Protecting-group-free catalytic asymmetric total synthesis of (-)-rosmarinecine

Article abstract of DOI:10.1016/j.tet.2012.06.095

The protecting-group-free asymmetric total synthesis of (-)-rosmarinecine was achieved in only four steps from the commercially available (±)-3-hydroxypyrrolidine hydrochloride (2a). The key steps include the direct oxidation of (±)-2a to (±)-3-hydroxy-1-pyrroline N-oxide (1a) using the Davis reagent and the domino reaction; viz., the lipase-catalyzed dynamic kinetic resolution of (±)-1a with 1-ethoxyvinyl ethyl maleate followed by the intramolecular [3+2] dipolar cycloaddition reaction of the generated optically active ester. Some insights into the mechanism of the racemization of the optically active 1a, observed during the enzymatic process, were also obtained.

Full text of DOI:10.1016/j.tet.2012.06.095

Tetrahedron 68 (2012) 7295e7301
Contents lists available at SciVerse ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Protecting-group-free catalytic asymmetric total synthesis of (ꢀ)-rosmarinecine
,
,
*
Hiroyuki Nemoto a, Kouichi Tanimoto b, Yukiko Kanao b, Sohei Omura b, Yasuyuki Kita b c, Shuji Akai a
a School of Pharmaceutical Sciences, University of Shizuoka, 52-1, Yada, Suruga-ku, Shizuoka, Shizuoka 422-8526, Japan
b Graduate School of Pharmaceutical Sciences, Osaka University, 1-6, Yamadaoka, Suita, Osaka 565-0871, Japan
c College of Pharmaceutical Sciences, Ritsumeikan University, 1-1-1 Nojihigashi, Kusatsu, Shiga 525-8577, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The protecting-group-free asymmetric total synthesis of (ꢀ)-rosmarinecine was achieved in only four
steps from the commercially available (ꢁ)-3-hydroxypyrrolidine hydrochloride (2a). The key steps in-
clude the direct oxidation of (ꢁ)-2a to (ꢁ)-3-hydroxy-1-pyrroline N-oxide (1a) using the Davis reagent
and the domino reaction; viz., the lipase-catalyzed dynamic kinetic resolution of (ꢁ)-1a with 1-
ethoxyvinyl ethyl maleate followed by the intramolecular [3þ2] dipolar cycloaddition reaction of the
generated optically active ester. Some insights into the mechanism of the racemization of the optically
active 1a, observed during the enzymatic process, were also obtained.
Received 8 June 2012
Accepted 24 June 2012
Available online 29 June 2012
Keywords:
Total synthesis
Protecting-group free
Rosmarinecine
Ó 2012 Elsevier Ltd. All rights reserved.
Dynamic kinetic resolution
[3þ2] Dipolar cyclization
Lipase
1. Introduction
sophisticated transformations using the intrinsic reactivity of
functional groups and the consecutive series of multi-reactions, the
Protecting groups have been widely used in organic synthesis to
suppress side reactions initiated by reactive functional groups.1
However, their use adds two extra steps (i.e., protection and
deprotection) to the synthetic sequence, and they sometimes pro-
duce problems in their removal and unexpected side reactions.2 For
example, it was reported that the removal of the THP group from
(S)-nitrone 1b (99% ee) caused partial or total loss of its optical
purity (Scheme 1).3
so-called domino reaction. In particular, the domino reactions en-
able the construction of complex multi-cyclic molecules from sin-
gle acyclic precursors or relatively simple moieties via highly
reactive intermediates.
Recently, we developed the lipase-catalyzed domino reactions
of racemic dienols with 1-ethoxyvinyl esters having dienophilic
moieties. In these reactions, the acyl moieties installed during the
enzymatic kinetic resolution were utilized as a part of the constit-
uent structure for the subsequent intramolecular DielseAlder re-
actions to directly give optically active cyclic compounds with
multi-stereogenic carbon centers.5 We subsequently extended
this methodology to the multi-substituted isoxazolidine synthesis
via an intramolecular [3þ2] dipolar cycloaddition reaction and
achieved the asymmetric total synthesis of (e)-rosmarinecine,
a typical necine-base-portion of some natural alkaloids, such as
rosmarinine,6 in 6 steps.7 This synthesis was efficient and featured
the fewest steps among the reported total syntheses;3a,c,6,8 how-
ever, it required a protecting group for preparing the key in-
termediate (ꢁ)-1a from the racemic 3-hydroxypyrrolidine (ꢁ)-2a.
We now report the direct oxidation of (ꢁ)-2a$HCl to (ꢁ)-1a and
the four-step protecting-group-free asymmetric total synthesis of
(e)-rosmarinecine. We also obtained some important insights into the
mechanism of the racemization of the optically active 1a, which was
observed during the enzymatic process. The full details are as follows.
Scheme 1. Racemization during the removal of the THP group of (S)-1b.
Under such circumstances, the protecting-group-free total
synthesis protocol has gained increasing attention and has ex-
panded its applicability to a wider range of compounds.4 The keys
to accomplishing the protecting-group-free total synthesis include
* Corresponding author. E-mail address: akai@u-shizuoka-ken.ac.jp (S. Akai).
0040-4020/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tet.2012.06.095

7296
H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
Table 1
2. Results and discussion
Optimization of the oxidation of 2a and 2ca
2.1. Protecting-group-free asymmetric total synthesis of
(e)-rosmarinecine
In our total synthesis of (e)-rosmarinecine,7 the preparation of
(ꢁ)-1a from the commercially available (ꢁ)-2a required 3 steps;
viz., the TBS protection of the hydroxyl group, the tungsten-
catalyzed oxidation of the secondary amino functionality,9 and
the removal of the TBS group (Eq. 1). This was because the direct
oxidation of (ꢁ)-2a to (ꢁ)-1a using conventional oxidants, such as
NaWO4eH2O2,9 mCPBA,10 MeReO3eH2O2,11 and dimethyldioxir-
ane,12 resulted in the formation of complex mixtures without
forming (ꢁ)-1a.
Entry
2
Additive
Solvent
NMR yield (%)b
1
4
1
2
3
2c
2c
2c
None
THF
1c <20 4c Trace
None
THFeH2O (50:1)
THFeH2O (50:1)
THFeH2O (50:1)
1c
1c
1c
18 4c Trace
28 4c 10
51 4c 11
iPr2NH (1 equiv)
4
2cꢃHCl iPr2NH (1 equiv)
5c
6
2cꢃHCl Pyridine (1 equiv) THFeH2O (50:1)
1c n.d.d 4c n.d.d
2cꢃHCl tBuNH2 (1.5 equiv) THFeH2O (50:1)
1c
1c
1c
69 4c 20
67 4c 24
68 4c 24
65 4c 26
75 4c 25
76 4c 24
We extensively reinvestigated the direct oxidation of (ꢁ)-2a to
find that the use of the Davis reagent 313 in CHCl3 at room tem-
perature (rt) for 30 min gave a complex mixture including (ꢁ)-1a
(16% NMR yield) and its regioisomer (ꢁ)-4a (14% NMR yield). A
similar reaction in THF provided a slightly better yield (22% NMR
yield) of (ꢁ)-1a (Eq. 2). Although similar oxidations of pyrrolidines
with 3 to give nitrones have been reported,14 there have been no
reports on the direct oxidation of secondary amines, including
pyrrolidines, having a hydroxyl group at the proximal position, to
the best of our knowledge.15
7
8
9
10
11
12
13
2cꢃHCl tBuNH2 (2 equiv)
THFeH2O (100:1)
2cꢃHCl tBuNH2 (2.5 equiv) THFeH2O (10:1)
2cꢃHCl tBuNH2 (2.5 equiv) TolueneeH2O (10:1) 1c
2cꢃHCl tBuNH2 (2.5 equiv) CH2Cl2eH2O (10:1) 1c
2cꢃHCl tBuNH2 (2.5 equiv) MeCNeH2O (10:1)
2cꢃHCl Aq NH3 (2.5 equiv) MeCNeH2O (10:1)
2aꢃHCl tBuNH2 (2.5 equiv) MeCNeH2O (10:1)
1c
1c 73e 4c 21e
1a 52e 4a 24e
a
Reaction conditions: an additive and 3 (2.5 equiv) were added to a 0.1 M so-
lution of 2, and the reaction mixture was stirred at rt for 30 min.
b
NMR yield determined by 1H NMR analysis using 1,1,2,2-tetrachloroethane as
the internal standard.
c
2 equiv of 3 was used.
Not detected.
Isolated yield.
d
e
were obtained under these conditions (Entries 10e12), which were
better than those of the related reactions.14 Finally, these results
were successfully fed back to the oxidation of the hydrochloride
(ꢁ)-2a$HCl to produce (ꢁ)-1a in 52% isolated yield along with
(ꢁ)-4a (24% isolated yield) (Entry 13). Our next attention focused
on the protecting-group-free asymmetric total synthesis of the
natural alkaloid, (e)-rosmarinecine. We examined the lipase-
catalyzed domino reaction of (ꢁ)-1a according to our model
study using an acyclic nitrone (for details, see: Ref. 7). Thus,
a mixture of (ꢁ)-1a, an acyl donor 6A having a (Z)-olefin moiety,
and commercially available immobilized Candida antarctica lipase,
B (CAL-B) was stirred in some typical organic solvents for the
lipase-catalyzed transesterification, such as toluene, THF, and iPr2O
at 30 ꢂC. The enzymatic kinetic resolution of (ꢁ)-1a gave an opti-
cally active ester (R)-7A, which immediately caused a intra-
molecular [3þ2] cycloaddition reaction. We were delighted to
With this faint hope in hand, we initiated the optimization of
the reaction conditions using 3; however, the high polarity and
water solubility of both 2a and 1a made their purification and the
evaluation of the reactions very laborious. Therefore, the TBS de-
rivative (ꢁ)-2c was employed as a surrogate of (ꢁ)-2a. Initially,
(ꢁ)-2c was treated with 3 (2.5 equiv) in THF at rt for 1 h to give
(ꢁ)-1a in less than 20% NMR yield (Table 1, Entry 1). The low yield
of (ꢁ)-1a was thought to be in part due to the amine exchange
reaction between (ꢁ)-2c and N-(phenylsulfonyl)benzimine, de-
rived from 3 during the oxidation reaction.14e Therefore, a similar
reaction was conducted in the presence of water (THFeH2O, 50:1);
however, no significant improvement was observed (Entry 2). The
obtain the expected tricyclic compound 8A as
a single di-
astereomer. However, the optical purities were at most 60% ee, and
the enantioselectivities, E values,16 of the kinetic resolution were
less than 10 (Table 2, Entries 1e3). The use of more polar solvents,
such as acetone and MeCN, produced better enantioselectivities
(Entries 4 and 5). In all these cases, the recovered amounts of the
less reactive enantiomer, (S)-1a, were much less than the theoret-
ical ones. This was probably due to the decomposition of 1a and/or
the side reactions, such as the intermolecular dipolar cycloaddition
of 1a to 6A. This problem was found to be solved by performing the
reactions at a lower temperature. Thus, the reaction in MeCN at 0 ꢂC
gave 8A (92% ee, 53% yield) and (S)-1a (73% ee, 43% yield) (Entry 7).
In a similar manner, the reaction of (ꢁ)-1a with the ethyl ester 6B
afforded 8B with a slightly better enantioselectivity (94% ee, 49%
yield) (Entry 8).
We noticed that the partial racemization of (S)-1a occurred
while the enzymatic domino reaction proceeded. The optical purity
of the recovered (S)-1a should be almost the same as that of the
products (8A and 8B) at about a 50% conversion; however, it was
always lower (Entries 7 and 8). The optically active 1a was reported
to cause the racemization,3 and a similar racemization must take
i
addition of 1 equiv of Pr2NH to the reaction slightly improved the
yield (Entry 3). After a significant number of unsuccessful trials,
a dramatic improvement in the yield of (ꢁ)-1c (51%) was observed
using the hydrochloride 2c$HCl as a substrate (Entry 4). To explore
the effect of the additive base, some other amines were examined.
t
The use of pyridine caused no reaction (Entry 5), while BuNH2
proved to be one of the most effective bases to produce (ꢁ)-1c (69%
NMR yield) with (ꢁ)-4c (20% NMR yield) (Entry 6). It is noteworthy
that the competitive oxidation of the additive amines was scarcely
observed in these reactions, and 2c was preferentially oxidized. The
ratio of water to the organic solvent had little effect on the yields
(Entries 6e8), and MeCN was the solvent of choice (Entries 8e11).
Aqueous ammonia also effectively provided (ꢁ)-1c and (ꢁ)-4c in
73% and 21% isolated yields, respectively (Entry 12). It is notewor-
thy that almost quantitative yields of mixtures of (ꢁ)-1c and (ꢁ)-4c

H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
7297
Table 2
Lipase-catalyzed domino reactions of (ꢁ)-1a and 6A, Ba
Entry
6
Solvent Reaction
conditions
8
(S)-1a
ee, % Yield, % ee, % Yield, %
E valueb
1
2
3
4
5
6
7
8
6A Toluene 30 ꢂC, 12 h 8A 50
65
64
45
21
27
44
53
49
69
60
60
58
ee
d
Trace
Trace
Trace
21
9.4
7.8
6.4
6A iPr2O
6A THF
30 ꢂC, 24 h 8A 49
30 ꢂC, 12 h 8A 60
d
6A Acetone 30 ꢂC, 12 h 8A 82
d
12.5
23.7
14.1
>100
>100
6A MeCN
30 ꢂC, 12 h 8A 89
d
Trace
52
6A Toluene 10 ꢂC, 12 h 8A 77
24
73
77
d
6A MeCN
6B MeCN
6B MeCN
6B MeCN
0
0
ꢂC, 12 h 8A 92
ꢂC, 8 h
8B 94
43c
33c
9d
10d
10 ꢂC, 17 h 8B 86
5 ꢂC, 12 h 8B 91
5 ꢂC, 11 h 8B 90
5 ꢂC, 11 h 8B 92
Trace
28
30
96
99
95
11d,e 6B MeCN
12d,e 6B MeCN
23
a
Reaction conditions: 6 (1.5 equiv) was added to a 0.1 M solution of (ꢁ)-1a
(1.0 equiv) and commercial immobilized CAL-B [150 wt % to (ꢁ)-1a] in an organic
solvent, and the reaction mixture was stirred at the indicated temperature for the
indicated time.
b
E value based on the yield and optical purity of 8 (for calculation, see: Ref. 16).
c
Yield determined by 1H NMR analysis.
d
300 wt % of CAL-B to (ꢁ)-1a was used.
e
The reaction was carried out by the one-pot procedure, for details, see the ex-
perimental section: substrate amount was 40 mg for Entry 11 and 1.0 g for Entry 12.
Scheme 2. Four-step protecting-group-free asymmetric total synthesis of (ꢀ)-ros-
marinecine 10. Reagents and conditions: (a) 3, tBuNH2, MeCN/H2O (10:1), rt, 0.5 h
(52%); (b) 6A or 6B, CAL-B, MeCN, see Table 2; (c) H2, Pd(OH)2, MeOH, rt, 20 h (55%); (d)
Red-Al, THF, reflux (90%).
place during the enzymatic reaction. Taking advantage of its nature,
we aimed at a dynamic kinetic resolution. After changing the
temperature, the amount of the lipase, and some additives, the
reaction using CAL-B (300 wt %) at 10 ꢂC for 17 h afforded 8B (86%
ee, 69% yield) (Entry 9) and that at 5 ꢂC for 12 h gave 8B (91% ee, 60%
yield) (Entry 10) (Note: the theoretical maximum optical purity of
8B by the standard kinetic resolution at 69% and 60% conversions
are 45% ee and 67% ee, respectively).
The above-mentioned studies were conducted using pure 6A
and 6B, while the more convenient one-pot procedure, including
the preparation of 6B from 5B and the enzymatic domino reaction
of (ꢁ)-1a, also gave 8B with the same optical and chemical yields
(Entry 11). This method was suitable for the multigram synthesis of
the chiral products. For instance, under identical reaction condi-
tions using the same ratio of (ꢁ)-1a, 5B, CAL-B, and the solvent, the
reactions of (ꢁ)-1a (40 mg and 1.0 g) afforded similar results (En-
tries 11 and 12).
2.2. Mechanistic elucidation of the racemization of the
nitrone 1a
To account for the racemization of (S)-1a, observed in the en-
zymatic domino reactions (vide supra), the three following mech-
anisms seem to be plausible (Scheme 3): The tautomerization of
(S)-1a to N-hydroxyenamine 11 followed by its reverse reaction
(path a),3d,18 the formation of the epoxide 12 followed by its
opening and dehydration (path b),3b and the elimination of the
hydroxyl group followed by the conjugate addition of water to the
pyrrolinium salt 13 (path c). In order to elucidate the possibility of
these candidates, optically pure 1a, 1d, and 1e were treated with
some reagents in MeCN.19
A recrystallization of 8B (92% ee) from a mixture of hexane and
EtOAc afforded the optically pure 8B [43% yield from (ꢁ)-1a]. The
hydrogenation of its NeO bond and subsequent Red-Al reduction3c
afforded (ꢀ)-rosmarinecine 10 in 50% yield for 2 steps. The physical
properties and the 1H and 13C NMR data of the synthesized 10
showed good agreement with those of the natural compound: mp
167e170 ꢂC, ½a 2D4
ꢄ
ꢀ117.7 (c¼0.98, EtOH) [lit.6 mp 168e170 ꢂC,
½ ꢄ ꢀ119.8 (c¼1.01, EtOH)] (Scheme 2).
a 2D1
Although the asymmetric total syntheses of (ꢀ)-10 had already
been reported by three groups, all of them used optically active
substrates3a,c,6,8 as either the starting material or a chiral auxiliary.
Tatsuta et al. succeeded in the first total synthesis of optically pure
(ꢀ)-10 in 18 steps and 7% overall yield from a
D-glucosamine
derivative.8a Denmark’s approach included their original sequen-
tial [4þ2]/[3þ2] cycloaddition of a nitroealkene to an olefin
having an optically active auxiliary to provide (ꢀ)-10 in 10 steps
and 9% overall yield from maleic anhydride.6 Goti’s route took 11
steps to afford (ꢀ)-10 in 9% yield from the commercially available
(S)-malic acid diethyl ester (>99% ee), in which they required
a temporary protection of the nitrone moiety of (S)-1a to avoid its
racemization (Scheme 1).3 Our method features the synthesis of
the optically pure (ꢀ)-10 from the commercially available race-
mate 2a$HCl, the fewest steps (4 steps), and the highest overall
yield (11.1%).17
Scheme 3. Possible mechanisms of the racemization of (S)-1a.
(S)-1a (99% ee) was first treated with CAL-B (300 wt %);
however, it remained intact at rt after 24 h (Table 3, Entry 1). The
ethoxyvinyl ester 6B (1 equiv) did not cause any reaction at 5 ꢂC
for 18 h (Entry 2), while a gradual decomposition of (S)-1a was
observed at rt.20 The monoethyl maleate 5B, which could be

7298
H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
Table 3
3. Conclusion
Investigation of the racemization of optically pure 1a, 1d, and 1ea
The protecting-group-free catalytic asymmetric total synthesis
of the natural alkaloid, (ꢀ)-rosmarinecine 10, was achieved, in
which the direct oxidation of 2a$HCl and the lipase-catalyzed
domino process were effectively employed as the key steps. Step
economy is another increasingly important measure of efficiency.21
Our revised route has accomplished the total synthesis of 10 in only
4 steps from the commercially available racemic 2a∙HCl, which is
the shortest and most efficient among the reported total
syntheses.3,6,8
Entry
1
Reagents
Reaction Products
time
In this study, we have discovered that the racemization-prone
nature of 1a allows the development of the dynamic kinetic reso-
lution, whereas such a nature sometimes causes serious problems
in the traditional chiral pool synthesis and requires extra
protection-deprotection steps. We have also proposed a plausible
mechanism for the racemization of 1a. The improvement of the
racemization process to achieve the more effective dynamic kinetic
resolution is currently in progress in our laboratory.
1
2
3
4
(S)-1a
(S)-1a
(S)-1a
(S)-1a
CAL-B (300 wt %)
6B (1 equiv)c
5B (1 equiv)
CAL-B (300 wt %), 24 h
5B (1 equiv)
CSA (1 equiv)
DBU (1 equiv)
CSA (1 equiv),
CD3OD (10 equiv)
24 h
18 h
24 h
(S)-1a (99% ee, 99%b)
(S)-1a (99% ee, 99%)
(S)-1a (99% ee, 32%b)
(S)-1a (99% ee, 30%b)
5
6
7
(S)-1a
(S)-1a
(S)-1a
24 h
24 h
24 h
(S)-1a (60% ee, 8%b)
(S)-1a (99% ee, 82%b)
(S)-1a (91% ee, 24%b)
(S)-1d0 (R¼CD3, 0% ee, 37%b)
(R)-1d (76% ee, 83%)
8
(R)-1d CSA (1 equiv)
MeOH (10 equiv)
24 h
24 h
3 h
4. Experimental section
4.1. General
9
(R)-1d CSA (1 equiv)d
1a (0% ee, 6%b)
(R)-1d (99% ee, 79%)
(R)-1a (99% ee, 90%b)
1d (0% ee, ꢅ5%)
10
11
(R)-1a CSA (1 equiv)
MeOH (10 equiv)
(R)-1e CSA (1 equiv)
MeOH (10 equiv)
3 h
1d (0% ee, 40%)
Melting points were determined by a Yazawa micro melting
point apparatus and are uncorrected. The infrared (IR) absorption
spectra were recorded by a SHIMADZU IRPrestige-21 spectropho-
tometer. The 1H and 13C NMR spectra were measured by a JEOL
JNM-AL 300 (1H: 300 MHz, 13C: 75 MHz), JEOL JMN-A500 (1H:
500 MHz, 13C: 125 MHz) or JNM-ECA500 (1H: 500 MHz, 13C:
125 MHz) instrument with chemical shifts reported in ppm relative
to the residual deuterated solvent or the internal standard tetra-
methylsilane. The mass spectra were measured on a Bruker
micrOTOF, JEOL JMS-700 MStation, or JMS-T100TD spectrometer.
The HPLC analyses were carried out using a JASCO LC-2000Plus
system (HPLC pump: PU-2080, UV detector: MD-2010). The yield
refers to the isolated yields of compounds greater than 95% purity
as determined by the 1H NMR analysis. The 1H NMR and melting
points (where applicable) of all known compounds were taken. All
new products were further characterized by elemental analysis or
(R)-1e (99% ee, 60%)
a
Reaction conditions: reagent(s) was added to a 0.05 M solution of the optically
pure substrate 1 in MeCN, and the reaction mixture was stirred at rt for the indicated
period.
b
The yield determined by 1H NMR analysis using 1,1,2,2-tetrachloroethane as the
internal standard.
c
The reaction was conducted at 5 ꢂC.
The reaction was conducted in MeCNewater (10:1).
d
generated from 6B during the enzymatic reaction, without or
with CAL-B, did not promote the racemization, but caused the
decomposition of 1a (Entries 3 and 4).
Because we could not obtain any useful information about the
effect of the reagents present in the reaction mixture on the ra-
cemization of (S)-1a, we next examined harsher conditions. In the
presence of camphorsulfonic acid (CSA) (1 equiv), the racemiza-
tion gradually proceeded at rt, while the decomposition of 1a also
took place (Entry 5). On the other hand, no racemization took
place with DBU (Entry 6). The reaction of (S)-1a with CSA
(1 equiv) and CD3OD (10 equiv) gave a mixture of (S)-1a (91% ee,
24% NMR yield) and (S)-1d0 (R¼CD3, 0% ee, 37% NMR yield), in
both of which no deuterium incorporation at their C3 position
was observed (Entry 7). These results agreed with the report by
Cordero et al.3b and ruled out the possibility of path a. To verify
paths b and c, the 3-methyl ether (R)-1d (99% ee) was treated
with CSA (1 equiv) and MeOH (10 equiv) for 24 h to give (R)-1d
(76% ee, 83% yield) (Entry 8). A similar treatment of (R)-1d with
CSA (1 equiv) in MeCNewater (10:1) produced 1a (0% ee, 6% NMR
yield) and (R)-1d (99% ee, 79% NMR yield) (Entry 9). The reaction
of the 3-acetyl derivative (R)-1e with CSA (1 equiv) and MeOH
(10 equiv) produced the racemic 1d (40% yield) (Entry 11), and its
solvolysis was faster than of (R)-1a (Entry 10). These results
suggest that path c is the most plausible among the three paths.
Although we have not ascertained the reactive species that pro-
moted the racemization of (S)-1a during the enzymatic domino
process, we have determined that the enzymatic dynamic kinetic
resolution of (ꢁ)-1a was produced by the concomitant racemi-
zation of (S)-1a while the optically active ester (R)-7 immediately
underwent the [3þ2] intramolecular cycloaddition to irreversibly
produce the optically active 8.
HRMS. CAL-B [ChirazymeÒ
L-2, c.-f. C2, dry (12,900 U/g lyo.)], a li-
pase immobilized on a polymer support, was used, which is com-
mercially available from Roche Diagnostics GmbH. The same lipase
is also on sale from Novozymes, Inc., by the name, NOVOZYM 435Ò.
The commercial lipase was dried (1 mmHg, rt, overnight) prior to
use. Kanto silica gel 60 N was used for the column chromatography.
Silica gel 60 (spherical) NH2 was obtained from the Kanto Chemical
Co. All other reagents were purchased from Tokyo Chemical In-
dustry Co., Aldrich Chemical Co., Kanto Chemical Co., Kishida
Chemical Co., Nacalai Tesque or Wako Pure Chemical Industries and
used without further purification. The known compounds (3,13
6A,5a 6B5a) were prepared according to the reported methods.
4.2. Preparation of ( )-3-(tert-butyldimethylsilyl)oxy-1-
pyrroline N-oxides [( )-1a and ( )-1c]
4.2.1. (ꢁ)-3-(tert-Butyldimethylsilyloxy)pyrrolidine [(ꢁ)-2c]. Under
a nitrogen atmosphere, TBSCl (2.1 g, 13.8 mmol) was added to
a solution of (ꢁ)-3-hydroxypyrrolidine 2a (1.0 g, 11.5 mmol) and
imidazole (1.2 g, 17.2 mmol) in anhydrous MeCN (20 mL). The re-
action mixture was stirred at rt for 12 h and concentrated in vacuo.
Brine (30 mL) was added, and the product was extracted ten times
with hexane. The combined organic layer was dried over MgSO4
and concentrated in vacuo to give 2.2 g of the crude (ꢁ)-2c as a pale
yellow oil. This product was used for the following reaction without
further purification: 1H NMR (CDCl3, 300 MHz)
d: 0.06 (6H, s), 0.88

H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
7299
(9H, s), 1.62e1.70 (1H, m), 1.80e1.92 (1H, m), 2.80e2.86 (3H, m),
3.06e3.15 (1H, m), 4.31e4.36 (1H, m).
HRMS Calcd for C4H7NNaO2 [(MþNa)þ] m/z: 124.0369, found:
124.0364.
4.2.2. (ꢁ)-3-(tert-Butyldimethylsilyloxy)pyrrolidine
hydrochloride
4.2.8. (ꢁ)-4-Hydroxy-1-pyrroline N-oxide [(ꢁ)-4a]. Colorless crys-
[(ꢁ)-2c$HCl]. Under nitrogen atmosphere, TBSCl (2.2 g,
a
tals, mp 92e94 ꢂC. 1H NMR (500 MHz, CD3OD)
d: 2.63e2.72 (1H, m),
14.4 mmol) was added to a solution of (ꢁ)-3-hydroxypyrrolidine
hydrochloride 2a$HCl (1.19 g, 9.6 mmol) and triethylamine
(3.3 mL, 24.1 mmol) in anhydrous MeCN (18 mL) The reaction
mixture was stirred at rt overnight, then concentrated in vacuo.
Hexane (100 mL) was added, and the layers were separated. The
organic layer was washed three times with water. The combined
aqueous layer was saturated with NaCl, and the product was
extracted ten times with EtOAc. The combined organic layer was
dried over MgSO4 and concentrated in vacuo to give 2.1 g of
(ꢁ)-2c$HCl (92%) as a colorless solid. This product was pure enough
for the subsequent reaction. Analytically pure sample was obtained
by recrystallization from EtOAc: A colorless solid, mp 98.5e100 ꢂC.
3.09e3.18 (1H, m), 3.71e3.78 (1H, m), 4.23e4.31 (1H, m),
4.58e4.64 (1H, m), 7.14e7.18 (1H, m). 13C NMR (125 MHz, CD3OD)
d:
40.6, 66.4, 71.3, 140.1. IR (CHCl3): 3405, 1607, 1258 cmꢀ1. HRMS
Calcd for C4H7NNaO2 [(MþNa)þ] m/z: 124.0369, found: 124.0365.
4.3. Total synthesis of (L)-rosmarinecine (10)
4.3.1. Domino reaction of (ꢁ)-1a with 6B (Table 2, Entry 10). Under
a nitrogen atmosphere, (ꢁ)-1a (50 mg, 0.49 mmol), 6B (159 mg,
0.74 mmol), CAL-B (150 mg) and anhydrous MeCN (5.0 mL) were
placed in a resealable flask, and the flask was sealed. The reaction
mixture was stirred at 5 ꢂC for 12 h, and filtered through a Celite
pad. The filtrate was concentrated in vacuo, and the residue was
purified by flash column chromatography (hexaneeEtOAc,
40:60/EtOAc/CH2Cl2eMeOH, 80:20) to give 67 mg of 8B (60%,
91% ee) and 14 mg of (S)-1a (28%, 96% ee). The optical purities of 8B
and (S)-1a were determined by Daicel CHIRALPAK AD-H (hex-
aneeiPrOH, 80:20, 1.0 mL/min; retention times 14.5, 15.5 min) and
Daicel CHIRALPAK AS-H [hexaneeiPrOH, 70:30, 1.4 mL/min; re-
tention times 15.8 (R), 21.7 min (S)], respectively.
1H NMR (500 MHz, CDCl3)
d: 0.08 (3H, s), 0.09 (3H, s), 0.88 (9H, s),
1.30e1.90 (2H, m), 1.95e2.08 (2H, m), 3.19 (1H, d, J¼12.0 Hz), 3.37
(1H, dd, J¼4.5, 12.0 Hz), 3.41e3.50 (2H, m), 4.51e4.55 (1H, m). 13C
NMR (125 MHz, CDCl3)
d
: ꢀ4.92, 17.9, 25.6, 34.2, 43.4, 52.5, 70.5. IR
(CHCl3): 1234 cmꢀ1. Anal. Calcd for C10H24ClNOSi$1/5H2O: C, 49.75;
H, 10.19; N, 5.80; Cl, 14.68. Found: C, 49.94; H, 10.23; N, 5.77; Cl,
14.79.
4.2.3. Oxidation of (ꢁ)-2c$HCl (Table 1, Entry 12). Under a nitrogen
atmosphere, 3 (137 mg, 0.53 mmol) was added an ice-cold solution
4.3.2. Ethyl (2aR,3S,6aR,6bR)-2-oxohexahydro-1,4-dioxa-4a-azacy-
of (ꢁ)-2c$HCl (50 mg, 0.21 mmol) and aq NH3 (30% w/w, 30
m
L,
clopenta[cd]pentalene-3-carboxylate (8B). Colorless crystals, mp
0.53 mmol) in MeCNewater (10:1, 2.1 mL). The reaction mixture
was stirred at rt for 30 min and concentrated in vacuo. The residue
was purified by flash column chromatography (EtOAceMeOH,
15:1/EtOAceMeOH, 5:1) to give 33.1 mg of (ꢁ)-1c (73%) and
9.6 mg of (ꢁ)-4c (21%).
117e118 ꢂC, ½a 2D6
ꢄ
ꢀ53.4 (c¼1.1, CHCl3). 1H NMR (CDCl3, 300 MHz)
d:
1.33 (3H, t, J¼7.0 Hz), 2.39e2.46 (2H, m), 2.99e3.10 (1H, m), 3.75
(1H, dd, J¼6.0 Hz, 8.5 Hz), 3.79e3.84 (1H, m), 4.29 (2H, q, J¼7.0 Hz),
4.43 (1H, dd, J¼7.0, 8.5 Hz), 4.70 (1H, d, J¼6.0 Hz), 5.12 (1H, dd,
J¼7.0, 13.5 Hz). 13C NMR (CDCl3, 75 MHz)
d: 13.9, 32.4, 52.1, 54.5,
61.9, 71.5, 79.1, 83.1, 165.9, 174.0. IR (KBr): 1755 cmꢀ1. Anal. Calcd for
C10H13NO5: C, 52.86; H, 5.77; N, 6.16. Found: C, 52.70; H, 5.68; N,
6.12.
4.2.4. (ꢁ)-3-(tert-Butyldimethylsilyl)oxy-1-pyrroline N-oxide [(ꢁ)-
1c].9a,17 A colorless solid, mp 74e75 ꢂC. 1H NMR (500 MHz, CDCl3)
d: 0.09 (3H, s), 0.10 (3H, s), 0.89 (9H, s), 2.02e2.12 (1H, m),
2.51e2.61 (1H, m), 3.80e3.90 (1H, m), 4.10e4.20 (1H, m),
4.99e5.02 (1H, m), 6.84 (1H, q, J¼2.0 Hz). 13C NMR (125 MHz,
4.3.3. (S)-3-Hydroxy-1-pyrroline N-oxide [(S)-1a]. Colorless crys-
tals, mp 104e106 ꢂC, ½a D24
ꢄ
ꢀ132 (c¼1.3, CHCl3) {lit.17 99% ee; mp
CDCl3)
d
: ꢀ4.83, ꢀ4.72, 18.0, 25.6, 30.9, 61.3, 72.2, 135.6. IR (CHCl3):
105e107 ꢂC; ½a 2D1
ꢄ
ꢀ137 (c¼0.94, CHCl3)}. The 1H NMR, 13C NMR data
1585, 1258 cmꢀ1. HRMS Calcd for C10H21NNaO2Si [(MþNa)þ] m/z:
and IR spectrum of (S)-1a were in good agreement with those of
(ꢁ)-1a.
238.1234, found: 238.1247.
4.2.5. (ꢁ)-4-(tert-Butyldimethylsilyl)oxy-1-pyrroline
N-oxide
4.3.4. Domino reaction of (ꢁ)-1a with 6B, in situ generated from 5B
(one-pot procedure: Table 2, Entry 12). Under a nitrogen atmo-
sphere, a solution of 5B (2.14 g, 14.8 mmol) in anhydrous acetone
(25 mL) was portion wise added to an ice-cold solution of ethox-
yacetylene (2.48 mL, 30 mmol) and [RuCl2(p-cymene)]2 (45 mg,
0.074 mmol) in anhydrous acetone (50 mL) over a period of 20 min.
The reaction mixture was stirred at rt for 34 h and concentrated in
vacuo. To the residue was added MeCN (100 mL), and the mixture
was cooled to 0 ꢂC. (ꢁ)-1a (1.00 g, 9.9 mmol) and CAL-B (3.0 g) were
added, and the reaction mixture was stirred at 5 ꢂC for 11 h. The
reminder of the procedure is the same as mentioned above. Flash
column chromatography of the crude product gave 1.31 g of 8B
(58%, 92% ee) and 0.23 g of (S)-1a (23%, 95% ee). The first re-
crystallization of 8B (1.31 g, 92% ee) from hexaneeEtOAc gave 8B
(0.66 g, >99% ee), and the second recrystallization of its mother
liquid from hexaneeEtOAc gave another crop of 8B (0.30 g, >99%
ee). Thus, a total of 0.96 g [43% yield from (ꢁ)-1a] of 8B (>99% ee)
was obtained.
[(ꢁ)-4c].9a A colorless solid, mp 45e52 ꢂC. 1H NMR (500 MHz,
CDCl3)
d
: 0.08 (6H, s), 0.88 (9H, s), 2.61 (1H, d, J¼19.0 Hz), 2.97e3.05
(1H, m), 3.80 (1H, dd, J¼1.0, 13.0 Hz), 4.08e4.16 (1H, m), 4.60e4.67
(1H, m), 6.85 (1H, s). 13C NMR (125 MHz, CDCl3)
d: ꢀ4.90, 17.9, 25.6,
39.9, 66.4, 70.5, 133.3. IR (CHCl3): 1597, 1258 cme1. HRMS Calcd for
C10H21NNaO2Si [(MþNa)þ] m/z: 238.1234, found: 238.1249.
4.2.6. Oxidation of (ꢁ)-2a$HCl (Table 1, Entry 13). Under a nitrogen
atmosphere, 3 (1.06 g, 4.1 mmol) was added an ice-cold solution of
(ꢁ)-2a$HCl (200 mg,1.62 mmol) and tBuNH2 (0.43 mL, 4.1 mmol) in
MeCNewater (10:1, 16 mL). The reaction mixture was stirred at rt
for 30 min and concentrated in vacuo. The residue was filtered
through a pad of NH2 supported silica gel. The filtrate was con-
centrated, and the residue was purified by flash column chroma-
tography (EtOAceMeOH, 3:1/EtOAceMeOH, 2:1) to give 85 mg of
(ꢁ)-1a (52%) and 39 mg of (ꢁ)-4a (24%).
4.2.7. (ꢁ)-3-Hydroxy-1-pyrroline N-oxide [(ꢁ)-1a]. Colorless crys-
tals. Mp 79e80 ꢂC. 1H NMR (500 MHz, CD3OD)
d
: 2.02e2.13 (1H, m),
4.3.5. Methyl (2aR,3S,6aR,6bR)-2-oxohexahydro-1,4-dioxa-4a-aza-
cyclopenta[cd]pentalene-3-carboxylate (8A) (Table 2, Entry
7). Similarly to the preparation of 8B, 22.4 mg of 8A (53%, 92% ee)
was obtained from ( )-1a (20 mg, 0.20 mmol) using 6A (59 mg,
2.62e2.73 (1H, m), 3.88e4.00 (1H, m), 4.10e4.20 (1H, m),
4.96e5.14 (1H, m), 7.17 (1H, q, J¼2.0 Hz). 13C NMR (125 MHz,
CD3OD) d .
: 31.0, 62.3, 72.4, 141.0. IR (KBr): 3415, 1599, 1241 cme1

7300
H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
0.30 mmol), CAL-B (30 mg), and MeCN (2.0 mL). Colorless crystals,
(2H, m), 3.31 (3H, s), 3.34e3.49 (4H, m), 3.91 (1H, brs). 13C NMR
(125 MHz, CDCl3) : 28.5, 30.0, 31.1, 43.5, 43.9, 50.3, 51.1, 56.5, 79.09,
mp 138e140 ꢂC, ½a D26
ꢄ
ꢀ48.7 (c¼0.40, CHCl3){lit.3c mp 138e139 ꢂC,
d
½
a 2D7
ꢄ
ꢀ53.5 (c¼0.39, CHCl3) for 8A with 99% ee}. The 1H NMR and 13
C
79.14, 79.9, 154.5, 154.6. IR (CHCl3): 1686, 1416 cme1. HRMS Calcd
for C10H19NNaO3 [(MþNa)þ] m/z: 224.1257, found: 224.1248.
Under a nitrogen atmosphere, 4 M HCl in EtOAc (1.2 mL) was
added to tert-butyl (R)-3-methoxypyrrolidine-1-carboxylate
(50 mg, 0.25 mmol) at 0 ꢂC. The solution was stirred at rt for
30 min and concentrated in vacuo. The residue was dissolved in
NMR data of 8A were in good agreement with those reported.3c
4.3.6. (2aR,3S,6aR,6bR)-3-Hydroxy-2,4-dioxohexahydrofuro-[2,3,4-
gh]pyrrolizine (9). Pd(OH)2/C (20% Pd, 40 mg) was added to a so-
lution of 8B (31 mg, 0.14 mmol, 99% ee) in MeOH (6 mL). The flask
was evacuated and back-filled with hydrogen. The reaction mixture
was stirred at rt for 20 h under a hydrogen atmosphere (1 atm) and
filtered through a Celite pad. The Celite pad was washed with
MeOH, and the combined organic layer was concentrated in vacuo.
The residue was purified by flash column chromatography (hex-
aneeEtOAc, 1:1/EtOAc/EtOAceMeOH, 90:10) to give 12.9 mg of
MeCNewater (10:1, 2.5 mL). Aqueous NH3 (30% w/w, 35
mL,
0.62 mmol) and 3 (162 mg, 0.62 mmol) were added to the solution
at 0 ꢂC. The reaction mixture was stirred at rt for 30 min and
concentrated in vacuo, and the residue was purified by flash col-
umn chromatography (CH2Cl2eMeOH, 15:1/10:1) to give 21 mg of
(R)-1d (75%, 99% ee) and 7.1 mg of (R)-4-methoxy-1-pyrroline N-
oxide (R)-4d (25%). The optical purity of (R)-1d was determined by
Daicel CHIRALPAK AD-3 [hexaneeiPrOH, 95:5, 2.0 mL/min; re-
tention times 20.3 (R), 24.6 min (S)].
9 (55%, 99% ee). Colorless crystals, mp 182e184 ꢂC, ½a D23
ꢄ þ89.4
(c¼0.36, abs EtOH){lit.3c mp 183e185 ꢂC, ½a 2D7
ꢄ þ94.2 (c¼0.31, abs
EtOH)}. The 1H NMR and 13C NMR data of 9 were in good agreement
with those reported.3c
(R)-1d. Pale yellow oil, ½a D20
ꢄ
þ113 (c¼0.85, CHCl3). 1H NMR
(500 MHz, CDCl3)
d
: 2.17 (1H, dddd, J¼3.5, 5.0, 9.0, 14.5 Hz),
4.3.7. (ꢀ)-Rosmarinecine (10). According to the reported proce-
dure,3c 17 mg of (ꢀ)-10 (90%, 99% ee) was obtained from 9 (20 mg,
2.48e2.57 (1H, m), 3.35 (3H, s), 3.87 (1H, dddd, J¼1.0, 6.5, 9.0,
15.5 Hz), 4.10e4.19 (1H, m), 4.56e4.61 (1H, m), 7.02 (1H, q,
a 24
0.11 mmol). Colorless crystals, mp 167e170 C, ½ ꢄD ꢀ117.7 (c¼0.98,
J¼1.5 Hz). 13C NMR (125 MHz, CDCl3)
d: 27.0, 56.5, 61.4, 80.0, 133.3.
ꢂ
EtOH){lit.6 mp 168e170 ꢂC, ½a 2D1
ꢄ
ꢀ119.8 (c¼1.01, EtOH)}. The 1H
IR (CHCl3): 1584, 1269, 1238 cme1. HRMS Calcd for C5H9NNaO2
[(MþNa)þ] m/z: 138.0526, found: 138.0534.
NMR and 13C NMR data of 10 were in good agreement with those
reported.6
(R)-4-Methoxy-1-pyrroline N-oxide [(R)-4d]. A pale yellow oil,
½
a 2D0
ꢄ
ꢀ22.5 (c¼0.66, CHCl3). 1H NMR (500 MHz, CDCl3)
d: 2.75 (1H, d,
4.4. Preparation of (R)-1a, (R)-1d and (R)-1e
J¼19.5 Hz), 2.94e3.03 (1H, m), 3.33 (3H, s), 3.94 (1H, d, J¼15.0 Hz),
4.08e4.15 (1H, m), 4.19e4.24 (1H, m), 6.84e6.87 (1H, m). 13C NMR
4.4.1. (R)-3-Hydroxy-1-pyrroline N-oxide [(R)-1a]. In the same
manner as the oxidation of (ꢁ)-2a$HCl, 84 mg of (R)-1a (52%) and
39 mg of (R)-4a (24%) were obtained from (R)-2a$HCl (200 mg,
1.62 mmol) with 3 (1.06 g, 4.1 mmol), tBuNH2 (0.43 mL, 4.1 mmol)
and MeCNewater (10:1, 16 mL).
(125 MHz, CDCl3) d: 36.1, 56.5, 67.3, 74.3, 133.1. IR (CHCl3): 1595,
1275, 1238 cmꢀ1. HRMS Calcd for C5H9NNaO2 [(MþNa)þ] m/z:
138.0526, found: 138.0533.
4.4.3. (R)-3-Acetoxy-1-pyrroline N-oxide [(R)-1e]. Crude tert-butyl
(R)-3-hydoxypyrrolidine-1-carboxylate, prepared from (R)-3-
hydroxylpyrrolidine hydrochloride (R)-2a$HCl (0.50 g, 4.1 mmol)
by the above-described procedure, was dissolved in anhydrous
CH2Cl2 (20 mL). Pyridine (0.98 mL, 12.1 mmol), Ac2O (0.77 mL,
8.1 mmol) and DMAP (25 mg, 0.20 mmol) were added, and the
reaction mixture was stirred at rt for 4 h before being quenched
with water. The layers were separated, and the aqueous layer was
extracted three times with CH2Cl2. The combined organic layer was
dried over Na2SO4, filtered, and concentrated in vacuo. The residue
was purified by flash column chromatography (hexaneeEtOAc, 3:1)
to give 0.87 g of tert-butyl (R)-3-acetoxypyrrolidine-1-carboxylate
(R)-1a. Colorless crystals, mp 107e110 ꢂC, ½a D20
ꢄ þ89.1 (c¼0.32,
MeOH). HRMS Calcd for C4H7NNaO2 [(MþNa)þ] m/z: 124.0369,
found: 124.0393. The 1H NMR and 13C NMR data and the IR spec-
trum of (R)-1a were in good agreement with those of (ꢁ)-1a.
(R)-4a. Colorless crystals, mp 116e118 ꢂC, ½a D20
ꢄ ꢀ57.9 (c¼1.96,
MeOH). HRMS Calcd for C4H7NNaO2 [(MþNa)þ] m/z: 124.0369,
found: 124.0393. The 1H NMR and 13C NMR data and the IR spec-
trum of (R)-4a were in good agreement with those of (ꢁ)-4a.
4.4.2. (R)-3-Methoxy-1-pyrroline N-oxide [(R)-1d]. Boc2O (1.02 mL,
4.5 mmol) was added to a solution of (R)-3-hydroxylpyrrolidine
hydrochloride (R)-2a$HCl (0.50 g, 4.1 mmol) in THF-satd NaHCO3
(1:1, 20 mL), and the reaction mixture was stirred at rt for 1.5 h.
EtOAc was added, and the layers were separated. The aqueous layer
was extracted three times with EtOAc. The combined organic layer
was washed with brine, dried over Na2SO4, filtered, and concen-
trated in vacuo to give tert-butyl (R)-3-hydoxypyrrolidine-1-
carboxylate, which was used for the following reaction without
further purification.
The above-described tert-butyl (R)-3-hydoxypyrrolidine-1-
carboxylate was dissolved in anhydrous DMF (20 mL), to which
was added NaH (55% oil suspension, 0.71 g, 16.2 mmol) at 0 ꢂC. The
ice-cold reaction mixture was stirred for 30 min, and Me2SO4
(0.77 mL, 8.1 mmol) was then added. The reaction mixture was
stirred overnight at 50 ꢂC before being quenched with water.
HexaneeEtOAc (1:1) was added, the layers were separated, and the
aqueous layer was extracted three times with hexaneeEtOAc (1:1).
The combined organic layer was washed two times with brine,
dried over Na2SO4, filtered, and concentrated in vacuo. The residue
was purified by flash column chromatography (hexaneeEtOAc, 2:1)
[93% from (R)-2a$HCl]. A colorless oil, ½a D20
ꢄ ꢀ16.1 (c¼0.90, CHCl3).
1H NMR (500 MHz, CDCl3)
d: 1.44 (9H, s), 1.94e2.05 (5H, m),
3.33e3.56 (4H, m), 5.24 (1H, br). 13C NMR (125 MHz, CDCl3)
d: 21.1,
28.4, 30.7, 31.5, 43.6, 43.9, 51.4, 51.8, 72.9, 73.8, 79.5, 154.3, 154.4,
170.5, 170.6. IR (CHCl3): 1735, 1689, 1685 cme1. HRMS Calcd for
C11H19NNaO4 [(MþNa)þ] m/z: 252.1206, found: 252.1183.
Similarly to the preparation of (R)-1d, 230 mg of (R)-1e (74%,
99% ee) and 76 mg of (R)-4-acetoxy-1-pyrroline N-oxide (R)-4e
(24%) were obtained from tert-butyl (R)-3-acetoxypyrrolidine-1-
carboxylate (500 mg, 2.2 mmol). The optical purity of (R)-1e was
determined by Daicel CHIRALCEL OZ-3 [hexaneeiPrOH, 80:20,
2.0 mL/min; retention times 34.5 (R), 41.4 min (S)].
(R)-1e. A pale yellow oil, ½a D20
ꢄ
þ215 (c¼0.42, CHCl3). 1H NMR
(500 MHz, CDCl3)
d
: 2.08 (3H, s), 2.25 (1H, dddd, J¼2.0, 4.0, 8.0,
14.5 Hz), 2.64e2.73 (1H, m), 3.93 (1H, dddd, J¼2.0, 4.0, 9.0, 14.0 Hz),
4.15e4.24 (1H, m), 5.72e5.77 (1H, m), 6.99 (1H, q, J¼2.0 Hz). 13C
NMR (125 MHz, CDCl3) d: 20.8, 26.8, 61.3, 73.8, 132.2, 170.4. IR
(CHCl3): 1744, 1582, 1242 cmꢀ1. HRMS Calcd for C6H9NNaO3
[(MþNa)þ] m/z: 166.0475, found: 166.0493.
to afford 0.69
g
of tert-butyl (R)-3-methoxypyrrolidine-1-
(R)-4-Acetoxy-1-pyrroline N-oxide [(R)-4e]. A pale yellow oil,
carboxylate [85% from (R)-2a$HCl]. A colorless oil,
½
a 2D0
: 1.44 (9H, s), 1.84e2.02
ꢄ
ꢀ8.4
½
a 2D0
ꢄ
ꢀ23.7 (c¼0.35, CHCl3).1H NMR (500 MHz, CDCl3)
d: 2.08 (3H, s),
(c¼0.52, CHCl3). 1H NMR (500 MHz, CDCl3)
d
2.78 (1H, d, J¼22.0 Hz), 3.17 (1H, ddd, J¼1.5, 7.0, 22.0 Hz), 3.92 (1H, d,

H. Nemoto et al. / Tetrahedron 68 (2012) 7295e7301
7301
5. (a) Akai, S.; Tanimoto, K.; Kita, Y. Angew. Chem., Int. Ed. 2004, 43, 1407; (b) Akai,
S.; Naka, T.; Omura, S.; Tanimoto, K.; Imanishi, M.; Takebe, Y.; Matsugi, M.; Kita,
Y. Chem.dEur. J. 2002, 8, 4255.
6. (a) Denmark, S. E.; Thorarensen, A.; Middleton, D. S. J. Am. Chem. Soc. 1996, 118,
8266; (b) Denmark, S. E.; Thorarensen, A.; Middleton, D. S. J. Org. Chem. 1995,
60, 3574.
J¼15.0 Hz), 4.30 (1H, ddd, J¼1.5, 7.0, 15.0 Hz), 5.42 (1H, t, J¼7.0 Hz),
6.87 (1H, t, J¼1.5 Hz). 13C NMR (125 MHz, CDCl3)
d: 20.8, 36.3, 67.4,
67.6, 134.2, 170.2. IR (CHCl3): 1743, 1595, 1238 cme1. HRMS Calcd for
C6H9NNaO3 [(MþNa)þ] m/z: 166.0475, found: 166.0480.
7. Akai, S.; Tanimoto, K.; Kanao, Y.; Omura, S.; Kita, Y. Chem. Commun. 2005, 2369.
8. (a) Tatsuta, K.; Takahashi, H.; Amemiya, Y.; Kinoshita, M. J. Am. Chem. Soc. 1983,
105, 4096 For reviews, see also: (b) Tatsuta, K.; Hosokawa, S. Chem. Rev. 2005,
105, 4707; (c) Tatsuta, K.; Hosokawa, S. Sci. Technol. Adv. Mater. 2006, 7, 397.
9. (a) Ohtake, H.; Imada, Y.; Murahashi, S. Bull. Chem. Soc. Jpn. 1999, 72, 2737; (b)
Murahashi, S.; Mitsui, H.; Shiota, T.; Tsuda, T.; Watanabe, S. J. Org. Chem. 1990,
55, 1736.
10. Tokuyama, H.; Kuboyama, T.; Amano, A.; Yamashita, T.; Fukuyama, T. Synthesis
2000, 1299.
11. (a) Goti, A.; Nannelli, L. Tetrahedron Lett. 1996, 37, 6025; (b) Murray, R. W.;
Iyanar, K.; Chen, J.; Wearing, J. T. J. Organomet. Chem. 1996, 61, 8099.
12. Murray, R. W.; Singh, M. J. Org. Chem. 1990, 55, 2954.
Acknowledgements
This work was supported by Grants-in-Aid for Scientific Re-
search from the Ministry of Education, Culture, Sport, Science, and
Technology, Japan. Roche Diagnostics K. K., Japan is thanked for the
generous gift of CAL-B.
Supplementary data
13. Davis, F. A.; Chattopadhyay, S.; Towson, J. C.; Lal, S.; Reddy, T. J. Org. Chem. 1988,
53, 2087.
Supplementary data related to this article can be found online at
http://dx.doi.org/10.1016/j.tet.2012.06.095.
14. (a) Brandi, A.; Cicchi, S.; Paschetta, V.; Pardo, D. G.; Cossy, J. Tetrahedron Lett.
2002, 43, 9357; (b) Schieweck, F.; Altenbach, H.-J. J. Chem. Soc., Perkin Trans. 1
2001, 3409; (c) Cicchi, S.; Numes, J., Jr.; Goti, A.; Brandi, A. Eur. J. Org. Chem.
1998, 419; (d) McCaig, A. E.; Wightman, R. H. Tetrahedron Lett. 1993, 34, 3939;
(e) Zajac, W. W., Jr.; Walters, T. R.; Darcy, M. G. J. Org. Chem. 1988, 53, 5856.
15. The oxidation of (S)-1,3-dihydroxypyrrolidine with HgO was reported to give
a 5:1 mixture of 1a and 4a Goti, A.; Cicchi, S.; Fedi, V.; Nannelli, L.; Brandi, A. J.
Org. Chem. 1997, 62, 3119.
16. Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem. Soc. 1982, 104, 7294.
17. For preparation of the rosmarinecine derivatives from natural rosmarinesine,
see: Percino, T. M.; Nieto-Alvarez, D. A. J. Mex. Chem. Soc. 2006, 50, 160.
18. Breuer, E. In Nitrones, Nitronates, and Nitroxides; Patai, S., Rappoport, Z., Eds.;
Wiley-Interscience: New York, NY, 1989; Chapter 2, pp 139e244.
19. A lipase-catalyzed deracemization of (ꢁ)-8B might be considered as another
possible path for producing 8B with a higher optical purity than the theoretical
ones (Table 2, entries 10e12). However, it was denied because the treatment of
(ꢁ)-8B with CAL-B (300% w/w) in MeCN at 5 ꢂC for 24 h did not cause any
reaction.
References and notes
1. Kocienski, P. J. In Protecting Groups Thieme: New York, NY, 2005.
2. For reviews, see: (a) Young, I. S.; Baran, P. S. Nat. Chem. 2009, 1, 193; (b) Hoff-
mann, R. W. Synthesis 2006, 3531.
3. (a) Pisaneschi, F.; Cordero, F. M.; Brandi, A. Synlett 2003, 1889; (b) Cordero, F.
M.; Pisaneschi, F.; Gensini, M.; Goti, A.; Brandi, A. Eur. J. Org. Chem. 2002, 1941;
(c) Goti, A.; Cacciarini, M.; Cardona, F.; Cordero, F. M.; Brandi, A. Org. Lett. 2001,
3, 1367; (d) Cordero, F. M.; Gensini, M.; Goti, A.; Brandi, A. Org. Lett. 2000, 2,
2475; (e) For the asymmetric synthesis of the rosmarinecine analogues, see:
Cicchi, S.; Marradi, M.; Vogel, P.; Goti, A. J. Org. Chem. 2006, 71, 1614.
4. For recent examples, see: (a) Xu, J.; Caro-Diaz, E. J. E.; Trzoss, L.; Theodorakis, E.
A. J. Am. Chem. Soc. 2012, 134, 5072; (b) Yu, F.; Cheng, B.; Zhai, H. Org. Lett. 2011,
13, 5782; (c) Herrmann, A. T.; Martinez, S. R.; Zakarian, A. Org. Lett. 2011, 13,
3636; (d) Lau, S. Y. W. Org. Lett. 2011, 13, 347; (e) Gu, Z.; Zakarian, A. Angew.
20. The use of an excess amount (3 equiv) of 6 for the domino reaction of (ꢁ)-1a
resulted in forming lower yields of 8, probably due to the intermolecular cy-
clization of 1ae6.
€
Chem., Int. Ed. 2010, 49, 9702; (f) Lobermann, F.; Mayer, P.; Trauner, D. Angew.
Chem., Int. Ed. 2010, 49, 6199; (g) Zhou, Q.; Chen, X.; Ma, D. Angew. Chem., Int. Ed.
2010, 49, 3513; (h) Li, C.; Yu, X.; Lei, X. Org. Lett. 2010, 12, 4284.
21. Wender, P. A.; Verma, V. A.; Paxton, T. J.; Pillow, T. H. Acc. Chem. Res. 2008, 41, 40.