Tandem double-intramolecular [4+2]/[3+2] cycloadditions of nitroalkenes. Studies toward a total synthesis of daphnilactone B: Piperidine ring construction

Article abstract of DOI:10.1021/jo052001l

Two model studies in support of a total synthesis of the complex polycyclic alkaloid daphnilactone B have been completed. The objectives of the models studies were to demonstrate the use of a tandem double-intramolecular [4+2]/[3+2] nitroalkene cycloaddit

Full text of DOI:10.1021/jo052001l

Tandem Double-Intramolecular [4+2]/[3+2] Cycloadditions of
Nitroalkenes. Studies toward a Total Synthesis of Daphnilactone B:
Piperidine Ring Construction
Scott E. Denmark* and Ramil Y. Baiazitov
Department of Chemistry, Roger Adams Laboratory, UniVersity of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
ReceiVed September 23, 2005
Two model studies in support of a total synthesis of the complex polycyclic alkaloid daphnilactone B
have been completed. The objectives of the models studies were to demonstrate the use of a tandem
double-intramolecular [4+2]/[3+2] nitroalkene cycloaddition for the stereocontrolled construction of four
of the rings in the core of the natural product. The first model study established the ability to create a
pyrrolidine ring corresponding to ring A of daphnilactone B through a modification of the dipolarophile
and subsequent functional group manipulations. The second model study required the modification of
the dienophile in the [4+2] cycloaddition to accommodate the formation of a piperidine ring (ring B of
daphnilactone B). Nitroalkene 26 containing a diene as the dienophile served well in the tandem
cycloaddition to afford the nitroso acetal 38a in 77% yield. Subsequent functional group manipulations
allowed for the high-yielding conversion to the core of daphnilactone B.
Introduction and Background
. Tandem Cycloaddition. The tandem [4+2]/[3+2] cy-
SCHEME 1
1
cloaddition of nitroalkenes has been extensively studied in these
laboratories and has been successfully applied to the syntheses
1
of many alkaloid natural products. Recently, a novel double-
intramolecular variant has been developed in which both the
2
dienophile and dipolarophile are tethered to the nitro olefin.
Because the cycloadditions require electronically complimentary
2π-partners and different reaction conditions, the double-
intramolecular [4+2]/[3+2] cycloaddition reaction can easily
operate as a tandem sequence. The example in Scheme 1
illustrates the C(5)/fused mode of tethering the reaction com-
ponents in nitro olefin 1. The more electron-rich alkene in the
vinyl ether first reacts with the electron-poor heterodiene
forming nitronate 2, which is an electron-rich dipole. This dipole
then can react with the terminal double bond (tethered through
C(5) in the nitro olefin) forming nitroso acetal 3.2
The synthetic utility of the tandem nitro olefin cycloaddition
is revealed by unmasking the product nitroso acetals through
hydrogenolysis or other transformations to afford nitrogen-
3
containing, highly functionalized polycyclic skeletons. The
reaction proceeds through cleavage of one or both N-O bonds
and, when appropriate, can involve further transformations such
as alkylation, reductive amination, or acylation of the nitrogen
(Scheme 2).
(1) (a) Denmark, S. E.; Cottell, J. J. In The Chemistry of Heterocyclic
Compounds: Synthetic Applications of 1,3-Dipolar Cycloaddition Chemistry
Toward Heterocycles and Natural Products; Padwa, A., Pearson, W. H.,
Eds.; Wiley-Interscience: New York, 2002; pp 83-167. (b) Denmark, S.
E.; Thorarensen, A. Chem. ReV. 1996, 96, 137-166.
(
2) (a) Denmark, S. E.; Gomez, L. J. Org. Chem. 2003, 68, 8015-8024.
(3) Denmark, S. E.; Vaugois, J.; Guagnano, V. Can. J. Chem. 2001, 79,
1606-1616.
(
b) Denmark, S. E.; Gomez, L. Org. Lett. 2001, 3, 2907-2910.
1
0.1021/jo052001l CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/17/2005
J. Org. Chem. 2006, 71, 593-605
593

Denmark and Baiazitov
SCHEME 2
SCHEME 3
Most applications of the tandem [4+2] cycloaddition have
employed an enol ether as the electron-rich dienophile. In this
2
case, the FG fragment on the nitroso acetal i is an alkoxy group.
Hydrogenolysis of this nitroso acetal produces intermediate
[
4+2] cycloaddition with 2-methyl-2-nitrostyrene (Chart 1).
These alkenes were either unreactive (under thermal conditions)
or unstable upon treatment with a Lewis acid (SnCl4, TiCl2(O-
i-Pr)2, AlMe3).
hemiacetal ii, which can act as a latent aldehyde and can undergo
1
reductive amination to form a pyrrolidine ring. Often, FG is a
7
carboxylic ester which, upon hydrogenolysis, can participate
Ultimately, it was found that allenylsilanes are viable dieno-
1
in cyclization to produce lactam vi. Alternatively, FG can be
8
philes in the cycloaddition (Scheme 4). The [4+2] cycloaddi-
an activated hydroxymethyl group.4 In this case, a piperidine
ring in indolizidine vii can be formed from the isoxazolidine
ring in i. Even though the hydrogenolysis process is a complex
sequence with multiple possibilities for side reactions, it usually
proceeds with high selectivity in good to excellent yields.
A consequence of the 1,5-functional disposition in the 1,2-
oxazines formed from enol ethers is that hydrogenolysis of enol
,5
tion step was successful; however, the regiochemical course of
the reaction was reversed, and the nitroso acetals formed could
not be directly used to construct piperidines.
CHART 1
ether-derived nitroso acetals (such as i) creates pyrrolidine rings
1
(
such as v or 4). Indeed, the majority of the natural and
6
unnatural products prepared by this sequence contain isolated,
fused, or embedded pyrrolidines. However, to construct a
piperidine by this process, a nitroso acetal leading to a 1,6-
functionality pattern is required. Therefore, the dienophile should
contain an additional carbon appropriately functionalized to
allow for six-membered ring synthesis. The sequence in Scheme
SCHEME 4
3
illustrates two proposed routes by which a piperidine ring
such as in xi can be created from the oxazine fragment in viii.
In both routes, the formation of the piperidine ring is initiated
by the hydrogenolytic cleavage of the nitroso acetal. In the upper
sequence, the piperidine is formed by alkylation of the amino
group in ix by a tosylate. This pathway finds analogy in the
formation of the piperidine ring in indolizidine vii from the
isoxazoline portion of the nitroso acetal i, Scheme 2. In the
lower sequence, the piperidine is formed by reductive amination
of the aldehyde xiii. This route resembles the well-developed
transformations leading from enol ether-derived nitroso acetals
to pyrrolidine rings upon hydrogenolysis (iii to v, Scheme 2).
The difference between these two routes is the oxidation state
of the additional carbon atom needed to make the piperidine
ring.
2. Plan for the Synthesis of Daphnilactone B Using a
Tandem Nitro Olefin Cycloaddition. The Daphniphyllum
alkaloids are a family of structurally complex polycyclic amines
that have been isolated from the fruit and bark of the Yuzuriha
9
tree native to Japan. Since antiquity, extracts of the bark and
leaves of Yuzuriha have been used as a folk remedy for asthma.
However, the active component remained elusive. The full
stereostructures of 34 members of this family have been
Recently, several dienophiles (at both oxidation states)
designed to be precursors to piperidine rings were tested in the
(
(
4) Denmark, S. E.; Cottell, J. J. J. Org. Chem. 2001, 66, 4276-4284.
5) Denmark, S. E.; Martinborough, E. A. J. Am. Chem. Soc. 1999, 121,
(7) Gomez, L. Postdoctoral Report, University of Illinois, Urbana, 2002.
(8) Denmark, S. E.; Gomez, L. Heterocycles 2002, 58, 129-136.
(9) Yamamura, S. Daphniphyllum Alkaloids. In The Alkaloids: Chemistry
and Pharmacology; Brossi, A., Ed.; Academic Press: Orlando, 1986; pp
265-286.
3
046-3056.
6) Denmark, S. E.; Montgomery, J. I. Angew. Chem., Int. Ed. 2005, 44,
732-3736.
(
3
594 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
established by X-ray crystal structure analysis and/or chemical
correlation.10
SCHEME 5
One of these components, daphnilactone B (5) (Figure 1),
was isolated by Hirata in 1972.11 The relative stereostructure
1
2
of the molecule was established by X-ray analysis. This
member of the class is unusual compared to other Daphniphyl-
lum alkaloids because of the two embedded seven-membered
rings (D and E). To date, there are no total syntheses or synthetic
studies toward daphnilactone B on record nor any reports of its
biological activity. The possibility of constructing the core of
daphnilactone B by tandem [4+2]/[3+2] cycloaddition and the
natural product therefrom made it a very interesting target.
addressed the task of incorporating the pyrrolidine ring A of
daphnilactone B with its attendant methyl group as embodied
in the target 6. The construction of 6 from precursor 9 was
envisioned by adapting the approach used in the synthesis of
simpler model 4 (Scheme 6). This plan represents only a minor
modification of the substrate, and both the [4+2] and [3+2]
steps leading to 8 are expected to proceed uneventfully if not
more readily than in 1. The hydrogenolysis step should lead to
pyrrolidinone 7 via lactamization following much precedent.
However, the chemical modification of the pyrrolidinone to the
methyl pyrrolidine will represent most of the new chemical
challenges.
FIGURE 1. Structure of daphnilactone B.
The core of the molecule consists of six cycles (Figure 1):
the pyrrolidine and piperidine rings (A and B) that are attached
to a cyclohexane ring (C), a seven-membered lactone (D), a
cycloheptane ring (E), and a cyclopentene (F). Previously, a
structure resembling the core of daphnilactone B was generated
using the [4+2]/[3+2] nitroalkene cycloaddition as the key step
(4, Scheme 1). The structure 4 bears the correct fusion of the
rings corresponding to B, C, and D of the natural product. This
suggests that a cycloaddition reaction of a substrate similar to
1
may be used as the key step for the total synthesis of
daphnilactone B. Obviously, compound 4 lacks rings A, E, and
F and any precursor for daphnilactone B would need to be
modified to accommodate introduction of those rings. In
addition, the ring B in daphnilactone B is a piperidine ring
whereas in 4 it is a pyrrolidine. Building on the use of a tandem
double intramolecular [4+2]/[3+2] cycloaddition as the key
strategic transformation, a retrosynthetic plan for daphnilactone
B was formulated and is shown in Scheme 5. This plan identifies
only the key elements of the strategy and is based on our current
knowledge of the tandem process. To accommodate the
incorporation of the rings and needed functional groups, several
questions need to be addressed: (1) how to create the piperidine
and pyrrolidine rings, (2) how to stereoselectively install the
vicinal stereogenic quaternary centers, (3) how to construct the
hydroazulene portion, (4) how to carry the necessary functional-
ity for creation of the lactone ring, and (5) how to provide a
means for absolute stereocontrol. Discussion of how this entire
plan will be reduced to practice is beyond the scope of this
paper. Our goals were to establish, in two model studies, the
modifications to the tandem cycloaddition and the chemical
transformations needed to incorporate the pyrrolidine and
piperidine rings A and B.
SCHEME 6
3.2. Model Study II: Building Piperidine Ring B. The goal
of the second model study is to build the ABCD-ring system
xv wherein the greatest challenge will be the construction of
the piperidine ring (Scheme 7). The problems to be solved are
as follows: (1) finding an appropriate functionalized dienophile
to form a piperidine ring, (2) controlling the exo/endo-selectivity
in folding the nitro olefin in the [4+2] cycloaddition step, and
3
. Objectives of the Model Studies. 3.1. Model Study I:
Building Pyrrolidine Ring A. The first part of this program
(10) (a) Toda, M.; Niwa, H.; Irikawa, H.; Hirata, Y.; Yamamura, S.
Tetrahedron 1974, 30, 2683-2688. (b) Heathcock, C. H. Angew. Chem.,
Int. Ed. Engl. 1992, 31, 665-681.
(3) construction of the pyrrolidine ring A along with the
(
11) Niwa, H.; Toda, M.; Hirata, Y.; Yamamura, S. Tetrahedron Lett.
stereoselective installation of the methyl group in this advanced
structure.
1
972, 13, 2697-2700.
(12) Sasaki, K.; Hirata, Y. Tetrahedron Lett. 1972, 13, 1891-1894.
J. Org. Chem, Vol. 71, No. 2, 2006 595

Denmark and Baiazitov
SCHEME 7
The use of dichloromethane as the solvent eliminated the
problem, and a mixture of the nitroso acetal 8 and the nitronate
1
6 was obtained in less than 30 min at -75 °C (Scheme 9).
Stirring the mixture in benzene at 60 °C for several hours
allowed the [3+2] cycloaddition to proceed to completion, and
nitroso acetal 8 was isolated in 92% yield. Apparently, the
nitronate 16 is not stable and undergoes room-temperature
cyclization to nitroso acetal 8. The stereostructure of 8 was
assumed for the nitroso acetal on the basis of previous
2
cycloaddition studies. The use of 2.5 equiv of SnCl4 was found
1
to be the optimal amount of Lewis acid. H NMR analysis of
the reaction mixture revealed that only one anomer of 8 was
formed. However, the configuration at this center was not
determined as it is irrelevant for the synthesis of daphnilactone
B, because both epimers converge to the same product 7 upon
hydrogenolysis.
Results
1
. Model Study I. Formation of the Pyrrolidine Ring A
and Diastereoselective Installation of the Methyl Group. 1.1.
Synthesis of Nitro Olefin 9. The preparation of the cycload-
dition precursor 9 followed the synthesis route used for
SCHEME 9
2
2
nitroalkene 1 (Scheme 8). Diene 11 was converted to alcohol
12 in excellent yield by a hydroboration/oxidation sequence with
9
-BBN-H and H2O2. Oxidation of 12 using the von Doering
1
3
protocol afforded the corresponding aldehyde 13, which was
directly subjected to E-selective olefination to afford the
unsaturated ester 14 in 66% yield for the two steps.
Hydrolysis of the TBS ether in 14 was accomplished by
treatment with HF/pyridine in 95% yield. von Doering oxida-
tion13 to the corresponding aldehyde was followed immediately
by Henry reaction with nitroethane to afford the nitro alcohol
1
5 in 72% yield. Dehydration of 15 with acetic anhydride/
DMAP afforded the nitro alkene 9 in 79% yield as a single
geometrical isomer. Although the elimination reaction was slow
(90 h), it gave better yields than attempts to affect the elimination
of the alcohol via a mesylate, tosylate, or trifluoroacetate.
SCHEME 8
The hydrogenolysis of 8 was one of the crucial steps in this
study because it was intended to construct the pyrrolidinone
ring. If the polycycle 7 is too strained, the lactam will not close.
Fortunately, hydrogenolysis of nitroso acetal 8 with Raney nickel
under 1 atm of hydrogen afforded a mixture of the expected
1
hydroxy lactam 7 (about 30% from H NMR integration),
uncyclized amino ester 17 (50%), and a side product (20%)
1
4
which was identified as lactam 21. The closure of 17 could
be accelerated by the action of NaOMe to produce hydroxy
lactam 7 in 79% overall yield.
3
To suppress the fragmentation leading to 21, the nickel
catalyst was washed with 0.1 M solution of NaH2PO4 (pH )
4.5) before use. This modification led to an improved ratio (10/1
compared to 4/1) of the desired product 7 to the undesired lactam
1
.2. Tandem Cycloaddition and Further Transformations.
2
On the basis of prior studies, we fully expected precursor 9 to
undergo smooth cycloaddition in the presence of Lewis acids.
Treatment of 9 with SnCl4 (2.0 equiv) in toluene at -78 °C
produced an insoluble dark precipitate and no detectable
cycloaddition product. It is likely that the complex of tin
tetrachloride and 9 is insoluble in toluene at low temperatures.
(13) Doering, W. v. E.; Parikh, J. R. J. Am. Chem. Soc. 1967, 89, 5505-
5507.
1
13
(
14) The structure of 21 was established by H and C NMR spectros-
-1
copy (two signals at 175 and 180 ppm), IR (1687 and 1738 cm ), as well
as HRMS (FAB, calcd for C15H23NO4 + H, 282.1705, found 282.1707).
596 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
2
1 after sodium methoxide treatment. The same improvement
2. Model Study II. Formation of Piperidine Ring B. The
objective of the second model study is to develop a variant of
the [4+2] component of the cycloaddition process that will lead
to formation of a piperidine ring (Scheme 7). The principal
modification of the dienophile in xvii (compared to the normally
employed enol ether dienophiles) is inclusion of an extra carbon
suitably functionalized to become incorporated into the piperi-
dine ring upon hydrogenolysis. Modified dienophiles designed
to accomplish that objective in three different ways are shown
in Chart 2. Because these substrates were only marginally useful
as cycloaddition substrates, their syntheses will not be presented.
However, the results of attempted cyclizations were instructive
in guiding the final choice of a suitable dienophile and will be
described briefly below.
(7/21,12/1) was obtained when the hydrogenolysis was per-
formed at 350 psi with non-NaH2PO4-washed catalyst. When
the reaction was performed under 350 psi of hydrogen and the
nickel catalyst was washed with 0.1 M solution of NaH2PO4,
the ratio was improved to 17/1 producing, nevertheless, lactam
7
in the same yield (80%), suggesting that some other side
reactions took place under these conditions.
To complete the objectives of the model study, the R-hydroxy
lactam 7 was to be converted to a methylated pyrrolidine. This
13
process began by von Doering oxidation of 7 to the keto lactam
8, which underwent Wittig olefination to form R-methylene
1
lactam 19 in 73% overall yield (Scheme 9). The carbon-carbon
double bond in unsaturated lactam 19 did not undergo hydro-
genation in the presence of Wilkinson catalyst with either
CHART 2
dihydrogen1 (at 340 psi, rt, overnight) or catecholborane (-20
5
16
°
C, THF); only the starting material was recovered. However,
17
nickel boride-catalyzed hydrogenation at ambient pressure of
hydrogen afforded a mixture of the two separable saturated
amides, 20a and 20b, in a 3.5/1 ratio. The stereochemical
assignment of the hydrogenation products was made on the basis
of the coupling constants between the introduced methine proton
(HC(2), Figure 2) and the adjacent methine proton HC(3), as
well as through NOE determination. Semiempirical calculations
1
8
(
PM5) showed that in the desired lactam 20a the dihedral angle
3
HC(2)C(3)H is 28° and J HC(2)C(3)H is expected to be 6-12 Hz
from the Karplus equation). In the undesired isomer (20b), the
dihedral angle HC(2)C(3)H is 112°, in which case J HC(2)C(3)H
(
3
1
should be less than 2 Hz. H NMR analysis showed that the
proton at the HC(2) methine is split into a quartet of doublets
3
in the major product (desired isomer, 20a, J HC(2)C(3)H ) 5.4
2
.1. Allylic Ether and Silanes as Dienophiles. The use of
Hz, Figure 2), whereas it is a quartet in the minor product
3
an allylic ether such as 22 is the most obvious way to add a
single carbon which can be activated for closure to a piperidine
ring. Cycloaddition would lead to a hydroxymethyl group at
C(6) of the oxazine which could be tosylated (see viii, Scheme
) or mesylated prior to hydrogenolysis. Unfortunately, all
attempts to effect [4+2] cycloaddition with a variety of Lewis
acids led to either desilylation (SnCl4, Ti(O-i-Pr)2Cl2) no reaction
(
undesired isomer, 20b, J HC(2)C(3)H is close to zero). An NOE-
experiment with the major product showed that there is an 11%
enhancement between HC(2) and H3C(5) as well as a 7%
enhancement between HC(2) and HC(3), also supporting
structure 20a (Figure 2).
3
2
2
23
(
MAD, MAPh) or undesired side reactions (Me3Al). Also,
purely thermal (199 °C, 14 h, p-cymene, CaCO3) activation
failed to afford cycloaddition products. Clearly, the modest
electrophilicity of the nitroalkene mandates a more electron rich
partner for successful inverse electron demand cycloaddition.
The allylic oxygen substituent must inductively or hypercon-
jugatively deactivate the dienophile beyond the range for
effective cycloaddition.
Toward that end, the allylic silanes 23 and 24 were identified
and prepared. The electron-donating nature of the silyl groups
should decrease the HOMO/LUMO gap for the [4+2] cyclo-
addition. Indeed, allyltrimethylsilane had been used successfully
in intermolecular nitroalkene cycloadditions previously.24 In the
presence of 2 equiv of TiCl4 in dichloromethane for 15 min at
FIGURE 2. Assignment of the relative configuration of the hydro-
genation products 20.
Attempts to convert lactam 20a into amine 6 using BH319 or
-BBN-H20 were unsuccessful, but the reduction was ultimately
9
2
1
accomplished by heating 20a with 2.0 equiv of DIBAL-H at
5 °C in THF/hexanes. The final product of the model study 6
6
was produced in 61% yield, proving that this is a viable strategy
to create the pyrrolidine ring A of daphnilactone B along with
the methyl bearing stereocenter.
-
70 °C, nitroalkene 23 did produce the corresponding nitroso
acetals 30 (as a 10/1 mixture of diastereomers) but only as a
2
5
(
15) Jardine, F. H.; Osborn, J. A.; Wilkinson, G. J. Chem. Soc. A 1967,
1
574-1578.
(21) Ojima, I.; Zhao, M.; Yamato, T.; Nakahashi, K.; Yamashita, M.;
Abe, R. J. Org. Chem. 1991, 56, 5263-5277.
(22) MAD: methyl(bis(2,6-di-tert-butyl-4-methylphenoxy))aluminum;
MAPh: methyl(bis(2,6-diphenylphenoxy))aluminum.
(23) Pecunioso, A.; Menicagli, R. J. Org. Chem. 1988, 53, 45-49.
(24) Hurd, A. R. Ph.D. Thesis, University of Illinois at Urbana-
Champaign, 2000.
(
(
16) Evans, D. A.; Fu, G. C. J. Org. Chem. 1990, 55, 5678-5680.
17) Russell, T. W.; Hoy, R. C.; Cornelius, J. E. J. Org. Chem. 1972,
3
7, 3552-3553.
18) Calculations were performed using a Cache Worksystem Pro Version
.1.12.33 software from Fujitsu Limited.
(
6
(
19) Brown, H. S.; Heim, P. J. Org. Chem. 1973, 38, 912-916.
(20) Collins, C. J.; Lanz, M.; Singaram, B. Tetrahedron Lett. 1999, 40,
(25) The relative configurations of 30a and 30b were not determined
because this route was abandoned.
3
673-3676.
J. Org. Chem, Vol. 71, No. 2, 2006 597

Denmark and Baiazitov
minor product (Scheme 10). The major product isolated was
SCHEME 11
2
9 (as a 4/1 mixture of diastereomers; (29/30 ) 2/1)) arising
from either fragmentation of the 1,2-oxazine or Sakurai-type
addition. The bulkier silane 24 was expected to be less prone
2
6
to desilylation and indeed led to an improved ratio of 29/31
1.5/1 under the same reaction conditions. Trimethylaluminum
at 0 °C led to slow 1,4-addition of a methyl group to the nitro
)
2
3
olefins.
SCHEME 10
step in the planned synthesis of nitro olefin 26 is an alkyne
carbocupration/cross-coupling sequence to create the 1,3-diene
28
in 35 stereoselectively (Scheme 12). Because it was not known
if the diene geometry would influence the outcome of the
cycloaddition, the Z-geometry was chosen for synthetic simplic-
ity. Alkyne 35 was prepared from the known aldehyde 34 by
29
acetalization in 76% yield. Carbocupration of 35 with a cuprate
3
0
reagent formed from CuBr and ClMg(CH2)3OMgCl followed
3
1
by Pd(0)-catalyzed cross-coupling with 1-iodo-2-methylpro-
pene afforded Z-diene 36 as a single geometrical isomer.
Optimization experiments demonstrated that it was necessary
to use several equivalents of the organocopper reagent. The yield
was improved from 30-40% to 62% if 3.0 equiv of the
alkylcopper reagent were employed instead of 1.1 equiv. It
seems that the carbocupration is relatively slow, and thermal
decomposition of the alkylcuprate intermediate competes.
Construction of the dipolarophile, involved a two-step sequence
beginning with Swern oxidation of the alcohol 36 in 88% yield
followed by treatment of the resulting aldehyde with car-
bomethoxymethylidenetriphenylphosphorane to provide the
unsaturated ester 37 in 99% yield (E/Z, 20/1). The synthesis
was completed by installation of the nitro olefin moiety.
SCHEME 12
2
.2. Conjugated Dienes as Dienophiles. The foregoing
results and much previous experience in nitroalkene [4+2]
cycloadditions support the notion of a highly asynchronous
process with a significant amount of charge separation. The
cationic character that accumulates on the dienophile must be
stabilized for cycloaddition to be successful (see Scheme 18).
A second solution for stabilizing the positive charge with
additional carbon substituents involves the delocalization of that
charge in a conjugated diene. Earlier studies in these laboratories
have shown that cyclic conjugated dienes function well as
dienophiles in intermolecular cycloadditions.27 Four nitro olefins
containing tethered dienes with different terminal groups 25-
2
8 were synthesized and evaluated for competence in the in
the tandem cycloaddition (Chart 2). Nitroalkene 25 bearing a
diene with no terminal substituent suffered decomposition in
the presence of SnCl4 or Ti(O-i-Pr)xCl4-x, (x ) 0, 2) at -70 °C
and was unaffected by trimethylaluminum at 0 °C. Apparently,
the unsubstituted diene polymerized under the strongly Lewis
acidic reaction conditions.
To address the problem of polymerization and provide
additional stabilization for positive charge, two other nitro-
alkenes, 27 and 28 bearing aromatic substituents were evaluated.
Under the most favorable conditions (2.0 equiv of SnCl4,
CH2Cl2, -65 °C, 25 min) an inseparable mixture of four
isomeric nitroso acetals, presumably 32a/32b and 33a/33b, was
isolated (Scheme 11). The use of Me3Al at -80 °C for 10 min
led to no conversion and ultimately to decomposition of the
substrate at room temperature. Similar results were obtained
with 28 containing a more electron-rich dienophile. The use of
conjugating aromatic substituents did allow for cycloaddition
to proceed, albeit in a nonselective fashion. Thus, a different
mode of stabilization of the charge was envisioned that
employed the electron-donating properties of simple methyl
groups as in 26.
2
.3. Synthesis of Nitroalkene 26. The electron-rich conju-
gated diene in 26 will serve a dual role. First, it will act as the
electron-rich 2π-component during the [4+2] cycloaddition.
Second, the isopropylidene group remaining attached to the
nitroso acetal serves as a masked aldehyde which will later
participate in the hydrogenolysis/reductive amination. The key
(
28) (a) Normant, J. F.; Alexakis, A. Synthesis 1981, 11, 841-871. (b)
Normant, J. F. Alkyne carbocupration and polyene synthesis. In Organo-
copper reagents, A practical approach; Taylor, R. J. K., Ed.; Oxford
University Press: New York, 1994. (c) Lipshutz, B. H.; Sengupta, S. Org.
React. 1992, 41, 135-631.
(
29) Hopf, H.; Kruger, A. Chem. Eur. J. 2001, 7, 4378-4385.
(
26) (a) Akiyama, T.; Kirina, M. Chem. Lett. 1995, 723-724. (b)
(30) Gardette, M.; Alexakis, A.; Normant, J. F. Tetrahedron 1985, 41,
5887-5899.
(31) (a) Jabri, N.; Alexakis, A.; Normant, J. F. Bull. Soc. Chim. Fr. 1983,
332-338. (b) Gardette, M.; Jabri, N.; Alexakis, A.; Normant, J. F.
Tetrahedron 1984, 40, 2741-2750.
Danheiser, R. L.; Dixon, B. R.; Gleasont, R. W. J. Org. Chem. 1992, 57,
6
094-6097.
27) Denmark, S. E.; Kesler, B. S.; Moon, Y.-C. J. Org. Chem. 1992,
7, 4912-4924.
(
5
598 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
Removal of the dimethoxy acetal with PPTS in aqueous acetone
provided the aldehyde, which was subjected to Henry reaction
with nitromethane to afford the nitro alcohol. Dehydration with
trifluoroacetic anhydride and triethylamine produced 26 in 66%
yield for the three steps as a single E-isomer.
oxidation with KMnO437 proceeded too slowly, probably due
to the hindered nature of the olefin. Ruthenium tetraoxide38
quickly destroyed the material.
Ultimately, it was found that the cleavage could be cleanly
accomplished by ozonolysis after careful optimization of the
reaction conditions. The cleavage could be successfully executed
in methanol containing 1 equiv of pyridine if 1 equiv of ozone
2
.4. Tandem Cycloaddition of Nitroalkene 26. Initial
attempts at performing the cycloaddition of 26 in the presence
of SnCl4 or Ti(O-i-Pr)xCl4-x, (x ) 0-2) again produced an
inseparable mixture of four isomeric nitroso acetals 38a/b-
3
9
was used. Upon reduction with trimethyl phosphite, the
aldehyde 40 could be isolated quantitatively (Scheme 14).
With the substrate required for closure by reductive amination
now available, we were poised to test the piperidine ring
formation. Exposure of the aldehyde to reducing conditions
presented a concern that conversion to an alcohol would be
competitive with the N-O bond cleavage. Gratifyingly, hydro-
genolysis of 40 in presence of Raney nickel in methanol under
1 atm of hydrogen afforded lactam 10 in 76% yield. As in model
study I, closure to the lactam required the assistance of sodium
methoxide. In forming the piperidine ring in 10 the major goal
of the model study II has been achieVed. The remaining
challenge was the transformation of lactam 10 into the final
product of the study 50a to evaluate possible intermediates in
the planned synthesis of daphnilactone B (Scheme 16).
39a/b as was seen with the aryl-substituted dienes (Scheme 13).
Much to our surprise, however, Me3Al (which had not been
useful with any of the previous substrates) promoted the
cycloaddition of 26 to afford two nitroso acetals. Optimization
of the conditions involved decreasing the reaction temperature
to -78 °C and increasing the amount of the Lewis acid to 3
equiv. Under these conditions, a 20/1 mixture of the isomeric
32
nitroso acetals was obtained in 77% yield. Fractional crystal-
lization allowed the isolation of the major isomer (59%) whose
full stereostructure was determined to be 38a by single-crystal
X-ray analysis.33 The crystal structure revealed that 38a arises
from an exo-fold (tether) of the dienophile and that the
configuration of the dienophile is preserved. The structures of
the byproducts 38b and 39a/b were assumed as shown on the
Oxidation of 10 according to the method of Swern afforded
diketone 43 in 75% yield. Unlike the related R-dicarbonyl
compound 18 from model study I, this intermediate was very
sensitive and prone to formation of hemiacetals and hydrates.
Chloroform solutions of chromatographically purified 43 dis-
played two ketone carbonyl resonances in the 1 C NMR
1
basis of their characteristic H NMR signals, but they were not
isolated and fully characterized.
SCHEME 13
3
spectrum at 201 and 207 ppm. However, in CD3OD, the 13
C
NMR spectrum showed signals at 212 and 102 ppm, the latter
of which is characteristic of an acetal or hemiacetal. Numerous
attempts to optimize this reaction and isolate the diketone in
analytically pure form failed.
SCHEME 14
2
.5. Transformation of 38a into the Tetracyclic Target of
Model Study II. The selective and high-yielding construction
of 38a showed that the electron-rich 1,1,4,4-tetraalkyl diene
served well in participating in the cycloaddition. The second
function of that unit is to serve as a latent aldehyde which must
be unmasked by oxidative cleavage. Several attempts to cleave
the carbon-carbon double bond were made to install an
appropriate functional group on the terminus of that fragment.
Exhaustive ozonolysis followed by reduction with dimethyl
3
4
35
sulfide or sodium borohydride led to recovery of uncharac-
3
6
terizable decomposition products in low yield. Osmylation or
(32) Nitroolefin 26 was formed as a mixture of isomers at the unsaturated
ester double bond (E/Z ∼ 20/1). It is possible that the cycloaddition of 26
was completely stereoselective and the 20/1 ratio of the diastereomeric
nitroso acetals reflects the presence of the two isomers in the starting
material. In this case, the minor cycloadduct may be the epimer of 38a at
the ester-group-bearing stereocenter.
Intermediate 43 also behaved very differently than 18 toward
olefination reagents. It was anticipated that the presence of the
lactam carbonyl group and absence of the angular methyl group
(33) The crystallographic coordinates of 38a have been deposited with
the Cambridge Crystallographic Data Centre; deposition no. CCDC 275186.
These data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/
retrieving.html (or from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, UK.; fax: (+44) 1223-336-033; or
deposit@ccdc.cam.ac.uk). For the Chem3D image, see the Supporting
Information.
(34) Pappas, J. J.; Keaveney, W. P.; Gancher, E.; Berger, M. Tetrahedron
Lett. 1966, 7, 4273-4278.
(35) Mukai, C.; Kataoka, O.; Hanaoka, M. J. Org. Chem. 1995, 60,
5910-5918.
J. Org. Chem, Vol. 71, No. 2, 2006 599

Denmark and Baiazitov
would render the ketone in the five-membered ring more reactive
than that in the six-membered ring. Under Wittig olefination
conditions, the starting material was quickly consumed. How-
ever, only approximately 30% of the unsaturated lactam 44
could be isolated (Scheme 15). The remaining material could
not be identified. Hydrogenation of 44 under the conditions used
in the first model study provided the lactam 45 as a single
diastereomer, whose configuration was assumed to be as shown
in Scheme 15 by analogy to 20.
SCHEME 16
In an attempt to induce Peterson olefination, 43 was combined
with Me3SiCH2MgCl to provide adduct 46 as a single diaste-
reomer in 60% yield. The relative configuration was assumed
as shown in Scheme 15 but is inconsequential because that
stereocenter will be destroyed later. Unfortunately, elimination
of trimethylsilanol could not be accomplished. When the
â-hydroxysilane 46 was treated with KHMDS, a complex
mixture of unidentified products was formed. Also different
acidic reagents did not effect conversion of the silane (AcOH,
HCl, TFA, BF3/Et2O, SnCl4).40 Because it was believed that
presence of the lactam carbonyl makes the olefin 44 unstable,
it was decided that the hydroxysilane should be subjected to
the elimination after the lactam is reduced to the amine.
7 (Figure 3) were calculated at the PM3 level.18 Accordingly,
4
the calculated dihedral angles for 47 are as follows: HaC(1)-
3
C(2)H ) 82° (expected JHaC(1)C(2)H is less than 1 Hz from the
3
Karplus equation), HbC(1)C(2)H ) 34° (expected JHbC(1)C(2)H
is 7-13 Hz). The calculated dihedral angles for epi-47 are: HaC-
3
(
(
1)C(2)H ) 51° (expected JHaC(1)C(2)H is 3-7 Hz), Hb(C1)C-
3
1
2)H ) 169° (expected JHbC(1)C(2)H is 8-14 Hz). In the H NMR
spectrum of the reduction product, HHC(1) is a doublet (3.86
2
3
ppm, J HaC(1)Hb ) 14.9, JHaC(1)C(2)H is close to 0 Hz), HHC(1)
2
3
is a doublet of doublets (3.24 ppm, JHbC(1)Ha ) 14.9, JHbC(1)C(2)H
3
)
6.1), HC(2) is a triplet (3.62 ppm, JHC(2)C(1)Hb ) 5.6) due to
SCHEME 15
a coupling to the hydroxyl proton. The observation of one
medium and one very small coupling between HC(2) and the
HC(1) protons is consistent with structure 47, and not epi-47.
The amide function in 47 was reduced with dihydrochloro-
2
1
alane to produce the amine 48 in 86% yield (Scheme 16).
Fortunately, elimination of trimethylsilanol under the strongly
Lewis acidic conditions of this reaction was not observed.
However, treatment of the amine 48 with KHMDS cleanly
provided the olefin 49 in 95% yield.
Attempts to add a nucleophile to the cyclohexanone moiety
in 46 failed. Addition of n-BuLi or n-BuMgCl led to enolization
(proved by deuterium incorporation upon quench with D2O).
A less basic reagent prepared by transmetalation of n-BuLi with
CeCl341 did not act on the starting material (and no D-
incorporation was noted either). Ultimately, this carbonyl group
was reduced stereoselectively with sodium borohydride to form
the secondary alcohol 47 (Scheme 16). Only one isomer was
observed which was assigned to be the equatorial alcohol from
the following considerations. First, the borohydride ion is
assumed to approach from the less hindered face, and second,
the 1H NMR coupling constants are consistent with this
configuration. Minimized energy conformations of 47 and epi-
(
36) Morikawa, K.; Park, J.; Andersson, P. G.; Hashiyama, T.; Sharpless,
K. B. J. Am. Chem. Soc. 1993, 115, 8463-8464.
37) (a) Lee, D. G.; Chang, V. S. J. Org. Chem. 1978, 43, 1532-1536.
b) Herriott, A. W.; Picker, D. Tetrahedron Lett. 1974, 15, 1511-1514.
38) Carlsen, P. H. J.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J.
Org. Chem. 1981, 46, 3936-3938.
39) (a) Liang, J.; Hoard, D. W.; Khau, V. V.; Martinelli, M. J.; Moher,
FIGURE 3. Configurational assignment of reduction product 47.
(
The last step of the sequence was the introduction of the
(
(
methyl group by hydrogenation of the exo-methylidene group.
42
Although homogeneous hydrogenation (RhCl(PPh3)3,1 atm)
(
43
was not effective, hydrogenation over Pt/C was successful but
afforded two diastereomers. The major isomer was separated
by column chromatography, and its structure was assigned to
be 50a by NOE spectroscopy (Figure 4). The fact that there is
E. D.; Moore, R. E.; Tius, M. A. J. Org. Chem. 1999, 64, 1459-1463. (b)
Slomp, G.; Johnson, J. L. J. Am. Chem. Soc. 1958, 80, 915-921.
(40) (a) AcOH: Boeckman, R. K.; Silver, S. M. Tetrahedron Lett. 1973,
1
4
4, 3497-3500. (b) TFA: Marchand, A. P.; Kaya, R. J. Org. Chem. 1983,
8, 5392-5395. (c) BF3/Et2O: Hudrlik, P. F.; Hudrlik, A. M.; Misra, R.
N.; Peterson, D.; Withers, G. P.; Kulkarni, A. K. J. Org. Chem. 1980, 45,
4
444-4448.
41) Imamoto, T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya,
Y. J. Am. Chem. Soc. 1989, 111, 4392-4398.
(42) Hussey, A. S.; Takeuchi, Y. J. Am. Chem. Soc. 1969, 91, 672-
675.
(43) Marx, J. N.; McGaughey, S. M. Tetrahedron 1972, 28, 3583-3586.
(
600 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
.8% NOE between HC(2) and HC(4) proves that they lie on
the same face of the pyrrolidine ring, which is consistent only
with structure 50a. The final product was converted into its
borane adduct 51 for full characterization.
3
increased hydrogen pressure (leading to faster N-O bond
cleavage) and that formation of 21 is a base-catalyzed frag-
mentation.
SCHEME 17
FIGURE 4. Proof of the structure of 50a by NOE and borane adduct
5
1.
Discussion
1
. Model Study I. The primary objective of model study I
The second goal of model study I was the stereoselective
installation of the methyl group on the pyrrolidine ring. Although
hydrogenation of lactam 19 provided a 3.5/1 mixture of two
diastereomers favoring the desired epimer 20a, this outcome
was viewed as satisfactory. The angular methyl group in 19 is
an artifact of the model system and is absent in the structure of
daphnilactone B. We expected that approach of hydrogen from
the convex face would be more facile and the overall selectivity
greater in a true precursor devoid of this methyl group. This
expectation was realized in model study II.
was to demonstrate that the pyrrolidine ring in 6 corresponding
to ring A of daphnilactone B can be introduced using the tandem
double-intramolecular [4+2]/[3+2] cycloaddition from nitro
olefin 9. This outcome was fairly assured as the only difference
between 9 and proven precursor 1 was the carbomethoxy group
attached to the dipolarophile. Because the [3+2] step prefers
an electron-poor dipolarophile, this change should have only
made the dipolar cycloaddition easier.
2
As was established in previous studies, the cycloaddition of
proceeded readily to afford the nitroso acetal 8 in very good
2
. Model Study II. 2.1. [4+2] Cycloaddition. The primary
9
objective of model study II was the development of a new
variant of the tandem cycloaddition that would lead to the
formation of the piperidine ring in 10 that corresponds to the
ring B of daphnilactone B. Subsequent introduction of the
pyrrolidine ring (ring A) and stereoselective introduction of the
ring A methyl group was also planned.
yield as a single isomer. It was assumed that the [4+2] step
proceeded through the exo-transition state by analogy to the
formation of 3. The reason for this preference is the presence
of the â-methyl group on the nitroalkene, which typically
disfavors the endo-transition structure through nonbonding
interactions.2
,44,45
The formation of 8 during workup at room
The electronic demands of the [4+2] component of the
cycloaddition required that a newly invented dienophile must
be (1) capable of introducing an additional carbon for the
piperidine ring, (2) highly electron rich, and (3) convertible to
an electrophilic function for subsequent ring closure. The
successful cycloadditions of 26, 27, and 28 showed that dienes
could serve as the requisite dienophiles, albeit with poor
stereoselectivity. Remarkably, reaction of 26 in the presence
of AlMe3 was not only high yielding but also very stereo-
selective.
temperature showed that the intermediate nitronate 16 did
manifest the dipolarophile activation as expected. Obviously,
the directing effect of the tether to the dipolarophile was strong
enough to lead to a single cycloadduct in this step as well.
The carbomethoxy group in 9 also served to provide the
additional carbon needed to produce the pyrrolidinone ring in
7
upon hydrogenolysis of nitroso acetal 8. Even though it was
2
not unprecedented, there was some concern whether the lactam
was not too strained to form spontaneously as had been seen
7
46
in previous pyrrolizidine ring assemblies. Indeed, the clean
From extensive prior experience with both simple alkenes
and enol ethers as dienophiles, we now view the [4+2] process
conversion to 7 required activation by NaOMe.
The formation of byproduct 21 during hydrogenolysis of
nitroso acetal 8 is intriguing. This lactam most likely arises by
1
as concerted but highly asynchronous. In most cases, the
geometry of the dienophile is preserved, but occasionally
secondary products arising from ionic pathways are isolated,
which suggests the intermediacy of a zwitterionic species similar
to xviii/xix (Scheme 18). The preservation of the dienophile
geometry in 38a implies that the [4+2] step is concerted and
not a stepwise process when promoted by AlMe3 (Scheme 18).
The nonpolar solvent toluene also disfavors ionic processes. The
formation of the four isomers 38a/38b and 39a/39b in the
presence of the stronger Lewis acids in dichloromethane as the
solvent can be explained by a combination of unselective side
chain folding in the [4+2] transition state (exo vs endo) and
the isomerization of the dienophile during the ionic [4+2]
cycloaddition.
(
acid- or base-catalyzed) fragmentation of the nitroso acetal with
concomitant N-O bond cleavage to afford the amino diester
2 (Scheme 17). The primary amino group in 52 can attack
5
either ester moiety to form 21 or 53. The reason for the preferred
formation of 21 may be the fixed, diaxial position of both the
amine and the carbonyl groups. The formation of 21 could be
suppressed by washing the Raney nickel to remove basic
contaminants or by executing the hydrogenation at 350 psi.
These observations suggest that 21 and 7 are formed via two
competing pathways and the formation of 7 is favored by
(44) Denmark, S. E.; Moon, Y.-C.; Cramer, C. J.; Dappen, M. S.;
Senanayake, C. B. W. Tetrahedron 1990, 46, 7373-7392.
The high stereoselectivity of the [4+2] cycloaddition of 26
was remarkable. The striking preference for reaction via an exo
(tether) transition structure demands explanation. Factors that
(
(
45) Moon, Y.-C. Ph.D. Thesis, University of Illinois, Urbana, 1991.
46) For a summary of recent strained and troublesome lactamizations,
see ref 4.
J. Org. Chem, Vol. 71, No. 2, 2006 601

Denmark and Baiazitov
SCHEME 18
must be considered are (1) the structures of the nitroalkene and
dienophiles, (2) the connecting tether, and (3) the nature of the
Lewis acid. In previous studies of intramolecular [4+2] cy-
cloadditions of nitroalkenes tethered by four methylene groups
to simple alkene dienophiles, the preference for exo-selectivity
level,18,48 and the results are presented in Figure 6. The
calculations suggest that in each case the [4+2] cycloaddition
is highly asynchronous, perhaps stepwise with ∼20 kcal/mol
activation energy barrier for the formation of the C-C bond
and very low (∼1 kcal/mol) barrier for the O-C bond-forming
step. The low activation energy for the latter step, leading from
xviii/xix to nitronates 55/56, means that the transient intermedi-
ate has a very short lifetime and quickly collapses to form the
cycloadduct. Nevertheless, this short lifetime may be long
enough to form minor amounts of the epimers via the C-C
bond rotation. The exo-transition structure is 2.5 kcal/mol lower
in energy than the endo-transition structure in accordance with
the observed stereoselectivity. In the exo-transition structure,
the C-C bond length is 1.98 Å and the N-O bond length is
2.86 Å. In the endo-structure, the C-C bond is 1.96 Å and the
N-O bond is 2.80 Å. Inspection of the two transition structures
reveals the following:
(with SnCl4 as the Lewis acid) was strongly dependent on the
substitution of the nitroalkene. In cases with no R-substituent,
44
the exo-transition structure was favored, but not strongly. The
role of Me3Al is difficult to evaluate because it has only been
employed for intermolecular cycloadditions with enol ethers as
the dienophiles. In these cases, it was found that the cyclo-
additions proceed via the endo (enol ether oxygen) transition
structure ascribable to secondary orbital interactions between
the heterodiene and the approaching dienophile and small size
and monodentate character of this Lewis acid.47 The cyclo-
addition of 26 proceeded in the same sense, placing the vinylic
substituent on the dienophile endo in the transition structure
(
Figure 5). The previously observed “endo-rule” for inter-
1. The formation of xix-endo proceeds through a slightly later
molecular [4+2] cycloadditions between nitro olefins and dienes
promoted by Lewis acids can also support the involvement of
the secondary orbital interactions in these systems.27 However,
in this case, in the absence of favorable interactions between
the dienophile oxygen and the complexed Lewis acid, the
preference for the exo (side chain)-transition structure is more
difficult to understand.
endo-transition structure (the C-C bond is 0.02 Å shorter),
suggesting that additional steric interactions must be overcome
to reach transition state compared to the exo-transition structure.
2
. The decisive steric difference in the two transition
structures can be seen by examining the interactions around the
incipient cyclohexane ring. In the exo-transition structure, the
forming cyclohexane and nitronate rings are trans-fused. Thus,
only the dipolarophile tether has 1,3-diaxial-type interactions
with neighboring hydrogens. However, in the endo-transition
FIGURE 5. Steric interactions during the [4+2] step in nitro olefin
2
6.
To gain insight into origins of the high exo (side chain)-
selectivity of the [4+2] step, transition structures leading from
6 to xviii-exo and xix-endo were calculated at the PM5
2
FIGURE 6. Calculated exo- and endo-transition structures for the C-C
bond-forming step leading from 26 to xviii-exo and xix-endo.
(47) Denmark, S. E.; Seierstad, M. J. Org. Chem. 1999, 64, 1610-1619.
602 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
structure, the cyclohexane and nitronate rings are cis-fused.
Thus, both the dipolarophile tether and the nitronate carbon are
axial in the cyclohexane ring. This arrangement creates an
additional nonbonding interaction in the endo-TS between the
R-hydrogen of the nitro group and a hydrogen on the tether to
the dienophile at the δ-position (the distance between them is
SCHEME 19
4
9
2
.13 Å). In the exo-TS, these interactions are absent because
these two hydrogen atoms lie on the opposite faces of the
forming ring.
To summarize, the exo-selectivity of the reaction may be
explained by the stabilizing secondary orbital interactions
between the 2π- and 4π-components in the exo-TS, as well as
destabilizing steric interactions present in the endo-transition
state.
hydrogenolysis of the aldehyde 40 provided diol 10 in very good
yield for such a complex series of transformations.
As expected, differentiation between the two keto groups in
2.2. [3+2] Cycloaddition. An interesting difference between
the cycloadditions in model studies I and II is the ease with
which the [3+2] step took place in the latter case. The
cycloaddition of 9 produced a mixture of the nitroso acetal 8
and the intermediate nitronate 16 (Scheme 9), which relatively
slowly collapsed into the acetal. The nitronates 55/56 (Scheme
4
3 was not problematic. The R-keto amide function was more
reactive than the simple ketone flanked by a quaternary carbon
center. The ease of hemiacetal and hydrate formation from 43
is a manifestation of its higher electrophilicity compared to 18.
This unusually high electrophilicity of the R-methylidene lactam
is manifested in its rapid decomposition in the presence of
nucleophilic species. This character required that the lactam
carbonyl be removed prior to introduction of the methylidene
function. The Petersen olefination served admirably in this
capacity because the hydroxy silane intermediate allowed
manipulation of the molecule elsewhere. For example deoxy-
genation of the lactam 47 could be achieved by the action of
chloroalane. This reagent is known to favor amine formation
in reduction of lactams as opposed to the formation of amino
1
8) were not observed and must have suffered the [3+2] step
at room temperature. This observation most likely arises from
the additional steric interactions provided by the methyl group
in 16.
2
.3. Elaboration of Rings A and B. The oxidative cleavage
of the isopropylidene group in nitroso acetal 38a to an aldehyde
took extensive experimentation. Exactly 1 equiv of ozone had
to be used which was also buffered by 1 equiv of pyridine.
Presumably, the pyridine forms a less electrophilic complex with
3
9
ozone, which makes the oxidant milder and more selective.
The nitroso acetal 38a has limited solubility in methanol at low
temperatures. However, when dichloromethane was used as the
reaction solvent, or even as a 1/1 cosolvent with methanol, the
aldehyde yield dropped. Cleavage of the primary ozonide xx
21
alcohols by C-N bond cleavage. Elimination of trimethylsi-
lanol under the acidic conditions of this reaction was not
observed. However, under basic conditions, the elimination was
fast and clean.
5
0
probably proceeds in the direction shown in Scheme 19. In
this case, aldehyde 40 and the carbonyl oxide xxi (which is
stabilized by the two alkyl groups and is trapped by methanol
forming xxii) are formed directly upon the fragmentation when
the reaction is done in methanol. It is not clear why more than
a stoichiometric amount of ozone was detrimental. Perhaps
oxidation of the nitroso acetal becomes competitive.
Finally, hydrogenation of 49 afforded the target structure 50a.
The diastereoselectivity of this reaction was higher than in the
similar reaction of 19; however, it was lower than in hydrogena-
tion of 44. The attenuated diastereoselectivity in the reduction
of 49 compared to 44 may be the directing effect of the hydroxy
51
group in 49. In the synthesis of daphnilactone B, this alcohol
can be functionalized to eliminate its directing effect and to
create more bulk on the concave face of the pyrrolidine ring,
which may further improve the diastereoselectivity.
The hydrogenolysis step for the formation of the key
piperidine ring was of greatest concern because many potential
complications could be envisioned. Most serious among these
concerns was the premature reduction of the aldehyde to the
corresponding alcohol before the amine 41 could be revealed
for reductive amination (Scheme 14). Second, even if the nitroso
acetal were cleaved first, we anticipated complications from the
revealed R-hydroxy aldehyde such as epimerization or tau-
tomerization to an R-hydroxy ketone. Finally, we also antici-
pated that the primary amino group in 41 could first react with
the ester instead of the aldehyde to form a pyrrolidinone which
would not undergo reductive amination. Remarkably, the
Conclusions
An extension of the tandem double intramolecular [4+2]/
[3+2] nitroolefin cycloaddition has been developed that allows
the generation of piperidines by the use of an electron-rich 1,3-
diene as a 2π-component in the cycloaddition. The crucial
tandem cycloaddition could be affected by trimethylaluminum
to afford nitroso acetal 38a in high yield and stereoselectively
via an exo-transition structure. Unmasking the isopropylidene
group in 38a to an aldehyde was possible by careful ozonolysis.
Upon hydrogenolysis of nitroso acetal aldehyde, a cascade of
steps generated the piperidine and pyrrolidinone rings in 10 by
intramolecular reductive alkylation/lactamization. This inter-
mediate was transformed into 50a, with formation of the fully
functionalized pyrrolidine ring (corresponding to ring A of
daphnilactone B).
(48) In each case, to locate the saddle point, an energy surface map was
calculated in the gas phase starting with the corresponding nitronate (bound
to a molecule of trimethylaluminum) and systematically varying the C-C
and N-O bond length. Frequency calculations were used to confirm the
nature of the saddle points. All transition structures had only one imaginary
frequency corresponding to the formation of the expected bonds. Reaction
coordinate calculations were performed to determine the connections of
the cycloaddition transition structures with the reactants and the products.
(
49) van der Waals radius of hydrogen is 1.20 Å (Bondi, A. J. Phys.
Chem. 1964, 68, 441-451.
50) Bunnelle, W. H. Chem. ReV. 1991, 91, 335-362.
(51) Thompson, H. W.; McPherson, E.; Lences, B. L. J. Org. Chem.
1976, 41, 2903-2906.
(
J. Org. Chem, Vol. 71, No. 2, 2006 603

Denmark and Baiazitov
Future challenges include further development of the tandem
intra/intramolecular [4+2]/[3+2] cycloaddition for a general
route to piperidine-containing natural products. In the context
of the total synthesis of daphnilactone B, the major challenges
are to build a more complex cycloaddition substrate that can
incorporate the functionality needed to introduce the two
adjacent quaternary stereocenters and provide the potential for
stereocontrol and the locus for elaboration of the hydroazulene
rings. These studies will be reported in due course.
became yellow in 5 min. After 25 min at -75/-76 °C, the reaction
mixture was quenched with methanol (1.0 mL) and diluted with
MTBE (20 mL). Brine (5 mL) was added, and the mixture was
warmed to rt and was washed with sodium hydroxide (10% aq
solution, 5 mL). The organic layer was washed with satd aq solution
2 4
of ammonium chloride, dried (Na SO ), filtered, and concentrated
in vacuo. The almost colorless oily material was purified by
chromatography (silica gel; hexanes/EtOAc, gradient, 7/1-3/1) to
afford 38a (327 mg, 1.02 mmol, 77%) as a white solid. Additional
purification by repeated crystallization (hexanes, three times)
provided 38a (248 mg, 0.77 mmol, 59%) as a white crystalline
1
material. Data for 38a: mp 99-100 °C (hexanes); H NMR (400
Experimental Section
MHz, CDCl ) 5.12 (doublet of apparent pentets, J ) 9.3, 1.5, 1 H,
3
General Experimental Procedures. See the Supporting Infor-
mation.
rel-(1S,2S,4R,6aR,10aS,10bR)-Methyl Octahydro-6-methoxy-
HC(13)), 4.71 (d, J ) 1.5, 1 H, HC(2)), 4.26 (dd, J ) 9.0, 1.5, 1
H, HC(6)), 3.74 (s, 3 H, H C(18)), 3.38 (dd, J ) 5.9, 2.4, 1 H,
HC(10b)), 2.55-2.63 (m, 1 H, HC(1)), 2.09-2.17 (m, 2 H, H C-
(11)), 2.03-2.09 (m, 1 H, HC(10a)), 1.88-1.99 (m, HHC(12)),
1.76-1.83 (m, 1 H, HHC(9)), 1.74 (s, 3 H, H C(15) or H C(16)),
1.70 (s, 3 H, H C(15) or H C(16)), 1.38-1.56 (m, 5 H, H C(8),
C(10) and HHC(12)), 1.24-1.34 (m, 1 H, HHC(9)), 1.16-1.24
3
2
1
0b-methyl-6H-1,6a-ethanoisoxazolo[2,3-c][2,3]benzoxazine-2-
carboxylate (8). A 250-mL, three-neck, round-bottom flask
equipped with a nitrogen inlet, a thermocouple, a rubber septum,
and a stir bar was charged with 9 (2.58 g, 8.30 mmol, 1.0 equiv)
3
3
3
3
2
H
2
1
3
and 80.0 mL of CH
temperature < -75 °C). SnCl
2.43 mL, 5.40 g, 20.7 mmol, 2.5 equiv), and the reaction mixture
2
Cl
2
and cooled in CO
2
/acetone bath (internal
(m, 1 H, HHC(7)), 0.85 (td, J ) 13.0, 4.8, 1 H, HHC(7)); C NMR
was added dropwise via syringe
(101 MHz, CDCl ) 171.2 (C(17)), 139.9 (C(14)), 118.3 (C(13)),
3
4
(
90.3 (C(2)), 84.9 (C(6)), 71.5 (C(10b), 52.3 (C(18)), 38.9 (C(10a)),
38.2 (C(1)), 32.9 (C(6a)), 32.0 (C(7)), 28.3 (C(11)), 26.1 (C(15)
or C(16)), 25.8 (C(9)), 25.1 (C(10)), 22.1 (C(12)), 20.7 (C(8)), 18.9
was stirred at this temperature for 30 min. The reaction was then
quenched by the addition of ethyl acetate (30 mL), followed by a
solution of Et
of NaHCO (120 mL). The contents were extracted with ethyl
acetate (3 × 100 mL), and the combined organic layers were washed
with a satd aq solution of NH Cl (100 mL) and brine (100 mL),
dried (MgSO ), filtered, and concentrated in vacuo. The colorless
3
N in methanol (1 M, 30 mL) and a satd aq solution
2 2
(C(15) or C(16)); IR (neat from CH Cl solution, salt plates) 2925
(s), 2869 (m), 1755 (s), 1748 (s), 1674 (w), 1440 (m), 1377 (w),
1274 (w), 1226 (m), 1199 (m), 1180 (m), 1141 (w), 1058 (w), 1038
(w), 995 (w), 971 (m), 907 (w); MS (EI, 70 eV) 321 (3), 291 (8),
262 (16), 237 (21), 205 (16), 203 (57), 178 (18), 175 (22), 161
(10), 150 (13), 149 (13), 148 (25), 147 (59), 145 (16), 135 (27),
134 (16), 133 (66), 131 (16), 121 (36), 119 (17), 109 (19), 107
3
4
4
oily residue (2.71 g) was placed into a one-neck, round-bottom
flask equipped with a nitrogen inlet and a stir bar, and then 60 mL
of benzene and 0.9 g of NaHCO
3
were added. The mixture was
(24), 105 (21), 100 (10), 95 (31); HRMS (ESI) calcd for C18
NO Na 344.1838, found 344.1850; TLC R 0.21 (hexanes/EtOAc
3/1) [silica gel, I , CAM]. Anal. Calcd for C17 (323.38):
27
H -
stirred for 6 h in an oil bath (60 °C, external temperature). The
reaction mixture was then filtered and concentrated in vacuo to
provide 2.63 g of greenish-yellow solid material, which was purified
by chromatography (silica gel, hexanes/EtOAc, gradient 3/1-1/1)
to afford 8 as a colorless oil (2.37 g, 7.61 mmol, 92%). Crystal-
lization from heptane yielded 8 as a white crystalline compound
4
f
2
H25NO
5
C, 67.26; H, 8.47; N, 4.36. Found: C, 67.54; H, 8.55; N, 4.42.
rel-(1S,2S,4R,6S,6aR,10aS,10bR)-1,6a-Ethano-2,6-dihydroxy-
decahydropyrrolo[2,1-a]isoquinolin-3(10bH)-one (10). A pea-
sized amount of Raney nickel was washed with water (3 × 20 mL)
and methanol (3 × 20 mL). The suspension of the washed catalyst
in methanol (50 mL) was transferred via pipet into a 100-mL, one-
neck, round-bottom flask equipped with a stir bar and charged with
40 (390 mg, 1.32 mmol). The flask was attached to a glass hydrogen
line, cooled in an ice bath (0 °C, external temperature), evacuated,
and flushed with hydrogen five times, and the mixture was stirred
vigorously under 1 atm of hydrogen for 15 h with warming to rt.
Sodium methoxide (230 mg, 4.25 mmol, 3.2 equiv) was added,
and the mixture was stirred under nitrogen for 50 min. After
acidification with acetic acid (0.3 mL, 5.0 mmol, 3.8 equiv), the
reaction mixture was diluted with MTBE (50 mL), filtered through
Celite, concentrated in vacuo, and dried by azeotropic distillation
with benzene in vacuo at rt to provide a white solid. This residue
was purified by chromatography (silica gel; EtOAc/methanol,
gradient, 20/1-10/1) to afford 10 (310 mg, 1.24 mmol, 93%) as a
light yellow solid. Additional purification by crystallization from
toluene/hexanes provided 10 (251 mg, 1.00 mmol, 76%) as a white
crystalline material (fine needles). Data for 10: mp 183-184 °C
(
2.13 g, 6.85 mmol, 82%). Data for 8: mp 89-90 °C (heptane);
1
H NMR (500 MHz, CDCl
1
2
3
) 4.85 (t, J ) 3.9, 1 H, HC(2)), 4.24 (s,
H, HC(6)), 3.77 (s, 3 H, H C(15)), 3.48 (s, 3 H, H C(16)), 2.44-
.48 (m, 1 H, HC(1)), 2.17-2.27 (m, 2 H, HC(10a) and HHC-
3
3
(
11)), 1.87-1.97 (m, 2 H, HHC(11) and HHC(12)), 1.76-1.82 (m,
H, HHC(9)), 1.69-1.75 (m, 1 H, HHC(10)), 1.53-1.60 (m, 2 H,
HHC(8) and HHC(7)), 1.15-1.45 (m, 4 H, HHC(12), HHC(7),
HHC(9), and HHC(10)), 1.12 (s, 3 H, H C(13)), 1.08-1.12 (m, 1
H, HHC(8)); 13C NMR (126 MHz, CDCl
) 171.1 (C(14)), 108.5
C(6)), 90.1 (C(2)), 74.2 (C(10b)), 55.1 (C(16)), 52.4 (C(15)), 43.1
1
3
3
(
(
(
C(1)), 36.2 (C(10a)), 34.1 (C(6a)), 31.2 (C(7)), 26.7 (C(11)), 25.6
C(12) or C(10)), 25.6 (C(12) or C(10)), 23.5 (C(13)), 23.4 (C(9)),
20.6 (C(8)); IR (neat) 3494 (w), 3005 (w), 2954 (s), 2935 (s), 2861
(s), 2833 (m), 1772 (s), 1452 (s), 1439 (s), 1369 (m), 1344 (m),
1
9
288 (m), 1273 (m), 1211 (s), 1138 (s), 1099 (s), 1007 (s), 972 (s),
01 (m), 825 (m), 800 (m); MS (FAB) 312 (99), 281 (13), 280
(
1
(
55), 252 (23), 249 (10), 195 (10), 189 (21), 161 (26), 159 (10),
55 (24), 153 (17), 152 (22), 149 (12), 147 (14), 137 (12), 135.0
44), 133 (28), 121 (19), 120 (10), 119.0 (100), 117 (16), 107 (12);
TLC R ) 0.16 (hexanes/EtOAc, 4/1) [silica gel, I , CAM]. Anal.
Calcd for C16 (311.17): C, 61.72; H, 8.09; N, 4.50.
1
(toluene/hexanes); H NMR (400 MHz, CD OD) 4.87 (s, 2 H, OH),
3
f
2
4.46 (dd, J ) 4.2, 0.7, 1 H, HC(2)), 4.13 (dd, J ) 13.4, 7.3, 1 H,
HHC(5)), 3.60 (dd, J ) 9.3, 7.6, 1 H, HC(6)), 3.52 (t, J ) 3.8, 1
H, HC(10b)), 2.77 (ddd, J ) 13.8, 9.3, 0.8, 1 H, HHC(5)), 2.43-
2.50 (m, 1 H, HC(1)), 1.95-2.02 (m, 1 H, HC(10a)), 1.90 (ddd, J
) 14.4, 10.3, 6.1, 1 H, HHC(12)), 1.77-1.85 (m, 1 H, HHC(8)),
1.56-1.70 (m, 3 H, HHC(7), HHC(9), and HHC(11)), 1.36-1.56
H25NO
5
Found: C, 61.74; H, 8.12; N, 4.63.
rel-(1S,2S,4S,6R,6aR,10aS,10bR)-Methyl Octahydro-6-(2-meth-
ylpropenyl)-6H-1,6a-ethanoisoxazolo[2,3-c][2,3]benzoxazine-
2
-carboxylate (38a). A 50-mL, one-neck, round-bottom flask
equipped with a nitrogen inlet and a stir bar was charged with 26
422 mg, 1.31 mmol, 1.0 equiv) and toluene (10 mL) and cooled
(m, 4 H, H
(m, 2 H, HHC(7) and HHC(8)), 1.02 (dt, J ) 14.4, 5.9, 1 H, HHC-
(12)); 13C NMR (101 MHz, CDCl
) 173.9 (C(3)), 75.4 (C(2)), 73.1
2
C(10), HHC(9) and HHC(11)), HHC(7)), 1.22-1.34
(
to -76 °C (external temperature). Trimethylaluminum (2.0 M
solution in toluene, 1.97 mL, 3.94 mmol, 3.0 equiv) was added at
once via syringe. A deep red solution was instantly formed, which
3
(C(6)), 56.9 (C(10b)), 43.4 (C(5)), 40.7 (C(1)), 36.1 (C(6a)), 34.1
(C(10a)), 33.0 (C(7)), 29.7 (C(12)), 26.6 (C(8)), 24.9 (C(10)), 22.4
604 J. Org. Chem., Vol. 71, No. 2, 2006

Double-Intramolecular [4+2]/[3+2] Cycloadditions of Nitroalkenes
(
C(9)), 17.4 (C(11)); IR (neat from methanol/CH
plates) 3363 (s, br), 2933 (s), 2864 (s), 1685 (s), 1458 (m), 1328
w), 1288 (w), 1237 (w), 1215 (w), 1194 (w), 1114 (w), 1063 (w),
2
Cl
2
solution, salt
1.45 (m, 4 H, HHC(9), H
H, HHC(8)), 0.99 (d, J ) 6.8, 3 H, H
4.8, 1 H, HHC(7)); 13C NMR (101 MHz, CDCl
2
C(10) and HHC(11)), 1.12-1.26 (m, 1
C(13)), 0.93 (td, J ) 13.1,
) 77.2 (C(6)), 64.9
3
(
3
1
2
025 (w); MS (EI, 70 eV) 252 (10), 251 (25), 224 (15), 223 (100),
22 (33), 209 (10), 208 (12), 206 (23), 205 (56), 204 (42), 194
(C(10b)), 62.8 (C(3)), 62.5 (C(5)), 38.5 (C(10a)), 38.3 (C(2)), 36.2
(C(1)), 35.4 (C(7)), 34.8 (C(6a)), 26.2 (C(8)), 25.5 (C(10)), 21.5
(C(9)), 21.2 (C(11)), 20.6 (C(12)), 14.5 (C(13)); MS (EI, 70 eV)
235 (15), 207 (21), 206 (16), 192 (10), 111 (12), 110 (100), 96
(
12), 192 (17), 190 (12), 176 (12), 175 (18), 163 (10), 149 (12),
48 (13), 147 (19), 145 (13), 136 (20), 135 (97), 134 (20), 133
28), 131 (12); HRMS (ESI) calcd for C14 252.1600, found
0.28 (EtOAc/methanol 9/1) [silica gel, I , CAM].
(251.32): C, 66.91; H, 8.42; N, 5.57.
Found: C, 66.84; H, 8.54; N, 5.65.
1
(
H22NO
3
(34); HRMS (ESI) calcd for C15
TLC R 0.14 (CH Cl /methanol/ammonium hydroxide ) 10/2/0.5)
[silica gel, I , CAM]. Data for 51: mp 91-92 °C (hexanes); H
NMR (500 MHz, CDCl ) 3.81 (ddd, J ) 10.5, 7.2, 0.9, 1 H, HC-
H26NO: 236.2014, found 236.2007;
252.1598; TLC R
f
2
f
2
2
1
Anal. Calcd for C14
H21NO
3
2
3
rel-(1S,6S,6aR,10aS,10bR)-1,6a-Ethano-2-methyldodecahy-
dropyrrolo[2,1-a]isoquinolin-6-ol (50a) and rel-(1S,6S,6aR,10aS,
(6)), 3.65 (dd, J ) 12.9, 10.7, 1 H, HHC(3)), 3.49 (dd, J ) 12.7,
7.1, 1 H, HHC(5)), 2.98 (dd, J ) 4.5, 3.2, 1 H, HC(10b)), 2.81 (t,
J ) 11.8, 1 H, HHC(5)), 2.71 (dd, J ) 12.9, 7.7, 1 H, HHC(3)),
2.45-2.55 (m, 1 H, HC(2)), 2.17-2.25 (m, 1 H, HC(1)), 2.10 (dt,
J ) 11.9, 3.5, 1 H, HC(10a)), 1.55-1.90 (m, 10 H), 1.25-1.5 (m,
10bR)-1,6a-Ethano-2-methyldodecahydropyrrolo[2,1-a]isoquin-
olin-6-ol Borane Adduct (51). In a 100-mL, one-neck, round-
bottom flask equipped with a stir bar were placed 49 (74 mg, 0.32
mmol), methanol (3.0 mL), and platinum on carbon (5% w/w, 20
mg). The flask was attached to a glass hydrogen line and then was
evacuated and flushed with hydrogen five times. The mixture was
stirred vigorously under 1 atm of hydrogen for 13.5 h, diluted with
MTBE (20 mL), filtered through Celite, and concentrated in vacuo.
The residue was purified by chromatography (silica gel; dichlo-
romethane/methanol/ammonium hydroxide, gradient, 10/1.5/0.2-
5 H), 1.06 (td, J ) 13.0, 4.5, 1 H), 1.02 (d, J ) 6.9, 3 H, H
3
C-
(13)); 13C NMR (126 MHz, CDCl
) 75.3 (C(6)), 73.6 (C(3)), 72.0
3
(C(10b), 65.8 (C(5)), 35.6 (C(1)), 35.0 (C(7)), 34.9 (C(2)), 34.8
(C(10a)), 34.7 (C(6a)), 25.9, 25.1, 21.3, 20.4, 20.0, 4.2 (C(13)); IR
(neat from dichloromethane) 3438 (w, br), 2925 (s), 2858 (m), 2370
(s), 2318 (s), 2270 (m), 1494 (w), 1463 (w), 1458 (w), 1447 (w),
1380 (w), 1231 (w), 1164 (s), 1084 (w), 1046 (w), 1025 (s), 956
(m), 939 (w), 867 (w), 831 (w); MS (EI, 70 eV) 249 (17), 248
(92), 247 (32), 246 (44), 245 (12), 235 (33), 234 (100), 232 (11),
217 (13), 207 (15), 110 (23), 108 (12), 107 (11), 105 (11), 96 (11),
1
0/2/0.4) to afford 50a as a colorless glassy solid (53 mg, 0.23
mmol, 72%). To purify the material further, 47 mg (0.20 mmol,
.0 equiv) of 50a was placed into a 10-mL, one-neck, round-bottom
1
flask equipped with a nitrogen inlet and a stir bar. THF (1.0 mL)
95 (13); HRMS (ESI) calcd for C15
248.2189; TLC R 0.18 (hexanes/EtOAc, 3/1) [silica gel, I
Anal. Calcd for C15 28BNO (249.20): C, 72.30; H, 11.33; N, 5.62.
Found: C, 72.49; H, 11.41; N, 5.77.
H
27NOB 248.2186, found
was added, the solution was cooled -62 °C (external temperature),
f
2
, CAM].
and BH
3
(1.0 M solution in THF, 0.60 mL, 0.6 mmol, 3.0 equiv)
H
was added via syringe. After 20 min at -62 °C, the reaction mixture
was quenched with 2 mL of satd aq solution of NaHCO , warmed
SO ),
3
to rt, extracted with dichloromethane (3 × 5 mL), dried (Na
2
4
Acknowledgment. We are grateful for the National Institutes
of Health (GM30938) for generous financial support. R.B.
thanks the Alumni Donors of the Chemistry Trust of UIUC and
Johnson & Johnson Pharmaceutical Research Institute for
graduate fellowships.
and concentrated in vacuo. The residue was purified by chroma-
tography (silica gel; hexanes/EtOAc, gradient, 4/1-2/1) to afford
51 as a colorless oily material (29 mg, 0.12 mmol, 60%).
Crystallization from heptane and then hexanes provided 51 (26 mg,
1
0.10 mmol, 50%) as a white crystalline solid. Data for 50a:
H
NMR (400 MHz, CDCl ) 5. 43 (s br, 1 H, OH), 3.40-3.48 (m, 2
3
Supporting Information Available: Full characterization of all
products, detailed experimental procedures, and crystallographic
information files (CIF). This material is available free of charge
via the Internet at http://pubs.acs.org.
H, HC(6) and HHC(5)), 3.37 (dd, J ) 11.1, 9.4, 1 H, HHC(3)),
2
2
.68 (t, J ) 2.9, 1 H, HC(10b)), 2.58 (t, J ) 13.7, 1 H, HHC(5)),
.14-2.28 (m, 2 H, HC(2) and HHC(3)), 1.90-1.99 (m, 1 H, HC-
(1)), 1.68-1.81 (m, 3 H, HHC(7), HHC(8) and HHC(12)), 1.45-
1.58 (m, 4 H, HHC(9), HC(10a), HHC(11) and HHC(12)), 1.32-
JO052001L
J. Org. Chem, Vol. 71, No. 2, 2006 605