Carbocyclic Isoadenosine Analogues of Neplanocin A

Article abstract of DOI:10.1021/ol035696q

(Equation presented) Isoadenosine (IsoA), a structural isomer of adenosine, was shown to possess interesting biological activity but was inherently unstable. In an effort to overcome this, we have designed a series of carbocyclic IsoA analogues, combining

Full text of DOI:10.1021/ol035696q

ORGANIC
LETTERS
2003
Vol. 5, No. 23
4401-4403
Carbocyclic Isoadenosine Analogues of
Neplanocin A
,†
Katherine L. Seley,* Sylvester L. Mosley,† and Fanxing Zeng
School of Chemistry and Biochemistry, Georgia Institute of Technology,
Atlanta, Georgia
kseley@umbc.edu
Received September 4, 2003
ABSTRACT
Isoadenosine (IsoA), a structural isomer of adenosine, was shown to possess interesting biological activity but was inherently unstable. In an
effort to overcome this, we have designed a series of carbocyclic IsoA analogues, combining the unique connectivity of IsoA with the structural
features of some biologically significant Neplanocin A analogues. Their design, synthesis, and structural elucidation is reported.
Isoadenosine (IsoA) (Figure 1) is a structural isomer of
adenosine where the purine base differs in its connectivity,
being coupled to the ribofuranose moiety at the N-3 rather
than the conventional N-9.
is the cross conjugation of the heterocyclic moiety and facile
migration of the glycosidic C-1/N-3 bond, which occurs in
an effort to restore the aromaticity of the purine ring, thereby
resulting in adenosine (Figure 2).
Figure 2.
Figure 1.
A second problem is that the glycosidic bond of IsoA is
highly susceptible to cleavage by both acids and bases,
thereby making it difficult to work with synthetically. In
contrast, the carbocyclic ring system is quite stable; replace-
ment of the furanose oxygen by a methylene group trans-
forms the unstable hemiaminal glycosidic bond to a more
stable tertiary amine. This isosteric change not only imparts
increased stability but also results in enhanced chemo-
IsoA initially showed promising chemotherapeutic proper-
ties; however, interest soon waned due to issues with
stability.1 IsoA is unstable for several reasons. One problem
† Present address: Department of Chemistry and Biochemistry, University
of Maryland, Baltimore County, Baltimore, MD.
(1) Leonard, N. J.; Laursen, R. A. J. Am. Chem. Soc. 1963, 85, 2026.
10.1021/ol035696q CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/18/2003

therapeutic effects for many modified nucleosides.2,3 As a
class, the carbocyclic nucleosides have exhibited potent
inhibitory activity against many biologically significant
enzymes, in particular, S-adenosyl-L-homocysteine hydrolase
(SAHase) and DNA methyltransferase (MeTase).4-6
Scheme 1 a
Two of the most well-known carbocyclic nucleosides are
Aristeromycin (Ari)7 and Neplanocin A (NpcA)8 (Figure 3).
a Reaction conditions: (a) NaBH4, CeCl3.7H2O, MeOH, 0 °C, 1
h. (b) for 5, (i) p-NO2BzOH, DIAD, PPh3, dry THF, rt, 2 days; (ii)
KOH, MeOH, H2O, 2 h; (iii) PPh3, NBS, DMF, 0 °C to rt, 3 h; for
6, MsCl, pyridine 0 °C, for 7, TsCl, pyridine, rt, 15 h. (c) 5, 6, or
7, adenine, dry N,N-dimethylacetamide, reflux, 2 days. (d) 2:1 TFA/
H2O.
(-)-3 to allylic alcohol 4 was carried out using Luche13
conditions employing NaBH4 and CeCl3‚7H2O (90%). Next,
conversion of the allylic alcohol of 4 into 5 was undertaken.
The hydroxyl group must first undergo inversion using
standard Mitsunobu conditions in order to ultimately achieve
the correct conformation for the bromine substituent. Once
inverted, standard hydrolysis of the ester, followed by
bromination using an SN2 displacement gave 5 in moderate
yield (60%). Subsequently, refluxing 5 and adenine in N,N-
dimethylacetamide in the absence of base afforded the
coupled product in 55% yield (under normal coupling
conditions employing a base, the major product is coupled
at N-9 instead of at N-3).14
In an attempt to increase the overall yield in addition to
alleviating the number of steps required when using the
bromide intermediate, we then tried converting the hydroxyl
group into the corresponding mesylate and tosylate. Treat-
ment of allylic alcohol 4 with methanesulfonyl chloride at 0
°C in pyridine provided 6 in 80% yield, while tosylation of
allylic alcohol 4 with p-toluenesulfonyl chloride in pyridine
at room temperature afforded tosylate 7 in an 82% yield.
The desired IsoA analogue was then realized by the
deprotection of the isopropylidene group using a 2:1 mixture
of TFA/H2O to give 1 (99%) in a 31% overall yield from
(-)-3. Enantiomer 2 was synthesized in an analogous manner
using the identical series of reactions used for 1, however,
starting with (+)-3 (43% in six steps). In both cases, the
N-9 product was formed in around 15%, as well as traces of
the N-7 product, but both were easily isolated during
purification by column chromatography. Unfortunately,
despite the shorter synthetic routes to the mesylate and
tosylate intermediates, a better overall yield for the coupling
was obtained using the bromide intermediate.
Figure 3.
Both have exhibited potent biological activity; however, due
to their close structural resemblance to adenosine, both are
readily phosphorylated to the corresponding mono-, di-, and
triphosphate forms, and this results in significant toxicity.
As a result, many laboratories, including ours, have pursued
additional structural modifications in an effort to overcome
this problem.
One of the most important of these modifications was the
elimination of the 4′-hydroxymethyl group of Ari and NpcA,
which provided the 4′,5′-enyl and tetrahydro analogues
(Figure 3). The 4′,5′-enyl analogues in particular have proven
to be very effective inhibitors of SAHase and are much less
toxic than their parent analogues.9,10
Using the lead provided by IsoA, combined with the potent
biological activity of the 4′,5′-enyl NpcA analogues, we have
designed the corresponding 4′,5′-enyl carbocyclic analogues
of IsoA.
The synthesis of 1 and 2 was first envisioned via
displacement of the C-1 bromide by adenine. To achieve
the requisite cyclopentenyl intermediate, our synthetic efforts
begin with the enantiospecific synthesis of cyclopentenones
(-)-3 and (+)-3 using literature procedures.11,12 As shown
in Scheme 1 on the next page, stereospecific reduction of
(2) Crimmins, M. T. Tetrahedron 1998, 54, 9229.
(3) Marquez, V. E. In AdVances in AntiViral Drug Design; De Clercq,
E., Ed.; JAI Press: Greenwich, CT, 1996; Vol. 2, pp 89-146.
(4) De Clercq, E. Biochem. Pharm. 1987, 36, 2567.
(5) Wolfe, M. S.; Borchardt, R. T. J. Med. Chem. 1991, 34, 1521.
(6) Liu, S.; Wolfe, M. S.; Borchardt, R. T. AntiVir. Res. 1992, 19, 247.
(7) Kishi, T.; Muroi, M.; Kusaka, T.; Nishikawa, M.; Kamiya, K.;
Mizuno, K. Chem. Pharm. Bull. 1972, 20, 940.
(8) Yaginuma, S.; Muto, N.; Tsujino, M.; Sudate, Y.; Hayashi, M.; Otani,
M. J. Antibiot. 1981, 34, 359.
(9) Wolfe, M. S.; Lee, Y.; Bartlett, W. J.; Borcherding, D. R.; Borchardt,
R. T. J. Med. Chem. 1992, 35, 1782.
Structural elucidation of 1 and 2 was then undertaken.
While the most direct way to accomplish this would be to
obtain a crystal structure, unfortunately both 1 and 2 proved
(10) Borcherding, D. R.; Scholtz, S. A.; Borchardt, R. T. J. Org. Chem.
1987, 52, 5457.
(11) Siddiqi, S. M.; Schneller, S. W.; Ikeda, S.; Snoeck, R.; Andrei, G.;
Balzarini, J.; De Clercq, E. Nucleosides Nucleotides 1993, 12, 185.
(12) Cyclopentenones (-)-3 and (+)-3 were synthesized with greater
than 94% ee as confirmed by optical rotation.
(13) Luche, J. L.; Rodriguez-Hahn, L.; Crabbe, P. J. Chem. Soc., Chem.
Commun. 1978, 601.
(14) Fujii, T.; Walker, Graham, C.; Leonard, N. J.; DeLong, D. C.;
Gerzon, K. J. Med. Chem. 1979, 22, 125.
4402
Org. Lett., Vol. 5, No. 23, 2003

to be difficult to crystallize, as they were both fine powders.
Assessment of the UV spectra for 1 and 2, however, provided
initial evidence that the coupling had occurred at N-3. Our
observed value of 274 nm fell perfectly in line with the
reported 277 nm for isoadenosine1 and 270 nm for the 2′,3′-
dideoxyisoadenosine analogue,15 as opposed to adenosine
analogues, which fall around 260 nm; as a result, we felt
reasonably confident in our structure and turned to NMR
for final verification.
which strongly suggested coupling at N-3 versus N-9;
however, the 13C spectra was inconclusive. Using two-
dimensional NMR techniques, however, the correct assign-
ments for 1 and 2 could be made. The HMQC spectra, which
correlates the 1H to the 13C spectrum, showed that the signal
for C-8 was shifted further downfield than that for C-2,
suggesting once again that the coupling had occurred at N-3.
This was confirmed by HMBC; if coupled at N-3, the signal
for H-2 must exhibit long-range coupling to C-1′, C-4, and
C-6, and H-8 would be coupled to C-4 and C-5. This was
indeed the case, thereby establishing that 1 and 2 were indeed
coupled at N-3.
In conclusion, we have synthesized and successfully
characterized the first two analogues in a series of carbocyclic
N-3-coupled IsoA nucleosides that we hope will prove to
be biologically interesting. Details of the biological inves-
tigation underway will be provided elsewhere soon.
All proton and carbon peaks could easily be identified
through comparison of various literature values of other
purine shifts, with the exception of H-2, H-8, C-2, and C-8.10
It has been noted in the literature that coupling at either
the N-9 or N-3 position results in a greater downfield shift
1
in the H NMR spectra for H-2 than is seen for H-8.16 In
cases involving N-3 coupled products, however, the differ-
ence in ppm between the two peaks is much larger than is
seen for N-9 coupled products, which often overlap. The
corresponding 13C NMR spectra for N-9- and N-3-coupled
nucleosides exhibit a different trend; C-2 for N-9-coupled
products is shifted further downfield than C-8, but in the
case of N-3 coupling, the order is reversed; the signal for
the C-8 is shifted further downfield than that for C-2.
The 1H NMR for both 1 and 2 showed a marked difference
of 0.20 ppm in the chemical shifts between H-2 and H-8,
Acknowledgment. The authors would like to thank Mr.
Joshua Sadler for his assistance with starting materials. In
addition, we appreciate the efforts of Dr. Leslie Gelbaum
with the two-dimensional NMR. Project support was pro-
vided by the National Institutes of Health, National Cancer
Institute RO1-CA97634.
Supporting Information Available: Experimental details
and two-dimensional NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
(15) Nair, V.; Buenger, G. S.; Leonard, N. J.; Balzarini, J.; DeClercq,
E. J. Chem. Soc., Chem. Commun. 1991, 1650.
(16) Iyer, R. S.; Voehler, M. W.; Harris, T. M. J. Am. Chem. Soc. 1994,
116, 8863.
OL035696Q
Org. Lett., Vol. 5, No. 23, 2003
4403