A stereodivergent cascade imine → azomethine ylide → 1,3-dipolar cycloadditive approach to α-chiral pyrrolidines

Article abstract of DOI:10.1016/j.tetlet.2005.05.119

Stereodivergent [3+2] cycloadditions of chiral α-amino azomethine ylides leading to highly functionalized pyrrolidines are reported. The marriage of substrate conformational preferences and either an inter- or intramolecular cycloaddition manifold leads t

Full text of DOI:10.1016/j.tetlet.2005.05.119

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 5181–5185
A stereodivergent cascade imine ! azomethine ylide ! 1,3-dipolar
cycloadditive approach to a-chiral pyrrolidines
*
Philip Garner and H. Umit Kaniskan
¨
Department of Chemistry, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106-7078, USA
Received 17 May 2005; revised 25 May 2005; accepted 26 May 2005
Available online 17 June 2005
Abstract—Stereodivergent [3+2] cycloadditions of chiral a-amino azomethine ylides leading to highly functionalized pyrrolidines are
reported. The marriage of substrate conformational preferences and either an inter- or intramolecular cycloaddition manifold leads
to either the l (syn) or u (anti) relationship between the pyrrolidine and a-stereocenters. The latter result may be applicable to a new
approach to the bioxalomycin family of antibiotics.
Ó 2005 Elsevier Ltd. All rights reserved.
Pyrrolidine rings are found in many bioactive natural
products and their synthetic analogs. The synthesis of
a-chiral pyrrolidine systems (see structure I in Scheme
1) is a particularly challenging goal. We became inter-
ested in this general problem in the context of a total
synthesis of complex bioactive alkaloids such as bioxalo-
mycin b11,2 using a strategy that employs a 1,3-dipolar
cycloaddition reaction to assemble the pyrrolidine D-
ring core of this molecule.3 In this particular scenario,
a complex pyrrolidine corresponding to I (X = NH2)
would make a very attractive structure-goal. The
required all cis-2,4,5-trisubstituted pyrrolidine I can be
assembled in one step by the endo-selective [3+2] cyclo-
addition of azomethine ylide II (M = metal or proton)
and dipolarophile III (acrolein equivalent).4 Both inter-
and intramolecular cycloaddition manifolds may be
envisioned for this process. This approach directly
addresses the significant challenge posed by the anti-dis-
posed C13c and C4a amino functions in the target mol-
ecule. The problem, then, reduces to controlling which
diastereoface of the dipole will be attacked by the chiral
azomethine ylide.
Before embarking on a program to explore such [3+2]
cycloadditions, the a-amino azomethine ylide corre-
sponding to II must be accessed via its precursor a-amino
imine IV. Although there was evidence that a-amino
iminium species could be trapped in an intramolecular,
nucleophilic sense without competing enamine forma-
tion,5 the issue was certainly not secure for the case at
hand. We were aware of a few examples of such a reac-
tion and none of them successfully addressed the stated
problem. Both Grigg et al.6 and Alcaide et al.7 had
reported cycloadditions of a-chiral azomethine ylides
derived from N-PMP-4-formyl-2-azetidinones to give
pyrrolidines related to I (X = N). These systems are
inherently resistant to enamine formation because of
the angle strain that would result due to the azetidinone
ring. Cheng et al. reported the microwave-induced intra-
Scheme 1. a-Chiral pyrrolidines via cascade imine-azomethine ylide-
1,3-dipolar cycloaddition.
Keywords: Pyrrolidine synthesis; Chiral azomethine ylides; Stereocon-
trolled [3+2] cycloadditions.
molecular [3+2] cycloaddition of
a serine-derived
azomethine ylide.8 Similarly, Chiacchio et al. reported
*
Corresponding author. Tel.: +1 216 368 3696; fax: +1 216 368
3006; e-mail: ppg@case.edu
the intramolecular cycloaddition of an a-sulfonamido
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.05.119

¨
P. Garner, H. U. Kaniskan / Tetrahedron Letters 46 (2005) 5181–5185
5182
azomethine ylide but found that the products were race-
mic.9 This result underscored the problem of facile and
reversible enamine formation. Recently, [3+2] cycload-
ditions of azomethine ylides corresponding to II
(X = O, R = riboside) were reported.10 More often than
not, the diastereoselectivity of these dipolar cycloaddi-
tions was not very good. We now report a solution to
the problem outlined in Scheme 1 that provides stereo-
controlled access to either syn- or anti-diastereomers
corresponding to I (X = N).
NMR spectra—even upon heating. This was presum-
ably due to rotamer populations with high interconver-
sion barriers. Somewhat surprisingly,12 the major
product of this reaction was found to possess the like
(l)13 or syn-stereochemistry. This conclusion was
reached after separately converting 11 and 12 into their
respective bicyclic c-lactams 13 and 14 with TFA and
then performing a series of NOE difference NMR experi-
ments on these conformationally constrained bicyclic
molecules. The diagnostic NOE correlations are
indicated in the scheme. Furthermore, Mosher amide
experiments on 13 indicated a 90% ee for this material,
demonstrating that the a-stereocenter had remained
largely intact throughout the process.
We first examined the model a-amino imine substrate 6,
the synthesis of which is shown in Scheme 2. The se-
quence began with the condensation of free phenylala-
nine methyl ester (1) with benzaldehyde to give imine
2 in 96% yield. The imine structure was clearly evident
from characteristic C13 (d 163.7) and H1 (d 7.90)
NMR signals. This compound was reduced with NaBH4
to give the N-benzylated derivative 3 in 94% yield. The
N-benzylamine function was then protected with a Boc
group to give 4 in 89% yield. Partial reduction of the
ester with DIBAL at ꢀ78 °C led to the corresponding
aldehyde 5 in 64% yield. This N-diprotected a-amino
aldehyde was then condensed with glycine methyl ester
to give desired a-amino imine 6 [NMR data: C13 (d
168.6) and H1 (d 7.69)]. In keeping with GriggÕs observa-
tions with a glycine-derived imine,11 disubstitution of
the a-amino group was found to be crucial for the suc-
cess of this condensation reaction.
The [3+2] cycloaddition reaction of a chiral azomethine
ylide derived from N-Boc-serinal acetonide14 was also
examined next (Table 1). The results of these experi-
ments with imine 16 and tert-butyl acrylate mirrored
those obtained with the phenylalaninal imine. Both
stoichiometric (entry 1) and achiral catalytic conditions
(entry 2) produced the undesired diastereomer 17 as the
major cycloadduct. The stereochemistry of 17 was
assigned based on a combination of J-coupling and
NOE data. The preference for this diastereomer with
this substrate is consistent with analogous nitrone-
derived azomethine ylide cycloadditions reported in
the literature.15 In an effort to overturn the substrateÕs
preference for the syn-diastereomer, we also tried the
asymmetric catalytic protocol of Schreiber using the
(R)-QUINAP ligand (entry 3).16 This resulted in a
definite increase in the percentage of the unlike (u) or
anti-diastereomer 18, but ligand control was still not
dominant in this stereochemically mismatched case.
Once again, no evidence of enamine formation or
b-elimination was observed.
As shown in Scheme 3, imine 6 reacted smoothly with
methyl acrylate in the presence of a stoichiometric
amount of AgOAc and Et3N (GriggÕs conditions) to give
a 6:1 mixture of cycloadducts 11 and 12 in 68% com-
bined yield. These isomeric (by HRMS) compounds
1
were isolated as oils but did not give very definitive H
Scheme 2. Synthesis of phenylalanine-derived imine substrates.

¨
P. Garner, H. U. Kaniskan / Tetrahedron Letters 46 (2005) 5181–5185
5183
Scheme 3. Intermolecular cascade 1,3-dipolar cycloaddition with phenylalanine-derived a-amino imine.
Table 1. Intermolecular cascade 1,3-dipolar cycloadditions with N-Boc-serinal acetonide imine
Entry
Reaction conditions
Combined yield (%)
Diastereomer ratio
1
2
3
AgOAc (1.2 equiv) Et3N (1.2 equiv) toluene, rt overnight
AgOAc (10 mol%) Ph3P (20 mol%) Et3N (10 mol%) toluene, rt overnight
AgOAc (3 mol%) (R)-QUINAP (3 mol%) Et3N (10 mol%) toluene, rt overnight
64
51
62
10:1
4.5:1
1:1
The diastereoselectivity of the intermolecular cycloaddi-
tion reactions of azomethine ylides derived from imines
6 and 16 can be rationalized by the pre-TS conforma-
tions A and B that minimize A-strain between the a-ste-
reocenter and the azomethine moiety (Fig. 1). In both
cases, this conformational preference forces the Boc
(or possibly the Bn group in the case of 6) to block
the Re-face of the azomethine ylide, leading to Si-attack.
We reasoned that one might be able to turn this confor-
mational preference to our advantage by tethering the
dipolarophile to the a-amino group as in pre-TS confor-
mation C (R = H). In such a case, the resulting (E,E)-
azomethine ylide would undergo intramolecular dipolar
cycloaddition to the acrylamide dipolarophile to give the
bicyclic lactam 14 directly.
Accordingly, we set out to examine the intramolecular
version of the cascade imine ! azomethine ylide ! 1,3-
dipolar cycloaddition with the phenylalanine-derived
model substrate 10 (Scheme 2). This compound was
readily prepared from amino ester 3 in four steps via a
sequence that began with reduction of the ester to give
the aminoalcohol 7. The amine was condensed with
acryloyl chloride to give the protected acrylamide 8,
and then subjected to Dess–Martin oxidation to pro-
duce the a-acrylamido aldehyde 9 in 45% overall yield.

¨
P. Garner, H. U. Kaniskan / Tetrahedron Letters 46 (2005) 5181–5185
5184
Figure 1. Rationale for stereodivergent intermolecular and intramolecular [3+2] cycloadditions.
Significantly, this sequence proceeded without interfer-
ence by the potentially reactive acrylamide moiety.
The aldehyde was condensed with glycine methyl ester
to give the imine 10, which was used directly. Exposure
of this compound to catalytic Ag(I) conditions resulted
in formation of two cycloadducts (Scheme 4). Gratify-
ingly, the major product was identical to compound 14
that had been obtained from the minor product in the
intermolecular cycloaddition sequence. A second iso-
meric compound, tentatively identified as the corre-
sponding epimeric methyl ester 19, was also obtained.
This cyclo- adduct could result from cycloaddition to
the isomeric (Z,E)-azomethine ylide or epimerization
of 14. Unfortunately, the combined yield of these two
cycloadducts was never above 30%. A compound that
we believe to be imidazolidine 20 (stereochemistry unas-
signed) was also found as a by-product and its forma-
tion provided us with a hint to the way around this
problem.
gap or increasing the concentration of the reactive
dipole. One or both of these conditions could be met
by replacing H–Gly–OMe with dimethyl 2-aminomalo-
nate (21).19 Note also that, in this case, the same product
would form irrespective of azomethine ylide geometry.
The optimal conditions for this [3+2] cycloaddition
involved simply mixing the aldehyde 9 and amine 21
together in THF. The reaction proceeded rapidly at
room temperature to give good yields of a single cyclo-
adduct, identified as compound 22 (Scheme 5). The u
(anti) stereochemistry of 22 was readily ascertained by
the absence of J-coupling between the pyrrolidine H-5
and a-proton. This situation is the consequence of a
90° dihedral angle, that is only possible with this diaste-
reomer. Thus, the desired cascade imine ! azomethine
ylide ! 1,3-dipolar cycloaddition sequence occurred
sequentially and selectively in one-pot.
In conclusion, we have defined a new approach to either
l (syn) or u (anti) configured a-chiral pyrrolidines based
on stereocontrolled [3+2] cycloadditions of chiral a-amino
azomethine ylides. Both inter- and intramolecular [3+2]
cycloadditions reactions of potentially labile a-amino
azomethine ylides proceed selectively without significant
a-racemization. The intermolecular [3+2] cycloaddition
leads to the l (syn) diastereomeric relationship whereas
the intramolecular cycloaddition manifold provides the
u (anti) diastereomeric series. This intramolecular cas-
cade imine ! azomethine ylide ! 1,3-dipolar cycload-
dition proceeds efficiently in one-pot under very mild
We reasoned that imidazolidine 20 was formed because
of ineffective trapping of the azomethine ylide C by the
acrylamide dipolarophile. One possible explanation for
attenuated reactivity was that the higher energies of
the acrylamide versus acrylate p molecular orbitals
resulted in a larger dipoleHOMO–dipolarophileLUMO fron-
tier molecular orbital (FMO) energy gap.17 This led to
competitive intermolecular [3+2] cycloaddition to the
imine double bond.18 The rate of azomethine cycloaddi-
tion could be accelerated by either reducing the FMO
Scheme 4. Intramolecular cascade [3+2] cycloaddition of glycine-derived azomethine ylide.

¨
P. Garner, H. U. Kaniskan / Tetrahedron Letters 46 (2005) 5181–5185
5185
4. Reviews of azomethine ylide cycloaddition chemistry: (a)
Grigg, R.; Sridharan, V. In Advances in Cycloaddition;
Curran, D. P., Ed.; JAI Press: Greenwich CT, 1993; Vol.
3, pp 161–204; (b) Kanemasa, S.; Tsuge, O. In Advances in
Cycloaddition; Curran, D. P., Ed.; JAI Press: Greenwich
CT, 1993; Vol. 3, pp 99–159; (c) Kanemasa, S. Synlett
2002, 1371; (d) Pearson, W. H.; Stoy, P. Synlett 2003,
903.
5. Myers, A. G.; Kung, D. W. J. Am. Chem. Soc. 1999, 121,
10828.
6. Grigg, R.; Thornton-Pett, M.; Xu, J.; Xu, L.-H. Tetra-
hedron 1999, 55, 13841.
7. Alcaide, B.; Almendros, P.; Alonso, J. M.; Aly, M. F.
J. Org. Chem. 2001, 66, 1351.
Scheme 5. One-pot intramolecular cascade [3+2] cycloaddition of
amino malonate derived azomethine ylide.
8. Cheng, Q.; Zhang, W.; Tagami, Y.; Oritani, T. J. Chem.
Soc., Perkin Trans. 1 2001, 452.
9. Chiacchio, U.; Antonino, C.; Rescinfina, A.; Bkaithan,
M.; Grassi, G.; Piperno, A.; Privitera; Romeo, G. Tetra-
hedron 2001, 57, 3425.
10. Dondas, H. A.; Fishwick, C. W. G.; Grigg, R.; Kilner, C.
Tetrahedron 2004, 60, 3473.
reaction conditions. The intramolecular [3+2] cycloaddi-
tion forms the basis for a new synthetic approach to the
bioxalomycin family of antibiotics.
11. Blaney, P.; Grigg, R.; Rankovic, Z.; Thornton-Pett, M.;
Xu, J. Tetrahedron 2002, 58, 1719.
Acknowledgment
12. Additions to analogous nitrones suggested that the anti-
selectivity should have been obtained with this substrate.
See: Merino, P.; Franco, S.; Merchan, F. L.; Tejero, T.
J. Org. Chem. 1998, 63, 5627.
Acknowledgment is made to the donors of the American
Chemical Society Petroleum Research Fund for support
of this research.
13. Seebach, D.; Prelog, V. Angew. Chem., Int. Ed. Engl. 1982,
21, 654.
Supplementary data
14. Garner, P.; Park, J. M. Org. Synth. 1991, 70, 18, The
D-serinal derivative was actually used for these experi-
ments. However, since we were only concerned with
diastereoselectivity in these model studies, the reaction is
depicted as if the L-antipode were used to facilitate
stereochemical comparisons with the natural product.
15. Hanessian, S.; Bayrakdarian, M. Tetrahedron Lett. 2002,
43, 9441. Also, see Ref. 12.
Supplementary data associated with this article can be
found, in the online version at doi:10.1016/j.tetlet.2005.
05.119. General experimental procedures and characteri-
zation for all new compounds are provided.
16. Chen, C.; Li, X.; Schreiber, S. L. J. Am. Chem. Soc. 2003,
125, 10174.
References and notes
17. Calculated LUMO energies for methyl acrylate and
acrylamide are ꢀ1.900 eV and ꢀ1.403 eV, respectively.
See: Zhan, C.-G.; Nichols, J. A.; Dixon, D. A. J. Phys.
Chem. A 2003, 107, 4184.
18. Groundwater, P. W.; Sharif, T.; Arany, A.; Hibbs, D. E.;
Hursthause, M. B.; Garnett, I.; Nyerges, M. J. Chem.
Soc., Perkin Trans. 1 1998, 2837.
19. (a) Joucla, M.; Hamelin, J. Tetrahedron Lett. 1978,
2885; (b) Amornraksa, K.; Grigg, R. Tetrahedron Lett.
1980, 21, 2197; (c) Husinec, S.; Savic, V.; Porter, A. E. A.
Tetrahedron Lett. 1988, 29, 6649; (d) Blazey, C. M.;
Heathcock, C. H. J. Org. Chem. 2002, 67, 298.
1. Zaccardi, J.; Alluri, M.; Ashcroft, J.; Bernan, V.; Kors-
halla, J. D.; Morton, G. O.; Siegel, M.; Tsao, R.; Williams,
D. R.; Maiese, W.; Ellestad, G. A. J. Org. Chem. 1994, 59,
4045.
2. For a comprehensive review of the bioxalomycins and
related tetrahydroisoquinoline antibiotics, see: Scott, J.
D.; Williams, R. M. Chem. Rev. 2002, 102, 166.
3. For an earlier, alternative 1,3-dipolar cycloadditive
approach, see: Garner, P.; Cox, P. B.; Anderson, J. T.;
Protasiewicz, J.; Zaniewski, R. J. Org. Chem. 1997, 62,
493.