Practical enantioselective synthesis of lamivudine (3TC) via a dynamic kinetic resolution

Article abstract of DOI:10.1016/j.tetlet.2005.10.002

A practical enantioselective synthesis of lamivudine 1 is described. A highly effective dynamic kinetic resolution of the 5-hydroxyoxathiolane, followed by chlorination and introduction of cytosine provides an efficient synthesis of lamivudine.

Full text of DOI:10.1016/j.tetlet.2005.10.002

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 8535–8538
TM
Practical enantioselective synthesis of lamivudine (3TC )
via a dynamic kinetic resolution
Michael D. Goodyear, Malcolm L. Hill, Jono P. West and Andrew J. Whitehead*
Synthetic Chemistry, GlaxoSmithKline Research and Development Limited, Gunnels Wood Rd, Stevenage,
Hertfordshire SG1 2NY, UK
Received 25 July 2005; revised 30 September 2005; accepted 4 October 2005
Available online 19 October 2005
Abstract—A practical enantioselective synthesis of lamivudine 1 is described. A highly effective dynamic kinetic resolution of the
5-hydroxyoxathiolane, followed by chlorination and introduction of cytosine provides an efficient synthesis of lamivudine.
Ó 2005 Elsevier Ltd. All rights reserved.
Lamivudine 1 also known as 3TCTM is an anti-HIV
agent, which has also been developed for the treatment
of hepatitis B.1
required an efficient route, which was suitable for
large-scale synthesis to support the development of this
compound.
In the first instance a route was developed, which uti-
O
lised a Vorbruggen reaction3 of the enantiomerically
N
¨
NH2
O
N
pure 5-acetoxy-oxathiolane 7 with pre-silylated cytosine
8 as the key convergent step as disclosed in Scheme 1.
This approach required a significant quantity of trimeth-
HO
S
1
ylsilyl iodide (TMSI) as the Lewis acid for the Vorbrug-
¨
gen reaction in order to achieve an acceptable yield of
the desired cytodine 9. Furthermore the approach
proved to be rather inefficient as the preparation of
Although there are several different methods reported2
for the synthesis of 1 as a single enantiomer, we
S
OH
Ph
O
O
O
HO2C
OH
(iv)
(ii)-(iii)
-O2C
OAc
HO
HO2C
OAc
(i)
S
+
HO
+
S
S
S
HO2CCHO
NH3+
(v)
3
4
2
5
(viii)-(ix)
NHTMS
NH2
N
N
N
TMSO
N
Yield = 16 %
O
O
O
O
8
O
(xi)
OAc
HO2C
(x)
(vi)-(vii)
O
1
OAc
O
S
O
S
7
S
6
9
Scheme 1. Reagents and conditions: (i) t-BuOMe, reflux, 70%; (ii) Ac2O, cat MeSO3H; (iii) NaHCO3, i-PrOAc, 70%; (iv) norephedrine, i-PrOAc,
35%; (v) 5 M HCl; (vi) (COCl)2, DMF; (vii) (À)-menthol, py; 80%; (viii) (COCl)2, DMF, CH2Cl2; (ix) (À)-menthol, py, 2,2,4-trimethylpentane, 16%;
(x) TMSI, CH2Cl2, 60%; (xi) NaBH4, EtOH, 83%.
Keywords: Practical; Enantioselective; Dynamic kinetic resolution; Lamivudine.
*
Corresponding author. Tel.: +44 01438 763738; fax: +44 01438 764869; e-mail: andrew.j.whitehead@gsk.com
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.10.002

8536
M. D. Goodyear et al. / Tetrahedron Letters 46 (2005) 8535–8538
O+
O
O
Mn
O+
O
O
O
O
SOCl2
OH
O
O
Cl
O
S
O
S
S
anchimeric
stabilisation
DMF, DCM
S
12
13
11
12
-
I
NHTMS
O
O
O
NH2
N
TMSI
OAc
O
I
O
N
O
O
O
S
TMSO
N
S
O
8
N
7
10
NHTMS
O
S
NH2
N
9
N
60%
O
O
TMSO
N
O
8
N
Scheme 3.
O
S
9
cytosine 8 to give the cytodine 9 as a 10:1 mixture of b
and a anomers in 70% yield (Scheme 2).
Scheme 2.
It is postulated that 5-a-iodooxathiolane 10 is formed
stereoselectively by reaction of the oxonium ion 11, gen-
erated by reaction of 7 with TMSI, undergoing stabilisa-
tion through anchimeric assistance from the C-2 ester
substituent. Iodide then attacks the stabilised oxonium
ion 12 to afford the a anomer 10, which then reacts with
pre-silylated cytosine 8 in an SN2 type nucleophilic sub-
stitution reaction to give predominantly the b-cytodine
adduct 9 (Scheme 2).4
the enantiomerically pure crystalline 5-acetoxy-oxathiol-
ane 7 was achieved either by its selective crystallisation
in the presence of the other three diastereoisomers,
albeit in only 16% yield, or by a classical resolution
via the norephedrine salt of acid 5. Therefore an alterna-
tive, more efficient route was sought.
Closer inspection of the reaction mechanism of the
Vorbruggen reaction revealed that when 5-acetoxyoxa-
¨
thiolane 7 was reacted with TMSI 5-a-iodoacetoxyoxa-
thiolane 10 was formed, which reacted with pre-silylated
We were interested in examining other leaving groups
than iodide in this substitution reaction. Consequently
O
O
OH
O
S
35
O
OH
OH
O
O
O
(iii) n-hexane
(i) toluene
O
O
OH
OH
14
O
O
S
OH
(ii)
S
S
12
35
HO
S
34% yield
2
O
O
OH
O
S
15
O
O
OH
O
S
15
Scheme 4.

M. D. Goodyear et al. / Tetrahedron Letters 46 (2005) 8535–8538
8537
O
O
O
O
OH
crystalline
OH
O
O
S
S
12
DCM, CHCl3
OH
O
O
O
Et3N
Et3N
OH
O
O
O
O
O
O
O
S
S
HS
17
18
16
DCM, CHCl3
O
O
O
OH
O
OH
O
O
S
S
15
Scheme 5.
5-a-chloro-oxathiolane 13 was prepared from 5-hydr-
oxy-oxathiolane 12 on reaction with thionyl chloride
and catalytic DMF in dichloromethane. This was also
found to react with the pre-silylated cytosine 8 to give
9 in good yield and selectivity (b:a, 10:1) (Scheme 3).
The desired b-anomer 9 could then be efficiently isolated
by crystallisation on addition of n-hexane to the crude
reaction mixture.
All that remained now was to identify an improved
synthesis of hydroxyoxathiolane 12. Initially, the
5-hydroxyoxathiolane 12 was prepared by heating a
mixture of menthyl glyoxylate hydrate 145 with dithiane
diol 2 in toluene. This procedure gave a mixture of four
diastereoisomers as a 35:35:15:15 ratio as depicted in
Scheme 4.
Crystallisation from n-hexane gave the desired diaste-
reoisomer 12 in 34% yield, which corresponds to
approximately all of that isomer in solution. To increase
the yield of the desired diastereoisomer 12 the intercon-
version of the other diastereoisomers was investigated.
O
O
O
OH
OH
(i)-(iii)
OH
O
O
S
14
It was initially found that in certain solvents, for exam-
ple, dichloromethane or chloroform, rapid equilibration
at C-5 occurred giving a mixture of b and a isomers
15 and 12, but there was no isomerisation at C-2. We
12
Scheme 6. Reagents and conditions: (i) toluene; (ii) dithiane diol; (iii)
n-hexane, Et3N, 80% overall yield.
O
O
O
(i)-(iii)
OH
(vi)
O
O
Cl
O
OH
O
O
OH
S
S
14
12
13
NH2
NH2
N
N
O
O
O
O
N
(v), (vi)
(vii)
N
O
O
HO
S
S
9
1
Scheme 7. Reagents and conditions: (i) toluene, reflux; (ii) dithiane diol; (iii) n-hexane, Et3N crystallisation (80% for three stages); (iv) SOCl2, DMF,
DCM; (v) 8, Et3N, toluene; (vi) Et3N, water, n-hexane (66% for three stages); (vii) NaBH4, EtOH, 83%.

8538
M. D. Goodyear et al. / Tetrahedron Letters 46 (2005) 8535–8538
Y.-C.; Chu, C. K. J. Org. Chem. 1992, 57, 2217–2219; (b)
Jeong, L. S.; Schinazi, R. F.; Beach, J. W.; Kim, H. O.;
Nampalli, S.; Shanmuganathan, K.; Alves, A. J.; McMil-
lan, A.; Chu, C. K. J. Med. Chem. 1993, 36, 181–185;
(c) Coates, J. A.; Cammack, N.; Jenkinson, H. J.; Mutton,
I. M.; Pearson, B. A.; Storer, R.; Cameron, J. M.; Penn, C.
R. Antimicrob. Agent Chemother. 1992, 36, 733–739; (d)
Chang, C. N.; Doong, S. L.; Zhou, J. H.; Beach, J. W.;
Jeong, L. S.; Chu, C. K.; Tasi, C. H.; Cheng, Y. C. J. Biol.
Chem. 1992, 267, 13938–13942; (e) Doong, S. L.; Tasi, C.
H.; Schinazi, R. F.; Liotta, D. C.; Cheng, Y. C. Proc. Natl.
Acad. Sci. U.S.A. 1991, 88, 8495–8499.
postulated that this equilibration could be further dis-
torted via cleavage of the hemithioacetal 16 to give men-
thyl glyoxalate 17 and mercaptoacetaldehyde 18.
Coupling together these two equilibria should then pro-
vide a method for converting the unwanted diastereoiso-
mers to the desired one, which would be removed from
the equilibrium by crystallisation (Scheme 5).
Hence, a reagent capable of effecting the equilibration at
C-2, but not detrimental to the crystallisation was
needed. It seemed likely that a base would carry out this
function. A number of bases were evaluated; pyridine
gave only a small amount of interconversion whereas
triethylamine caused rapid interconversion. Further-
more it was necessary to add only a catalytic amount
of triethylamine to achieve rapid interconversion and
crystallisation of 12 in 80% yield as depicted in Scheme
6.
2. (a) Li, J.; Lain-xun, G.; Meng-xian, D. Synth. Commun.
2002, 32, 2355–2359; (b) Cousins, R. P.; Mahmoudian, M.;
Youds, P. M. Tetrahedron: Asymmetry 1995, 6, 393–396; (c)
Milton, J.; Brand, S.; Jones, M. F.; Raynor, C. M.
Tetrahedron: Asymmetry 1995, 6, 1903–1906; (d) Humber,
D. C.; Jones, M. F.; Payne, J. J.; Ramsay, M. V.; Zacharie,
B. Tetrahedron Lett. 1992, 33, 4625–4628; (e) Jin, H.;
Siddiqui, A.; Evans, C. A.; Tse, A.; Mansour, T. S.;
Goodyear, M. D.; Ravenscroft, P.; Beels, C. D. J. Org.
Chem. 1995, 60, 2621–2623; (f) Milton, J.; Brand, S.; Jones,
M. F.; Rayner, C. M. Tetrahedron Lett. 1995, 36, 6961–
6964; (g) Mahmoudian, M.; Baines, B. S.; Drake, C. S.;
Hale, R. S.; Jones, P.; Piercy, J. E.; Montgomery, D. S.;
Purvis, I. J.; Storer, R. Enzyme Microb. Technol. 1993, 15,
749–755; (h) Storer, R.; Clemens, I. R.; Lamont, B.; Noble,
S. A.; Williamson, C.; Belleau, B. Nucleosides Nucleotides
1993, 12, 225–236.
This highly effective crystallisation-induced dynamic
kinetic resolution then provided an efficient synthesis
of the 5-hydroxy-oxathiolane 12, which in turn was used
to prepare lamivudine 1 in an efficient process (Scheme
7).6
In conclusion, an efficient and enantioselective, synthesis
of lamivudine has been developed, which utilises a
highly effective dynamic kinetic resolution as the key step.
3. Vorbruggen, H.; Ruh-Pohlenz, C. Org. React. 2000, 55, 1–
¨
630.
4. Choi, W. B.; Wilson, L. J.; Yeola, S.; Liotta, D. C. J. Am.
Chem. Soc. 1991, 113, 9377–9379.
5. Fernandez, F.; Garcia, G.; Rodriguez, J. E. Synth. Com-
mun. 1990, 20, 2837–2847.
References and notes
6. Goodyear, M. D.; Dwyer, O.; Hill, M. L.; Whitehead, A. J.;
Hornby, R.; Hallet, P. PCT Int. Appl. WO 9529174 A1
19951102; Chem. Abstr. 1995, 124, 146759.
1. (a) Beach, J. W.; Jeong, L. S.; Alves, A. J.; Pohl, D.; Kim,
H. O.; Chang, C. N.; Doong, S. L.; Schinazi, R. F.; Cheng,