of five
sugar protons, and no isopropylidene group protons. On
the basis of these data, the compound was tentatively
assigned as a 4,5-dihydrofuran 12 (Scheme 4). It was
likely that a deprotonation at the anomeric position
The retrosynthetic design required an oxidation of
compound 8 directly to a phenone nucleoside. However,
the secondary benzyl bromide seemed to resist a number
of oxidizing conditions, including heated DMSO, bis-
(tetrabutylammonium) dichromate,24 and 4-(dimethyl-
amino)pyridine N-oxide.25 We then attempted to displace
(25) Mukaiyama, S.; Inanaga, J .; Yamaguchi, M. Bull. Chem. Soc.
J pn. 1981, 54, 2221-2222.
(26) Larock, R. J . Org. Chem. 1974, 39, 3721-3727.
(27) J ones, G. Quinolines: Part I; J ones, G., Ed.; J ohn Wiley &
Sons: New York, 1977; pp 93-318.
(24) Landini, D.; Rolla, F. Chem. Ind. 1979, 46, 213-213.
J . Org. Chem, Vol. 68, No. 11, 2003 4173

Chen et al.
SCHEME 5
(TBAF) to afford 6,7-dichloro-4-(2,3-O-isopropylidene-R-
D-ribofuranosyl)quinolin-2-one (14). The values of the
isopropylidene groups of 13 and 14 were both smaller
than 0.15 ppm, which indicated that both nucleosides
were the R-anomers. Compound 14 was treated with
trifluoroacetic anhydride (TFAA) to protect the 5′-hy-
droxy group in situ, and the protected nucleoside was
chlorinated with POCl3/DMF. The crude chlorinated
product was stirred in MeOH overnight to remove the
labile trifluoroacetyl protecting group and resulted in the
formation of 4-(2,3-O-isopropylidene-R-D-ribofuranosyl)-
2,6,7-trichloroquinoline (15), along with a small amount
(2%) of the â-anomer 16. The â-anomer (16) was most
likely obtained from the minor â-anomers of 8-14.
Treatment of compounds 15 and 16 with aqueous TFA
furnished the desired target compounds 4-(R-D-ribofura-
nosyl)-2,6,7-trichloroquinoline and 4-(â-D-ribofuranosyl)-
2,6,7-trichloroquinoline (17 and 18), respectively.
The anomeric configurations of the quinoline C-nucleo-
sides were based on their 1H NMR spectral data. For
example, the ∆δ is equal to 0.00 ppm for the R-anomer
15 and 0.36 ppm for the â-anomer 16. Anomeric assign-
ments based on these observations agreed with the
Imbach rules.29 The H-1’s of 15 and 17 are also at lower
field than those of 16 and 18, respectively. This chemical
shift difference is consistent with the general observa-
tions for â- and R-nucleosides. A nuclear Overhauser
enhancement (NOE) study involving the irradiation of
H-1′ of 16 gave 4.2% of NOE at H-4′. This enhancement
was only possible with a â-nucleoside.
The R-homo-C-nucleoside 11 should anomerize easily
due to the acidic anomeric proton. However, treatment
of 11 with strong bases, e.g., NaOH, K2CO3, i-PrNH2, or
DBU at room temperature, only afforded the elimination
product 12. Treatment of 11 with weak bases, such as
NaHCO3, pyridine, or Et3N, did not effect any conversion
at temperatures up to 60 °C. The treatment of 11 with
LDA only gave unidentified decomposed products. After
a number of trials, we discovered that a small amount
of the â-anomer 19 could be obtained with excess (>20
equiv) DMAP in benzene at reflux temperature. It was
quite surprising that the Rf value between 11 and 19 was
greater than 0.3 on TLC (30% EtOAc/Hex). This dramatic
difference afforded a facile chromatographic separation
of the two anomers after the anomerization. However,
the combined recovery of both anomers under the DMAP
conditions was only 50%. Thus, we reinvestigated other
weak bases, i.e., pyridine and Et3N. Pyridine did provide
almost a 100% recovery; however, the conversion was
only 15% at reflux temperature after 2 days. Pure Et3N
gave mostly the elimination product 12 at reflux tem-
perature, while a 25% conversion and a 96% combined
recovery can be achieved in a 20% Et3N solution in
benzene at reflux temperature for 36 h (Scheme 6). The
elimination product 12 started to accumulate if the
reaction was kept longer and resulted in a lower com-
bined recovery.
induced a fragmentation and this reaction resulted in a
loss of the isopropylidene group. Compound 12 was found
to be highly susceptible for further elimination to a furan
derivative. Mass spectroscopy did not show the expected
peaks for the proposed structure of compound 12. How-
ever, a major peak corresponding to a [M - H2O]+ for a
furan analogue was observed.
Another possible route to construct a quinolin-2-one
was to apply a Wittig reaction. An ylide, such as meth-
oxycarbonylmethylene(triphenyl)phosphorane (MCMTP),
could couple with the ketone functionality of 11 to provide
a two-carbon unit extension. A subsequent amide forma-
tion, under acid-catalyzed conditions, could provide the
desired quinolin-2-one nucleoside. However, several trials
at elevated temperatures did not yield any of the desired
product. It was likely that MCMTP was not a strong
enough ylide to attack a sterically hindered ketone, since
this ylide was deactivated by the electron-withdrawing
carboxylate group. Keteneylidene(triphenyl)phosphorane,
derived from MCMTP,28 has been used in the preparation
of a simple quinolin-2-one from an 2-aminophenylcar-
boxylate.14 We found that when compound 11 was treated
with this ketene ylide in benzene at reflux temperature,
a 50% yield of the desired quinolin-2-one nucleoside, 4-[5-
O-(tert-butyldimethylsilyl)-2,3-O-isopropylidene-R-D-ribo-
furanosyl]-6,7-dichloroquinolin-2-one (13) (Scheme 5),
was obtained. It would appear that the amino group
attacked the ketene first to afford an amide ylide, and
then the resulting ylide underwent an intramolecular
Wittig reaction to furnish the quinolin-2-one. Compound
13 was then treated with tetrabutylammonium fluoride
Treatment of 19 with the ketene ylide afforded 4-[5-
O-(tert-butyldimethylsilyl)-2,3-O-isopropylidene-â-D-ribo-
furanosyl]-6,7-dichloroquinolin-2-one (20) in a 52% yield.
Conversion of 20 to the target compounds was simplified
(28) Bestman, H. J . Angew. Chem., Int. Ed. Engl. 1977, 16, 349-
364.
(29) Chu, C. K.; El-Kabbani, F. M.; Thompson. Nucleosides Nucle-
otides 1984, 3, 1-31.
4174 J . Org. Chem., Vol. 68, No. 11, 2003

Synthesis of Polyhalogenated Quinoline C-Nucleosides
phorus pentoxide), toluene (phosphorus pentoxide), dimethyl-
formamide (calcium oxide), and tetrahydrofuran (sodium/
benzophenone) were distilled from the indicated drying agent
and stored under a positive pressure of argon prior to use (if
not used immediately). The phrases “evaporated in vacuo” and
“concentrated” were meant to imply the use of a rotary
evaporator with a bath temperature not exceeding 40 °C using
a water aspirator. Thin-layer chromatography (TLC) was
carried out on Analtech 60F-254 silica gel plates, and detection
of components on TLC was made by UV light absorption at
254 nm, 365 nm, staining with iodine vapor, or heating to a
char following treatment with 10% sulfuric acid in methanol.
Solvent systems are expressed as a percentage of the more
polar component with respect to total volume (v/v %). Flash
column chromatography refers to the chromatography tech-
nique described by Still (J . Org. Chem. 1978, 43, 2923-2925).
(X% EtOAc/Hex, Y cm × Z cm): the solvent system that was
used in the column, the diameter of the column size, and the
height of silica gel. Mallinckrodt SilicAR 230-400 mesh (40-
63 µm) was used for chromatography. Melting points were
determined and are uncorrected. The 1H (300, 360, or 500
MHz) and 13C (67.5, 90, or 125 MHz) NMR spectra were
recorded with the chemical shift values being expressed in ppm
(parts per million) relative to tetramethylsilane as an internal
SCHEME 6
to a four-step procedure without isolation of any of the
intermediates. Both the TBDMS and the isopropylidene
groups of 20 were first removed under strong acidic
conditions. The resulting nucleoside was then protected
with the trifluoroacetyl groups. The protected nucleoside
was subjected to halogenation reactions, and this was
followed by a removal of the trifluoroacetyl groups to
afford the target compounds 18 and 21 in 64% and 49%
yields, respectively.
In conclusion, we have developed a multistep synthesis
of the targeted 2-halo-6,7-dichloro-4-(â-D-ribofuranosyl)-
quinolines. In the process, we discovered that an elec-
trophile-mediated intramolecular cyclization of an alkene
intermediate afforded an nucleoside with an R-configu-
ration, which was contrary to the literature precedence.
We established that a double-bond isomerization was the
cause of the unexpected stereochemistry. A key Wittig
reaction was used to construct a quinolin-2-one nucleo-
side with a ketene ylide from a 2-aminophenone nucleo-
side, which was susceptible for elimination by conven-
tional construction methods under basic conditions.
1
standard or the standard chemical shift of the solvents for H
NMR, and relative to the standard chemical shift of the solvent
for 13C NMR. The coupling constant values are expressed in
Hz. Mass spectroscopy and elemental analyses were performed
by the University of Michigan Chemistry Department.
4,5-Dich lor o-2-n itr oben zyl Br om id e (2). 4,5-Dichloro-2-
nitrotoluene15 1 (20.2 g, 98.1 mmol), NBS (20.9 g, 117 mmol),
and benzoyl peroxide (1.19 g, 4.9 mmol) were suspended in
120 mL of benzene. The mixture was heated at reflux tem-
perature for 14 h. The resulting orange suspension was cooled
to room temperature, and water (50 mL) was added, followed
by sodium bisulfite (4 g). The resulting mixture was stirred
at room temperature for 2 h to destroy any of the remaining
benzoyl peroxide. The mixture was then extracted with toluene
(3 × 100 mL). The combined toluene extracts were washed
with 200 mL of brine, dried over MgSO4, and concentrated to
dryness. The residue was purified by flash column chroma-
tography (from hexane to 3% EtOAc/Hex, then to 5% EtOAc/
Hex, 5 cm × 20 cm) to give 19.7 g (60%) of 2 as a yellow solid.
Mp: 57-58 °C. Rf (10% EtOAc/Hex) ) 0.45. 1H NMR
(CDCl3): δ 4.78 (s, 2H), 7.70 (s, 1H), 8.20 (s, 1H). 13C NMR
(CDCl3): δ 27.6, 127.8, 133.0, 134.1, 134.2, 138.9, 146.2. Anal.
Calcd for C7H4BrCl2NO2: C, 29.51; H, 1.42; N, 4.92. Found:
C, 29.65; H, 1.59; N, 4.94.
Biologica l Resu lts
Compounds 18 and 21 were evaluated for antiviral
activity against human cytomegalovirus (HCMV) and
herpes simplex virus type 1 (HSV-1) as well as for
cytotoxicity to uninfected cells using techniques we have
described previously.1 Neither compound was active
against HSV-1 but both had modest activity against
HCMV. Compounds 18 and 21 were active in a plaque
reduction assay (IC50 ) 30 µM) and in a yield reduction
assay (IC90’s ) 31 and 37 µM, respectively). However, 18
produced cytotoxicity in stationary human foreskin fi-
(4,5-Dich lor o-2-n it r ob en zyl)t r ip h en ylp h osp h on iu m
Br om id e (3). To a solution of compound 2 (18.2 g, 64.0 mmol)
in 200 mL of toluene was added dropwise a solution of
triphenylphosphine (16.8 g, 64.0 mmol) in 100 mL of toluene
over a period of 30 min at room temperature. The resulting
pale-yellow mixture was stirred at room temperature for 4
days to give a milky suspension. The precipitate was separated
by filtration and dried in vacuo to give 31.0 g (89%) of 3 as a
pale-yellow powder. Mp: 250 °C (dec started from 180 °C). 1H
NMR (CDCl3): δ 6.18 (d, 2H, J ) 15.1), 7.6-7.8 (m, 15H), 8.01
(s, 1H), 8.23 (d, 1H, J ) 2.7). 13C NMR (CDCl3): δ 116.6, 117.3,
124.5, 127.3, 130.7, 130.8, 134.4, 135.7, 135.8, 136.5, 136.6,
139.9, 146.8. Anal. Calcd for C25H19BrCl2NO2P‚0.9toluene
(amount of toluene was verified by 1H NMR): C, 59.70; H, 4.05;
N, 2.22. Found: C, 59.71; H, 4.31; N, 2.25.
broblasts (IC50 ) 32 µM) and in growing KB cells (IC50
)
70 µM) suggesting the apparent antiviral activity could
be a manifestation of cytotoxicity. In contrast, compound
21 did not produce significant cytotoxicity at the highest
concentration tested (100 µM) indicating this compound
selectively inhibited HCMV replication.
(4,5-Dich lor o-2-n itr oben zyl)tr ip h en ylp h osp h or a n e (4).
Compound 3 (29.7 g, 54.2 mmol) was partitioned in 150 mL of
water and 150 mL of CH2Cl2, and NaOH (31 mL, 3.5 N, 108
mmol) was added dropwise at room temperature. The mixture
was stirred vigorously for 1 h and extracted with CH2Cl2 (3 ×
100 mL). The combined CH2Cl2 extracts were washed with 300
mL of brine, dried over MgSO4, and concentrated to dryness
Exp er im en ta l Section
Gen er a l Meth od s. Unless otherwise noted, materials were
obtained from commercial suppliers and were used as pro-
vided. Acetonitrile (calcium hydride), dichloromethane (phos-
J . Org. Chem, Vol. 68, No. 11, 2003 4175

Chen et al.
to give 25.3 g (99%) of 4 as a deep-purple foam. This compound
was used in the next reaction directly without further purifica-
tion.
p r op ylid en e-D-a llitol (8). Meth od 1. To a solution of com-
pound 5 (2.28 g, 4.63 mmol) in 50 mL of CH3CN was added
NBS (0.91 g, 5.11 mmol). The mixture was heated at 50 °C
for 5 h. The resulting yellow solution was cooled and quenched
with 20 mL of saturated NaHCO3. The mixture was concen-
trated to remove most of the organic solvents and then
extracted with EtOAc (3 × 50 mL). The combined EtOAc
extracts were washed with 100 mL of brine, dried over MgSO4,
and concentrated to dryness. The residue was subjected to
flash column chromatography (5% EtOAc/Hex, 4 cm × 10 cm)
to give 2.30 g (87%) of 8 as a yellow oil (contained a small
amount of inseparable impurities).
Meth od 2. The procedure is the same as that described in
method 1, except that 6 (100 mg, 0.40 mmol) was used instead
of 5. This reaction provided compound 8 (without any of the
impurities that were observed in method 1) as a yellow oil in
a 95% yield.
(Z)-6-O-(ter t-Bu t yld im et h ylsilyl)-1-(4,5-d ich lor o-2-n i-
tr op h en yl)-1,2-d id eoxy-3,4-O-isop r op ylid en e-D-a llo-1-en -
itol (5) an d (E)-6-O-(ter t-Bu tyldim eth ylsilyl)-1-(4,5-dich lo-
r o-2-n it r op h en yl)-1,2-d id eoxy-3,4-O-isop r op ylid en e-D-
a llo-1-en itol (6). Compound 4 (25.3 g, 54.2 mmol) and 5-(tert-
butyldimethylsilyl)-2,3-O-isopropylidene-D-ribose11 (6.20 g, 20.4
mmol) were dissolved in 300 mL of CH3CN to provide a deep-
purple solution. The solution was heated at reflux temperature
for 16 h and then concentrated to dryness. The residue was
subjected to a short column (50% EtOAc/Hex, 4 cm × 5 cm) to
remove most of the triphenylphosphine oxide. Fractions con-
taining the alkene products were combined and concentrated
to dryness. The residue was subjected to flash column chro-
matography (from 5% EtOAc/Hex to 10% EtOAc/Hex, then to
15% EtOAc/Hex, 4 cm × 15 cm) to give 8.53 g (85%) of a
mixture of 5 and 6 as a yellow oil in a 12:1 ratio (based on 1H
NMR, peaks used for integration are underlined below). Pure
samples of each compound were obtained by using preparative
TLC.
5. Rf (10% EtOAc/Hex) ) 0.40. 1H NMR (DMSO-d6): δ 0.04
(s, 6H), 0.87 (s, 9H), 1.21 (s, 3H), 1.36 (s, 3H), 3.57 (m, 1H),
3.61 (dd, 1H, J ) 10.5, 5.0), 3.72 (dd, 1H, J ) 10.5, 2.1), 4.05
(dd, 1H, J ) 9.2, 6.0), 4.65 (dd, 1H, J ) 10.0, 5.9), 5.03 (d, 1H,
J ) 5.1, D2O exchangeable), 5.99 (dd, 1H, J ) 11.6, 10.2), 6.78
(d, 1H, J ) 11.7), 8.00 (s, 1H), 8.40 (s, 1H). 13C NMR (DMSO-
d6): δ -5.3, -5.2, 18.2, 25.4, 25.9, 28.1, 65.6, 69.2, 73.2, 76.8,
108.3, 126.0, 126.5, 131.0, 131.1, 131.2, 133.2, 136.2, 146.4.
Anal. Calcd for C21H31Cl2NO6Si: C, 51.22; H, 6.34; N, 2.84.
Found: C, 51.25; H, 6.32; N, 2.96.
6. Rf (10% EtOAc/Hex) ) 0.35. 1H NMR (DMSO-d6): δ 0.02
(s, 3H), 0.03 (s, 3H), 0.86 (s, 9H), 1.30 (s, 3H), 1.42 (s, 3H),
3.48 (m, 1H), 3.57 (m, 1H), 3.70 (m, 1H), 4.12 (m, 1H), 4.82
(pseudo t, 1H), 4.98 (d, 1H, J ) 6.1, D2O exchangeable), 6.65
(dd, 1H, J ) 15.7, 6.4), 6.84 (d, 1H, J ) 15.7), 8.10 (s, 1H),
8.30 (s, 1H). 13C NMR (DMSO-d6): δ -5.3, -5.2, 18.2, 25.3,
25.9, 27.5, 65.6, 69.6, 77.1, 77.3, 108.3, 122.6, 126.2, 129.7,
130.4, 131.6, 135.1, 136.2, 146.2. Anal. Calcd for C21H31Cl2-
NO6Si‚1/2H2O: C, 50.29; H, 6.43; N, 2.79. Found: C, 50.47; H,
6.16; N, 2.73.
(2S)-2,5-An h ydr o-6-O-(ter t-bu tyldim eth ylsilyl)-1-deoxy-
1-(4,5-d ich lor o-2-n itr op h en yl)-3,4-O-isop r op ylid en e-D-a l-
litol (7). Meth od 1. To a solution of compound 5 (140 mg,
0.28 mmol) in 2 mL of MeOH was added aqueous NaOH (0.1
mL, 3.5 N, 0.35 mmol) at room temperature. The mixture was
stirred at 50 °C for 2 h. The resulting yellow solution was
diluted with 5 mL of water and extracted with EtOAc (3 × 5
mL). The combined EtOAc extracts were washed with 10 mL
of brine, dried over MgSO4 and concentrated to dryness. The
residue was subjected to flash column chromatography (20%
EtOAc/Hex, 2 cm × 10 cm) to give 127 mg (98%) of 7 as a
yellow oil.
From Method 2. Rf (10% EtOAc/Hex) ) 0.50. 1H NMR
(DMSO-d6): δ -0.01 (s, 3H), 0.02 (s, 3H), 0.85 (s, 9H), 1.30 (s,
3H), 1.36 (s, 3H), 3.57 (m, 2H), 3.92 (m, 1H), 4.71 (dd, 1H, J )
9.7, 3.8), 4.75 (d, 1H, J ) 6.0), 4.86 (dd, 1H, J ) 5.8, 3.8), 5.55
(d, 1H, J ) 9.7), 8.00 (s, 1H), 8.33 (s, 1H). HRMS for C21H30
-
BrCl2NO6Si [M + H]+: calcd 570.0681, found 570.0480.
(1R,2R)-2,5-An h yd r o-6-O-(ter t-b u t yld im et h ylsilyl)-1-
(4,5-d ich lor o-2-n itr op h en yl)-3,4-O-isop r op ylid en e-D-a lli-
tol (9R) (1S,2R)-2,5-An h ydr o-6-O-(ter t-bu tyldim eth ylsilyl)-
1-(4,5-d ich lor o-2-n it r op h en yl)-3,4-O-isop r op ylid en e-D-
a llitol (9S). To a solution of compound 8 (4.00 g, 7.01 mmol)
in 100 mL of acetone-water (9:1) was added aqueous NaOH
(4.0 mL, 3.5 N, 14 mmol) at room temperature. The resulting
mixture was stirred at room temperature for 16 h. The
resulting yellow solution was diluted with 20 mL of H2O. The
mixture was concentrated to remove most of the organic
solvents and then extracted with EtOAc (3 × 100 mL). The
combined EtOAc extracts were washed with 100 mL of brine,
dried over MgSO4, and concentrated to dryness. The residue
was subjected to flash column chromatography (20% EtOAc/
Hex, 4 cm × 15 cm) to give 1.17 g (33%) of a mixture of 9R
and 9S as a yellow oil (1:1). Rf (20% EtOAc/Hex) ) 0.30. 1H
NMR (DMSO-d6): δ -0.17 (s, 3H), -0.16 (s, 3H), 0.03 (s, 3H),
0.05 (s, 3H), 0.73 (s, 9H), 0.84 (s, 9H), 1.06 (s, 3H), 1.24 (s,
3H), 1.29 (s, 3H), 1.43 (s, 3H), 3.40 (m, 2H), 3.62 (m, 2H), 3.72
(m, 2H), 3.96 (m, 1H), 4.11 (m, 1H), 4.19 (m, 1H), 4.55 (d, 1H,
J ) 6.1), 4.61 (d, 1H, J ) 6.0), 4.74 (m, 1H), 5.25 (m, 1H), 5.57
(m, 1H), 6.06 (d, 1H, J ) 5.2, D2O exchangeable), 6.17 (d, 1H,
J ) 5.2, D2O exchangeable), 7.86 (s, 1H), 7.89 (s, 1H), 8.15 (s,
1H), 8.19 (s, 1H). 13C NMR (DMSO-d6): δ -5.8, -5.7, -5.6,
-5.5, 17.7, 24.5, 25.1, 25.4, 25.6, 25.7, 25.8, 26.4, 63.0, 63.7,
63.8, 66.3, 80.4, 80.5, 82.4, 82.5, 84.0, 84.3, 84.6, 85.8, 111.4,
111.6, 125.9, 126.1, 129.7, 130.2, 130.8, 131.2, 135.5, 135.7,
136.7, 139.2, 147.6, 147.9. HRMS for C21H31Cl2NO7Si [M +
NH4]+: calcd 525.1590, found 525.1566.
(2S)-2,5-An h yd r o-6-O-(ter t-b u t yld im et h ylsilyl)-1-(4,5-
dich lor o-2-n itr oph en yl)-3,4-O-isopr opyliden e-D-allose (10).
To a solution of DMSO (1.16 mL, 16.4 mmol) in 60 mL of CH2-
Cl2 at -78 °C was added dropwise oxalyl chloride (0.72 mL,
8.2 mmol). The mixture was stirred at -78 °C for 15 min, and
then a solution of a mixture of compounds 9R and 9S (1.39 g,
2.73 mmol) in 10 mL of CH2Cl2 was added dropwise. The
resulting mixture was stirred at -78 °C for 40 min, and then
Et3N (3.42 mL, 24.6 mmol) was added. The reaction was
warmed to room temperature and stirred for 40 min. After
being quenched with 30 mL of saturated NH4Cl, the mixture
was extracted with CH2Cl2 (3 × 30 mL). The combined CH2-
Cl2 extracts were washed with 100 mL of brine, dried over
MgSO4, and concentrated to dryness. The residue was sub-
jected to flash column chromatography (10% EtOAc/Hex, 3 cm
× 15 cm) to give 1.10 g (80%) of 10 as a yellow solid. Mp: 100-
Meth od 2. The procedure is the same as that described in
method 1, except that 6 (40 mg, 0.16 mmol) was used instead
of 5. This reaction provided a mixture of compound 7 and its
R-anomer as an oil in a combined 95% yield and 1:1 ratio.
From Method 1. Rf (10% EtOAc/Hex) ) 0.45. 1H NMR
(DMSO-d6): δ 0.03 (s, 3H), 0.04 (s, 3H), 0.85 (s, 9H), 1.25 (s,
3H), 1.41 (s, 3H), 3.19 (m, 1H), 3.28 (m, 1H), 3.60 (m, 2H),
3.88 (m, 1H), 4.04 (m, 1H), 4.46 (m, 1H), 4.54 (m, 1H), 7.84 (s,
1H), 8.27 (s, 1H). 1H NMR (CDCl3): δ 0.09 (s, 3H), 0.10 (s,
3H), 0.91 (s, 9H), 1.35 (s, 3H), 1.53 (s, 3H), 3.08 (m, 1H), 3.34
(m, 1H), 3.72 (m, 2H), 4.07 (m, 1H), 4.15 (m, 1H), 4.39 (m,
1H), 4.68 (m, 1H), 7.60 (s, 1H), 8.04 (s, 1H). 13C NMR
(CDCl3): δ -5.2, -5.1, 18.6, 25.8, 26.2, 27.7, 36.4, 63.8, 82.1,
83.9, 85.0, 85.1, 114.2, 126.7, 131.8, 133.7, 134.6, 137.6, 148.0.
HRMS for C21H31Cl2NO6Si [M + H]+: calcd 492.1376, found
492.1375.
1
111 °C. Rf (20% EtOAc/Hex) ) 0.60. H NMR (DMSO-d6): δ
0.03 (s, 6H), 0.85 (s, 9H), 1.22 (s, 3H), 1.24 (s, 3H), 3.57 (m,
2H), 3.90 (m, 1H), 4.68 (d, 1H, J ) 5.9), 4.94 (d, 1H, J ) 4.3),
5.05 (pseudo t, 1H), 7.61 (s, 1H), 8.49 (s, 1H). 13C NMR (DMSO-
(1S,2S)-2,5-An h yd r o-1-br om o-6-O-(ter t-bu tyld im eth yl-
silyl)-1-d eoxy-1-(4,5-d ich lor o-2-n it r op h en yl)-3,4-O-iso-
4176 J . Org. Chem., Vol. 68, No. 11, 2003

Synthesis of Polyhalogenated Quinoline C-Nucleosides
d6): δ -5.8, -5.6, 17.8, 23.9, 25.4, 25.7, 63.6, 81.6, 82.7, 85.2,
85.8, 112.0, 125.6, 129.9, 133.8, 134.5, 137.8, 145.4, 198.0.
Anal. Calcd for C21H29Cl2NO7Si: C, 49.80; H, 5.77; N, 2.77.
Found: C, 49.65; H, 5.80; N, 2.85.
7.94 (s, 1H), 11.9 (broad s, 1H, D2O exchangeable). 13C NMR
(DMSO-d6): δ -5.5, -5.4, 17.9, 24.7, 25.8, 25.9, 62.6, 78.7, 81.4,
82.8, 83.5, 111.9, 116.9, 117.3, 120.2, 123.9, 125.3, 132.6, 138.3,
145.7, 161.1. HRMS for C23H31Cl2NO5Si [M + H]+: calcd
500.1427, found 500.1447. Anal. Calcd for C23H31Cl2NO5Si: C,
55.20; H, 6.24; N, 2.80. Found: C, 54.91; H, 6.22; N, 2.70.
(2S)-1-(2-Am in o-4,5-d ich lor op h en yl)-2,5-a n h yd r o-6-O-
(ter t-b u t yld im et h ylsilyl)-3,4-O-isop r op ylid en e-D-a llose
(11). To a solution of compound 10 (1.10 g, 2.17 mmol) in 120
mL of MeOH-H2O (9:1) were added Fe (1.2 mg, 21.7 mmol)
and FeSO4 (0.60 g, 2.17 mmol). The mixture was heated at
reflux temperature for 7 h. The resulting suspension was
filtered through Celite. The Celite and the solid residue were
washed with hot MeOH. The combined MeOH solutions were
concentrated to dryness. The resulting yellow foam was
subjected to flash column chromatography (from 20% EtOAc/
Hex to 40% EtOAc/Hex, then to 50% EtOAc/Hex, 3 cm × 10
cm) to give 0.95 g (92%) of 11 as a yellow foam. Mp: 63-66
6,7-Dich lor o-4-(2,3-O-isopr opyliden e-r-D-r ibofu r an osyl)-
qu in olin -2-on e (14). To a solution of compound 13 (160 mg,
0.32 mmol) in 10 mL of THF was added dropwise TBAF (0.38
mL, 0.38 mmol, 1.0 M solution in THF) at room temperature.
The mixture was stirred for 1 h and then concentrated to
dryness. The residue was subjected to flash column chroma-
tography (from 5% MeOH/CH2Cl2 to 10% MeOH/CH2Cl2, 2 cm
× 10 cm) to give 110 mg (89%) of 14 as a yellow solid. Mp:
234-236 °C. Rf (60% EtOAc/Hex) ) 0.15. 1H NMR (DMSO-
d6): δ 1.16 (s, 3H), 1.19 (s, 3H), 3.53 (t, 2H, J ) 5.7), 4.16 (t,
1H, J ) 6.0), 4.81 (d, 1H, J ) 5.9), 4.97 (t, 1H, D2O
exchangeable, J ) 5.5), 5.16 (pseudo t, 1H), 5.40 (d, 1H, J )
3.9), 6.58 (s, 1H), 7.52 (s, 1H), 8.06 (s, 1H), 11.9 (broad s, 1H,
D2O exchangeable). 13C NMR (DMSO-d6): δ 24.7, 25.8, 60.3,
78.3, 81.4, 82.7, 83.7, 111.7, 116.8, 117.4, 120.2, 123.9, 125.7,
132.5, 138.2, 146.1, 161.2. Anal. Calcd for C17H17Cl2NO5‚1/
4H2O: C, 52.26; H, 4.51; N, 3.58. Found: C, 52.02; H, 4.48; N,
3.50.
1
°C. Rf (30% EtOAc/Hex) ) 0.35. H NMR (DMSO-d6): δ 0.06
(s, 3H), 0.07 (s, 3H), 0.89 (s, 9H), 1.18 (s, 3H), 1.23 (s, 3H),
3.68 (m, 2H), 4.15 (m, 1H), 4.73 (d, 1H, J ) 6.1), 5.13 (pseudo
t, 1H), 5.37 (d, 1H, J ) 4.8), 7.03 (s, 1H), 7.36 (broad s, 2H,
D2O exchangeable), 7.95 (s, 1H). 13C NMR (DMSO-d6): δ -5.5,
-5.4, 17.9, 24.7, 25.8, 63.2, 82.4, 82.5, 83.5, 84.0, 112.3, 115.2,
115.3, 117.8, 131.7, 136.4, 150.6, 193.6. Anal. Calcd for C21H31
-
Cl2NO5Si: C, 52.94; H, 6.56; N, 2.94. Found: C, 52.74; H, 6.43;
N, 2.74.
4-(2,3-O-Isopr opyliden e-r-D-r ibofu r an osyl)-2,6,7-tr ich lo-
r oqu in olin e (15) a n d 4-(2,3-O-Isop r op ylid en e-â-D-r ibo-
fu r a n osyl)-2,6,7-tr ich lor oqu in olin e (16). To a solution of
compound 14 (107 mg, 0.277 mmol) in 6 mL of CH2Cl2 was
added TFAA (78 µL, 0.54 mmol). The mixture was stirred at
room temperature for 1 h and then concentrated to dryness.
The residue was kept in vacuo for 1 h and then dissolved in
0.5 mL of DMF. In a separate flask, POCl3 (0.20 mL, 2.2 mmol)
was added to DMF (2 mL) at approximately 10 °C. After the
POCl3/DMF solution was stirred at room temperature for 5
min, the trifluoroacetyl compound in DMF was added at room
temperature. The resulting yellow solution was stirred at 60
°C for 3 h. The mixture was cooled to room temperature and
diluted with 5 mL of EtOAc and 3 mL of cold water. The
resulting mixture was stirred at room temperature for 30 min
and then extracted with EtOAc (3 × 5 mL). The combined
EtOAc extracts were washed with 10 mL of brine, dried over
MgSO4, and concentrated to dryness. The residue was dis-
solved in 20 mL of MeOH and stirred for 24 h. The solvents
were removed in vacuo. The residue was subjected to prepara-
tive TLC (40% EtOAc/Hex, 2 plates) to give 43 mg (41%) of 15
as a white solid and 2 mg (2%) of 16 as a white solid.
15. Mp: 187-188 °C. Rf (40% EtOAc/Hex) ) 0.30. 1H NMR
(DMSO-d6): δ 1.13 (s, 6H), 3.59 (t, 2H, J ) 5.8), 4.26 (t, 1H, J
) 5.9), 4.87 (d, 1H, J ) 5.9), 5.01 (t, 1H, D2O exchangeable, J
) 5.4), 5.25 (pseudo t, 1H), 5.76 (d, 1H, J ) 4.3), 7.57 (s, 1H),
8.30 (s, 1H), 8.51 (s, 1H). 13C NMR (DMSO-d6): δ 24.4, 25.6,
60.4, 78.6, 81.7, 82.9, 84.0, 111.7, 120.8, 123.7, 125.7, 129.7,
130.3, 133.6, 145.9, 148.0, 151.3. Anal. Calcd for C17H16Cl3-
NO4: C, 50.46; H, 3.99; N, 3.46; Found: C, 50.22; H, 3.99; N,
3.36.
1-(2-Acet yla m in o-4,5-d ich lor op h en yl)-2,5-a n h yd r o-6-
O-(ter t-b u t yld im et h ylsilyl)-3-d eoxy-2,3-d id eh yd r o-D-a l-
lose (12). To a solution of compound 11 (40 mg, 0.084 mmol)
in 0.5 mL of pyridine at 0 °C was added dropwise AcCl (30
µL, 0.42 mmol). The resulting suspension was warmed to room
temperature and stirred for 10 min. The resulting mixture was
diluted with H2O (5 mL) and extracted with CH2Cl2 (3 × 5
mL). The combined CH2Cl2 extracts were washed with 10 mL
of brine, dried over MgSO4, and concentrated to dryness. The
residue was subjected to a short column (40% EtOAc/Hex, 1
cm × 4 cm) to give 42 mg of the acetyl intermediate as a yellow
oil. This intermediate was used in the following reaction
directly without further purification. (HRMS of the acetyl
intermediate C23H33Cl2NO6Si [M + H]+ calcd 518.1532, found
518.1547). To a solution of this intermediate (20 mg) in 1.5
mL of EtOH at room temperature was added NaOMe (3 mg).
The resulting yellow solution was stirred for 30 min. The
mixture was then diluted with H2O (5 mL) and extracted with
EtOAc (3 × 5 mL). The combined EtOAc extracts were washed
with 10 mL of brine, dried over MgSO4, and concentrated to
dryness. The residue was subjected to preparative TLC (40%
EtOAc/Hex, 1 plate) to give 10 mg of compound 12 as a yellow
1
oil. Rf (40% EtOAc/Hex) ) 0.30. H NMR (DMSO-d6): δ 0.03
(s, 3H), 0.05 (s, 3H), 0.84 (s, 9H), 2.00 (s, 3H), 3.70 (m, 2H),
4.31 (m, 1H), 4.78 (m, 1H), 5.51 (d, 1H, D2O exchangeable, J
) 6.0), 5.77 (d, 1H, J ) 3.0), 7.74 (s, 1H), 7.89 (s, 1H), 10.2
(broad s, 1H, D2O exchangeable). This compound was found
to be very unstable. Decomposition was observed immediately
after purification based on TLC and NMR. Thus, we were only
1
able to obtain an accurate H NMR spectrum, which indicated
16. Rf (40% EtOAc/Hex) ) 0.35. 1H NMR (DMSO-d6): δ 1.30
(s, 3H), 1.66 (s, 3H), 3.65 (m, 2H), 4.21 (m, 1H), 4.57 (m, 1H),
4.77 (m, 1H), 5.10 (t, 1H, D2O exchangeable, J ) 5.5), 5.52 (d,
1H, J ) 5.4), 7.76 (s, 1H), 8.32 (s, 1H), 8.39 (s, 1H). 13C NMR
(DMSO-d6): δ 25.4, 27.2, 61.2, 81.6, 81.8, 84.6, 85.0, 113.9,
119.3, 123.6, 126.1, 129.6, 130.0, 133.6, 146.1, 149.9, 151.7.
HRMS for C17H16Cl3NO4 [M + H]+: calcd 404.0223, found
404.0205.
the proposed structure. Mass spectroscopy did not show the
expected peaks for the proposed 4,5-dihydrofuran 12. However,
a major peak was [M - H2O]+, which indicated that the
expected decomposition product was a furan.
4-[5-O-(ter t-Bu tyld im eth ylsilyl)-2,3-O-isop r op ylid en e-
r-D-r ibofu r a n osyl]-6,7-d ich lor oqu in olin -2-on e (13). Com-
pound 11 (300 mg, 0.630 mmol) and keteneylidene(triphenyl)-
phosphorane28 (0.90 g, 3.0 mmol) were suspended in 30 mL of
benzene. The mixture was heated at reflux temperature for
14 h. The resulting dark orange solution was directly subjected
to flash column chromatography (from 30% EtOAc/Hex to 60%
EtOAc/Hex, then to 80% EtOAc/Hex, 3 cm × 10 cm) to give
158 mg (50%) of 13 as a yellow solid. Mp: 213-215 °C (EtOAc/
Hex). Rf (40% EtOAc/Hex) ) 0.30. 1H NMR (DMSO-d6): δ 0.07
(s, 3H), 0.08 (s, 3H), 0.90 (s, 9H), 1.17 (s, 3H), 1.20 (s, 3H),
3.76 (m, 2H), 4.21 (pseudo t, 1H), 4.82 (d, 1H, J ) 6.0), 5.14
(pseudo t, 1H), 5.42 (d, 1H, J ) 4.0), 6.57 (s, 1H), 7.52 (s, 1H),
4-(r-D-Ribofu r an osyl)-2,6,7-tr ich lor oqu in olin e (17). Com-
pound 15 (32 mg, 0.08 mmol) was dissolved in 1 mL of TFA/
H2O (9:1). The mixture was stirred at room temperature for
30 min and then concentrated to dryness. The residue was
subjected to preparative TLC (10% MeOH/CH2Cl2, 2 plates)
to give 28 mg (97%) of compound 17 as a white solid. Mp: 245-
1
246 °C. Rf (10% MeOH/CH2Cl2) ) 0.50. H NMR (DMSO-d6):
δ 3.50 (m, 1H), 3.72 (m, 1H), 3.95 (m, 1H), 4.17 (m, 1H), 4.40
(m, 1H),4.80 (m, 2H, D2O exchangeable), 4.96 (d, 1H, D2O
J . Org. Chem, Vol. 68, No. 11, 2003 4177

Chen et al.
exchangeable, J ) 7.3), 5.72 (d, 1H, J ) 3.3), 7.56 (s, 1H), 8.29
(s, 1H), 8.51 (s, 1H). 13C NMR (DMSO-d6): δ 61.4, 72.4, 73.0,
78.5, 82.3, 121.1, 124.3, 125.7, 129.5, 130.0, 133.3, 145.9, 150.3,
151.4. Anal. Calcd for C14H12Cl3NO4: C, 46.12; H, 3.32; N, 3.84.
Found: C, 46.16; H, 3.72; N, 3.49.
4-(â-D-R ib ofu r a n osyl)-2,6,7-t r ich lor oq u in olin e (18).
Meth od 1. Compound 16 (2 mg, 0.005 mmol) was dissolved
in 0.2 mL of TFA/H2O (9:1). The mixture was stirred at room
temperature for 30 min and then concentrated to dryness. The
residue was subjected to preparative TLC (10% MeOH/CH2-
Cl2, 1 plate) to give 2 mg (100%) of compound 18 as a white
solid. This solid has the same 1H NMR spectrum as that
obtained from method 2.
4-[5-O-(ter t-Bu tyld im eth ylsilyl)-2,3-O-isop r op ylid en e-
D-r ibofu r a n osyl]-6,7-d ich lor oqu in olin -2-on e (20). Com-
â-
pound 19 (220 mg, 0.462 mmol) and keteneylidene(triphenyl)-
phosphorane (0.42 g, 1.4 mmol) were suspended in 8 mL of
benzene. The mixture was heated at reflux temperature for 8
h. The resulting dark orange solution was directly subjected
to flash column chromatography (from 30% EtOAc/Hex to 60%
EtOAc/Hex, then to 80% EtOAc/Hex, 1 cm × 10 cm) to give
120 mg (52%) of 20 as a yellow solid. Rf (40% EtOAc/Hex) )
0.40. 1H NMR (DMSO-d6): δ 0.04 (s, 6H), 0.80 (s, 9H), 1.30 (s,
3H), 1.62 (s, 3H), 3.76 (m, 2H), 4.21 (m, 1H), 4.50 (pseudo t,
1H), 4.68 (m, 1H), 5.23 (d, 1H, J ) 5.5), 6.67 (s, 1H), 7.50 (s,
1H), 7.94 (s, 1H), 11.9 (broad s, 1H, D2O exchangeable). 13C
NMR (DMSO-d6): δ -5.6, -5.5, 17.9, 25.4, 25.7, 27.3, 63.2,
81.7, 81.9, 84.1, 84.7, 113.6, 116.8, 117.0, 118.6, 123.6, 126.2,
132.6, 138.6, 148.2, 161.2. HRMS for C23H31Cl2NO5Si [M + H]+:
calcd 500.1427, found 500.1411.
Meth od 2. Compound 20 (150 mg, 0.30 mmol) was dissolved
in 1.0 mL of TFA/H2O (9:1). The mixture was stirred at room
temperature for 1 h and then concentrated to dryness and kept
in vacuo for 1 h to give a yellow solid. This solid was then
suspended in 7 mL of CH2Cl2, and TFAA (0.33 mL, 2.4 mmol)
was added. The mixture was stirred at room temperature for
2 h, concentrated to dryness, and kept in vacuo for 1 h to give
a yellow solid. This solid was dissolved in 4 mL of DMF, and
POCl3 (0.22 mL, 2.4 mmol) was added at 0 °C. The resulting
yellow solution was stirred at 60 °C for 3 h. The mixture was
cooled to room temperature and diluted with 10 mL of EtOAc
and 10 mL of cold water, followed by NaHCO3 (0.5 g). The
resulting mixture was stirred at room temperature for 15 min
and then extracted with EtOAc (3 × 10 mL). The combined
EtOAc extracts were washed with 15 mL of brine, dried over
MgSO4, and concentrated to dryness. The residue was dis-
solved in 5 mL of MeOH and stirred for 24 h. The solvents
were evaporated in vacuo. The residue was subjected to flash
column chromatography (10% MeOH/CH2Cl2, 1 cm × 10 cm)
to give a crude product, which was crystallized from MeOH/
EtOAc to give 42 mg of 18 as a pale yellow solid. The filtrate
was subjected to preparative TLC (10% MeOH/CH2Cl2, 2
plates) to give an additional 28 mg (total 64%) of 18. Mp: 195-
2-Br om o-6,7-d ich lor o-4-(â-D-r ib ofu r a n osyl)q u in olin e
(21). Compound 20 (167 mg, 0.334 mmol) was dissolved in 1.0
mL of TFA/H2O (9:1). The mixture was stirred at room
temperature for 1 h, concentrated to dryness, and kept in vacuo
for 1 h to give a yellow solid. This solid was then suspended
in 7 mL of CH2Cl2, and TFAA (0.37 mL, 2.6 mmol) was added.
The mixture was stirred at room temperature for 2 h. It was
then concentrated to dryness and kept in vacuo for 1 h to give
a yellow solid. The solid was dissolved in 4 mL of DMF, and
POBr3 (0.76 g, 2.6 mmol) was added at 0 °C. The resulting
yellow solution was stirred at 60 °C for 3 h. The mixture was
cooled to room temperature and diluted with 10 mL of EtOAc
and 10 mL of cold water, followed by NaHCO3 (0.5 g). The
resulting mixture was stirred at room temperature for 15 min
and then extracted with EtOAc (3 × 10 mL). The combined
EtOAc extracts were washed with 15 mL of brine, dried over
MgSO4, and concentrated to dryness. The residue was dis-
solved in 5 mL of MeOH and stirred for 24 h. The solvents
were evaporated in vacuo. The residue was subjected to flash
column chromatography (10% MeOH/CH2Cl2, 1 cm × 10 cm)
to give a crude product, which was crystallized from MeOH/
EtOAc to give 38 mg of 21 as a pale-yellow solid. The filtrate
was subjected to preparative TLC (10% MeOH/CH2Cl2, 2
plates) to give an additional 28 mg (total 49%) of 21. Mp: 171-
1
196 °C. Rf (10% MeOH/CH2Cl2) ) 0.50. H NMR (DMSO-d6):
δ 3.61 (m, 1H), 3.72 (m, 1H), 3.84 (m, 1H), 3.94 (m, 2H), 5.01
(t, 1H, D2O exchangeable, J ) 5.5), 5.12 (d, 1H, D2O exchange-
able, J ) 5.1), 5.34 (d, 1H, J ) 5.7), 5.56 (d, 1H, D2O
exchangeable, J ) 6.3), 7.86 (s, 1H), 8.29 (s, 1H), 8.52 (s, 1H).
13C NMR (DMSO-d6): δ 60.9, 70.8, 77.3, 79.7, 84.8, 119.6,
124.1, 126.3, 129.4, 129.9, 133.5, 146.1, 151.4, 151.8. HRMS
for C14H12Cl3NO4 [M]+: calcd 362.9832, found 362.9834. Anal.
Calcd for C14H12Cl3NO4‚1/2H2O: C, 45.00; H, 3.51; N, 3.75.
Found: C, 45.20; H, 3.35; N, 3.75.
1
172 °C. Rf (10% MeOH/CH2Cl2) ) 0.50. H NMR (DMSO-d6):
δ 3.61 (m, 1H), 3.71 (m, 1H), 3.82 (m, 1H), 3.96 (m, 2H), 5.02
(t, 1H, D2O exchangeable, J ) 5.5), 5.12 (d, 1H, D2O exchange-
able, J ) 4.1), 5.29 (d, 1H, J ) 5.5), 5.56 (d, 1H, D2O
exchangeable, J ) 6.3), 7.94 (s, 1H), 8.26 (s, 1H), 8.48 (s, 1H).
13C NMR (DMSO-d6): δ 60.9, 70.8, 77.3, 79.6, 84.7, 123.0,
124.3, 126.4, 129.5, 130.0, 133.5, 143.6, 146.6, 150.6. Anal.
Calcd for C14H12BrCl2NO4: C, 41.11; H, 2.96; N, 3.42. Found:
C, 41.00; H, 3.08; N, 3.34.
(2R)-1-(2-Am in o-4,5-d ich lor op h en yl)-2,5-a n h yd r o-6-O-
(ter t-b u t yld im et h ylsilyl)-3,4-O-isop r op ylid en e-D-a llose
(19). A solution of 11 (866 mg, 1.82 mmol) in 2 mL of Et3N
and 8 mL of benzene was kept at reflux temperature for 36 h,
and the yellow solution was then concentrated to dryness. The
resulting yellow foam was subjected to flash column chroma-
tography (from 15% EtOAc/Hex to 30% EtOAc/Hex, then to
40% EtOAc/Hex, 2 cm × 10 cm) to give 200 mg (23%, 96%
based on recovered starting material) of 19 as a yellow foam
and 660 mg (76%, 99% combined recovery) of 11 as a yellow
foam. Rf (30% EtOAc/Hex) ) 0.65. 1H NMR (DMSO-d6): δ
-0.07 (s, 6H), 0.77 (s, 9H), 1.30 (s, 3H), 1.50 (s, 3H), 3.50 (m,
2H), 4.14 (m, 1H), 4.59 (m, 1H), 5.04 (m, 1H), 5.19 (d, 1H, J )
3.6), 7.03 (s, 1H), 7.41 (broad s, 2H, D2O exchangeable), 8.03
(s, 1H). 13C NMR (DMSO-d6): δ -5.6, -5.5, 18.0, 25.3, 25.7,
27.0, 63.3, 81.5, 81.8, 85.5, 86.1, 112.6, 114.5, 115.0, 117.7,
133.6, 136.8, 151.2, 196.5. HRMS for C21H31Cl2NO5Si [M + H]+
calcd 476.1427, found 476.1411.
Ack n ow led gm en t. We thank Mr. J ack M. Hinkley
for large-scale preparation of starting materials, Roger
G. Ptak and J ulie M. Breitenbach for expert perfor-
mance of the biological assays, and Ms. Kimberly
Barrett for her assistance with the preparation of this
manuscript. This research was supported by NCDDG
research grants U19-AI31718 and 5-PO1-AI46390 from
the National Institute of Allergy and Infectious Disease
of the National Institutes of Health.
J O020643S
4178 J . Org. Chem., Vol. 68, No. 11, 2003