Solid-phase synthesis of carbocyclic nucleosides.

Article abstract of DOI:10.1021/ol005614n

[formula: see text] An efficient solid-phase synthesis of carbocyclic nucleosides has been developed. The key step is the palladium-catalyzed coupling of a purine derivative to a resin-bound allylic benzoate. The resulting products may be further functionalized on the solid phase. Acidic cleavage affords carbocyclic nucleosides, a class of compounds with demonstrated biological activity and substantial current interest.

Full text of DOI:10.1021/ol005614n

ORGANIC
LETTERS
2000
Vol. 2, No. 8
1065-1067
Solid-Phase Synthesis of Carbocyclic
Nucleosides
Michael T. Crimmins* and William J. Zuercher
Venable and Kenan Laboratories of Chemistry, The UniVersity of North Carolina at
Chapel Hill, Chapel Hill, North Carolina 27599-3290
crimmins@email.unc.edu
Received February 2, 2000
ABSTRACT
An efficient solid-phase synthesis of carbocyclic nucleosides has been developed. The key step is the palladium-catalyzed coupling of a
purine derivative to a resin-bound allylic benzoate. The resulting products may be further functionalized on the solid phase. Acidic cleavage
affords carbocyclic nucleosides, a class of compounds with demonstrated biological activity and substantial current interest.
The development of agents that function as nontoxic,
selective inhibitors of kinases and polymerases for the control
of viral diseases has been the focus of intense research.1
However, despite significant progress, the need for new
inhibitors of HIV, herpes simplex virus (HSV), Epstein-
Barr virus (EBV), human cytomegalovirus (CMV), and
hepatitis B virus (HBV) continues, and the ongoing problem
of drug resistance adds to this need.2 Carbocyclic nucleosides
have demonstrated significant potential in this regard as new
antiviral (as well as antitumor) agents.3 For example, carbovir
(1)4 and abacavir (Ziagen) (2)5 have both displayed inhibitory
activity toward HIV.
Despite the success of this approach, there remained several
areas in need of improvement. The yield of the coupling
between the pseudosugar and the purine derivative was
moderate (61-68%) and plagued by contamination with a
minor amount of the N7 isomer (up to 26% of the product).
Functionalization of the purine C6 also proceeded in moder-
ate yield (60-70%) and was limited to volatile amines due
to purification issues.
To expand the applicability of our synthesis of carbocyclic
nucleosides 1 and 2, we recognized that solid-phase organic
synthesis provides certain potential advantages. A large
excess of reagents may be used to drive reactions to
completion, and product purification is facilitated by the
attachment of the substrate to a polymer resin. Additionally,
it was envisioned that by changing the environment around
the pseudosugar, it might be possible to improve the N9:N7
product ratio in the coupling of the base to the sugar.
Accordingly, we set out to develop a solid-phase synthesis
(2) Kavlick, M. F.; Shirasaka, T.; Kojima, E.; Pluda, J. M.; Hiu, F., Jr.;
Yarchoan, R.; Mitsuya, H. AntiViral Res. 1995, 28, 133-146.
(3) Crimmins, M. T. Tetrahedron 1998, 54, 9229-9272.
(4) Vince, R.; Hua, M. J. Med. Chem. 1990, 33, 17-21.
We recently reported an efficient solution-phase enantio-
(5) Abacavir (Ziagen) is the carbocyclic nucleoside formerly known as
1592U89: (a) Daluge, S. M.; Good, S. S.; Faletto, M. B.; Miller, W. H.;
St. Clair, M. H.; Boone, L. R.; Tisdale, M.; Parry, N.; Reardon, J. E.;
Dornsfife, R. E.; Averett, D. R. Krenitsky, T. A. Antimicrob. Agents
Chemother. 1997, 41, 1082-1093. (b) Daluge, S. M. U.S. Patent 5,034,-
394, 1991.
selective synthesis of carbocyclic nucleosides 1 and 2.6
(1) (a) Borthwick, A. D.; Biggadike, K. Tetrahedron 1992, 48, 571-
623. (b) Agrofoglio, L.; Suhas, E.; Farese, A.; Condom, R.; Challand, S.
R.; Earl, R. A.; Guedj, R. Tetrahedron 1994, 50, 10611-10670. (c) Huryn,
D.; Okabe, M. Chem. ReV. 1992, 92, 1745-1768.
(6) Crimmins, M. T.; King, B. W. J. Org. Chem. 1996, 61, 4192-4193.
10.1021/ol005614n CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/17/2000

Scheme 1
of these biologically important structures based on our
acrolein9 formed adduct 9 which was elaborated to cyclo-
pentene 10 by catalytic ring-closing metathesis (RCM).
Reductive removal of the chiral auxiliary with LiBH4 and
MeOH formed diol 11. Selective silylation of the primary
alcohol in 11 with TBSCl and Et3N was followed by
acylation of the secondary alcohol in 12 with Bz2O, Et3N,
and 4-DMAP to provide 13. Desilylation of 13 with 5% HF
in CH3CN afforded 3 in 85% yield over the final three
steps.Allylic benzoate 3 was loaded onto p-nitrophenyl Wang
carbonate resin10 in the presence of EtN-i-Pr2 and 4-DMAP
in CH2Cl2 at 40 °C overnight to produce resin 4. Quantitation
of the loading of 3 onto the resin was measured with
attenuated total reflection (ATR) IR spectroscopy, and
complete incorporation of the allylic benzoate was realized
with 2 equiv of 3. Excess 3 can be recycled by chromato-
graphic purification of the reaction filtrate.
successful solution-phase strategy. The approach (Scheme
1) involves palladium-catalyzed coupling of the resin-bound
carbocyclic pseudosugar 4 with a purine derivative to form
nucleoside resin 5; this is followed by further functional-
ization of the purine, affording resin 6 which, upon cleavage
from the resin, yields purine-substituted carbocyclic nucleo-
sides 7.
Allylic benzoate 37 served as the pseudosugar portion of
the carbocyclic nucleoside (Scheme 2). Diol 11, which is
Scheme 2
As in the solution-phase synthesis of carbocyclic nucleo-
sides, palladium catalysis was utilized in the attachment of
a chloropurine to the pseudosugar moiety.11 Several reaction
variables were investigated (Table 1), including palladium
Table 1. Palladium-Catalyzed Purine Coupling Reaction
entry
Y
Pd source
loading, mol %
base
convn, %
1
2
3
4
5
6
7
8
Cl
15
15
10%
100%
50
50
50%
31
16
10
31
31
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
EtN-i-Pr2
Et3N
<5
30
70
60
>95
80
60
45
55
65
85
85
>95
>95
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
16/P(O-i-Pr)3
16/PPh3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/P(O-i-Pr)3
14/PPh3
9
10
11
12
13
14
NMM
31
31
10
12
Pempidine
EtN-i-Pr2
Pempidine
Pempidine
readily accessible in racemic8 or enantiomerically pure6 form,
served as a key intermediate in the synthesis of 3. Asym-
metric aldol reaction of 4-pentenoyloxazolidinethione 8 and
14/PPh3
NH2 14/PPh3
(7) All new compounds were fully characterized spectroscopically (1H
and 13C NMR, IR, and HRMS).
(8) MacKeith, R. A.; McCague, R.; Olivo, H. F.; Palmer, C. F.; Roberts,
S. M. J. Chem. Soc., Perkin Trans. 1 1993, 313-314.
source, catalyst loading, ligand, and base. Conversion of the
reaction was measured by ATR IR.12 The solid-phase
1066
Org. Lett., Vol. 2, No. 8, 2000

coupling reaction was found to require a much higher catalyst
loading than the analogous solution-phase reaction (entry 1),
and optimized conditions for the coupling utilized 10 mol
% of Pd2(dba)3 (14), 0.4 equiv of PPh3, 6.0 equiv of
pempidine (1,2,2,6,6-pentamethylpiperdine), and 1.5 equiv
of 2,6-dichloropurine in 1:1 THF:DMSO at 45 °C for 16 h
(entry 13). Both Pd(PPh3)4 (15) and allylpalladium chloride
dimer (16) were much less effective in this coupling, and
both led to lower conversion at identical catalyst loading.
The more electron rich PPh3 effected a higher conversion
than did P(O-i-Pr)3. Bulkier tertiary amine bases led to the
highest conversion (entries 6 and 9-11), a trend observed
previously by Trost and co-workers.13 Both 2,6-dichloropu-
rine (entry 13) and 2-amino-6-chloropurine (entry 14) were
effective nucleophiles in the coupling reaction to produce
chloropurine resin 5a (Y ) Cl) and 5b (Y ) NH2),
respectively.
Table 2. Results of Purine Functionalization with Nitrogen
Nucleophiles
No product containing an isomeric N7 purine linkage was
1
observed by H NMR of the resin 5a (Y ) Cl) or 5b (Y )
NH2) or in the products after cleavage with 5% TFA in CH2-
Cl2, and this constitutes an advantage over the solution-phase
synthesis where N7 contamination can be significant. The
isomeric N7 product results from a more hindered trajectory
of the nucleophile, and it is presumed that the resin amplifies
this effect; Trost has noted the steric protection of the metal
center and a resulting steering of the incoming nucleophile
in similar solid-phase reactions.11
Functionalization of the purine system at the 6-position
with nitrogen-based nucleophiles was accomplished via a
thermal SNAr reaction to produce resins of type 6.14 Standard
conditions were 5 equiv each of the amine and EtN-i-Pr2 in
BuOH at 80 °C for 4 h. A diverse range of functionality
was incorporated, including primary and secondary amines,
anilines, hydrazides, and alkoxyamines. ATR IR spectra
indicated complete conversion from 5 to the derivatized resin
6.
Cleavage of the product from the resin was accomplished
by exposure to 5% TFA in CH2Cl2 for 1.5 h. The initially
cleaved product was impure by TLC and by 1H NMR, which
showed signals for both the 5′-alcohol and its trifluoroacetate
ester. Exposure of the mixture to a primary or secondary
amine cleaved the TFA ester, and purification of the resulting
alcohol by silica gel chromatography provided the pure
purine-substituted carbocyclic nucleosides. Yields of the
isolated, chromatographically purified products15 were in the
range 60-82% (Table 2).
In conclusion, we have presented an efficient method for
the solid-phase synthesis of carbocyclic nucleosides. This
synthesis is amenable to combinatorial library synthesis, with
the potential for diversity in the pseudosugar, the aromatic
base, and the substitution of the base. Such investigations
will be reported in due course.
Acknowledgment. The generous support of Glaxo-
Wellcome is acknowledged. We thank Susan Daluge, Sam
Gerritz, Doug Minick, Wendy White, and Andrea Sefler for
helpful discussions and analytical support. W.J.Z. is grateful
to the NIH for a Postdoctoral Fellowship.
Supporting Information Available: Complete spectral
data (1H and 13C NMR, IR, and HRMS) for compounds 3
and 7a-7f. This material is available free of charge via the
Internet at http://pubs.acs.org.
(9) Crimmins, M. T.; King, B. W.; Tabet, E. A. J. Am. Chem. Soc. 1997,
119, 7883-7884.
(10) Dixit, D. M.; Leznoff, C. C. J. Chem. Soc., Chem. Commun. 1977,
798-799.
(11) Although Pd(0) is commonly employed in the deprotection of resin
bound allylic esters (see: Guibe, F. Tetrahedron 1998, 54, 2967-3042 and
selected references therein), reactions involving a resin-bound π-allylpal-
ladium intermediate are rare. One example by Trost involves a supported
Pd(0) catalyst: Trost, B. M.; Keinan, E. J. Am. Chem. Soc. 1978, 100,
7779-7781.
(12) Conversion was determined by measuring the amount of unreacted
benzoate (1716 cm-1) remaining. The dialkyl carbonate (1745 cm-1) was
used as an internal standard.
OL005614N
(15) Cleaved products 7a-7f are carbocyclic analogues of 2-chloro-6-
substituted adenosine analogues, structures which have demonstrated
biological activity. For example, NNC 53-0017 is a potent inhibitor of TNFR
production but does not interfere with the A3 receptor. Bowler, A. N.; Raven,
A.; Olsen, U. B.; Thomsen, C.; Knutsen, L. J. Drug. DeV. Res. 1998, 43,
26.
(13) Trost, B. M.; Madsen, R.; Guile, S. G.; Elia, A. E. H. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 1569-1572.
(14) Nugiel, D. A.; Cornelius, L. A. M.; Corbett, J. W. J. Org. Chem.
1997, 62, 201-203.
Org. Lett., Vol. 2, No. 8, 2000
1067