Synthesis of (-)-(R)-pyrrolam A and studies on its stability: A caveat on computational methods

Article abstract of DOI:10.1021/jo049460r

The asymmetric synthesis of (-)-(R)-pyrrolam A was achieved in three operations from N-Boc pyrrolidine via α-(N-carbamoyl)alkylcuprate vinylation reaction followed by N-Boc deprotection and cyclization. One-pot deprotection-cyclization procedures led to mixtures of pyrrolam A and its double bond isomers. These isomerization events could be circumvented by use of a two-step procedure. To guide aspects of the experiments, a series of computational energy evaluations and chemical shift predictions were performed with molecular mechanics, semiempirical, ab initio, and density functional methods. The relative stabilities of the double bond isomers, as estimated by experiment, challenged a number of computational methods, and only the MP2 model with its moderate degree of electron correlation came close to matching the experimental data. The MP2 method was further applied to an unusual aspect of the double bond migration between pyrrolam A and its isomer 9. The reaction (1 to 9) on neat samples is irreversible without racemization, and the alumina-mediated equilibration is accompanied by complete loss of enantiomeric excess. The source of the irreversibility was traced to asymmetric charge distribution in the intermediate dienolate anion. The analysis ultimately led to a semiquantitative sketch of the pyrrolam energy surface.

Full text of DOI:10.1021/jo049460r

Syn th esis of (-)-(R)-P yr r ola m A a n d Stu d ies on Its Sta bility:
A Ca vea t on Com p u ta tion a l Meth od s
Rhett T. Watson,† Vinayak K. Gore,† Kishan R. Chandupatla,† R. Karl Dieter,*,† and
J ames P. Snyder*,‡
Department of Chemistry, Hunter Laboratory, Clemson University, Clemson, South Carolina 29634, and
Department of Chemistry, Emory University, 1515 Dickey Drive, Atlanta, Georgia 30322
dieterr@clemson.edu; snyder@euch4e.chem.emory.edu
Received April 1, 2004
The asymmetric synthesis of (-)-(R)-pyrrolam A was achieved in three operations from N-Boc
pyrrolidine via an R-(N-carbamoyl)alkylcuprate vinylation reaction followed by N-Boc deprotection
and cyclization. One-pot deprotection-cyclization procedures led to mixtures of pyrrolam A and
its double bond isomers. These isomerization events could be circumvented by use of a two-step
procedure. To guide aspects of the experiments, a series of computational energy evaluations and
chemical shift predictions were performed with molecular mechanics, semiempirical, ab initio, and
density functional methods. The relative stabilities of the double bond isomers, as estimated by
experiment, challenged a number of computational methods, and only the MP2 model with its
moderate degree of electron correlation came close to matching the experimental data. The MP2
method was further applied to an unusual aspect of the double bond migration between pyrrolam
A and its isomer 9. The reaction (1 to 9) on neat samples is irreversible without racemization, and
the alumina-mediated equilibration is accompanied by complete loss of enantiomeric excess. The
source of the irreversibility was traced to asymmetric charge distribution in the intermediate
dienolate anion. The analysis ultimately led to a semiquantitative sketch of the pyrrolam energy
surface.
In tr od u ction
The pyrrolizidine skeleton is found in many alkaloids
isolated from plant extracts1 and from insects2 that feed
upon the plants. These alkaloids and their metabolized
products serve as pheromones and defensive agents in
insect biology and display a range of biological activities
in mammals.3 Pyrrolams A-D (Figure 1, 1-4) were
isolated by Zeeck and co-workers in 1990 from the
bacterial strain Streptomyces olivaceus (strain Tu¨ 3082)
by elution with methanol from Amberlite XAD-16 resin
followed by further purification via column chromatog-
raphy (silica gel, CH2Cl2/MeOH).4 Bioassays revealed
modest biological activity. Pyrrolam A was shown to have
low herbicidal activity and to damage fertilized fish
(Brachydanio rerio) eggs but was inactive against Gram-
positive and Gram-negative bacteria, fungi, molds, vi-
ruses, worms, protozoa, and tumor cell lines.4 Whereas
pyrrolam A was isolated as a pure enantiomer, pyrrolams
B-D were obtained as racemates. The chemical inter-
F IGURE 1. Pyrrolams A-D and dimer 5.
conversion of pyrrolams A and C was considered unlikely,
although the conversion of C to B was not ruled out.4
Crystalline 1 experienced dimerization to 5 when allowed
to stand in open vessels for extended periods of time.
Since the initial isolation, pyrrolam A has been synthe-
sized several times. Its conversion to pyrrolam C (3) has
not been observed in these synthetic endeavors.5-8
Carbinol amine 3 was nonetheless synthesized9 prior to
its reported isolation, and in one instance9c minor amounts
of 1 were isolated as a byproduct.
(5) Aoyagi, Y.; Manabe, T.; Ohta, A.; Kurihara, T.; Pang, G.-L.;
Yuhara, T. Tetrahedron 1996, 52, 869.
† Clemson University.
‡ Emory University.
(6) For syntheses of (S)-(+)-pyrrolam, A see: (a) Arisawa, M.;
Takezawa, E.; Nishida, A.; Mori, M.; Nakagawa, M. Synlett 1997, 1179
(b) Arisawa, M.; Takahashi, M.; Takezawa, E.; Yamaguchi, T.; Torisa-
wa, Y.; Nishida, A.; Nakagawa, M. Chem. Pharm. Bull. 2000, 48, 1593.
(c) Murray, A.; Proctor, G. R.; Murray, P. J . Tetrahedron 1996, 52,
3757.
(7) Giovenzana, G.; Sisti, M.; Palmisano, G. Tetrahedron: Asym-
metry 1997, 8, 515.
(8) Huang, P. Q.; Chen, Q. F.; Chen, C. L.; Zhang, H. K. Tetrahe-
dron: Asymmetry 1999, 10, 3827.
(1) (a) Wro´bel, J . T. In The Alkaloids: Chemistry and Pharmacology;
Brossi, A., Ed.; Academic Press: San Diego, 1985; Vol. 26, Chapter 7,
p 327 and references therein. (b) Robins, D. J . Fortschr. Chem. Org.
Naturst. 1982, 41, 115.
(2) Boppre´, M. Naturwissenschaften 1986, 73, 17.
(3) Mattocks, A. R. Chemistry and Toxicology of Pyrrolizidine
Alkaloids; Academic Press: London, 1986.
(4) Grote, R.; Zeeck, A.; Stu¨mpfel, J .; Za¨hner, H. Liebigs Ann. Chem.
1990, 525.
10.1021/jo049460r CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/05/2004
J . Org. Chem. 2004, 69, 6105-6114
6105

Watson et al.
SCHEME 1
F IGURE 2. Asymmetric syntheses of (-)-(R)-pyrrolam A (1)
and en t-1.
metric deprotonation11 in Et2O. Treatment of this orga-
nolithium reagent with a THF solution of CuCN‚2LiCl
at -78 °C afforded the scalemic stereogenic lithium
dialkylcuprate reagent, which gave vinylation product 7
upon quenching with ethyl (Z)-3-iodopropenoate.12 The
reaction afforded excellent enantioselectivity on both a
1.0 mmol (95:5-96:4 er) and a 10 mmol scale (90:10 er).
Deprotection and cyclization of 7 could be achieved in a
single operation. Utilization of trifluoroacetic acid gave
low yields of pyrrolam A (1), whereas good chemical yields
could be obtained with either PhOH/Me3SiCl13 or tri-
methylsilyltriflate on a 1-2 mmol scale (Scheme 2, Table
1, entries 1, 2-3, and 5-6). Efforts to scale-up the one-
pot N-Boc deprotection-cyclization protocol resulted in
decreasing amounts of 1 and an increased yield of the
Asymmetric synthetic routes to pyrrolam A (1) and its
enantiomer (en t-1) have all relied upon the chiral pool
(Figure 2) for the pyrrolidine ring or for its construction.
The first synthesis of 1 utilized expensive nonnatural
D-proline (A)5 and involved an intramolecular SmI2-
promoted cyclization, and a second approach from D-
proline relied on intramolecular lactam formation (C).7
The conversion of (S)-malic acid to 1 (B)8 required the
longest linear sequence but gave a reasonable overall
yield. Naturally occurring L-proline was converted into
en t-1 via key cyclization steps involving either olefin
metathesis (D)6a-b or Dieckmann cyclization (E).6c The
low yield in the olefin metathesis reaction was attributed
to the instability of the natural product under the
reaction conditions. We now report a three-pot asym-
metric synthesis of (R)-(-)-pyrrolam A. Observations
from our synthetic endeavor suggest that pyrrolams B-C
may well be artifacts of the isolation process arising from
pyrrolam A via its ∆1,7a-isomer (i.e., 9). The propensity
for this double bond migration prompted computational
studies on the relative stabilities of the two double bond
isomers, which revealed an unusual combination of
conjugation and strain effects that challenged a number
of computational methods. Experimental verification of
the most accurate calculations provides an assessment
of the ability of various modeling procedures to handle
strain and conjugation in small ring systems.
∆
1,7a-enamide 9 (entries 3-4 and 6-7). The distribution
of products 1 and 9 was a function of the experimental
conditions (Scheme 2, Table 1). Attempted purification
of the mixture of 1 and 9 via silica gel column chroma-
tography (Scheme 2) gave pyrrolam A (1) and carbinol
amide 3 (pyrrolam C). All efforts to cleanly effect the one-
pot conversion of 7 to 1 proved unsuccessful. We therefore
turned our attention to the HCl salt 8 (Scheme 2), which
was readily prepared in quantitative yield from 7 with
methanolic HCl (Me3SiCl/MeOH). Initial efforts to neu-
tralize 8 with various bases led to mixtures of bicyclic
lactams 1, 9, 3, and 10 (Scheme 2, Table 1).
Addition of a saturated aqueous NaHCO3 solution to
neat samples of 8 led to mixtures of pyrrolam A (1) and
its double bond isomer 9 in various amounts depending
upon the scale of the reaction (Table 1, entries 8-12).
As the scale increased from 1 to 8 mmol, the ratio 1:9
changed from 88:12 to 71:29. Removal of solvent and
exposure of the neat material to air, in one instance,
Resu lts a n d Discu ssion
Syn th esis. Pyrrolam A was readily synthesized in a
three-pot sequence employing R-(N-carbamoyl)alkyl-
cuprate methodology (Scheme 1).10 Pyrrolidine amine
protection gave N-Boc carbamate 6, which was converted
to a scalemic stereogenic organolithium reagent by asym-
(11) Beak, P.; Kerrick, S. T.; Wu, S.; Chu, J . J . Am. Chem. Soc. 1994,
116, 3231.
(12) (a) Dieter, R. K.; Topping, C. M.; Chandupatla, K. R.; Lu, K. J .
Am. Chem. Soc. 2001, 123, 5132. (b) Dieter, R. K.; Oba, G.; Chandu-
patla, K. R.; Topping, C. M.; Lu, K.; Watson, R. T. J . Org. Chem. 2004,
69, 3076.
(9) (a) Yoshifuji, S.; Arakawa, Y.; Nitta, Y. Chem. Pharm. Bull. 1985,
33, 5042. (b) Ikeda, M.; Harada, S.; Yamasaki, A.; Kinouchi, K.;
Ishibashi, H. Heterocycles 1988, 27, 943. (c) Flitsch, W.; Hampel, K.
Liebigs Ann. Chem. 1988, 387.
(10) For
a review see: Dieter, R. K. Heteroatomcuprates and
R-Heteroatomalkylcuprates in Organic Synthesis. In Modern Organo-
copper Chemistry; Krause, N., Ed.; J ohn Wiley & Sons: New York,
2002; p 79.
(13) (a) Kaiser, E., Sr.; Picart, F.; Kubiak, T.; Tam, J . P.; Merrifield,
R. B. J . Org. Chem. 1993, 58, 5167. (b) Dieter, R. K.; Lu, K. J . Org.
Chem. 2002, 67, 847.
6106 J . Org. Chem., Vol. 69, No. 18, 2004

Synthesis of (-)-(R)-Pyrrolam A
TABLE 1. Distr ibu tion of P yr r ola m A (1) a n d P r od u cts Der ived fr om It a s a F u n ction of Exp er im en ta l Con d ition s
productsa
entry
sample (mmol)
reaction conditions
CF3COOH, CH2Cl2
product sample
1
9
3
10
1
2
3
4
5
6
7
8
7 (1)
7 (1)
7 (2)
7 (5)
7 (0.5)
7 (2)
7 (7)
8 (1)
8 (3)
8 (5)
8 (5)
8 (8)
8
8 (8)
8 (1)
8 (1)
A
D
E
F
C
C
8
I
J
K
L
25
50c
48c
20a
62c
70
10b
30
PhOH, Me3SiCl, CH2Cl2
40
60c
TMSOTf, CH2Cl2
30
70c
12
15
19
20
29
30
33
90
99
77
80
80
16
80
21
0
20c
88
85
81
80
79
60
66
10
1
23
20
20
14
20
28
aqueous NaHCO3 (addition to 8)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
aqueous NaHCO3, 8 (THF suspension) inverse addition
Hexamethyldisilazane
LiHMDS (1.1 equiv), THF
LiHMDS (2.0 equiv), THF
cold neat sample, standing 12 h
10
13
A
B
C
D
E
F
G
H
(i) add Et2O, filter ppt; (ii) Al2O3 purification
CDCl3 NMR sample, 12 h, 25 °C
neat, open to air, solvent removal
purification on neutral alumina, EtOAc
purification on basic alumina, EtOAc
aqueous NaHCO3, CH2Cl2, inverse addition of 8
neat, under Ar, refrigeration, 24 h
neat, under Ar, refrigeration, 48 h
neat, under Ar, refrigeration, 72 h
neat, under Ar, 25 °C, 96 h
57
51
I
J
100d
90d
85d
77d
51d
55d
8
10
15
23
17
K
L
M
N
Oe
32
M
N
CDCl3 solution, refrigeration
45
(i) neutral alumina, EtOAc, Ar, 25 °C, 21 h;
(ii) purification on silica gel (CH2Cl2/MeOH)
purification on neutral alumina, EtOAc
18
54c
30
K
25f
75
a
b
Determined from 1H NMR ratios unless otherwise noted. Monocyclic amines observed (approximately 35%). c Yields based upon
isolated products purified by column chromatography. Enantiomeric ratio ) 90:10 as determined by chiral phase HPLC. e Dimer 5 was
d
formed (20% molar ratio) as a 53:47 mixture of diastereomers. f Racemic as determined by chiral phase HPLC.
SCHEME 2
respectively, indicating the greater stability of 10 relative
to 9. Although 10 has been reported in the literature,9c,14
no spectral data is available, and we were unable to
separate 9 and 10 by silica gel or alumina chromatogra-
phy. Nonetheless, GC-MS analysis of a mixture of 1:9:
10 (determined by NMR analysis) showed two separate
ion-current peaks, each of which displayed a similar mass
spectrum indicative of isomers. The 1H and 13C NMR
absorptions due to 1, 3, or 9 could be identified from the
NMR spectra of pure samples of 1, 3, or 9. The remaining
absorptions attributed to 10 include a triplet (δ 4.59,
1
calcd 5.4) in the H NMR spectrum for an enamide olefin
proton. 13C absorptions at δ 97.4 and 147.2 were assigned
to the enamide double bond, and that at δ 170.9 corre-
sponds to the amide carbonyl. To more fully characterize
10, we performed a series of calculations to predict the
NMR chemical shifts for the compound. Figure 3 includes
results for the latter and a comparison with compounds
1 and 9 as standards. It can be seen that density
functional (mPW1PW9115,16/6-311G*//Becke3LYP/6-311G*)
and fragment database (ACD)17 evaluations readily as-
sign the spectrum for 1 and 9 and equally accurately
account for the 1H and 13C spectra for structure 10.
Although the quantum chemical approach is more faith-
ful to experiment, both methodologies leave little doubt
caused isomerization to 10 (57%) and capture of water
to give carbinol 3 (13%) in addition to enamides 1 (14%)
and 9 (16%) (Table 1, entry 20). Although this event was
irreproducible, attempted purification of a mixture of 9:1
(99:1) on basic alumina gave 10:1:9 in a 51:28:21 ratio
as determined by NMR analysis (entry 22). These two
experiments gave a 78:22 and a 71:29 ratio of 10:9,
(14) Ha, D.-C.; Yun, C.-S.; Yu, E. Tetrahedron Lett. 1996, 37, 2577.
(15) Adamo, C.; Barone, V. J . Chem. Phys. 1998, 108, 664-675.
(16) Cimino, P.; Gomez-Paloma, L.; Duca, D.; Riccio, R.; Bifulco, G.
Magn. Res. Chem. 2004, in press.
(17) The Advanced Chemistry Development (ACD) Labs CNMR
Predictor [http://www.acdlabs.com/].
J . Org. Chem, Vol. 69, No. 18, 2004 6107

Watson et al.
1
F IGURE 3. Comparison of H and 13C NMR chemical shifts for pyrrolam isomers 1, 9, and 10 (ppm relative to TMS). The first
value is experimental, the second parenthetical value is mPW1PW91/6-31G*//Becke3LYP/6-311G*, and the third is the ACD
database fragment prediction.
as to the location of the double bond. A comparison of
several quantum chemical methods for calculating 13C
chemical shifts in the present series is discussed in
Molecular Modeling below. Subsequently, enamide 10
was synthesized via an intramolecular Wittig reaction
(Scheme 2) affording 13C NMR absorptions identical to
those abstracted from the spectrum for the mixture of
1:9:10 and to those calculated (Figure 3).
Neutralization of 8 with hexamethyldisilazane afforded
1 and 9 in a 66:33 (Table 1, entry 14) ratio. Manipulation
of this material (i.e., A) during NMR sample preparation
and recovery resulted in isomerization of 1 to 9 to afford
a 1:9 ratio of roughly 1:4 (entries 17-19), indicating
greater thermodynamic stability for enamide 9. At-
tempted neutralization of 8 with aqueous sodium meth-
oxide gave an intractable mixture by TLC analysis (20
spots). Neutralization of 8 with lithium hexamethyldisi-
lazide afforded largely enamide 9 (entries 15 and 16) with
small amounts of 1 obtained when a slight excess of
LHMDS was employed and only a trace of 1 when 2 equiv
of LHMDS were used. In general, the mixtures rich in 1
range from a ratio of 7.3:1 (entry 8) to 2:1 (entries 13
and 14) for an energy spread of ∆G298 ) 0.4-1.2 kcal/
mol. Those rich in 9 range from a ratio of 99:1 (entry 16)
to 1:1 (entry 20) for a similar inverse energy spread of
-0.1 to -2.7 kcal/mol. Entry 20 furthermore suggests a
10/1 ratio of 4:1.
Attempted purification of 9 (1:9 ) 1:99) by neutral
alumina flash column chromatography gave a 20:80
mixture of 1:9, and attempted purification of a mixture
of 9:1 (99:1) on basic alumina gave 10:1:9 in a 51:28:21
ratio (Table 1, entries 21-22). Although we were unable
to achieve careful equilibration between, 1, 9, and 10,
the relative stability of these double bond isomers ap-
pears to be in the order of 10 > 9 > 1 with the 9:1
stability difference reflected in an 80:20 ratio and the
10:9 stability difference represented by ratios of 71:29
to 78:22. Averaging leads to an approximate 10:9:1 ratio
of 12:4:1 (i.e., 69%, 25%, and 6%). At 25 °C these
distributions correspond to ∆G298 values of 0.0, 0.6, and
1.4 kcal/mol, respectively. All efforts to effect equilibra-
tion between 1, 9, and 10 under acid catalysis (e.g.,
p-TsOH/CH2Cl2, pyrridinium p-toluenesulfononate/CH2-
Cl2, etc.) resulted in decomposition of the material and
formation of additional products as evidenced by analyti-
cal TLC.
Although insoluble in THF, addition of a suspension
of 8 in THF dropwise to a cold rapidly stirring biphasic
mixture of CH2Cl2 and an aqueous NaHCO3 solution gave
rise to 1 and 9 (60:30) along with minor amounts of 3
(Table 1, entry 13). Dissolution of 8 in dichloromethane
followed by dropwise addition to a cold rapidly stirring
biphasic mixture of methylene chloride and aqueous
NaHCO3 gave clean pyrrolam A uncontaminated with its
double bond isomers (entry 23). This procedure was
reproducible and on a 3 mmol scale gave excellent yields
(90%) of pyrrolam A with 90:10 er as measured by chiral
stationary phase HPLC on a CHIRALCEL OD column
[cellulose tris(3,5-dimethylphenylcarbamate) on silica
gel]. An NMR sample of this material (i.e., 1) in CDCl3
did not undergo isomerization to 9 when stored in the
refrigerator for 96 h. A neat sample of crude 1 stored in
a round-bottom flask in the refrigerator slowly isomerized
1
to 9. H NMR and chiral stationary phase HPLC mea-
surements at 24 h [1:9 ) 90:10, er ) 90:10], 48 h [1:9 )
85:15, er ) 90:10], and 72 h [1:9 ) 77:23, er ) 90:10]
revealed that isomerization of 1 to 9 occurred without
racemization of the remaining 1, ruling out an equilib-
rium process (Table 1, entries 23-26). Eventual inad-
vertent addition of water to the sample led to formation
of 3 (entry 27) and eventually to a mixture of 1 and 3
(entry 28). Exposure of this sample to EtOAc/neutral
alumina for 21 h led to formation of dimer 5 as a 53:47
mixture of diastereomers [a set of doublets at δ 7.02 and
6.05 and a second set at δ 6.82 and 5.85 with J ) 6.0
Hz] (entry 29). Interestingly, the isolation study reported
the formation of 5 as a single diastereomer [δ 7.08 (d, J
) 5.8 Hz, 1H) and 6.06 (d, J ) 5.8 Hz, 1H)] when 1 was
allowed to stand neat for 3 weeks in an open vessel.4
Purification of the NMR sample (32 mg) by silica gel
column chromatography [MeOH/CHCl3 (1:9, v/v)] gave a
white solid whose melting point and specific rotation were
6108 J . Org. Chem., Vol. 69, No. 18, 2004

Synthesis of (-)-(R)-Pyrrolam A
TABLE 2. Exp er im en ta l a n d Ca lcu la ted Rela tive
En er gy Differ en ces betw een Dou ble Bon d Isom er s 1, 9,
a n d 10a
method
1
9
10
experimental
Beck3LYP/6-311G*
MP2/6-311G*
MP2/6-311+G*
MM3* (gas)
(CHCl3)
0.8 (21)
3.6 (0.2)
0.8 (15)
0.0(52)
0.0 (91)
0.0 (96)
0.0 (96)
0.0 (100)
0.0 (100)
0.0 (100)
0.0 (100)
2.5 (1)
3.3 (0.2)
3.1 (0.4)
3.3 (0.3)
2.1 (3)
8.4 (0.0)
7.4 (0.0)
0.0 (79)
0.7 (23.4)
0.0 (56)
0.3(32)
1.4 (9)
1.9 (4)
0.0 (76.4)
0.4 (29)
0.7(16)
4.1 (0.1)
4.5 (0.05)
4.0 (0.1)
24.0 (0.0)
24.7 (0.0)
25.3 (0.0)
15.9 (0.0)
1.4 (9)
0.0 (58)
0.0 (69.6)
0.9 (18)
1.2 (11)
0.0 (54)
(H2O)
1.9 (4)
MMFF (gas)b
(CHCl3)
32.5 (0.0)
33.3 (0.0)
34.3 (0.0)
4.5 (0.0)
0.0 (90)
0.2 (41.8)
0.5 (30)
0.0 (81.7)
0.0 (86)
0.1 (46)
0.4 (34)
(H2O)
Sybylc
Sybyld
AM1d (gas)
(H2O)e
PM3d (gas)
(H2O)e
MNDOd (gas)
(H2O)e
0.0 (66)
aCalculated populations at 298 K are given in parentheses, kcal/
b
mol. A nearly identical result is obtained from MMFF in Mac-
Spartan-02 and Spartan-04.27 c MacSpartan, Tripos force field.
d
Spartan-04, Tripos force field. e SM5.4/p23 as implemented in
Spartan-04.
close [26.5 mg, mp 56-58 °C, [R]D -21.7 (c ) 1.06,
CHCl3)] to those reported for the natural product.4 Thus,
conditions are available for the preparation of pyrrolam
A on a large scale.
MP 2 Eva lu a tion of th e P yr r ola m Isom er En er gy
Su r fa ce. Our goals in applying modeling to the pyrrolam
problem were 3-fold. First, given the variability of product
composition and ratios listed in Table 1, we sought an
accurate assessment of the relative energies for 1, 9 and
10 independent of the estimates from the tabulated
mixtures. Second, while we initially anticipated that
tautomeric exchange among the latter compounds could
be described in terms of a straightforward equilibrium
process, the rearrangement of neat 1 to 9 without
racemization and the infrequent appearance of 10 in the
mixtures suggested otherwise. Consequently, in an at-
tempt to understand the pyrrolam isomer potential
energy surface, we examined the possible role of the base-
promoted anionic intermediates 13 and 14 (Scheme 4 and
Figure 6). Third, the literature has been unclear on the
possibility of forming pyrrolam C (3) from the A form (1)
in contrast to its being a functional component of the
natural product isolate. We felt that evaluation of the
consequences of acid-promoted formation of an interme-
diate cation would assist interpretation of the experi-
mental outcomes. Although a variety of methods were
used to estimate relative energies for the pyrrolam
tautomers (Table 2), the most reliable protocols involved
optimizing molecular geometries with density functional
theory and a triple-ú basis set (Becke3LYP/6-311G*) in
some cases augmented with a diffuse function (Becke-
3LYP/6-311+G*) followed by energy evaluation with MP2
at the same level. Thus, energies and electronic proper-
ties were derived by MP2/6-311G*//Becke3LYP/6-311G*
and MP2/6-311+G*//Becke3LYP/6-311+G*.18 The diffuse
(i.e., “+”) function was included to accommodate the
negatively charged intermediates 13 and 14.
F IGURE 4. Geometries of the density functional Becke3LYP/
6-311+G*//Becke3LYP/6-311+G* optimized conformations of
1, 9, and 10, top and side views. Pyrrolam 1 exists in two
conformations, 1 and 1′.
P yr r ola m Isom er Rela tive En er gies. Table 1 il-
lustrates that the ratio of 1:9 formed depends on the
reaction microenvironment but that both acid and base
conditions lead to mixtures. Analysis of the tabulated
experiments above led us to conclude that the stability
order is most likely 10 > 9 > 1 with a ∆G298 energy
spread of 0.0, 0.6, and 1.4 kcal/mol (i.e., 12:4:1), respec-
tively. Clearly, the three isomers are similar in energy.
The MP2/6-311+G*//Becke3LYP/6-311+G* level predicts
the dominant isomer at room temperature to be 1 (52%
population; 0.0 kcal/mol) with 9 and 10 appearing in
smaller but observable amounts (32% and 16% popula-
tions, ∆Erel ) 0.3 and 0.7 kcal/mol, respectively). The
corresponding 10:9:1 ratio of 1:2:3.3 is to be compared to
the Table 1 analysis estimate of 12:4:1. The source of this
discrepancy is not apparent, though it could lie in the
fact that thermodynamic equilibration has not been
achieved in the isomerization reactions or with inaccura-
cies in the MP2 relative energies. Recalculation of the
latter with a solvent correction (polarizable continuum
model (PCM)19) at a dielectric constant of 6.89 (ethyl
acetate, ꢀ ) 6.02) gave nearly identical energy differences.
The major point of agreement is that energy differences
in both estimates fall within the rather narrow 1.4 kcal/
mol window. Structurally, compound 1 exists as a folded
structure in two conformations (Figure 4). The chairlike
(19) (a) Cance`s, M. T.; Mennucci, B.; Tomasi, J . J . Chem. Phys. 1997,
107, 3032. (b) Cossi, M.; Barone, V.; Mennucci, B.; Tomasi, J . Chem.
Phys. Lett. 1996, 286, 253. (c) Cossi, M.; Scalmani, G.; Rega, N.; Barone,
V. J . Chem. Phys. 2002, 117, 43.
(18) Gaussian 03, Revision A.1; Frisch, M. J . et al.; Gaussian, Inc.:
Pittsburgh, PA, 2003.
J . Org. Chem, Vol. 69, No. 18, 2004 6109

Watson et al.
SCHEME 3.
Isom er s of P yr r ola m A
∆
5,6- (11) a n d ∆6,7(12)-Dou ble Bon d
form is more stable than the boat form 1′ by 1.1 kcal/
mol. Isomer 9 is nearly flat but sustains a slight envelope
shape in the saturated five-membered ring. Compound
10 is completely planar.
All efforts to effect controlled acid-catalyzed intercon-
version of 1 and 9 or equilibration between 1, 9, and 10
led to complex mixtures of numerous products as evi-
denced by TLC analysis. The identities of the minor
components in these complex mixtures were not estab-
lished. In two instances (neat sample left standing,
treatment with basic alumina; Table 1, entries 20 and
22), putative enamide 10 was obtained as the major
component of the mixture. There are, of course, two other
double bond isomers within the pyrrolam family, namely,
1120 and 1221 (Scheme 3), both of which are known
compounds. At the MP2/6-311G*//Becke3LYP/6-311G*
and MP2/6-311+G*//Becke3LYP/6-311+G* levels we cal-
culate the energies relative to global minimum 1 to be
2.2 and 2.3 kcal/mol, respectively. At 25 °C these values
would correspond to populations of 1.2% and 1.0%,
respectively. If the calculations closely represent the
equilibration experiments, these compounds could be
present in small quantities as components of the complex
mixture of products. We have not, however, observed 11
or 12 in the mixtures.
Ra cem iza tion -F r ee Rea r r a n gem en t of 1 to 9.
Subsequent experiments with pyrrolam samples enriched
in 1 (i.e., 9:1 ) 15:85) and 9 (i.e., 9:1 ) 99:1) led to nearly
identical 75:25 and 80:20 mixtures of 9:1, respectively,
when either sample was subjected to column chromatog-
raphy on alumina with EtOAc at room temperature
(Table 1, entries 21 and 30). Accordingly, one is led to
postulate that the 75:25 to 80:20 value represents the
mixture at thermodynamic equilibrium under these
conditions. The striking observation that the neat rear-
rangement of 1 to 9 proceeds with a constant er of 90:10
for 1 suggests that equilibrium concentrations are ap-
proached without equilibrium. The implication is that
transformation of chiral 1 to achiral 9 is irreversible and
that thermodynamic conditions are not achieved. One
interpretation of the results assumes that isomerization
proceeds through the highly delocalized and planar anion
13 (Scheme 4).
F IGURE 5. Selected MP2/6-311+G* NBO bond orders and
atomic charges for intermediate anion 13 (a) and the corre-
sponding carbon analogue (b) geometry optimized with the
DFT Becke3LYP/6-311+G* model.
SCHEME 4
SCHEME 5
character is found between C-1 and C-7a (i.e., NC7a-C1).
Second, the three bonds to nitrogen are all single,
implying that the π-type lone electron pair on N has been
essentially isolated from the delocalized anion except for
its repulsive character. In resonance theory terms, the
relative contribution of the structures is thereby pre-
dicted to be in the order a > b > c (Scheme 5) with no
significant contribution from amide resonance.
The NBO charge distribution (Figure 5a) is in line with
this view but adds an additional insight. Whereas sizable
negative charge is located at carbon R to the carbonyl
(-0.46), the γ carbon is neutral or even slightly positive
(+0.04). Thus, nitrogen adjacent to C-γ not only repels
electrons through the π-framework but withdraws them
through the N-C σ-bond framework as well. Conse-
quently, once a proton is removed from 1 to give 13,
reprotonation is predicted to take place exclusively at O
or C-R in the neat reaction. If this view is accurate, then
the corresponding all-carbon anion would be expected to
possess electronic properties more conducive to protona-
tion at both R- and γ-carbons. Indeed, the calculations
displayed in Figure 5b illustrate negative charge at both
centers albeit polarized by the conjugated carbonyl.
The electronic characteristics of 13 at the MP2/6-
311+G* level of theory as expressed by natural bond
orbital (NBO) analysis22 are depicted in Figure 5a. The
Wiberg bond orders reveal two interesting features. First,
the π-bond with the greatest degree of double bond
Rea r r a n gem en t of 9 to 1. We speculated that under
conditions where achiral 9 rearranges to 1, reprotonation
should occur from both faces of planar anion 13 leading
to racemization. This hypothesis was tested with a
mixture of approximately 15:85 for 9:1 and a 90:10 er
for 1. The blend was chromatographed on an Al2O3
column in EtOAc to give a 75:25 mixture of 9:1 with an
er of 0. The latter result demonstrates that both 1 and 9
are deprotonated and confirms the predicted racemiza-
tion. In addition, the 75:25 ratio is a close match of the
(20) Thomas, E. W.; Rynbrandt, R. H.; Zimmermann, D. C.; Bell, L.
T.; Muchmore, C. R.; Yankee, E. W. J . Org. Chem. 1989, 54, 4535.
(21) (a) Martin, S. F.; Chen, H.-J .; Courtney, A. K.; Liao, Y.; Pa¨tzel,
M.; Ramser, M. N.; Wagman, A. S. Tetrahedron 1996, 52, 7251. (b)
Andersson, P. G.; Ba¨ckvall, J . E. J . Am. Chem. Soc. 1992, 114, 8696.
(22) (a) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988,
88, 899-926. (b) Weinhold, F. Natural Bond Orbital Methods. In
Encyclopedia of Computational Chemistry; Schleyer, P. v.R., Allinger,
N. L., Clark, T., Gasteiger, J ., Kollman, P. A., Schaefer, H. F., III,
Schreiner, P. R., Eds.; J ohn Wiley & Sons: Chichester, UK, 1998; Vol.
3, pp 1792-1811. (c) cf. http://www.chem.wisc.edu/∼nbo5/index.htm.
6110 J . Org. Chem., Vol. 69, No. 18, 2004

Synthesis of (-)-(R)-Pyrrolam A
F IGURE 6. (a) Anion 14 connecting pyrrolam isomers 9 and
10; (b) selected MP2/6-311+G* NBO bond orders and atomic
charges for 14.
80:20 proportion derived from the 99:1 sample of the two
isomers described above as an equilibrium value.
Elu siven ess of 10 a n d th e P yr r ola m En er gy Su r -
fa ce. Table 1 reveals that pyrrolam 10 rarely appears
under conditions where 1 and 9 can be interconverted.
Given that all three isomers are calculated to lie within
0.7 kcal/mol of one another (see above), this result is
surprising. Accordingly, we optimized the geometry of
anion 14 (Figure 6a), the putative intermediate that
connects 9 and 10. Anion 14 is characterized by ap-
proximate symmetry around the allylic fragment both in
terms of bond orders and atomic charges (Figure 6b). The
system can be expected to protonate readily at either of
the allylic carbons, facilitating a true equilibrium be-
tween 9 and 10. Unlike anion 13, 14 shows the classical
amide resonance as indicated by the enhanced and
reduced N-C(dO) and CdO bond orders, respectively.
Structure 14 with its planarity and 6-π electrons located
around the bridgehead is a candidate for Y-aromatic
character.23 However, the calculated single bond order
for the bridging N-C bond makes it clear that nitrogen
experiences little productive interaction with the es-
sentially π-isolated all-carbon allylic anion center.
In the two instances of the appearance of compound
10 in mixtures reported above, one of which involved
exposure to basic alumina, both 1 and 9 were likewise
present in near 1:1 ratios to the extent of 25-50% of 10.
This is consistent with the notion that under conditions
sufficient to produce the latter, the other isomers are
generated and interconverted as well. Assuming a base-
catalyzed process, this requires formation of both anions
13 and 14, the latter of which is predicted to be less stable
than the former by 19.9 kcal/mol (MP2/6-311+G*). The
considerable energy gap can be attributed to differences
in conjugation, strain, and internal electron repulsion.
Importantly, it explains the only occasional appearance
of 10 from the multitude of acid-base experiments
carried out to form the pyrrolam structure (Table 1). It
would appear that rather strong basic conditions are
required to strip the proton from C-7 in 9 to generate
anion 14. Once formed, however, the intermediate most
certainly resides in an environment energetically suf-
ficient to deliver all three double bond isomers.
F IGURE 7. Pyrrolam energy relationships based on a com-
bination of experiment and MP2/6-311+G*//Becke3LYP/6-
311+G* energies (kcal/mol), calculated energies in parenthe-
ses. Slightly nonplanar anion 13 (pictured upper left) is
surrounded by two transition states, the lower energy one
leading directly to either 9 or 15 or both. Anion 14 (upper right)
is planar.
ships in the pyrrolam system. Figure 7 provides the
graphic. Isomers 1, 9, and 10 are within 1 kcal/mol of
one another. Enol 15, not observed in the present work,
lies 13 kcal/mol above 1, and dienolate 13 is 20 kcal/mol
more stable than allyl anion 14. The irreversible and
racemization-free rearrangement of 1 to 9 is captured by
the intersecting dotted lines representing two transition
states of unspecified energies bracketing 13. These
indicate a strong kinetic preference for proton capture
by 13 to give achiral 9 or enol 15. We do not suggest that
15 is an obligatory intermediate on the path to 9, only
that the bias for protonating either C-2 or O versus C-7a
in 13 is highly favored (cf. Figure 5). Excessive quantities
of 9 in the presence of base (Al2O3) lead irreversibly to
the 80:20 ratio of 9:1, an approximation to thermody-
namic equilibrium. On the other hand, excessive quanti-
ties of 1 under similar conditions generate a comparable
75:25 ratio, this time establishing a true equilibrium.
P yr r ola m Ca p t u r e of Nu cleop h iles. Nucleophilic
capture during preparation of 1 and 9 most certainly
involves ammonium cation 16. Such a species is an
obvious intermediate to both the oxygenated substances
2, 3, and 4 and the dimer 5. To develop a perspective on
the ability of 16 to divert the neutral pyrrolam species
by nucleophile capture, we have constructed the reactions
of Scheme 6 at the MP2/6-311G*//Becke3LYP/6-311G*
level.
With the hydronium ion as the general acid proton
donor in eq 1, the reaction is strongly exothermic.
Capture of a water molecule leading to carbinol 3 is
estimated to be uphill in energy by 26.7 kcal/mol. The
overall result of hydrating enamide 9 to 3 corresponds
to an energy drop of 26.1 kcal/mol at the MP2 level. In
the context of equilibrating pyrrolam isomers, which
differ in energy by no more than 2-3 kcal/mol by the
same method, if 16 is a genuine intermediate, its interac-
The experimental observations and accompanying
calculations permit a rough sketch of the energy relation-
(23) Gobbi, A.; MacDougall, P. J .; Frenking, G. Ang. Chem. Int. Ed.
Engl. 1991, 103, 1001-1003. (b) Skancke, A. J . Phys. Chem. 1994, 98,
5234-5239.
J . Org. Chem, Vol. 69, No. 18, 2004 6111

Watson et al.
SCHEME 6
including polarization (6-311G*) is insufficient to improve
the energy of 1 relative to 9 so as to model the experi-
mental equilibrium ratio of 9:1, i.e., 80:20. It would
appear that the present pyrrolam series (1, 9, and 10)
presents a particularly challenging relative energy esti-
mation problem for the methods routinely applied in
synthetic chemistry. For the peculiar blend of delocal-
ization and strain across the trio of double-bond isomers,
the relative energetics is poorly handled by all parametric
methods and by a rather robust density functional
approach as well. Only the MP2 model, with its capacity
for a moderate degree of electron correlation, comes close
to matching the experimental data in terms of predicting
relative energies and populations. Even here, the ques-
tion of the relative energy of 10 remains an open
question. It might be possible to resolve this experimen-
tally by seeking more controlled conditions that allow the
range of isomers to be fully equilibrated. Similarly, higher
levels of theory that expand the basis set and include
additional layers of electron correlation would provide
additional confidence in the MP2 estimates for 1, 9, and
10. We see these issues as the subject of attention in a
future report.
SCHEME 7
1
With respect to the 13C and H chemical shift calcula-
tions, we employed single point ab initio (RHF and MP2)
and density functional (Becke3LYP and mPW1PW91)
methodologies with 6-31G* and 6-311G* basis sets ap-
plied to Becke3LYP/6-311G* optimized geometries for 1,
9 and 10. Table S-1 (Supporting Information) illustrates
that the MP2/6-311G* and mPW1PW91/6-31G* ap-
proaches provide superior agreement with experiment
and are comparable in accuracy. However, the MP2
calculations require much more computer memory (see
Experimental Section) and are approximately 20 times
slower. As a result the density functional method is
preferred for molecules in the size range of natural
products and drug-like molecules.16
tion with an oxygen-containing nucleophile in the pres-
ence of acid is predicted to be highly favored. To test this
idea we treated 9 separately with MeOH, H2O, or EtOAc
under acidic conditions, and in all instances obtained a
complex mixture of products. However, treatment of 9
dissolved in ethyl acetate with wet silica gel cleanly
afforded 3 in high yield, which in turn could be converted
to 2 by treatment with methanol and dry silica gel
(Scheme 7). Although the formation of 4 was not achieved,
these observations strongly suggest that pyrrolams B-D
are artifacts of the isolation procedure and not organism
metabolites.
Su m m a r y a n d Con clu sion s
Molecu la r Mod elin g. Prior to determination of the
identity and ratio of products from the experiments
reported in Table 1, we attempted to evaluate the relative
stabilities of pyrrolam isomers with various molecular
mechanics and semiempirical quantum chemical meth-
ods. The results were uneven and contradictory (Table
2). With one exception, all the force field variations in
MacroModel and different versions of Spartan delivered
structure 1 as the global minimum with virtually no
contribution from 9 or 10. Application of the continuum
solvation model had little impact on this outcome. The
MMFF method was particularly unfaithful to experiment,
suggesting abnormally high energies for the latter two
isomers. It is clear that the MMFF force field parameters
for the fused and unsaturated five-membered rings apply
unevenly across the range of double bond isomers. An
older version of Sybyl molecular mechanics accurately
predicts 9 to be the most stable species but seriously
underestimates the relative proportion of 1.
In summary, the rapid three-pot asymmetric synthesis
of (R)-(-)-pyrrolam A illustrates the synthetic potential
of scalemic stereogenic R-(N-carbamoyl)alkylcuprates for
the synthesis of enantiopure alkaloids. Conditions for
avoiding the isomerization of pyrrolam A (1) to enamide
9 have been developed, permitting the larger scale
synthesis of the natural product. A surprising result is
the observation that a neat sample of a 90:10 mixture of
1:9 with an er of 90:10 for chiral 1 underwent isomer-
ization in the cold without racemization. The process was
judged to be irreversible and kinetically controlled. NBO
analysis with the MP2/6-311+G* model reveals a highly
asymmetric charge distribution at C-2 and C-7a for the
dienolate anion 13. The corresponding negative and
neutral/positive atomic charges, respectively, appear to
lead to exclusive proton capture at C-2 to give 9. A
parallel set of experiments on a 15:85 mixture of 9:1 (er
of 90:10 for 1) treated to chromatography over alumina
led to a 75:25 equilibrium mixture with racemization.
Further geometry optimization using density functional
theory followed by MP2 energy evaluation (MP2/6-
311+G*//Becke3LYP/6-311+G*) allowed the construction
The semiempircal methods all predict 9 to be low in
energy and of somewhat comparable stability to 10.
However, the same methodology significantly underes-
timates the stability of 1. The results are not altered by
including the SM5.4 aqueous solvation model of Hawkins,
Cramer, and Truhlar.24 In this respect, the Becke3LYP
density functional approach even with a triple-ú basis
(24) Hawkins, G. D.; Cramer, C. J .; Truhlar, D. G. J . Phys. Chem.
1996, 100, 19824-19839; http://comp.chem.umn.edu/amsol/.
6112 J . Org. Chem., Vol. 69, No. 18, 2004

Synthesis of (-)-(R)-Pyrrolam A
1H), 5.94 (dd, J ) 1.46 Hz, 4.28 Hz, 1H), 7.13 (dd, J ) 1.68
Hz, 4.0 Hz, 1H); 13C NMR (CDCl3) δ 28.9, 29.8, 41.7, 67.8,
128.2, 148.8, 175.9; MS m/z (rel intensity) 123 (61, M+), 95
(72), 67 (100), 55 (10).
of a partial potential energy surface for the pyrrolam
isomers and their interconversion through dienolate and
allyl anion intermediates (Figure 7). A similar analysis
for iminium cation 16 permits the deduction that alcohol
3 and dimer 5 are most likely artifacts of the natural
product isolation procedure and not natural agents
themselves.
The facile rearrangement of 1 to 9 was not predicted
by molecular mechanics calculations based on the relative
stabilities of the two double bond isomers. It illustrates
the difficulties that small strained cyclic molecules with
conjugated functionality can pose for widely used molec-
ular mechanics programs. Although semiempirical meth-
ods fared better, they significantly underestimated the
stability of pyrrolam A. Only the MP2 model with a
moderate degree of electron correlation comes close to
matching the experimental data. Strain and conjugation
in small rings challenges many commonly used compu-
tational tools, and ab initio calculations should be
employed to confirm or deny the reliability of molecular
mechanics calculations on such molecules.
The enantiomeric purity of (R)-(-)-pyrrolam A (1) was
determined by chiral stationary phase HPLC on a CHIRACEL
OD column [cellulose tris(3,5-dimethylphenylcarbamate) on
silica gel] to be a 90:10 er [hexane/iPrOH, 90:10 (v/v), flow rate
) 0.5 mL/min, detection at λ ) 254 nm], which compares well
with the 90:10 er determined for the starting ester 7. The (S)-
enantiomer (minor) eluted first with a retention time of 22.14
min followed by the (R)-isomer (major) at 24.64 min. Purifica-
tion [silica gel chromatography, MeOH/CHCl3 (1:9, v/v)] of the
NMR sample (32 mg) of 1 gave a white solid [26.5 mg, mp )
56-58 °C, [R]D -21.7] comparable to the literature reports [mp
) 62 °C, [R]24 -29.3 (c 1.0, CHCl3);4 mp ) 60-62 °C, [R]D
D
-26.3 (c 0.31, CHCl3);5 mp ) 59 °C, [R]24D -26.3 (c 0.8, CHCl3);7
mp ) 61.5 °C, [R]24 -26.3 (c 0.4, CHCl3)8].
D
2,5,6,7-Tetr a h yd r o-3H-p yr r olizin -3-on e (9). The HCl salt
8 (205 mg, 1.0 mmol) was suspended in THF (5.0 mL) and
added to a rapidly stirring, cold (0 °C) solution of lithium
hexamethyldisilazide (LiHMDS, 2.0 equiv) in THF (10 mL).
After 20 min of stirring, H2O (10 mL) was added, and the
layers were separated. The organic layer was dried (MgSO4)
and concentrated in vacuo to afford enamide 9 as a light
yellow, clear oil (118 mg, 96%): 1H NMR (CDCl3) δ 2.20-2.40
(m, 2H), 2.45-2.55 (m, 2H), 3.25 (dt, J ) 2.7 Hz, 2.9 Hz, 2H),
3.43 (t, J ) 6.8 Hz, 2H), 4.79 (t, J ) 2.1 Hz, 1H); 13C NMR
(CDCl3) δ 23.6, 27.5, 40.2, 43.2, 93.3, 149.2, 174.9.
Exp er im en ta l Section
1,1-Dim et h ylet h yl E st er 2(R)-[(1Z)-3-E t h oxy-3-oxo-1-
p r op en yl]-1-p yr r olid in eca r boxyla te (7).12 N-Boc-pyrroli-
dine (342 mg, 2.0 mmol) was dissolved in freshly distilled Et2O
(6.0 mL) along with (-)-sparteine (468 mg, 2.0 mmol). The
reaction mixture was cooled to -78 °C under an argon
atmosphere, and s-BuLi (1.25 mL, 2.0 mmol) was added
dropwise by syringe. The resultant solution was stirred at -78
°C for 1 h. Then a solution containing CuCN (90 mg, 1.0 mmol)
and LiCl (90 mg, 2.0 mmol) in THF (6.0 mL) was added
portion-wise by syringe. The mixture was allowed to stir at
-78 °C for 30 min before the addition of ethyl (Z)-3-iodopro-
penoate (226 mg, 1.0 mmol). The reaction was allowed to
slowly warm to room temperature overnight, diluted with Et2O
(20 mL), and quenched with HCl (5% aqueous, 15 mL). Upon
separation of the layers, the organic layer was dried (MgSO4)
and concentrated in vacuo to give an oil, which was purified
by column chromatography (Rf 0.3, petroleum ether/EtOAc,
90:10, v/v) to afford carbamate 712 as a clear colorless oil (229
mg, 85%).
Eth yl 3-[(2R)-2-P yr r olid in yl]-2-(Z)-p r op en oa te Hyd r o-
ch lor id e (8). To a solution of pyrrolidinyl carbamate 7 (269
mg, 1.0 mmol) in MeOH (5.0 mL) was added chlorotrimeth-
ylsilane (216 mg, 2.0 mmol). The mixture was allowed to stir
overnight at room temperature and then concentrated in vacuo
to afford a thick oil, which crystallized upon addition of THF
to 8 as a white, highly hygroscopic solid (200 mg, 97%): 1H
NMR (CDCl3) δ 1.23 (t, J ) 7.1 Hz, 3H), 1.70-1.95 (m, 1H),
1.95-2.20 (m, 2H), 2.40-2.60 (m, 1H), 3.25-3.55 (m, 2H), 4.12
(q, J ) 7.1 Hz, 2H), 5.11 (br s, 1H), 5.94 (d, J ) 11.2 Hz, 1H),
6.68 (m, 1H), 9.78 (br s, 2H); 13C NMR (CDCl3) δ 14.0, 23.6,
31.3, 45.2, 57.6, 60.7, 123.6, 141.6, 165.1.
1,2,5,6-Tetr a h yd r o-3H-p yr r olizin -3-on e (10).14 Enamide
9 was left exposed to the laboratory atmosphere for 20 min
with only residual traces of solvent (EtOAc) remaining.
Concentration in vacuo afforded the isomeric enamide 10 as
the major component of a 78:22 mixture of 10:9. A mixture of
1:9:10 enriched in 10 was also obtained by attempted purifica-
tion of 9 by column chromatography on basic alumina: 1H
NMR (CDCl3) δ 2.50-2.65 (m, 2H), 2.65-2.80 (m, 2H), 2.80-
3.00 (m, 2H), 3.58 (t, J ) 8.7 Hz, 2H), 4.59 (t, J ) 1.9 Hz, 1H);
13C NMR (CDCl3) δ 18.6, 33.6, 34.9, 40.0, 97.4, 147.2, 170.9
Alter n a tive Syn th esis of (10). To a suspension of the
phosphonium iodide (529 mg, 1.0 mmol, Scheme 2, see Sup-
porting Information for preparation) in anhydrous THF (30
mL) was added potassium tert-butoxide (136 mg, 1.2 mmol)
at 0 °C. The reaction mixture was stirred at room temperature
for 18 h and then refluxed for 3 h. After cooling to room
temperature, the mixture was poured into ice-water (50 mL)
and extracted with diethyl ether (3 × 20 mL). The combined
extracts were dried over anhydrous MgSO4 to give the crude
material (100 mg) contaminated with small amounts of Ph3-
PO: 1H NMR (CDCl3) δ 2.58 (br s, 2H), 2.77 (br s, 2H), 2.92
(br s, 2H), 3.62 (t, J ) 8.6, 2H), 4.55 (s, 1H); 13C NMR (CDCl3)
δ 18.7, 33.7, 35.0, 40.1, 97.6, 146.6, 170.9.
P yr r ola m C or (Hexa h yd r o-7a -h yd r oxy-3H-p yr r olizin -
3-on e (3). Enamide 9 (123 mg, 1 mmol) was dissolved in
EtOAc (5 mL), and wet silica gel (500 mg, 400 mesh) was
added. The slurry was stirred at room temperature for 12 h
and then filtered. Concentration in vacuo followed by chro-
matography on alumina (Rf ) 0.3, 100% EtOAc) afforded
pyrrolam C (3) as a slightly yellow solid (127 mg, 90%): mp
71-72 °C; IR (film) 3378(br), 2981 (m), 1678 (s) cm-1; 1H NMR
(CDCl3) δ 1.52-1.64 (m, 1H), 1.98-2.25 (m, 4H), 2.25-2.42
(m, 2H), 2.92 (dt, J ) 17 Hz, 9.6 Hz, 1H), 3.10-3.20 (m, 1H),
3.38 (dt, J ) 5 Hz, 3.2 Hz, 1H), 5.00 (s, 1H); 13C NMR (CDCl3)
δ 25.6, 33.9, 34.0, 37.8, 40.5, 97.8, 174.9; MS m/z (rel intensity).
Anal. Calcd for C7H11NO2: C, 59.57; H, 7.80; N, 9.93. Found:
C, 59.74; H, 7.84; N, 9.73.
H exa h yd r o-7a -m et h oxy-3H -p yr r olizin -3-on e [P yr r o-
la m B] (2). Pyrrolam C (3) (127 mg, 0.9 mmol) was dissolved
in MeOH (5.0 mL), and dry silica gel (500 mg, 400 mesh) was
added. The slurry was stirred at room temperature for 12 h
and then filtered. Concentration in vacuo afforded a mixture
of 3 and pyrrolam B (2) as a clear, yellow oil (98 mg): 1H NMR
(R)-(-)-P yr r ola m A or (7a R)-5,6,7,7a -Tetr a h yd r o-3H-
p yr r olizin -3-on e (1). The HCl salt 8 (620 mg, 3.0 mmol) was
dissolved in CH2Cl2 (15 mL) and added portionwise at room
temperature to a rapidly stirring, biphasic mixture containing
saturated aqueous NaHCO3 (10 mL) and CH2Cl2 (40 mL). It
is imperative to use a biphasic mixture of methylene chloride
and aqueous sodium bicarbonate and to proceed with an
inverse addition. This mixture was allowed to stir for 12 h at
room temperature, and then the layers were separated. The
organic layer was dried (MgSO4) and concentrated in vacuo
to afford (R)-(-)-pyrrolam A (1) as a clear, light yellow oil (336
1
mg, 91%): IR(film) 1678 (s) cm-1; H NMR (CDCl3) δ 0.95-
1.25 (m, 1H), 1.80-2.05 (m, 1H), 2.05-3.00 (m, 2H), 3.10-
3.25 (m, 1H), 3.25-3.45 (m, 1H), 4.18 (dd, J ) 6.0 Hz, 4.7 Hz,
J . Org. Chem, Vol. 69, No. 18, 2004 6113

Watson et al.
(CDCl3) δ 1.54-1.61 (m, 1H), 2.04-2.17 (m, 3H), 2.20-2.32
(m, 2H), 2.44-2.53 (m, 1H), 2.80-2.89 (m, 1H), 3.06-3.14 (m,
1H), 3.21 (s, 3H), 3.55-3.63 (m, 1H); 13C NMR (CDCl3) δ 25.4,
28.2, 34.7, 37.3, 41.2, 49.8, 101.7, 175.4.
Thus, the carbon in question was evaluated at the MP2/6-
311G* level, whereas the remainder of the molecule was
treated at the MP2/3-21G* level.
Com p u ta tion a l Meth od s. All ab initio, density functional
calculations, as well as geometry optimizations, chemical shift
evaluations, and solvation calculations were performed with
Gaussian 03.18 MM3* and MMFF molecular mechanics geom-
etries and energy evaluations using the GBSA continuum
solvation method25 were carried out with MacroModel.26 AM1,
PM3, and MNDO results were obtained with various versions
of Spartan27 as indicated in the footnotes of Table 2. 13C NMR
chemical shifts using the MP2 model (Table S-1, Supporting
Information) required large memory requirements. As a result,
each 13C value was computed individually one job at a time
using a mixed basis set, the locally dense basis set method.28
Ack n ow led gm en t. This work was generously sup-
ported by the National Institutes of Health (GM-60300-
01) and the National Science Foundation (CHE-0132539).
Support of the NSF Chemical Instrumentation Program
for purchase of a J EOL 500 MHz NMR instrument is
gratefully acknowledged (CHE-9700278). We are grate-
ful to Professor Luigi Gomez-Paloma (University of
Salerno, Italy) for alerting us to the usefulness of the
mPW1PW91 DFT method.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra for compounds 1, 9, 10, 3, and the 1 + 9 and the 1 +
9 + 10 mixtures; 1H NMR spectrum for 9; chiral HPLC trace
for 1; procedure for preparation of the phosphonium salt used
in the synthesis of 10; and table of experimental and calculated
13C NMR chemical shifts for 1, 9, and 10. This material is
available free of charge via the Internet at http://pubs.acs.org.
(25) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J .
Am. Chem. Soc. 1990, 112, 6127-6129.
(26) (a) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liscamp,
R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.
J . Comput. Chem. 1990, 11, 440. (b) Cf. http://www.schrodinger.com/
(27) http://www.wavefun.com/
(28) Chesnut, D. B.; Moore, K. D. J . Comput. Chem. 1989, 10, 648-
659.
J O049460R
6114 J . Org. Chem., Vol. 69, No. 18, 2004