A highly diastereoselective vinylogous Mannich condensation and 1,4-conjugate addition of (Z)-propenyl cuprate in the synthesis of an influenza neuraminidase inhibitor

Article abstract of DOI:10.1016/S0957-4166(03)00597-4

A practical synthesis of neuraminidase influenza inhibitor, A-322278, has been developed. Asymmetry is introduced into the synthesis by an enzyme mediated ester hydrolysis. A highly diastereoselective vinylogous Mannich condensation reaction of N-Boc-2-te

Full text of DOI:10.1016/S0957-4166(03)00597-4

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 14 (2003) 3541–3551
A highly diastereoselective vinylogous Mannich condensation and
1,4-conjugate addition of (Z)-propenyl cuprate in the synthesis of
an influenza neuraminidase inhibitor
David M. Barnes, Lakshmi Bhagavatula, John DeMattei,† Ashok Gupta, David R. Hill,
Sukumar Manna, Maureen A. McLaughlin, Paul Nichols,† Ramiya Premchandran,‡
Michael W. Rasmussen, Zhenping Tian and Steven J. Wittenberger*
Abbott Laboratories, Global Pharmaceutical Research and Development, Process Research and Development,
1401 Sheridan Road, North Chicago, IL 60064-4000, USA
Received 21 March 2003; accepted 13 June 2003
Abstract—A practical synthesis of neuraminidase influenza inhibitor, A-322278, has been developed. Asymmetry is introduced
into the synthesis by an enzyme mediated ester hydrolysis. A highly diastereoselective vinylogous Mannich condensation reaction
of N-Boc-2-tert-butyldimethylsilyloxypyrrole (TBSOP) and an N-(triphenylmethylsulfenyl)imine proceeds under thermodynamic
control to assemble the framework. A significant temperature dependent rate difference for the transfer of (Z)- and (E)-propenyl
moieties from a cuprate reagent during a 1,4-conjugate addition was observed. A very selective addition of cyanide to an
N-acyliminium intermediate was employed to control the final stereocenter.
© 2003 Elsevier Ltd. All rights reserved.
1. Introduction
and duration of uncomplicated influenza infection by
viral strains against which they are active.3
Influenza (the ‘flu’) is a respiratory tract infection
caused by the influenza virus. Between 25 and 75
million people in the United States are infected annu-
ally and cost estimates for heath care and lost produc-
tivity are in excess of 10 billion dollars per year.1
Although most people recover in 1 to 2 weeks, this
illness can be quite serious. In the United States, influ-
enza is associated with more than 20,000 deaths and
more than 100,000 hospitalizations in an average year.2
At the present time, there are four approved medica-
tions available in the US for the control and prevention
of influenza: amantadine, rimantadine, zanamivir, and
oseltamivir. Amantadine and rimantadine are active
only against influenza type A viruses. Zanamivir and
oseltamivir are influenza neuraminidase inhibitors (NI)
with activity against both of the clinically important
influenza viruses, A and B. Inhibition of neuraminidase
impairs the ability of the newly produced virus to be
released from the infected cell, thereby disrupting viral
replication. All of these agents can reduce the severity
Abbott
Laboratories’
influenza
neuraminidase
inhibitor, A-315675 1, is a potent inhibitor of influenza
A and B viral replication.4 The initial synthetic route
provided this material in racemic form; gram quantities
of compound for early biological evaluation were
resolved by preparative HPLC on a chiral stationary
support. As greater quantities of material would be
required for development, it was obvious that a more
efficient, asymmetric synthesis was necessary. Such an
approach was demonstrated that allowed for multi-
gram quantities of drug to be prepared.5 Our studies
targeted several key stereochemical and practical issues
left unresolved from this initial work. Herein the pro-
cess research and initial scale up to prepare bulk quan-
tities of A-322278 2, a pro-drug of A-315675, required
for further pre-clinical evaluation is described (Fig. 1).6
2. Results and discussion
The starting point for our research was the asymmetric
route previously demonstrated by our colleagues.5 This
approach is shown retrosynthetically in Scheme 1 to
highlight the key transformations. The C-2 carboxylate
* Corresponding author. E-mail: steve.j.wittenberger@abbott.com
† Present address: Array Biopharma, 3200 Walnut Street, Boulder,
CO 80301.
‡ Present address: MediChem Life Sciences, Woodridge, IL 60517.
0957-4166/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0957-4166(03)00597-4

3542
D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
Early supplies were prepared through the use of the
Sharpless asymmetric epoxidation reaction, however,
throughput was low and an alternate method was
developed to prepare 5. Reaction of 2-pentanone with
bromoform in methanolic potassium hydroxide solu-
tion provided carboxylic acid 7 in 80% yield.8 Attempts
to accomplish an efficient classical resolution of ( )-7
with commercially available enantiomerically pure
amines failed. Therefore, a series of simple carboxylic
esters were prepared and screened versus a battery of
readily available enzymes in an effort to achieve resolu-
tion by kinetic hydrolysis. The combination of methyl
ester ( )-8 and lipase from Candida rugosa proved to be
very effective; 49% yield (98% of a theoretical 50%) of
]97%ee carboxylic acid (S)-7 was routinely obtained
on multi-hundred gram scale. Reduction of acid (S)-7
with BH3·SMe2 provided alcohol 9 in near quantitative
yield. Oxidation followed by condensation with
triphenylmethanesulfenamide gave 5 in 89% yield.
Figure 1.
The method9 reported by Casiraghi for the preparation
of
N-Boc-2-(tert-butyldimethylsiloxy)pyrrole
6
(TBSOP) was unsuitable for large-scale use (Scheme 2,
Eq. (1)). The oxidation of pyrrole with hydrogen perox-
ide posed safety concerns if carried out on large scale
and the purification was tedious.10 An alternate synthe-
sis was developed in order to safely and reliably pro-
duce multi-kilogram quantities of 6. We were prompted
by a paper from the Pifferi group to explore the dehy-
drative cyclization of ( )-4-amino-3-hydroxybutyric
acid 11 employing HMDS and TMSCl in refluxing
xylene.11 In our laboratory the product yield was quite
variable (60–95%) with seemingly uncontrollable
amounts of des-silyl by-product as the remainder; how-
ever, exposure of 11 to HMDS and pyridine in reflux-
ing xylene led to a high yield of 12 (91–98%).
Concentration of the solvent removed the volatile reac-
tion by-products. Crude 12 was dissolved in THF and
treated initially with di-tert-butyl dicarbonate and
DMAP to introduce the Boc group and provide ( )-N-
Scheme 1. Reagents and conditions: (a) CHBr3, KOH,
MeOH; (b) MeOH, H2SO4 (cat.); (c) lipase Candida rugosa,
NaOH (aq), acetone; (d) BH3-SMe2; (e) DMSO, SO3–pyridine
complex, Et3N; (f) H2NSCPh3, Na2SO4.
of 1 (or 2) was derived from the corresponding C-2
nitrile which was prepared by a stereoselective addition
of cyanide to the N-acyliminium species generated from
reduction of N-Boc lactam 3 followed by activation
with a Lewis acid. The (Z)-propenyl side chain at C-4
was incorporated via conjugate addition of a cuprate
reagent to N-Boc-D3-2-pyrrolidinone 4. In the pivotal
reaction of this sequence, 4 was formed by the stereo-
selective condensation of imine 5 and N-Boc-2-(tert-
butyldimethylsilyloxy)pyrrole (TBSOP) 6. This reaction
is an extension of the studies of Casiraghi et al. on the
Lewis acid promoted condensations of pyrrole-, furan-,
and thiophene-based 2-siloxydienes with aldehydes and,
to a lesser extent, imines.7 Thus, efficient preparations
of these two starting materials, imine 5 and TBSOP 6,
became early goals.
Scheme 2. Reagents and conditions: (a) HMDS, C5H5N,
xylene, D; (b) Boc2O, DMAP; (c) Et3N-3HF; (d) MsCl, Et3N;
(e) TBSOTf, Et3N.

D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
3543
Boc-4-(trimethylsiloxy)-2-pyrrolidinone 13. Upon reac-
tion completion, triethylamine trihydrofluoride
(Et3N·3HF) was added to the mixture to remove the
trimethylsilyl moiety. As the deprotection proceeded,
the product ( )-N-Boc-4-hydroxy-2-pyrrolidinone 14
crystallized from the reaction mixture. This one-pot
three-step sequence provided 14 in 97% overall yield in
the laboratory and 89% yield on a 15 kg scale in the
pilot plant.12
The stable crystalline 14 was a convenient intermediate
to stockpile for ready access to TBSOP. Dehydration
was carried out by treatment with methanesulfonyl
chloride and triethylamine to give the N-Boc-D3-pyrro-
lidinone 15. To a mixture of crude 15 and heptane was
added triethylamine and diatomaceous earth. After
cooling to 0°C the mixture was treated with tert-
butyldimethylsilyl trifluoromethanesulfonate (TBSOTf).
Scheme 3. Reagents and conditions: (a) Optimized conditions:
TfOH, THF, −40°C.
the presence of small amounts of residual water might
be responsible for the better results on small scale.
Replacing TMSOTf with trifluoromethanesulfonic acid
resulted in a robust and reproducible procedure where 5
and 6 reacted to produce a 19:1 ratio of 4 and 16 in
90–95% HPLC assay yield. Following workup, the
major product 4 was isolated by crystallization in 80–
85% yield in >99:1 diastereomeric purity. Model studies
and calculations provided evidence to understand the
origins of the high selectivity observed; under the reac-
tion conditions an intramolecular hydrogen bond
between the C-6 S-triphenylmethyl ammonium sub-
stituent and the C-7 methoxy group stabilizes the major
diastereomer 4 relative to the minor diastereomer 16.13
The
triethylammonium
trifluoromethanesulfonate
formed as a by-product was absorbed onto the filter
agent as the reaction progressed. At the end of reaction,
the darkened diatomaceous earth was removed by
filtration. All other reaction by-products were volatile
and were removed upon concentration of the filtrate.
For use in our process, the residue was reconstituted in
heptane and held for next step; the yield of TBSOP 6
was 95–98% (85–95% overall).
From the outset, it was very apparent that achieving an
efficient and selective coupling of imine 5 and TBSOP 6
was key to the success of this strategy. Our colleagues
had found that reaction of 5 and 6 in the presence of
BF3·Et2O resulted in the formation of two products, 4
and 16, in a 3–7:1 ratio (75% combined yield) along
with some unreacted starting materials.5 Screening of
Lewis acids and solvents resulted in modest improve-
ment in diastereoselectivity with TMSOTf in THF at
−78°C (6:1, 4:16).
At this stage, a number of preliminary investigations of
strategies to incorporate the (Z)-1-propenyl moiety
were conducted. Our endeavors quickly focused upon
the use of [(Z)-1-propenyl]2CuMgBr derived from the
corresponding Grignard reagent and CuBr·SMe2 com-
plex due to the high isomeric purity of commercially
available (Z)-1-bromo-1-propene14 and the high yield
of the 1,4-conjugate addition products (Scheme 4).
Early experiments showed that removing the (E)-1-pro-
penyl adduct 17 from the desired (Z)-isomer 3 would be
extremely difficult either by crystallization or chro-
matography after the 1,4-conjugate addition or later in
the synthesis, so a goal was established to develop
conditions to control this impurity to <1% in the
cuprate addition reaction.
During optimization experiments, two important obser-
vations were made. First, the ratio of product
diastereomers was observed to improve during the
course of some reactions. By extending reaction times
(15 h versus 3–4 h) or increasing the reaction tempera-
ture (to −40°C) the product diastereomeric ratio
improved markedly (11:1, 83% assay yield and 19:1,
95%, respectively). Re-subjecting purified 4 and 16 to
the reaction conditions provided additional evidence for
equilibration of the products under the reaction condi-
tions. In the case of 4, the ratio (4:16) changed from
>99:1 to 93:7, which was consistent with previous
experiments. Exposure of 16 (2:98) led to a ratio of
26:74. In both experiments there was substantial degra-
dation of the products due to loss of the N-Boc group
which may explain why equilibrium ratios were not met
(especially in the latter case) (Scheme 3).
(Z)-1-Bromo-1-propene is sensitive to thermal and
acid-catalyzed isomerization.15 In contrast to earlier
reports15,16 however, we observed complete retention of
configuration on preparation of the Grignard reagent in
THF.17 A brief survey of copper(I) salts showed that
CuBr·SMe2, CuI, or CuCN could be used to prepare
cuprate reagents that produced the 1,4-conjugate addi-
tion products in high yield. The mixed higher order
cuprate reagent18 derived from CuCN, 2-thienyllithium
and (Z)-1-propenylmagnesium bromide also yielded the
desired product, however, it was contaminated by sev-
eral low level impurities that were not observed with
the other reagents. CuBr·SMe2 complex was chosen for
scale up because of its relative ease in handling and
reduced toxicological liability relative to copper(I)
cyanide.
The second observation was that reactions on small
scale (50–100 mg) proceeded to completion and pro-
vided high yields of product. Upon increasing to gram
scale however, reactions would tend to stall and could
not be pushed to completion even with additional
aliquots of starting materials. This led to the idea that

3544
D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
unit versus the (E)-1-propenyl unit was judged to be
2–4:1 at −78°C but increases to 6–12:1 at −20 to
−25°C.22 Reactions that were run at temperatures
higher than −20°C required additional Grignard
reagent to go to completion and provided lower
product yield with increased levels of trans-isomer 17.
Conjugate addition product 3 was carried on without
purification since (a) the crude product was devoid of
any significant impurities and (b) there were concerns
about the stability of the triphenylmethylsulfenamine
moiety.23 Investigations into the selective removal of
the triphenylmethylsulfenyl group24 were not promis-
ing; under a variety of conditions, cleavage of the
N-Boc group was competitive. Selectivity could be
achieved, but the required conditions would be chal-
lenging to control on large-scale (i.e. iodotrimethylsi-
lane at −78°C) or difficult to purify from the
triphenylmethylsulfenyl by-products (i.e. zinc in acetic
acid). The most pragmatic solution was a global depro-
tection–reprotection sequence (Scheme 4). Exposure of
the organic phase following work up of the cuprate
reaction to gaseous HCl resulted in rapid and complete
removal of both the triphenylmethylsulfenyl and N-Boc
groups. Addition of water partitioned ammonium salt
18 into the aqueous phase and the triphenylmethyl-
sulfenyl by-products remained in the organic layer.
Following pH adjustment, the solution was treated with
acetic anhydride to provide acetamide 19. This com-
pound was sufficiently hydrophobic to be extracted into
isopropyl acetate after salting the aqueous phase. After
azetropic drying and switching the solvent to THF, 19
was treated with di-tert-butyl dicarbonate, triethy-
lamine and catalytic DMAP. The product imide 20 was
isolated by crystallization in 87% yield over the four
steps from 4 (Scheme 5).
Scheme 4. Reagents and conditions: (a) (E)-propenyl cuprate,
TMSCl, THF/toluene, −20°C; (b) HCl (g); (c) NaOH (aq),
Ac2O; (d) Boc2O, DMAP.
Early in the development of this reaction, the cuprate
reagent was prepared in excess from CuBr·SMe2 (5
equiv.) and (Z)-CH3CHꢀCHMgBr (10 equiv.) in THF
at −78°C. Chlorotrimethylsilane (3 equiv.) was added,
followed by a THF solution of 4. Typically a high yield
(]95%) of desired adduct 3 was obtained along with
2–4% of isomer 17.5,19 The quantities of reagents could
be reduced (CuBr·SMe2 (0.5 equiv.) and (Z)-
CH3CHꢀCHMgBr (3 equiv.)) without effect; lower
amounts gave incomplete conversion. We investigated
alternate modes of addition in an effort to conserve
(Z)-1-bromo-1-propene because it was an expensive
reagent. By adding the Grignard reagent to a mixture
of 4, CuBr·SMe2, and TMSCl only 2.1 equiv. (Z)-
CH3CHꢀCHMgBr was required to consume 4, how-
ever, the level of trans-product 17 was high (]3%).
We were pleasantly surprised on investigating the effect
of reaction temperature. With our preferred stoi-
chiometries for the pre-formed cuprate reagent20 (0.5
equiv. CuBr·SMe2, 3 equiv. (Z)-CH3CHꢀCHMgBr, and
1.2 equiv. TMSCl in THF) at −78°C, 1.6% of trans-
product 17 was observed, but at −20 to −25°C only
0.5–0.7% 17 was detected! It is important to note that
this was the last point in the synthetic scheme that a
sensitive analytical method was in place for determina-
tion of the (Z):(E)-olefin ratio prior to final product.
The selective transfer of the (Z)-1-propenyl moiety
from [(Z)-1-propenyl]2CuMgBr in a 1,4-conjugate addi-
tion reaction has been previously observed.21 Consider-
ing that the Grignard reagent (and the derived cuprate
reagent, measured by assay of vinyl iodides following
iodine quench) contained 3% (E)-1-propenyl ligand, the
rate difference for the transfer of the (Z)-1-propenyl
Scheme 5. Reagents and conditions: (a) LiB(Et3)H, THF; (b)
CH3C(OCH3)3, CSA; (c) TMSCN, TfOH, CH3CN, −42°C.

D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
3545
Imide reduction of 20 could be accomplished with
either DIBAL-H or lithium triethylborohydride to pro-
duce 21a as a mixture of diastereomers. LiB(Et)3H was
selected because the crude reaction mixture appeared
slightly cleaner and the work up was simpler due to the
absence of aluminum salts. Exposure of crude 21a to
trimethyl orthoacetate and methanol with catalytic
camphorsulfonic acid yielded N-Boc-methoxyaminal
22a in 96% yield from imide 20. As noted above, a
sensitive assay to measure the (Z):(E)-olefin ratio had
not been developed between cuprate adduct 3 and final
product 2. After completing our initial scale up cam-
paign, it was discovered that some isomerization (0.7%
(E)“1.7% (E)) of the (Z)-propenyl moiety had
occurred. A thorough investigation of intermediate
samples and reaction conditions determined that the
loss of stereochemistry had occurred in the LiB(Et)3H
reduction step! With this finding, we suspected that
B(Et)3 in the hydride reagent might be responsible as it
is well known that trialkylboranes and oxygen are a
source of radical species.25 In fact, in parallel reactions
20 (0.5% trans-olefin isomer) was treated with B(Et)3
(0.26 equiv.) in the presence and absence of oxygen
under conditions similar to the reduction reaction (0.13
M in THF, −10 to −15°C). With oxygen, 2.3% (E)-pro-
penyl isomer was detected, while the control reaction
showed 0.5% (E)-isomer.26
the N-Boc group.27 Early studies conducted on the
triphenylmethylsulfenyl protected derivative 22b
showed that the choice of Lewis acid and solvent28 had
a significant impact the reaction yield and selectivity
(Table 1, entries 1–7). This proclivity was encountered
again with 22a, which was pursued for development
after encountering stability problems with the triphenyl-
methylsulfenyl moiety.23 Nevertheless, the optimized
conditions for 22a (3.0 equiv. TMSCN, 1.5 equiv.
TfOH, CH3CN, −40°C) were reproducible in providing
2-cyanopyrrolidine adducts 23a and 24a in 85–90%
yield in a 96:4 diastereomeric ratio. Approximately
5–8% yield of the des-Boc product was also typically
observed.29 Results with the 2-ethoxy analogue of 22a
were similar (23a in 80% yield), but the 2-isopropoxy
ether and N-Boc-hydroxyaminal 21a were inferior (18
and 48%, respectively) with significant amounts of the
elimination side product 25 formed. In practice the
crude 23a containing ꢀ4% epimeric 24a was carried
into the final reaction sequence.
The transformation of 23a to the target A-322278 2
could be accomplished in three steps through the inter-
mediacy of the amino acid 1 by exposure to 6N HCl
(0.8 mM, 60°C, 16 h, 100% yield)5,6 followed by treat-
ment with thionyl chloride and i-PrOH (rt, 48 h, 2, 31%
yield, unoptimized) and tosylate salt formation.30 From
a processing perspective, however, a more direct con-
version that avoided chromatographic purification was
desired. After several iterations of optimization studies,
during which time our analytical methods steadily
improved, we came to the following conclusions. First,
the hydrolysis/alcoholysis of the nitrile would have to
be carried out under acidic conditions as exposure to
base caused rapid epimerization of the nitrile to the
thermodynamically favored b-isomer. Second, HCl was
the acid of choice as a host of others resulted in
significant amounts of diketopiperazine 26 (Fig. 3). The
empirical observations were that build up of the amino
nitrile intermediate 27 correlated with formation of 26
and that HCl facilitated rapid alcoholysis at low tem-
perature. Finally, incorporation of HCl into the
molecule,31 presumably by addition across the alkene,
Incorporation of the C2-carboxylate established the
final stereogenic center. Nucleophilic addition of cya-
nide anion to the N-acyliminium ion (Fig. 2) generated
from 22a or 22b was expected to proceed anti to the
bulky C5-substituent that was envisioned to adopt a
pseudo-axial conformation to minimize A1,2 strain with
Figure 2.
Table 1. Optimization of nitrile addition
Entry
Substrate
Lewis acid
Solvent
Temp. (°C)
% Yield 30
a:b ratio
1
2
3
4
5
6
7
8
22b
22b
22b
22b
22b
22b
22b
22a
22a
22a
22a
22a
22a
22a
22a
BF3·OEt2
BF3·OEt2
BF3·OEt2
BF3·OEt2
SnCl4
TMSOTf
TMSOTf
BF3·OEt2
BF3·OEt2
BF3·OEt2
TMSOTf
TfOH
CH2Cl2
CH3CN
THF
Toluene
Toluene
CH2Cl2
Toluene
CH2Cl2
Toluene
THF
−78
−30
−78
−78
−78 to −45
−78
−78
−78
−78
−78
−42
−42
−42
−42
−42
89
ꢀ11:1
ꢀ1:1
ꢀ10:1
ꢀ1:1
1.5:1
6:1
ꢀ3:1
ꢀ2:1
ꢀ1:1
–
Nd
Nd
Nd
32
Nd
68
Nd
Nd
NR
67
9
10
11
12
13
14
15
CH3CN
CH2Cl2
Toluene
THF
93:7
50
60
19
90
90:10
88:12
87:13
96:4
TfOH
TfOH
TfOH
CH3CN

3546
D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
toluenesulfonic acid monohydrate resulted in the pre-
cipitation of 2·TsOH, which was isolated in 84% yield.
3. Conclusions
In the course of developing a practical synthesis of
influenza neuraminidase inhibitor A-322278, a number
of interesting observations were made and several syn-
thetic challenges were overcome. A new preparation of
N-Boc-2-tert-butyldimethylsilyloxypyrrole (TBSOP) 6
Figure 3.
that is amendable to large-scale was demonstrated.12
A
could occur at and above ambient temperature to give
28 (Fig. 3, up to several percent in some cases) under
the very acidic conditions necessary to efficiently con-
vert 23a into 2, via imidate 29.
short route to (2S)-2-methoxy-2-methyl-pentylidene-
(triphenyl-methylsulfenyl)amine that featured an
5
extremely efficient enzymatic resolution established
asymmetry in the synthesis. These subunits were cou-
pled in a remarkably diastereoselective vinylogous
Mannich reaction under thermodynamic control.13 The
(Z)-1-propenyl substituent was introduced by a 1,4-
conjugate addition reaction to 4. A temperature depen-
dent kinetic preference for transfer of the (Z)-isomer
relative to the (E)-1-propenyl isomer from the
organocopper reagent was observed and the relative
rate difference was calculated. We unexpectedly experi-
enced ꢀ1% erosion in alkene geometry of the (Z)-1-
propenyl substituent due to presence of triethylborane
and oxygen during the lithium triethylborohydride
reduction of an imide. The C-2 carboxylate functional-
ity was incorporated by the nucleophilic addition of
cyanide ion to an N-acyliminium ion with a high level
of stereochemical control. The end game of N-Boc
deprotection and nitrile alcoholysis was carried out in
one pot under conditions that minimized the formation
of diketopiperazine and HCl olefin addition side prod-
ucts. This synthetic route, which proceeds in 16 steps by
the longest linear sequence (including an enzymatic
resolution) and 12.9% overall yield, was used to suc-
cessfully deliver more than 1 kg A-322278·TsOH to
support pre-clinical biological, toxicological and drug
safety evaluations.
The optimized procedure for this Pinner reaction
entailed treatment of a cold solution of 23a in i-PrOH
with anhydrous HCl (Scheme 6). Cleavage of the N-
Boc group was very fast. Upon warming to about rt,
HPLC analysis of the reaction mixture showed that 27
was consumed. The reaction mixture was diluted with
i-PrOH. In effect this reduced the HCl concentration
and minimized formation of 28. The reaction mixture
was warmed to reflux for several hours to convert the
intermediate imidate 29 to the isopropyl ester 2. There
was some erosion in the C-2 a:b diastereomer ratio,
96:4 in 23a to 92:8 in 2. Typically a small amount of the
amino acid 1 was produced in the reaction and it was
not converted to 2 under these conditions. After cool-
ing to rt, the reaction mixture was concentrated to
half-volume to remove most of the HCl, diluted with
isopropyl acetate (i-PrOAc), and washed with 15% wt
aqueous potassium bicarbonate solution. This wash
removed the small amount of amino acid 1 and pro-
vided an organic solution of 2 free base. Following a
brine wash, the wet organic solution of 2 (ca. 2–3% wt
water by KF titration) was azeotropically dried and
filtered to remove precipitated inorganic salts that
deposited. Addition of an i-PrOAc solution of para-
4. Experimental
4.1. General procedures
Starting materials, reagents, and solvents were pur-
chased from commercial suppliers and were used with-
out further purification. 1H NMR spectra were
recorded at 300, 400 or 500 MHz and 13C NMR spectra
were recorded at 75 or 100 MHz with chemical shifts (l
ppm) reported relative to tetramethylsilane (TMS) as
an internal standard. Melting points were determined
on a Thomas–Hoover capillary melting point apparatus
and were uncorrected. Elemental analyses were per-
formed by Robertson Microlit Laboratories. Column
chromatography was carried out on silica gel 60 (230–
400 mesh). Thin-layer chromatography (TLC) was per-
formed using 250 mm silica gel 60 glass-backed plates
with F254 as indicator. Visualization of TLC was car-
ried out by UV light, KMnO4 or phosphomolybdic acid
spray reagent.
Scheme 6. Reagents and conditions: (a) HCl (g), i-PrOH,
−30°C to rt; (b) add i-PrOH, D; (c) KHCO3 (aq), i-PrOAc;
(d) TsOH·H2O, i-PrOAc.

D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
3547
4.2. ( )-Methyl-2-methoxy-2-methyl-pentanoate, 8
55°C was added a solution of SO3·pyridine complex
(35.0 g, 220 mmol) in DMSO (160 mL) over ca. 2.5 h to
maintain the reaction temperature <10°C. Following
the addition, the reaction was monitored by GC. The
reaction was quenched by the addition of 2 M H3PO4
(325 mL) at a rate to maintain internal temperature
<20°C. The layers were separated and the upper
aqueous phase was extracted with CH2Cl2 (160 mL).
The combined organics were washed with 2 M H3PO4
(160 mL) and then dried over silica gel (9.5 g). Filtra-
tion and careful partial solvent distillation gave 60.4 g
of a solution that contained 10 (17.97 g, 137.6 mmol,
95.3% yield) by GC assay: DB-1 column, 30 m×1.5 mm
film×0.53 mm; temperature gradient: 40°C for 3 min
then ramp to 190°C at 15°C/min, hold at 190°C 3 min
(16 min run time); FID detection, 9 (8.6 min), aldehyde
10 (7.5 min). 10: 1H NMR (300 MHz, CDCl3) l 9.58 (s,
1H), 3.28 (s, 3H), 1.69–1.51 (m, 2H), 1.45–1.25 (m, 2H),
1.22 (s, 3H), 0.92 (t, J=7.0 Hz, 3H). To this solution
was added Ph3CSNH2 (39.30 g, 134.9 mmol), PPTS
(0.69 g, 2.7 mmol), and Na2SO4 (9.81 g, 69 mmol); the
mixture was stirred at rt ca. 12 h. The disappearance of
10 was monitored by GC. After the reaction was judged
to be complete, the mixture was diluted with heptane
(32 mL) and the solid Na2SO4 was filtered, rinsing with
heptane. The solution was concentrated at reduced
pressure to remove residual CH2Cl2 and obtain a
ꢀ50% wt solution of imine 5. The solution was filtered
through a pad of silica gel (64 g); the pad was rinsed
with 5% MTBE/heptane (500 mL). The filtrate was
concentrated in vacuo to a 40–50% wt solution that
contained imine 5 (50.6 g, 125 mmol, 93.0% yield) by
HPLC assay: Zorbax SB-C8 250 mm×4.6 mm column,
gradient 50% CH3CN/0.1% H3PO4 aq to 90% CH3CN/
0.1% H3PO4 aq over 15 min, 1.5 mL/min, column temp.
35°C, UV at 210 nm. A sample was purified by silica
A solution of ( )-78 (2.5 kg, 17.1 mol) in methanol (12
L) was treated with concentrated H2SO4 (200 mL) and
heated to reflux for 16 h. After cooling to ambient
temperature, the reaction mixture was treated with
NaHCO3 (1.5 kg) and concentrated. The residue was
dissolved in MTBE (8 L) and extracted with 0.5%wt
NaHCO3 solution (4 L). The organic layer was concen-
trated and the residue distilled at 50 mmHg; the frac-
tion boiling at 90–95°C was collected to provide 8 (2.1
1
kg, 77% yield). H NMR (300 MHz, CDCl3) l 3.7 (s,
3H), 3.25 (s, 3H), 1.7–1.6 (m, 2H), 1.35 (3, 3H), 1.3–
1.15 (m, 2H), 0.89 (t, J=7.0 Hz, 3H).
4.3. (2S)-2-Methoxy-2-methyl-pentanoic acid, (S)-7
A reactor equipped with a mechanical stirrer and tem-
perature probe was charged with Candida rugosa lipase
(800 g, 875 units/mg), 0.05 M phosphate buffer (pH 7,
16 L), 8 (0.8 kg, 5 mol) and acetone (1.6 L). The
mixture was warmed to 35–37°C for ꢀ120 h with pH
adjustment to 6 with 50% NaOH every 24 h. Upon
completion, the conversion was 51.2% and (S)-7
formed was 97.8% ee. The pH of the reaction mixture
was adjusted to 3 with concentrated HCl and extracted
with EtOAc (20 L). The organic layer was extracted
with 10% wt Na2CO3 solution (2×5 L). The combined
aqueous portions were acidified with concentrated HCl
to pH 3 and were extracted with MTBE (24 L). The
organic phase was dried over Na2SO4, filtered and
concentrated to provide (S)-7 (339 g, 46% yield, 97%
ee). Chiral GC analysis conditions: J&W CyclodexB
column, 30 m×0.32 mm id, 0.25 mm df; He at 2.5
mL/min (ꢀ44 cm/s); oven temperature: 100°C for 14
min then 10°C/min to 120°C then hold 5 min; 1 mL
injection at 2–3 mg/mL, 50:1 split, 250°C; FID detec-
tion, 300°C w/He at 30 mL/min; enantiomer peaks at
1
gel chromatography for characterization. 5: H NMR
(300 MHz, CDCl3) l 7.57 (s, 1H), 7.40–7.20 (M, 15H),
2.97 (s, 3H), 1.42–1.33 (m, 2H), 1.17–1.07 (m, 2H), 1.00
(s, 3H), 0.81 (t, J=7.3 Hz, 3H).
1
11.4 min (desired) and 11.8 min. H NMR (300 MHz,
CDCl3) l 11.6 (s, 1H), 3.26 (s, 3H), 1.8–1.23 (m, 4H),
1.39 (s, 3H), 0.89 (t, J=7.0 Hz, 3H).
4.4. (2S)-2-Methoxy-2-methyl-hexan-1-ol, 9
4.6. tert-Butyl (2R)-2-{(1R,2S)-2-methoxy-2-methyl-1-
[(triphenylmethylsulfenyl)amino]pentyl}-5-oxo-2,5-dihy-
dro-1H-pyrrole-1-caboxylate, 4
A solution of (S)-7 (550 g, 3.77 mole) in CH2Cl2 (2.75
L) at 0°C was treated with 10 M BH3-SMe2 (0.68 L,
1.8 equiv.). The mixture was stirred 5 h at 20–25°C then
2 M NaOH (3.02 L) was added. The layers were
separated; the aqueous layer was extracted with
CH2Cl2. The organic layer was assayed by GC for 9
(423 g, 85% yield). GC analysis conditions: HP-5
(crosslinked 5% PH ME Siloxane, 30 m×0.32 mm×0.25
mm); column pressure 25 psi; oven temperature: 29°C 2
min, increase 10°C/min to 100°C then hold 2 min; 1 mL
injection at 2–3 mg/mL, 100:1 split, 250°C; FID detec-
tion, 300°C w/He at 7.7 mL/min; retention time is 4.9
min.
To a mixture of imine 5 solution (22.25 g by assay, 55.1
mmol), TBSOP12 6 solution (23.66 g by assay, 79.6
mmol) and THF (350 mL) at −40 to −45°C was added
TfOH (3.74 mL, 42.3 mmol) dropwise such that the
internal temp was maintained 5−35°C. The reaction
progress was monitored by HPLC. After 2.5 h, ꢀ5% 5
remained so an additional aliquot of TfOH (0.12 mL,
1.2 mmol) was added. After an additional 1 h the
reaction was quenched with 0.5 M NaHCO3 solution
(350 mL) and diluted with heptane (84 mL). The mix-
ture was stirred while it warmed to rt, then the layers
were separated and the organic portion was washed
with 25% wt NaCl solution (84 mL). HPLC assay of
the solution showed 4 (31.47 g, 97% yield) and a 4:16
isomer ratio of 19.3:1. The solvent was switched to
heptane (<1A% THF by GC) by distillation in vacuo
while maintaining the volume at ꢀ400 mL until distil-
4.5. (2S)-2-Methoxy-2-methyl-pentylidene-N-triphenyl-
methylsulfenamine, 5
To a solution of alcohol 9 (19.21 g, 145 mmol) and
Et3N (43.5 g, 430 mmol) in CH2Cl2 (100 mL) chilled to

3548
D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
1.58 (d, J=5.1 Hz, 3H), 1.49 (m, 1H), 1.46 (m, 1H),
1.44 (s, 9H), 1.12 (m, 1H), 1.05 (m, 1H), 0.86 (t, J=6.6
Hz, 3H), 0.52 (s, 3H). 13C NMR (125 MHz, CDCl3) l
173.3, 150.9, 144.6, 133.0, 130.0, 127.7, 126.5, 123.1,
82.5, 79.7, 71.7, 66.4, 65.2, 48.8, 39.7, 37.8, 28.0, 27.4,
19.8, 16.9, 14.8, 12.9.
lation was complete. The product began to crystallize
during distillation. The suspension was distilled to a
volume of ꢀ240 mL and allowed to cool to rt and stir
overnight. The solid was filtered and washed with hep-
tane (2×30 mL). The solid was dried in vacuo at 40°C
to give 4 (27.6 g, 85.2% yield, de >99%). Mp 150–
1
152°C. H NMR (400 MHz, CDCl3) l 7.32 (dd, J=2.0,
6.1 Hz, 1H), 7.29–7.19 (m, 15H), 6.02 (dd, J=1.4, 6.1
Hz, 1H), 4.83 (m, 1H), 3.86 (dd, J=3.1, 11.5 Hz, 1H),
3.05 (s, 3H), 2.62 (d, J=11.2 Hz, 1H), 1.62–0.98 (m,
4H), 1.36 (s, 9H), 0.92 (t, J=6.66 Hz, 3H), 0.43 (s, 3H).
13C NMR (100 MHz, CDCl3) l 148.2, 144.8, 130.1,
127.6, 126.7, 126.6, 82.4, 79.2, 66.9, 6436, 48.9, 38.7,
27.9, 19.3, 17.1, 14.9. HRMS m/z: [M+H] calcd for
C35H43N2O4S, 587.2944; found, 587.2953. Minor
4.8. tert-Butyl (2R,3S)-2-{(1R,2S)-2-methoxy-2-methyl-
1-[(acetylamino)pentyl]-5-oxo-3-[(1Z)-prop-1-enyl]-
pyrrolidine-1-caboxylate, 20
The solution of crude 3 (50.68 g by assay, 80.6 mmol)
was chilled to −20°C and anhydrous HCl (g) (90–100 g)
was slowly charged into the solution. The disappear-
ance of 3 was monitored by TLC and HPLC. Upon
completion, diatomaceous earth (50 g) was added and
the mixture was concentrated in vacuo to ca. one-quar-
ter volume. The mixture was filtered and the vessel and
cake were rinsed with water (200 mL) and toluene (200
mL). After separating layers, the organics were
extracted with water (100 mL). The combined aqueous
portions were washed with toluene (200 mL) and then
assayed by HPLC for 18 (20.47 g, 80.5 mmol, 100%
1
diastereomer 16 (oil): H NMR (400 MHz, CDCl3) l
7.28–7.18 (m, 16H), 6.02 (dd, J=1.7, 6.1 Hz, 1H), 4.90
(m, 1H), 3.89 (dd, J=2.0, 11.5 Hz, 1H), 3.06 (s, 3H),
2.42 (d, J=11.5 Hz, 1H), 1.54 (s, 9H), 1.28–0.85 (m,
4H), 1.16 (s, 3H), 0.57 (t, J=7.1 Hz, 3H). 13C NMR
(100 MHz, CDCl3) l 148.2, 129.9, 129.3, 128.1, 127.9,
127.8, 127.7, 127.3, 126.5, 67.2, 65.3, 49.2, 37.3, 28.3,
28.1, 20.9, 17.5, 14.7.
1
yield); H NMR (300 MHz, CDCl3) l 6.78 (br s, 1H),
4.7. tert-Butyl (2R,3S)-2-{(1R,2S)-2-methoxy-2-methyl-
1-[(triphenylmethylsulfenyl)amino]pentyl}-5-oxo-3-[(1Z)-
prop-1-enyl]pyrrolidine-1-caboxylate, 3
5.60–5.45 (m, 2H), 3.37 (dd, J=7.7, 8.8 Hz, 1H),
3.25–3.15 (m, 1H), 3.16 (s, 3H), 2.80 (d, J=8.8 Hz,
1H), 2.42 (dd, J=8.8, 16.9 Hz, 1H), 2.10 (dd, J=10.3,
16.9 Hz, 1H), 1.70 (d, J=5.5 Hz, 3H), 1.65–1.45 (m,
3H), 1.35–1.20 (m, 3H), 1.09 (s, 3H), 0.92 (t, J=7.3 Hz,
3H). MS (APCI) m/z 255 [M+1]+. After chilling the
aqueous solution to ꢀ10°C, 50% wt NaOH solution
(17 mL) was added to adjust the pH to ꢀ12. Ac2O
(16.62 g, 163 mmol) was added; the pH dropped to 6.5
and TLC indicated remaining 18. Additional 50% wt
NaOH solution (4 mL) and Ac2O (1.48 g, 14.5 mmol)
were charged to the reaction mixture; TLC indicated
complete consumption of 18. The pH was raised to
ꢀ12 with 50% NaOH (2 mL) and NaCl (7 g) was
added. The mixture was extracted with i-PrOAc (2×200
mL). HPLC analysis of the combined organics for 19
showed 24.85 g (83.8 mmol, 104% yield); 1H NMR (300
MHz, CDCl3) l 6.50 (s, 1H), 5.81 (d, J=10.6 Hz, 1H),
5.58–5.35 (m, 2H), 4.25 (dd, J=8.5, 10.3 Hz, 1H), 3.52
(dd, J=7.0, 8.5 Hz, 1H), 3.29 (m, 1H), 3.18 (s, 3H),
2.44 (dd, J=9.2, 16.9 Hz, 1H), 1.98 (dd, J=9.2, 16.9
Hz, 1H), 1.95 (s, 3H), 1.62 (dd, J=1.5, 6.6 Hz, 3H),
1.50–1.20 (m, 4H), 1.13 (s, 3H), 0.88 (t, J=6.6 Hz, 3H).
13C NMR (75 MHz, CDCl3) l 175.8, 169.8, 132.0,
124.9, 79.6, 60.3, 55.4, 48.8, 36.8, 36.7, 36.3, 23.2, 18.8,
15.8, 14.4, 12.9. The solution was azeotropically dried
with i-PrOAc (3×200 mL) and THF (200 mL) was
added. The solution was concentrated and the reconsti-
tuted with THF (250 mL). KF analysis showed 510
mol% water versus 19. Et3N (23.5 mL, 169 mmol),
DMAP (0.51 g, 4.2 mmol) and Boc2O (27.43 g, 125.6
mmol) in THF (20 mL) were added. The reaction was
complete after 3.5 h; it was chilled to −10°C and 0.5 M
HCl solution (310 mL) was added over 25 min to keep
the internal temperature 55°C. The mixture was
extracted with i-PrOAc (2×200 mL). The combined
organics were washed with 23% wt NaCl solution (400
mL) and then assayed by HPLC for 20 (28.7 g, 88%
A three neck round bottom flask with overhead
mechanical stirrer, nitrogen inlet, and thermocouple
was evacuated and flushed with nitrogen three times.
THF (750 mL) was charged and then the mixture was
sparged with nitrogen. CuBr·SMe2 (8.74 g, 42.5 mmol)
was added and the mixture was chilled to ca. −30°C. A
solution of (Z)-CH3CHꢀCHMgBr (0.485 M, 526 mL,
255.3 mmol) was added via cannula maintaining an
internal temperature <−20°C (ca. 15 min). TMSCl
(11.09 g, 102 mmol) was added and the mixture was
stirred 10–15 min while warming to ca. −25°C. A
solution of 4 (50 g, 85.1 mmol) in toluene (750 mL) was
sparged with nitrogen and then added via cannula to
the cuprate reagent solution at such a rate that the
internal temperature was 5−20°C (ca. 20 min). After
stirring 5 min, TLC and HPLC analysis showed the
consumption of 4. The reaction was quenched by the
addition of NH4Cl/NH4OH buffer (600 mL 25% wt
NH4Cl solution+75 mL conc NH4OH+75 mL water).
The nitrogen inlet was removed and the mixture was
allowed to warm to rt exposed to air for at least 1 h.
The light yellow organic phase was separated from the
dark blue aqueous phase. The organics were succes-
sively washed with a second portion of the NH4Cl/
NH4OH buffer (750 mL) and 25% NaCl solution (650
mL). HPLC analyses showed 3 (52.23 g, 83.1 mmol,
97.6% yield) and a Z:E isomer ratio of 99.4:0.6. This
solution was carried on to the next step. A sample was
taken for characterization. 1H NMR (500 MHz,
CDCl3) l 7.3–7.1 (m, 15H), 5.42–5.36 (m, 2H), 4.00 (s,
1H), 3.69 (t, J=9.1 Hz, 1H), 3.59 (dd, J=2.2, 10.3 Hz,
1H), 3.06 (d, J=10.7 Hz, 1H), 3.03 (s, 3H), 2.81 (dd,
J=9.2, 17.6 Hz, 1H), 1.95 (dd, J=1.1, 17.6 Hz, 1H),

D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
3549
centration/dilution sequence to azeotropically dry the
solution (KF <10 mol% water). HPLC assay indicated
22a, 49.23 g (119.5 mmol, 94.8% yield from 20). ꢀ1:1
yield). The solution was concentrated in vacuo to ꢀ100
mL, i-PrOAc (300 mL) was added to the suspension
and the mixture was concentrated again to about one-
half volume. Heptane (200 mL) was added to the thick
slurry and the mixture was stirred at rt ca. 20 min. The
product imide 20 was collected by filtration and dried
in vacuo at 50°C to yield 28.24 g (94.9% wt potency,
67.7 mmol, 82% potency adjusted yield from 4). A
sample was recrystallized from MTBE: mp 203–204°C.
[h]2D5 +48.8 (c 1.04 CHCl3). 1H NMR (300 MHz,
CDCl3) l 5.92 (d, J=9.6 Hz, 1H), 5.50–5.40 (m, 2H),
4.54 (dd, J=2.6, 9.6 Hz, 1H), 4.01 (d, J=2.7 Hz, 1H),
3.80 (t, J=9.2 Hz, 1H), 3.23 (s, 3H), 2.76 (dd, J=9.6,
18.0 Hz, 1H), 2.06 (dd, J=1.5, 18.0 Hz, 1H), 1.98 (s,
3H), 1.77 (m, 1H), 1.64 (dd, J=1.1, 6.6 Hz, 3H), 1.63
(m, 1H), 1.56 (s, 9H), 1.31–1.22 (m, 2H), 1.14 (s, 3H),
0.93 (t, J=7.0 Hz, 3H). 13C NMR (75 MHz, CDCl3) l
173.5, 170.1, 149.8, 132.5, 123.5, 83.1, 78.7, 64.7, 54.9,
49.0, 39.6, 36.9, 28.1, 27.6, 23.1, 20.5, 17.2, 14.7, 12.9.
Anal. calcd for C21H36N2O5: C, 63.61; H, 9.15; N, 7.06.
Found: C, 63.41; H, 9.00; N, 6.95.
1
mixture of anomers: H NMR (300 MHz, CDCl3) l
5.98 (d, J=10.3 Hz, 1H), 5.49–5.38 (m, 1H), 5.35–5.28
(m, 1H), 5.17 (d, J=5.5 Hz, 0.5H), 5.01 (d, J=5.1 Hz,
0.5H), 4.72 (dd, J=2.6, 9.9 Hz, 1H), 3.93 (dd, J=2.6,
7.7 Hz, 0.5H), 3.82 (dd, J=3.0, 7.7 Hz, 0.5H), 3.50–
3.10 (m, 2H), 3.33 (s, 1.5H), 3.30 (s, 1.5H), 3.06 (s, 3H),
2.05 (s, 1.5H), 2.03 (s, 1.5H), 2.01–1.85 (m, 2H), 1.63
(dd, J=1.5, 6.6 Hz, 3H), 1.60–1.40 (m, 3H), 1.56 (s,
4.5H), 1.49 (s, 4.5H), 1.12 (s, 3H), 0.95 (m, 3H). MS
(ESI) m/z 413 [M+1]+.
4.10. tert-Butyl (2R,3S,5R)-5-cyano-2-{(1R,2S)-2-
methoxy-2-methyl-1-(acetylamino)-pentyl}-3-[(1Z)-prop-
1-enyl]pyrrolidine-1-carboxylate, 23a
CAUTION: TMSCN is readily hydrolyzed to HCN.
Procedures such as described here must be carried out
in a well-ventilated fume hood with appropriate safety
precautions; see Ref. 29. A crude acetonitrile solution
of 22a (177.5 g solution, 23.83 g by assay, 48.5 mmol)
was chilled to −20°C. TMSCN (15.68 g, 158 mmol) was
added and the internal temp was lowered to −40°C.
TfOH (10.91 g, 72.7 mmol) was added over 30 min.
The reaction was stirred at −40°C for 2.5 h at which
point HPLC analysis of a sample indicated that the
reaction was complete. The reaction was quenched into
a mixture of 10% wt K2CO3 (400 mL) with NaOH (7.8
g) and i-PrOAc (400 mL); pH of aqueous ꢀ11.2. The
organic portion was washed with 23% wt NaCl solution
(200 mL), pH of aq ꢀ9.6; then with 23% wt NaCl
solution (200 mL) containing KH2PO4 (2.8 g) and
Na2HPO4·7H2O (7.1 g), pH of aq ꢀ6.2. After solvent
switching to i-PrOH, HPLC assay showed the desired
a-isomer 23a (16.49 g, 40.5 mmol, 83.5% yield) and
undesired b-isomer 24a (0.67 g, 1.6 mmol, 3.4% yield);
a:b ratio 24.6:1. 23a: 1H NMR (500 MHz, C6D6) l 5.77
(m, 1H), 5.38 (d, J=9.3 Hz, 1H), 5.28 (dq, J=13.7, 7.1
Hz, 1H), 4.86 (dd, J=1.3, 9.4 Hz, 1H), 4.07 (d, J=9.3
Hz, 1H), 4.04 (s, 1H), 3.56 (dd, J=8.1, 10.4 Hz, 1H),
2.79 (s, 3H), 2.12 (m, 1H), 1.98 (ddd, J=8.2, 9.3, 12.4
Hz, 1H), 1.63 (d, J=13.4 Hz, 1H), 1.60 (s, 9H), 1.55–
1.48 (m, 4H), 1.46 (s, 3H), 1.26 (m, 1H), 1.07 (m, 1H),
0.90 (t, J=7.1 Hz, 3H), 0.85 (s, 3H). 13C NMR (125
MHz, C6D6) l 168.5, 152.8, 132.7, 123.3, 120.5, 81.2,
79.3, 65.2, 54.2, 48.5, 46.8, 37.9, 37.4, 36.1, 28.4, 22.8,
4.9. tert-Butyl (2R,3S)-5-methoxy-2-{(1R,2S)-2-
methoxy-2-methyl-1-(acetylamino)pentyl}-3-[(1Z)-prop-
1-enyl]pyrrolidine-1-carboxylate, 22a
To a solution of 20 (50.0 g, 126 mmol) in THF (500
mL) at 5−10°C was added a THF solution of LiEt3BH
(1 M, 157 mL, 157 mmol) over ca. 30 min. After
stirring briefly, analysis of a sample showed consump-
tion of 20. MeOH (75 mL) was carefully added and the
mixture was then allowed to warm to rt and stir 30 min.
The reaction mixture was concentrated to ꢀ200 mL,
THF (500 mL) was added, and the solution reconcen-
trated to ꢀ100 mL. The solution was poured into 5%
wt KH2PO4 solution (750 mL) and extracted with i-
PrOAc (750 mL). The layers were separated and the
organic portion was washed with ꢀ12% wt NaCl solu-
tion (750 mL). HPLC assay showed 21a 48.42 g (96.6%
1
yield) in solution. ꢀ3:1 mixture of anomers: H NMR
(300 MHz, CDCl3) l 5.92 (d, J=9.6 Hz, 1H), 5.75 (dt,
J=1.8, 10.7 Hz, 1H), 5.45 (dd, J=1.8, 6.6 Hz, 0.75H),
5.35–5.41 (m, 1.25H), 4.59 (dd, J=1.8, 9.6 Hz, 0.25H),
4.52 (dd, J=2.6, 9.6 Hz, 0.75H), 4.18 (m, 1H), 3.90 (m,
0.25H), 3.80–3.70 (m, 1.5H), 3.61 (m, 0.25H), 3.21 (s,
2.25H), 3.20 (s, 0.75H), 2.35–2.20 (m, 1H), 2.00 (s,
2.25H), 1.95 (s, 0.75H), 1.85–1.60 (m, 3H), 1.64 (dd,
J=1.8, 7.0 Hz, 3H), 1.55 (s, 6.75H), 1.53 (s, 2.25H),
1.35–1.20 (m, 2H), 1.14 (s, 0.75H), 1.12 (s, 2.25H), 0.93
(t, J=7.3 Hz, 3H). MS (ESI) m/z 397 [M−1]−. Anal.
calcd for C21H38N2O5: C, 63.29; H, 9.61; N, 7.03; O,
20.07. Found C, 63.45; H, 9.63; N, 7.01; O, 19.88. The
solution was concentrated to ꢀ100 mL, diluted with
MeOH (500 mL) and reconcentrated (twice). MeOH
(90 mL) and CH3CH(OCH3)3 (50 mL) were added. The
solution was treated with camphorsulfonic acid (1.42 g,
6.1 mmol) and stirred at rt ca. 1 h (no 21a by TLC).
The reaction mixture was poured into 5% wt KHCO3
solution (800 mL) and extracted with i-PrOAc (1 L).
The organics were washed with 23% wt NaCl solution
(500 mL). The organics were concentrated to ꢀ150 mL
and diluted with CH3CN (500 mL). Repeated the con-
1
20.2, 17.1, 15.1, 12.8. 24a: H NMR (500 MHz, C6D6)
l 5.85 (d, J=9.8 Hz, 1H), 5.26 (m, 1H), 5.06 (dd,
J=2.3, 9.8 Hz, 1H), 4.92 (m, 1H), 4.09 (dd, J=4.3, 8.5
Hz, 1H), 3.83 (dd, J=2.5, 5.0 Hz, 1H), 3.53 (m, 1H),
2.80 (s, 3H), 2.03 (m, 1H), 1.92 (s, 3H), 1.88 (m, 1H),
1.55 (s, 9H), 1.48 (dd, J=1.7, 6.8 Hz, 3H), 1.42 (m,
1H), 1.30–1.18 (m, 3H), 0.96 (s, 3H), 0.91 (t, J=7.0 Hz,
3H). 13C NMR (125 MHz, C6D6) l 169.8, 153.0, 132.3,
124.1, 119.3, 81.2, 79.2, 65.0, 53.3, 48.1, 47.2, 37.7, 37.2,
36.1, 28.2, 23.4, 20.2, 16.9, 14.8, 12.7. Elimination side
product, 25. ꢀ1:1 N-Boc rotomers: 1H NMR (300
MHz, CDCl3) l 6.61 (dd, J=1.1 Hz, 4.0, 0.5H), 6.43
(dd, J=1.5 Hz, 4.4, 0.5H), 5.88 (d, J=9.9 Hz, 1H),
5.47–5.33 (m, 1H), 5.26 (t, J=10.6, 1H), 4.73 (dd,

3550
D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
J=2.9 Hz, 4.4, 0.5H), 4.68 (d, J=2.2 Hz, 0.5H), 4.64
(dd, J=2.9 Hz, 4.0, 0.5H), 4.60 (d, J=2.6 Hz, 0.5H),
4.29 (d, J=9.6 Hz, 0.5H) 4.13 (d, J=9.9 Hz, 0.5H),
4.09 (t, J=2.6, 0.5H), 3.95 (t, J=2.6 Hz, 0.5H), 3.19 (s,
1.5H), 3.15 (s, 1.5H), 1.96 (s, 1.5H), 1.95 (s, 1.5H),
1.77–1.60 (m, 2H), 1.68 (dd, J=1.5 Hz, 6.6, 3H), 1.55
(s, 4.5H), 1.49 (s, 4.5H), 1.44–1.20 (m, 2H), 1.14 (s,
1.5H), 1.15 (s, 1.5H), 0.95 (dt, J=1.1 Hz, 6.6, 3H). MS
(ESI) m/z 381 [M+1]+.
MHz, D6-DMSO) l 169.6, 168.3, 145.7, 137.6, 130.3,
128.0, 125.5, 124.8, 79.4, 70.0, 63.0, 58.9, 52.3, 48.7,
35.9, 33.8, 22.4, 21.4, 21.3, 20.8, 18.0, 15.1, 14.3, 12.7.
Anal. calcd for C27H44N2O7S: C, 59.97; H, 8.20; N,
5.18; O, 20.71; S, 5.93. Found: C, 59.91; H, 8.08; N,
5.15; S, 5.94. Diketopiperazine, 26. 1H NMR (300 MHz,
CD3OD) l 5.38 (s, 2H), 5.29 (dq, J=6.6, 10.6 Hz, 2H),
5.20 (ddq, J=8.8, 10.6, 1.5 Hz, 2H), 3.98 (d, J=8.8 Hz,
2H), 3.55 (t, J=7.7 Hz, 2H), 3.10 (s, 6H), 3.02 (dd,
J=7.0, 8.5 Hz, 2H), 2.83 (m, 2H), 2.19 (dt, J=12.5, 7.7
Hz, 2H), 1.80 (s, 6H), 1.48 (dd, J=1.5, 6.6 Hz, 6H),
1.45–1.20 (m, 6H), 1.12 (s, 6H), 1.10–0.95 (m, 4H), 0.77
(t, J=7.0 Hz, 6H). MS (ESI) m/z 617.6 [M+1]+.
4.11. Isopropyl (2R,4S,5R)-5-[(1R,2S)-1-(acetylamino)-
2-methoxy-2-methylpentyl}-4-[(1Z)-prop-1-enyl]-
pyrrolidine-2-carboxylate, A-322278·TsOH 2·TsOH
An i-PrOH solution of crude 2-cyanopyrrolidines (231
g solution containing 23a (46.7 g, 114.7 mmol) and 24a
(2.1 g, 5.2 mmol)) was chilled to 5−30°C with a dry
ice/MeOH bath and carefully treated with anhydrous
HCl (g) at such a rate to maintain an internal temp of
<0°C until the exotherm ceased (ca. 100 g HCl added).
After the addition was complete, the reaction was
allowed to warm while stirring to ꢀ+15°C. HPLC
analysis showed that starting materials 23a/24a and
intermediate 29 had been consumed. The reaction mix-
ture was diluted with i-PrOH (235 g) and warmed to
reflux for 8 h. After cooling to rt, the reaction was
concentrated in vacuo to about one-half volume. It was
replenished with i-PrOAc and reconcentrated. The mix-
ture was poured into 15% wt KHCO3 solution (690 g)
and the reaction flask rinses, i-PrOAc (250 g) and water
(50 g), were also added. After through mixing, the
layers were separated. The organic portion was washed
with 15% wt KHCO3 solution (380 g) and the com-
bined aqueous portions were back-extracted with i-
PrOAc (160 g). The combined organics were washed
with 20% wt NaCl solution (300 g); KF of the organic
portion was ꢀ2.4% wt water. The organics were
azeotropically dried with i-PrOAc (to KF <0.05% wt
water). The mixture was filtered to removed precipi-
tated inorganic salts and the filtrate was assayed by
HPLC for 2 (37.2 g, 101 mmol, 88% yield) and the a/b
ratio was 12.0:1. The filtrate solution was diluted with
i-PrOAc to give 1500 g solution and then it was
warmed to 50–60°C. TsOH·H2O (24.07 g, 126.6 mmol)
was dissolved in i-PrOAc (256 g) by warming to 50–
60°C; this solution was added to the solution of 2
through an in-line filter. Within minutes, a white pre-
cipitate began to form; after ca. 1 h at 50–60°C, the
mixture was cooled to rt at 5°C/h and allowed to stir at
rt several h. The product was collected by filtration and
dried in vacuo to provide 2·TsOH (53.78 g, 96.7%
potency, 96.2 mmol, 84% yield from 23a). mp 207–
208°C. [h]2D5 −31.4 (c 1.04, MeOH). 1H NMR (300
MHz, D6-DMSO) l 9.08 (br s, 1H), 8.74 (br s, 1H),
7.66 (d, J=10.3 Hz, 1H), 7.48 (d, J=8.1 Hz, 2H), 7.12
(d, J=8.1 Hz, 2H), 5.45 (dq, J=11.0, 6.6 Hz, 1H), 5.28
(ddq, J=8.8, 11.0, 1.5 Hz, 1H), 5.01 (spt, J=6.3 Hz,
1H), 4.40–4.28 (m, 2H), 3.60 (t, J=9.6 Hz, 1H), 3.36
(br s, 1H), 3.20–3.10 (m, 1H), 3.13 (s, 3H), 2.43 (dt,
J=12.9, 8.1 Hz, 1H), 2.29 (s, 3H), 1.79 (s, 3H), 1.60–
1.40 (m, 2H), 1.52 (dd, J=1.5, 6.6 Hz, 3H), 1.30–1.10
(m, 2H), 1.25 (d, J=6.3 Hz, 3H), 1.23 (d, J=6.3 Hz,
3H), 1.18 (s, 3H), 0.80 (t, J=7.0 Hz, 3H). 13C NMR (75
Acknowledgements
Mr. Kenneth Engstrom prepared samples of racemic
intermediates for assay development and Ms. Rebecca
Dorbin, a summer intern student, assisted in the cou-
pling reaction model studies. Dr. Kent Stewart and Dr.
Tetsuro Oie performed the Mannich reaction calcula-
tions. Mr. Ravi Kurukulasariya, Dr. Jianjun Jiang and
Dr. Anthony Haight prepared some intermediates for
the team. Mr. Michael Fitzgerald developed and per-
formed many of the HPLC analytical measurements
described, and Mr. Steven Glenn and Dr. James Mor-
ley of Abbott Laboratories Pharmaceutical Analytical
R&D area provided additional analytical chemistry
support. Dr. Steven Hollis, Dr. Ian Marsden, and Dr.
Casey Zhou measured and interpreted many of the
NMR spectra reported. Dr. David DeGoey, Dr. Hui-Ju
Chen, and Dr. Larry Klein pioneered the synthetic
strategy and engaged in useful discussions.
References
1. Service, R. F. Science 1997, 275, 756.
2. Centers for Disease Control and Prevention Website.
http://www.cdc.gov/ncidod/diseases/flu/fluinfo.htm (acc-
essed Jan., 2003).
3. Centers for Disease Control and Prevention Website.
Morbidity and Mortality Weekly Report (MMWR) 2000,
49,
RR-3,
1.
http://www.cdc.gov/mmwr/preview/
mmwrhtml/rr4903a1.htm (accessed Dec 2001).
4. Maring, C.; McDaniel, K.; Krueger, A.; Zhau, C.; Sun,
M.; Madigan, D.; DeGoey, D.; Chen, H.-J.; Yeung, M.
C.; Flosi, W.; Grampovnik, D.; Kati, W.; Klein, L.;
Stewart, K.; Stoll, V.; Saldivar, S.; Montgomery, D.;
Carrick, R.; Steffy, K.; Kempf, D.; Molla, A.; Kohlbren-
ner, W.; Kennedy, A.; Herrin, T.; Xu, Y.; Laver, W. G.
Presented at 14th International Conference on Antiviral
Research, Antiviral Research 2001, 50, A76; Abstract 129.
5. DeGoey, D. A.; Chen, H.-J.; Flosi, W. J.; Grampovnik,
D. J.; Yeung, C. M.; Klein, L. L.; Kempf, D. J. J. Org.
Chem. 2002, 67, 5445.
6. A unique asymmetric total synthesis of 1 has recently
been accomplished in the laboratories of Professor
Stephen Hanessian. Hanessian, S.; Bayrakdarian, M.;
Luo, X. J. Am. Chem. Soc. 2002, 124, 4716.

D. M. Barnes et al. / Tetrahedron: Asymmetry 14 (2003) 3541–3551
3551
and
7. Reviews: (a) Casiraghi, G.; Zanardi, F.; Appendino, G.;
Rassu, G. Chem. Rev. 2000, 100, 1929; (b) Rassu, G.;
Zanardi, F.; Battistini, L. Casiraghi, G. Chem. Soc. Rev.
2000, 29, 109; (c) Rassu, G.; Zanardi, F.; Battistini, L.;
Casiraghi, G. Synlett 1999, 1333; (d) Casiraghi, G.;
Rassu, G. Synthesis 1995, 607.
type
products,
i.e.
(Z,Z)-,
(Z,E)-,
(E,E)-2,4-hexadienes at both temperatures and (Z)- and
(E)-1-trimethylsilyl-1-propene at −20 to −25°C were vari-
able but measurable side products and are largely respon-
sible for the range in observed rate differences, but not
the enhanced selectivity.
8. Yabuuchi, T.; Kusumi, T. Chem. Pharm. Bull. 1999, 47,
684.
9. Casiraghi, G.; Rassu, G.; Spanu, P.; Pinna, L. J. Org.
Chem. 1992, 57, 3760.
10. Bocchi, V.; Chierici, L.; Gardini, G. P.; Mondelli, R.
Tetrahedron 1970, 26, 4073.
23. Decomposition of triphenylmethylsulfenyl amine contain-
ing compounds (such as 3 and 22b) was observed when
solutions had been left standing at ambient temperature.
24. Kessler, W.; Iselin, B. Helv. Chim. Acta 1966, 49, 1330.
25. Brown, H. C.; Kabalka, G. W. J. Am. Chem. Soc. 1970,
92, 714.
11. Pellegata, R.; Pinza, M.; Pifferi, G. Synthesis 1978, 614.
12. For a full account of the preparation of the TBSOP
intermediate, see: Tian, Z.; Rasmussen, M. W.; Witten-
berger, S. J. Org. Process Res. Dev. 2002, 6, 416.
13. For a full account of the details of this reaction and
associated studies, see: Barnes, D. M.; McLaughlin, M.
A.; Oie, T.; Rasmussen, M. W.; Stewart, K. D.; Witten-
berger, S. J. Org. Lett. 2002, 4 (9), 1427.
26. After completing the campaign, an HPLC assay was
developed to measure the (Z)- and (E)-isomer ratio at
compound 19. Following the triethylborane experiment,
the samples were treated with trifluoroacetic acid to
remove the N-Boc group and then assayed.
27. Tanaka, K.; Sawanishi, H. Tetrahedron 1998, 54, 10029.
28. Other Lewis acids screened with 22b include: Zn2Cl2,
TiCl4, Et2AlCl, SnCl4, BF3·OEt2, TMSOTf, and
Et2AlCN; solvents included CH2Cl2, toluene, THF, and
CH3CN. See also Ref. 5.
14. (Z)-1-Bromo-1-propene or cis-1-bromo-1-propene [590-
13-6] is available in 97% purity (3% (E)-isomer) from the
Aldrich Chemical Company.
29. CAUTION: It is worth noting the safety precautions
taken when conducting this reaction. The reaction and
work up were executed with a nitrogen sweep through the
headspace; the outlet was connected to a caustic (NaOH)
scrubber solution with an intervening trap to protect
against the scrub solution backing-up into the vessel.
After the reaction, the mixture was quenched into a 10%
wt potassium carbonate solution containing sodium
hydroxide (ca. 1.4 equiv. versus TMSCN used) to main-
tain a pH >11 (the pKa of HCN is 9.2, the pH of a 0.1N
solution of HCN is 5.1). The aqueous waste steam and
scrubber solutions were treated with bleach (sodium
hypochlorite) to oxidize excess cyanide (see Lunn, G.;
Sanaone, E. B. Destruction of Hazardous Chemicals in the
Laboratory; John Wiley & Sons: New York, 1994, 133–
138.) The cyanide levels were tested with ion specific test
strips (EM QUANT by E-Merck, item number EM-
10044-1 from VWR Scientific Products) prior to disposal
to insure complete cyanide ion oxidation.
30. Maring, C. J.; Giranda, V. L.; Kempf, D. J.; Stoll, V. S.;
Sun, M.; Zhao, C.; Gu, Y. G.; Hanessian, S.; Wang, G.
T.; Krueger, A. C.; Chen, H.-J.; Chen, Y.; DeGoey, D.
A.; Flosi, W. J.; Grampovnik, D. J.; Kati, W. M.;
Kennedy, A. L.; Klein, L. L.; Lin, Z.; Madigan, D. L.;
McDaniel, K. F.; Muchmore, S. W.; Sham, H. L.; Stew-
art, K. D.; Tu, N. P.; Wagenaar, F. L.; Wang, S.;
Wiedeman, P. E.; Xu, Y.; Yeung, M. C.; Bayrakdarian,
M.; Luo, X. Preparation of pyrrolidine neuraminidase
inhibitors. PCT Int. Appl. (2001), WO Pat. Appl.
2001028996, April 26, 2001.
15. Linstrumelle, G.; Alami, M. In Encyclopedia of Reagents
for Organic Synthesis; Paquette, L. A., Ed.; John Wiley &
Sons: New York, 1995; p. 4321.
16. Mechin, B.; Naulet, N. J. Organomet. Chem. 1972, 39,
229.
17. The isomeric purity of the 1-propenylmagnesium bromide
reagent was determined by reaction with benzaldehyde
and measurement of the ratio of (Z)- and (E)-carbinol
1
products by HPLC and H NMR.
18. Lipshutz, B. H.; Koerner, M.; Parker, D. A. Tetrahedron
Lett. 1987, 28, 945.
19. Accurate assessment of the ratio of 3 to 17 was a difficult
measurement. Higher levels of 17 (]2%) were judged by
proton-NMR integration. For more precise determina-
tions of lower levels, a SFC-HPLC assay on a proprietary
chiral stationary phase was developed.
20. Although with these ratios of CuBr·SMe2 (0.5 equiv.) and
(Z)-CH3CHꢀCHMgBr (3 equiv.) higher order cuprate
species are expected to be present in the reaction mixture,
we observed similar results when 4 was exposed to an
excess of [(Z)-1-propenyl]2CuMgBr. See Lipshutz, B. H.;
Sengupta, S. Org. React. 1992, 41, 135 for a comprehen-
sive review of organocopper reagents and Lipshutz, B.
H.; Wilhelm, R. S.; Kozlowski, J. A. Tetrahedron 1984,
40, 5005 for a review of higher order organocuprates.
21. Kant, J. J. Org. Chem. 1993, 58, 2296.
22. The increase in (Z)-propenyl product 3 was shown not to
be due to selective isomerization or destruction of the
(E)-1-propenyl moiety under the reaction conditions. The
mass balance of the (Z)- and (E)-1-propenyl ligands was
accounted for by considering products 3 and 17 and by
an iodine quench of cuprate reactions (to react with
remaining organocopper species) to give (Z)- and (E)-1-
propenyl iodide which were assayed by HPLC. Wurtz-
31. HCl addition product 28 was identified using HPLC/MS
techniques. Sufficient quantities for full characterization
were not available and the regiochemistry of the HCl
addition is unassigned.