Enantioselective syntheses of carbocyclic ribavirin and its analogs: Linear versus convergent approaches

Article abstract of DOI:10.1016/S0040-4039(00)01704-4

The first enantioselective syntheses of carbocyclic ribavirin by both convergent and linear approaches are described. The linear approach from chiral nonracemic 2-azabicyclo[2.2.1]hept-5-en-3-one proves to be a highly efficient route to carbocyclic analog

Full text of DOI:10.1016/S0040-4039(00)01704-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 9575–9579
Enantioselective syntheses of carbocyclic ribavirin and its
analogs: linear versus convergent approaches†
Rongze Kuang*,a Ashit K. Ganguly*,b Tze-Ming Chan,a Birendra N. Pramanik,a
David J. Blythin,a Andrew T. McPhailc and Anil. K. Saksenaa,*
aSchering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
bDepartment of Chemistry, Stevens Institute of Technology, Castle Point on Hudson, Hoboken, NJ 07030, USA
cPaul M. Gross Chemical Laboratories, Duke University, Durham, NC 27706, USA
Received 29 August 2000; accepted 26 September 2000
Abstract
The first enantioselective syntheses of carbocyclic ribavirin by both convergent and linear approaches
are described. The linear approach from chiral nonracemic 2-azabicyclo[2.2.1]hept-5-en-3-one proves to be
a highly efficient route to carbocyclic analogs of ribavirin. © 2000 Elsevier Science Ltd. All rights
reserved.
Keywords: nucleoside; antiviral; carbocyclic ribavirin; enantioselection; azide; triazole; olefin metathesis.
Ribavirin (1) is a broad-spectrum antiviral agent, and the combination of ribavirin and
interferon-a has been approved by the FDA as a new therapy for hepatitis C.1 Structural
modification of ribavirin represents a promising approach in the search for new antiviral agents.
One strategy is to replace the furanose oxygen in ribavirin with a methylene group, which gives
the carbocyclic version of ribavirin. Such a modification is of particular interest, since the
resulting carbocyclic ribavirin (2) may possess greater metabolic stability to the phosphorylase
enzymes which cleave the N-glycosidic linkage in normal nucleosides.2 This communication
describes our effort to develop an efficient enantioselective synthesis of chiral carbocyclic
ribavirin analogs.
* Corresponding authors. Fax: 908-740-7152; e-mail: anil.saksena@spcorp.com
†
Dedicated to Professor Harry H. Wasserman on the happy occasion of his 80th birthday. It is a privilege to
recognize his brilliant contributions to organic chemistry.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01704-4

9576
Strategically, carbocyclic ribavirin can be synthesized by both convergent and linear
approaches.3,4 In seeking convergent routes to 2 and its analogs, the key cyclopentyl alcohol 3
was readily synthesized in six steps from ribonolactone.5 However, attempted nucleophilic
displacement of the triflate derived from 3 with the Na-salt of ethyl 1,2,4-triazole-3-carboxylate
provided to our surprise 5 as a major product resulting from the alkylation of the more hindered
nitrogen. The desired product 4 was estimated to be less than 5% by NMR of the crude reaction
mixture. The structure of the major product was confirmed by single-crystal X-ray crystallo-
graphy of the amide 5.13 Deprotection then provided 7, the 5-carboxamido isomer of 2.
Attempted direct coupling of 3 with 1H-1,2,4-triazol-3-carboxylate under Mitsunobu conditions
also failed to provide the desired product 4. (Scheme 1).
Scheme 1. (i) (CF3SO2)2O; (ii) 1,2,4-triazol-carboxylic acid ethyl ester, NaH, DMF, −10ꢀ0°C; (iii) NH3/MeOH,
100%; (iv) CF3CO2H/H2O, 87%
An alternative, convergent strategy required regioselective ring-opening of the chiral epoxide
12. Since 12 was available from earlier syntheses only in the racemic form,6 a new enantioselec-
tive synthesis of this epoxide was developed based on an anti-aldol condensation7 and ring-clos-
ing olefin metathesis. In a similar approach Crimmins et al. have used syn-aldol adducts for the
construction of other carbocyclic structures.8 An efficient synthesis of epoxide 12 was achieved
in four steps beginning with chiral oxazolidinone 8. While the anti-aldol condensation gave only
a moderate yield, all of the other steps proceeded in high yields to provide the non-racemic
epoxy diol 11.13 After protecting the hydroxyl groups, the resulting epoxide 12 was reacted with
the triazole in the presence of NaH to give the desired ring-opening product 13 as the major
isomer in 40% yield. Further treatment with ammonia followed by hydrogenolytic deprotection
provided carbocyclic ribavirin 2 in 14% overall yield in nine steps. This epoxide route represents
a highly concise approach to carbocyclic ribavirin 2, the major deficiency being the yield in the
crucial triazole displacement on 12. (Scheme 2).
Our linear approach to carbocyclic ribavirin 2 focuses on the application of the commercially
available chiral nonracemic building block, 2-azabicyclo[2.2.1]hept-5-en-3-one 14 (Scheme 3).
This bicyclic lactam has been used in a number of syntheses of carbocyclic nucleosides, most of
which involve the construction of heterocycles from cyclopentylamine intermediates.3 However,
transformations of the lactam 14 into a cyclopentylhydrazine or cyclopentylazide, possible
precursors for 2, have not been reported.
The starting 2-azabicyclo[2.2.1]hept-5-en-3-one 14 was first converted into the protected
cyclopentylamine 15 according to a published method.9 A high-yielding transformation of 15
into the cyclopentylhydrazine 17 was accomplished by modification of a two-step protocol
developed by Vilarrasa.10,11 The t-Boc protected amine 16 was first reacted with nitrosonium

9577
Scheme 2. (i) 2 equiv. N-Bu2BOTf, iPr2NEt, acrolein, −78°C, 53%; (ii) Grubb’s catalyst, 96%; (iii) LiBH4, MeOH,
THF, 0°C, 83%; (iv) mCPBA, 97%; (v) PhCH2Br, NaH, THF, 74%; (vi) 1,2,4-triazol-3-carboxylic acid ethyl ester,
NaH, DMF, 110°C, 40%; (vii) NH3/MeOH, 100%; (viii) H2, 10% Pd/C
Scheme 3. (i) Four steps, Ref. 9; (ii) Ac2O, TEA, DMAP, CH3CN, rt, 1.5 h,100%; (iii) NOBF4, pyridine, CH3CN,
−30°C, 0.5 h, then 0°C, 2 h, 100%; (iv) Zn, HOAc, 8–10°C, 20 min, 95%; (v) TFA/H2O, 1 h; (vi) ethyl
carboethoxyformimidate, NaHCO3, MeOH, 0°C, 0.5 h; (vii) excess HC(OEt)3, toluene, 110°C, 4 h, 82% from 17; (viii)
5% HCl/THF(1:1), 50°C, overnight, 93%; (ix) saturated NH3/MeOH (−15°C), sealed flask, rt, overnight, 100%
tetrafluoroborate in the presence of pyridine in acetonitrile, and the resulting N-nitroso
compound was then reduced to 17 with zinc powder in glacial acetic acid. After the TFA
deprotection, the hydrazine TFA salt was coupled with ethyl carboethoxyformimidate,12 and the
intermediate 18 was heated with triethyl orthoformate in toluene at 110°C for 5 h to give the
carbocyclic triazole intermediate 19 with concomitant protection of the 1,2-diol functionality as
a cyclic orthoformate. Heating this intermediate in 5% HCl/THF solution provided 20 which
upon further treatment with ammonia/methanol gave the carbocyclic ribavirin 2 in 12 steps
from the bicyclic lactam 14 in 46% overall yield.

9578
In a further extension of the above linear approach, a high yielding synthesis of the
cyclopentylazide 22 was developed by heating the N-Boc,N%-acetyl cyclopentylhydrazine (21)
with isoamyl nitrite in CCl4.10 This efficient approach to the cyclopentylazide now allows us to
access structurally diverse analogs of carbocyclic ribavirin through 1,3-dipolar cycloaddition
reactions. For example, reaction of 22 with methyl propiolate provided 23, the 1,2,3-triazolo
analog of 2, in excellent yield (Scheme 4).
Scheme 4.
Compared to the convergent approaches examined in this study, the linear approach is
preferred. It is noteworthy that chromatographic purification is necessary only for two interme-
diates (19 and 20), and that the steps (ii)–(v) can be performed on the same day.
In summary, we have developed a highly efficient enantioselective synthesis of carbocyclic
ribavirin 2. The facile syntheses of both the cyclopentylhydrazine and cyclopentylazide interme-
diates 17 and 22 from the bicyclic lactam 14 attest to the versatility of this readily available
chiral building block in the synthesis of carbocyclic nucleosides and other biologically useful
compounds. We shall report on the biological activity of these compounds in a future
communication.
Spectroscopic data for key compounds
Compound 2: identical NMR and HRMS data were obtained for 2 from both Schemes 2 and
3. Compound 2 is highly hygroscopic. It became a viscous oil upon exposure to the atmosphere.
1H NMR (CD3OD): 1.91 (ddd, J=8.2, 9.9, 13.2 Hz, 1H), 2.20 (m, 1H), 2.40 (dt, J=8.6, 13.2
Hz, 1H), 3.63 (d, J=6.0 Hz, 2H), 3.98 (dd, J=3.1, 5.3 Hz, 1H), 4.21 (dd, J=5.3, 8.4 Hz, 1H),
4.77 (dt, J=8.5, 9.5 Hz, 1H), 8.53 (s, 1H); HRMS [M+1]+ found 243.1096, calcd for C9H15N4O4
243.1093. [h]2D0=−60.2 (c 0.5, MeOH).
1
Compound 7: H NMR (CD3OD): 1.78 (ddd, J=8.6, 9.5, 13.0 Hz, 1H), 2.23 (m, 1H), 2.39 (dt,
J=8.5, 13.0 Hz, 1H), 3.63 (m, 2H), 4.04 (dd, J=4.0, 5.4 Hz, 1H), 4.38 (dd, J=5.6, 7.8 Hz, 1H),
4.89 (s, br, 5H), 5.81 (dt, J=8.4, 9.4 Hz, 1H), 7.98 (s, 1H); HRMS [M+1]+ found 243.1098, calcd
for C9H15N4O4 243.1093.
Compound 22: 1H NMR (CDCl3): 1.30 (s, 3H), 1.46 (s, 3H), 1.64 (dt, J=4.3, 13.9 Hz, 1H), 2.08
(s, 3H), 2.30 (ddd, J= 6.0, 7.8, 14.0 Hz, 1H), 2.43 (m, 1H), 3.98–4.14 (m, 3H), 4.48 (m, 2H).
FTIR: 2104.74 (ꢀN3) cm−1.
1
Compound 23: H NMR (CDCl3): 1.30 (s, 3H), 1.54 (s, 3H), 2.07 (s, 3H), 2.32 (dt, J=11.0, 12.6
Hz, 1H), 2.55 (m, 1H) 2.63 (dt, J=7.7, 12.7 Hz, 1H), 3.93 (s, 3H), 4.18 (d, J=5.9 Hz, 2H), 4.59
(dd, J=4.6, 7.0 Hz, 1H), 4.77 (m, 1H), 4.85 (dd, J=5.4, 7.0 Hz, 1H), 8.19 (s, 1H); HRMS
[M+1]+ found 340.1516, calcd for C15H21N3O6 340.1509. [h]2D0=−50.7 (c 1.1, CH2Cl2).

9579
Acknowledgements
We thank Drs. Jinping McCormick, Frank Bennett, and V. M. Girijavallabhan for helpful
discussions and Dr. Xian Liang for his help in providing a key intermediate.
References
1. Glue, P. Semin. Liver Dis. 1999, 19 (Suppl. 1), 17–24; Battaglia, A. M.; Hagmeyer, K. O. Ann. Pharmacother.
2000, 34(4), 487–94.
2. Noble, S. A.; Beddall, N. E.; Beveridge, A. J.; Marr, C. L. P.; Mo, C. L.; Myers, P. L.; Penn, C. R.; Storer, R.;
Woods, J. M. Nucleosides & Nucleotides 1991, 10, 478–490. Racemic carbocyclic ribavirin and its activity as a
weak inhibitor of IMPDH (inosine monophosphate dehydrogenase) was reported in this publication. However,
no synthetic details were provided.
3. Recent review on carbocyclic nucleosides: Crimmins, M. T. Tetrahedron 1998, 54, 9229–9272.
4. Kapeller, H.; Marschner, C.; Weißenbacher, M.; Griengl, H. Tetrahedron 1998, 54, 1439–1456.
5. Wolfe, M. S.; Borcherding, D. R.; Borchardt, R. T. Tetrahedron Lett. 1989, 30, 1453.
6. Hutchison, A.; Grim, M.; Chen, J. J. Heterocyclic Chem. 1989, 26, 451–452; Palmer, C. F.; Parry, K. P.; Roberts,
S. M. Tetrahedron Lett. 1990, 31, 278–282.
7. Raimundo, B. C.; Heathcock, C. H. Synlett. 1995, 1213–1214; Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am.
Chem. Soc. 1981, 103, 2127–2129.
8. Crimmins, M. T.; King, B. W. J. Org. Chem. 1996, 61, 4192–4193; for a review on ring-closing olefin metathesis
reactions (RCM), see: Grubbs, R. H.; Chang, S. Tetrahedron 1998, 371.
9. Hutchinson, E. J.; Taylor, B. F.; Blackburn, G. M. Chem. Commun. 1996, 2765–2766.
10. Garcia, J.; Vilarrasa, J. Tetrahedron Lett. 1987, 28, 341–342.
11. Romea, P.; Urpi, F.; Vilarrasa, J. J. Org. Chem. 1989, 54, 3209–3211.
12. McKillop, A.; Chattopadhyay, S. K.; Henderson, A.; Avendano, C. Synthesis 1997, (3), 301–304.
13. X-Ray crystallographic data for compounds 6 and 11 have been submitted for deposition at the Cambridge
Crystallographic Data Centre.
.