Approaches to installing a N-gem-dimethylmethylene-2-oxazolyl group and application to the synthesis of a second generation HIV protease inhibitor

Article abstract of DOI:10.1016/j.tetlet.2005.01.103

Two syntheses of the title compound 1 were developed based on different approaches for installing the oxazole ring moiety. Formation and dehydration of ketoamide was initially used and scaled up to afford 1 on several kilogram scale, then oxazolyl anion/i

Full text of DOI:10.1016/j.tetlet.2005.01.103

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 1867–1871
Approaches to installing a
N-gem-dimethylmethylene-2-oxazolyl group and application to
the synthesis of a second generation HIV protease inhibitor
a,
a,
a
a
Norihiro Ikemoto,a,* Ross A. Miller, Fred J. Fleitz, Jinchu Liu, Daniel E. Petrillo,
Joseph F. Leone,b Joseph Laquidara,b Benjamin Marcune,a Sandor Karady,a
Joseph D. Armstrong, IIIa and Ralph P. Volantea
*
*
aDepartment of Process Research, Merck Research Laboratories, Rahway, NJ 07065, USA
bDepartment of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 5 November 2004; revised 19 January 2005; accepted 20 January 2005
Abstract—Two syntheses of the title compound 1 were developed based on different approaches for installing the oxazole ring moi-
ety. Formation and dehydration of ketoamide was initially used and scaled up to afford 1 on several kilogram scale, then oxazolyl
anion/iminium coupling reaction was developed for a more convergent approach.
Ó 2005 Elsevier Ltd. All rights reserved.
The HIV protease inhibitors, such as indinavir sulfate,
N
O
have offered valuable therapies in controlling the pro-
gression of AIDS in HIV infected patients.1–3 However,
due to the fast emergence of resistant viral strains of
HIV4,5 and the rapid metabolism and clearance of drug
from the body, more effective second generation analogs
were sought with improved potency against resistant
strains and improved pharmacokinetic profiles. Com-
pound 1 showed superior properties versus indinavir
and was chosen as a developmental candidate.6–9 A scale-
able synthesis of 1 was thus needed to prepare several
kilograms of bulk drug to support a highly accelerated
program. For greatest convergency, the disconnection
in Scheme 1 was adopted involving coupling of a previ-
ously reported epoxide 310 and the oxazole ring contain-
ing biarylpiperazine 2. The synthesis of the latter posed
a challenge due to sterically demanding linkage of the
piperazine and oxazole moieties via a gem-dimethyl-
methylene group.11 Ultimately, two disconnections of
2 were realized (at bonds A and B), which are summa-
rized in this report.
OH
N
O
OH
MeO
H
N
N
O
N
O
NH
HCl
CF3
1
N
N
O
MeO
O
O
NH
N
B
A
N
O
O
NH
CF3
O
3
2
Scheme 1. Retrosyntheses of 1.
Disconnection A started with conversion of 3,5-dibro-
mopyridine 4 to 3-bromo-5-methoxypyridine 5 at
90–100 °C in DMF with 1.2 equiv of NaOCH3. The
reaction mixture after aqueous workup was filtered
through a pad of silica gel, and the solution in MTBE
was solvent switched to THF and treated with 2M
i-PrMgCl/THF to generate Grignard 6, which was
converted to Boc-aminoketone 8 via reaction with
Keywords: Oxazole; Piperazine; Epoxide opening; Protease inhibitor.
*
Corresponding authors. Tel.: +1 7325945132; fax: +1 7325945170;
e-mail: fred_fleitz@merck.com
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.01.103

1868
N. Ikemoto et al. / Tetrahedron Letters 46 (2005) 1867–1871
Boc-glycine Weinrebamide reagent 7 in 75% overall iso-
lated yield. Pre-treatment of 7 with i-PrMgCl to deprot-
onate the acidic NH group12 allowed for the use of only
1 equiv of Grignard 6 versus 2 equiv without pre-treat-
ment for this coupling. The Boc-aminoketone 8 was
crystallized after workup and isolated in high purity.
A total of 9 kg of 8 was prepared using this sequence
(Scheme 2).
although P2O5/H2SO4 did produce some oxazole in
lower yields. A significant racemization problem was ob-
served, however, with 30% oleum, affording oxazole 2
with typically <90% ee at full conversion. Halting the
reaction at lower conversions gave less racemization
but at the expense of lower yields.
It was observed that the product 2 from the P2O5/H2SO4
reaction condition was obtained with no loss in optical
purity (ꢀ98.7% ee), although the yield was only 50–
60% due to formation of impurities. Addition of P2O5
to the 30% oleum reaction mixture did not suppress
racemization and 2 was isolated with only 87% ee at
81% yield. Suspecting that more water or a different,
more hydrated form of P2O5 was required in the reac-
tion mixture to suppress racemization, polyphosphoric
acid (PPA)/30% oleum combination was tried instead,
and happily 2 was obtained with 96% ee at >85% yield.
Scale-up under these conditions (2.5 vol of 30% oleum,
3 equiv of PPA, 6–7 h at 40–45 °C), in two batches gen-
erated 5 kg of oxazole 2 in 84% yield with 95% ee purity.
(Caution: Oleum is highly corrosive. Also, the aqueous
quench is highly exothermic and needs to be performed
slowly and with adequate cooling.)
This ketone was used to prepare the ketoamide 13 pre-
cursor to the oxazole 2. The previously reported chiral
Boc-piperazine amide 9 (Ref. 10) was converted to the
N1-Alloc derivative 10, and the gem-dimethylmethylene
moiety was installed via a silver mediated13 N-alkylation
with bromo-gem-dimethylacetic acid.9 Ketone 8 was
deprotected with aqueous HCl to afford aminoketone
salt 12 and was coupled with acid 11 via EDC/HOBT
activation. Due to the large steric bulk of the gem-di-
methyl group, the coupling to 13 was slow and incom-
plete using the standard catalytic HOBT procedure.
However, by using stoichiometric HOBT to first form
the activated acid/HOBT adduct (rt, 1 h) and then add-
ing the aminoketone salt 12 and HunigÕs base, good con-
¨
version and yield of 13 were achieved. The ketoamide 13
was isolated as the bis-sulfate salt in 84% overall yield
from BOC-aminoketone 8 (98.5% ee).
An upgrade in optical purity for 2 was, however, still re-
quired, and various salt forms of 2 were screened for
crystallization since 2 as the free base could not be crys-
tallized. A crystalline form of tris-2-naphthalenesulfonic
acid (NSA)14 salt 14 was identified that was capable of
upgrading 2 to 99% ee from 87% ee via crystallization
from acetonitrile/water. By forming this salt, 4.4 kg of
2 was purified to 99% ee from the initial 95% ee. A salt
break with KOH/methanol and removal of K-NSA salt
then afforded the free base 2 with 97% recovery without
loss of optical purity (Scheme 3).
Cyclization and deprotection of the ketoamide 13 to
afford oxazole 2 was best accomplished using a strongly
dehydrating reagent, neat 30% fuming sulfuric acid
(oleum) at 40 °C. Milder reagents (e.g., POCl3, POCl3/
PCl5, TFA/TFAA, PCl5/MSA, POCl3/MSA, POCl3/
H2SO4, or PCl5/H2SO4) did not effect this cyclization,
Br
Br
NaOMe/MeOH MeO
Br
iPrMgCl
DMF, 90-100°C
(86%)
With pure biarylpiperazine 2 in hand, coupling with
epoxide 3 was investigated. Among a wide range of sol-
vents (alcohols, amides, DMSO, HOAc, water, etc.,
including mixtures) the optimized yield of 80% was
achieved by heating in tert-amyl alcohol at 55 °C15 for
4 days. Variation of reactant stoichiometries and con-
centrations had a negligible effect on the yield. Addition
of a wide variety of Lewis acids16 also failed to improve
the yield (and in most cases was detrimental).
N
N
4
5
O
O
O
NHBOC
N
MeO
MgX
MeO
NHBOC
7
N
N
(87%)
8
6
Scheme 2. Synthesis of Boc-aminoketone 8.
BOC
HO
1) allylchloroformate
N
HN
O
HO
O
N
NaHCO3, toluene, H2O
Br
NH
N
MeO
NH2
HCl
N
O
O
alloc
NH
CF3
alloc
NH
CF3
+
2) HCl, toluene, 40 °C
Na2CO3, THF
(87%)
.
Ag2O, Et3N
(62%)
NH
N
O
O
CF3
12
9
10
11
O
N
1) SO3/H2SO4
H
N
MeO
N
SO3H
polyphosphoric acid
MeO
N
1) EDC, HOBT
O
.
NH
NH
3
N
O
SO3H
N
alloc
NH
CF3
2)
N
2) H2SO4, MTBE/IPAC
O
2 H2SO4
O
(84% from 3)
CF3
(82%)
14
13
Scheme 3. Synthesis of oxazole intermediate 14.

N. Ikemoto et al. / Tetrahedron Letters 46 (2005) 1867–1871
1869
While the above coupling gave an acceptable yield it
also generated several low level impurities, one of which
was difficult to reject via crystallization. A highly pro-
ductive HPLC separation17 was employed to remove
this impurity and afford 5.6 kg of purified product 15
in 80% overall yield from 2. Subsequent removal of
the acetonide was best performed with gaseous HCl in
methanol (TFA/THF, HCl/IPA, MeOH/concd HCl
were all notably slower). Isolation of 1 by crystallization
as an HCl salt (isopropylacetate, IPA, concd HCl) pro-
vided the desired product with high purity in 79% yield
from 15. Therefore, to support this highly accelerated
program, the development and execution of this scale-
able synthesis was achieved in six months to prepare sev-
eral kilograms of the candidate HIV protease inhibitor 1
(Scheme 4).
iminium 20 afforded the protected biarylpiperazine 21
in good yield. Deprotection of the acetonide group
and formation of the tris-salt 14 could then be achieved
in one pot via heating in NSA/acetonitrile in 90% yield.
This process, which was demonstrated successfully on
100 g scale, has the advantage of greater convergency
and efficiency and is likely the more scaleable and prac-
tical route.
The development of a more scaleable synthesis of 1 also
needed to address the requirement for chromatographic
purification. The problematic impurity was identified as
an oxidative byproduct17 and therefore the addition of
antioxidants/reductants to the coupling reaction was
screened. BHT, BHA, various silanes, sodium formate,
sodium thiosulfate, potassium sulfite, and potassium
thiosulfate were all ineffective. In contrast addition of
0.20 equiv of sodium metabisulfite almost completely
suppressed the formation of the oxidative byproduct.
Epoxide opening under these conditions, followed by
the previously described deprotection and HCl salt form-
ation, gave a 64% overall yield of 1 from 2 in the re-
quired purity without the need for chromatographic
purification.
However, for a longer term synthesis of 1, a more effi-
cient synthesis of the biarylpiperazine 2 was desired that
did not rely on a stoichiometric silver catalyzed N-alkyl-
ation and dehydration of ketoamide 13 using a very cor-
rosive reagent. Thus, in a parallel study with the
development of the above process, an alternative route
to 2 was investigated and was ultimately successful
(Scheme 5).
Thus, the synthesis of a second generation HIV protease
inhibitor was achieved employing two different
approaches to install the challenging N-gem-dimethyl-
methylene-2-oxazolyl group. The initial approach
involving formation and dehydration of the ketoamide
enabled a rapid synthesis of several kilograms of inhi-
bitor 1 to support the rapid-paced program. The utility
of polyphosphoric acid to suppress racemization during
the oleum-mediated dehydration was demonstrated and
two novel and possibly general purification protocols of
2-naphthalenesulfonic acid14 were devised. As a success-
ful strategy to rapid drug development for application to
longer term synthesis, a more convergent disconnection
was pursued in parallel and resulted in a highly efficient
process involving oxazole anion addition to the pipera-
zine iminium salt.
In this approach, the Grignard 6 was prepared and
quenched with DMF to afford the aldehyde 16, which
was then converted to the oxazole 17 with tosylmethyl
isocyanide reagent (TosMIC) in 82% overall yield.18
Then, the acetonide of piperazine amide 19 was pre-
pared from piperazine amide salt 18 (a precursor to 9)
(Ref. 9) and converted to the acetone iminium salt 20
by heating in dimethoxypropane in 90% overall yield.
Replacement with anions other than triflate gave non-
crystalline salts of 19. Formation of the C2 anion of
oxazole 17 with i-PrMgCl followed by the addition of
N
N
O
OH
O
Characterization data—8: mp 112 °C. 1H NMR
(400 MHz, CDCl3): d 1.46 (s, 9H), 3.90 (s, 3H), 4.65
(d, 4.6 Hz, 2H), 5.50 (br s, 1H), 7.68 (dd, 2.9, 1.7 Hz,
1H), 8.50 (d, 2.9 Hz, 1H), 8.75 (d, 1.0 Hz, 1H). 13C
NMR (100 MHz, CDCl3): d 28.2, 47.8, 55.7, 80.0,
117.5, 130.5, 141.2, 143.5, 155.6, 155.9, 193.7.
MeO
1. HCl/ MeOH
2. HCl/ IPA-IPAC
(77%)
O
N
N
1
N
O
O
NH
CF3
15
Scheme 4. Penultimate and conversion to product 1.
TosMIC
NaOMe
N
H
O
iPrMgCl
MeO
CHO
DMF
MeO
5
N
(91%)
N
(90%)
N
N
-
MeO
O
N
16
17
iPrMgCl
N
N
(83-97%)
O
O
MeO OMe
HN
O
+
CF3
21
14
TfOH
HN
K2CO3
N
1.
NH
NH
N
N
TfO
N
2 CSA
2. TfOH
(98%)
NSA
Acetonitrile
(90%)
MeO
OMe
N
O
O
CF3
CF3
CF3
(92%)
18
19
20
Scheme 5. Alternative route to oxazole intermediate 14.

1870
N. Ikemoto et al. / Tetrahedron Letters 46 (2005) 1867–1871
25
D
13 (free base): ½aꢁ ꢂ44 (c 0.41, MeOH); 1H NMR
References and notes
(400 MHz, DMSO): d 8.76 (s, 1H), 8.65 (t, 1H), 8.53
(d, 2.3 Hz, 1H), 7.79 (s, 1H), 7.61 (t, 1H), 5.91 (m,
1H), 5.22 (m, 2H), 4.80 (dd, 6.2, 18.6 Hz, 1H), 4.55
(m, 5H), 3.90 (s, 3H+2H), 3.83 (d, 12.4 Hz, 1H), 3.38
(m, 2H), 2.81 (m, 1H), 2.41 (d, 11 Hz, 1H), 2.20 (t,
11 Hz, 1H), 1.12 (s, 3H), 1.08 (s, 3H).
1. Vacca, J. P.; Condra, J. H. Drug Discov. Today 1997, 2,
261.
2. Ghosh, A. K.; Bilcer, G.; Schiltz, G. Synthesis 2001, 2203–
2229.
3. Askin, D. Curr. Opin. Drug Discov. Develop. 1998, 1, 338–
348.
4. Lu, Z.; Raghavan, S.; Bohn, J.; Charest, M.; Stahlhut, M.
W.; Rutkowski, C. A.; Simcoe, A. L.; Olsen, D. B.; Schleif,
W. A.; Carella, A.; Gabryelski, L.; Jin, L.; Lin, J. H.;
Emini, E.; Chapman, K.; Tata, J. R. Bioorg. Med. Chem.
Lett. 2003, 13, 1821–1824.
25
D
14: mp 201 °C; ½aꢁ ꢂ8 (c 0.51, MeOH); 1H NMR
(400 MHz, DMSO): d 1.57 (s, 6H), 2.49–2.51 (m, 2H),
3.00–3.03 (m, 2H), 3.35 (d, 12.6 Hz, 1H), 3.46 (d,
11.0 Hz, 1H), 3.97 (s, 3H), 3.97–4.07 (m, 3H), 7.49–
7.54 (m, 6H), 7.71 (dd, 1.4, 8.5 Hz, 3H), 7.85 (d,
8.5 Hz, 3H), 7.86–7.93 (m, 3H), 7.94–7.97 (m, 3H),
7.99 (s, 1H), 8.05 (t, 2.1 Hz, 1H), 8.1–8.4 (broad, 3H),
8.50 (d, 2.6 Hz, 1H), 8.73 (d, 1.5 Hz, 1H), 8.8–8.9
(broad, 1H), 9.1–9.2 (broad, 1H), 9.29 (t, 6.3 Hz, 1H).
13C NMR (100 MHz, DMSO): d 24.3, 24.7, 43.0, 43.7,
47.1, 56.5, 57.7, 59.3, 123.2, 124.2, 124.7, 124.9 (q,
279 Hz), 126.9, 127.1, 127.3, 127.4, 127.9, 128.0, 128.9,
131.1, 131.8, 132.6, 133.4, 145.4, 146.4, 157.8, 166.4,
167.5.
5. Schock, H. B.; Garsky, V. M.; Kuo, L. C. J. Biol. Chem.
1996, 271, 31957–31963.
6. Tata, J. R.; Chapman, K. T.; Duffy, J. L.; Kevin, N. J.;
Cheng, Y.; Rano, T. A.; Zhang, F.; Huening, T.; Kirk, B.
A.; Lu, Z.; Raghavan, S.; Fleitz, F. J.; Petrillo, D. E.;
Armstrong, J. D., III; Varsolona, R. J.; Askin, D.;
Hoerrner, R. S.; Purick, R. Preparation of c-hydroxy-2-
fluoroalkylaminocarbonyl-1-piperazine-pentanamides as
HIV protease inhibitors. PCT Int. Appl. WO 0138332,
2001.
7. Tata, J. R.; Charest, M.; Lu, Z.; Raghavan, S.; Huening,
T.; Rano, T. Preparation of a-hydroxy-c-[[(carbocyclic-or
heterocyclic-substituted)amino]carbonyl]alkanamides as
HIV protease inhibitors. PCT Int. Appl. WO 0105230,
2001.
8. Tata, J. R.; Raghavan, S.; Lu, Z.; Zhang, F.; Cheng, Y.;
Chang, J.; Kim, R. M.; Bohn, J. M.; Rano, T.; Shen,
D.-M.; Shu, M. Preparation of piperazine pentanamide
HIV protease inhibitors and methods of treating AIDS.
PCT Int. Appl. WO 0296359, 2002.
9. Zhang, F.; Chapman, K. T.; Schleif, W. A.; Olsen, D. B.;
Stahlhut, M.; Rutkowski, C. A.; Kuo, L. C.; Jin, L.; Lin,
J. H.; Emini, E. A.; Tata, J. R. Bioorg. Med. Chem. Lett.
2003, 13, 2573–2576.
17: 1H NMR (400 MHz, CDCl3): d 8.55 (m, 1H), 8.25 (s,
1H), 7.95 (s, 1H), 7.45 (s, 1H), 7.38 (m, 1H), 3.9 (s, 3H);
13C NMR (100 MHz, CDCl3): d 155.7, 151.1, 148.7,
138.0, 137.6, 124.4, 123.0, 115.6, 55.7.
19: 1H NMR (400 MHz, CD3CN): d 7.2 (br s, 2H), 3.93
(q, J = 9.4 Hz, 2H), 2.7–3.7 (m, 7H), 1.38 (s, 3H), 1.23
(s, 3H); 13C NMR (100 MHz, CD3CN) d 168.8, 124.1
(q, J = 279.5 Hz), 121.1 (q, J = 319.5 Hz), 78.9, 55.0,
45.0, 44.6, 41.0 (q, J = 35.7 Hz), 40.4, 24.1, 16.9.
1
10. Gauthier, D. R.; Ikemoto, N.; Fleitz, F. J.; Szumigala, R.
H.; Petrillo, D.; Liu, J.; Armstrong, J. D.; Yehl, P. M.;
Wu, N.; Volante, R. P. Tetrahedron: Asymmetry 2003, 14,
3557–3567.
20: H NMR (500 MHz, CD2Cl2): d 4.61 (m, 1H), 4.45
(m, 1H), 3.7–3.95 (m, 2H), 3.45–3.65 (m, 3H), 3.17 (dt,
J = 11.3, 3.2 Hz, 1H), 3.01 (td, J = 11.3, 2.8 Hz, 1H),
2.65 (s, 3H), 2.64 (s, 3H), 1.45 (s, 3H), 1.28 (s, 3H);
13C NMR (126 MHz, CD2Cl2): d 191.9, 169.4, 124.2
(q, J = 279.9 Hz), 79.5, 57.3, 53.9, 53.8, 42.7, 41.9 (q,
J = 36 Hz), 26.5, 26.4, 24.7, 19.3.
11. Syntheses of compounds containing a N-gem-dimethyl-
methylene-2-oxazolyl or benzoxazolyl group: (a) Wipf, P.;
Miller, C. P. J. Org. Chem. 1993, 58, 3604–3606; (b)
Huang, J.; Haigh, d. S. J. Heterocycl. Chem. 1990, 27,
331–334; (c) Katritzky, A. R.; Aslan, D.; Shcherbakova, I.
V.; Chen, J.; Belyakov, S. A. J. Heterocycl. Chem. 1996,
33, 1107–1114; (d) Villalgordo, J. M.; Vincent, B. R.;
Heimgartner, H. Helv. Chim. Acta 1990, 73, 959–974.
12. Liu, J.; Ikemoto, N.; Petrillo, D.; Armstrong, J. D.
Tetrahedron Lett. 2002, 43, 8223–8226.
1
1: mp 205 °C; H NMR (400 MHz, CD3OD): d 8.50 (d,
1.5 Hz, 1H), 8.25 (d, 2.7 Hz, 1H), 7.70 (s, 2H), 7.32–7.17
(m, 5H), 7.16 (d, J = 7.6 Hz, 1H), 7.09 (m, 1H), 6.86 (m,
1H), 6.40 (d, 8.1 Hz, 1H), 5.17 (d, 4.0 Hz, 1H), 4.84 (s,
4H), 4.21 (br s, 1H), 4.17–3.98 (m, 4H), 3.96 (s, 3H),
3.97–3.75 (m, 3H), 3.40 (br s, 1H), 3.32–3.13 (m, 2H),
3.07 (br s, 1H), 3.03–2.93 (m, 2H), 2.85–2.62 (m, 3H),
1.90 (m, 1H), 1.68 (s, 6H), 1.53 (m, 1H). 13C NMR
(100 MHz, CD3OD): d 177.8, 169.0, 167.5, 158.0,
155.7, 150.1, 140.5, 138.0, 137.9, 130.4, 130.0, 129.8,
129.6, 127.6, 126.6, 125.8 (q, 278 Hz), 125.0, 122.4,
122.0, 117.9, 117.4, 69.4, 66.6, 65.5, 62.2, 59.9, 56.8,
54.0, 49.7, 49.1, 46.1, 45.5, 41.5 (q, 34.9 Hz), 40.5,
38.6, 24.6, 24.3.
13. The silver containing waste stream was forwarded to an
outside vendor for metal recovery.
14. Since bulk commercial NSA was only available with low
purity (ꢀ65–80 wt % and <90 A% purity), novel and
practical purification procedures were devised to upgrade
this acid to the required >99 A% purity. Method 1: Crude
NSA is dissolved in toluene/acetonitrile at 80 °C, the
bottom dark layer is cut, and the top layer containing
the NSA is saturated with water and cooled to crystallize
the NSA hydrate in 98–99 A% purity. Recrystallization
from acetonitrile/toluene affords >99 A% purity (57%
overall recovery). Method 2: The toluene layer after the
cut is extracted with water. The aqueous layer is evapo-
rated and solvent switched to acetonitrile. Toluene is
added and heated to form a clear solution. Seeding with
NSA and cooling generates a slurry, which is filtered,
rinsed with toluene and dried to afford NSA with 99.5 A%
purity (66% recovery).
Acknowledgements
We thank Dr. Tom Novak for HRMS data and
Lisa DiMichele for NMR identification of key
impurities.

N. Ikemoto et al. / Tetrahedron Letters 46 (2005) 1867–1871
1871
15. Higher temperatures lead to lower product yields and
purity while lower temperatures unacceptably increased
the time required for complete conversion.
17. Welch, C. J.; Fleitz, F.; Antia, F.; Yehl, P.; Waters, R.;
Ikemoto, N.; Armstrong, J. D.; Mathre, D. Org. Proc.
Res. Devel. 2004, 8, 186–191.
16. Al(O-iPr)3, LiClO4, LiCl, LiBF4, NaClO4, Ti(O-iPr)4,
MgBr2, Mg(OH)2, Mg(ClO4)2, Sc(Otf)3, Cu(Otf)2,
Zn(Otf)2, La(O-iPr)3.
18. Silylation of oxazoles: Miller, R. A.; Smith, R. M.;
Karady, S.; Reamer, R. A. Tetrahedron Lett. 2002, 43,
935–938.