An efficient, enantioselective synthesis of branched polyhydroxylated pyrrolidines

Article abstract of DOI:10.1016/S0040-4039(00)01452-0

The enantioselective synthesis of branched polyhydroxylated pyrrolidines from a novel α-methylene bicyclic lactam (4b) is described. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01452-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 9377–9381
An efficient, enantioselective synthesis of branched
polyhydroxylated pyrrolidines†
Mark F. Mechelke and A. I. Meyers*
Department of Chemistry, Colorado State University, Fort Collins, CO 80523, USA
Received 2 August 2000; revised 28 August 2000; accepted 29 August 2000
Abstract
The enantioselective synthesis of branched polyhydroxylated pyrrolidines from a novel a-methylene
bicyclic lactam (4b) is described. © 2000 Elsevier Science Ltd. All rights reserved.
Polyhydroxylated pyrrolidines have generated a great deal of synthetic interest in the past
decade due to their structural resemblance to the sugar moiety of monosaccharides.1 Because of
their ability to mimic carbohydrates, they function as potent inhibitors of glycosidases; enzymes
that are involved in a wide range of important biological processes.2 The discovery that these
sugar mimetics display considerable activity against cancer,3 diabetes,1 and viral infections
(anti-HIV behavior)4 has led to an enhanced interest in them as potential therapeutic agents.
Recently, the syntheses of novel polyhydroxylated piperidines, 1, and pyrrolidines, (+)-2,
containing branched skeletons5 (Scheme 1) have been reported in an attempt to develop more
selective and potent glycosidase inhibitors. Due to our previous success in preparing azasugars
from bicyclic lactams6 and our continued effort in expanding the versatility of this chiral
molecule, we anticipated that the synthesis of branched polyhydroxylated pyrrolidines, 2, could
be achieved in high enantiomeric purity from an appropriately functionalized 5,5-bicyclic lactam
(4b).
It was envisioned that bicyclic lactam 4b, possessing an a-methylene substituent, could be
accessed from the known lactam 3b.6 The olefin, 4b, would be subjected to dihydroxylation
conditions to afford diol 5b—hopefully with some degree of exo–endo selectivity. Reductive
cleavage of 5b, after diol protection, would complete an efficient synthesis of pyrrolidine 6. The
latter could be further elaborated to azasugars such as 2, with the option of placing a variety of
substituents (R2) on the nitrogen atom.
* Corresponding author.
†
This paper is dedicated with respect and admiration to Professor Harry Wasserman on his 80th birthday.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01452-0

9378
Scheme 1.
The route to 2 and 6 began with the preparation of known bicyclic lactam 3b6 from
2,3-dihydrofuran in 47% yield over three steps. Readily available lactam 3a,7 containing an
angular methyl substituent, was also prepared and employed as a model. Attention was then
turned toward the introduction of the a-methylene substituent. Although a variety of methods
to prepare a-methylene lactones are known,8 all attempts to use these met with failure or limited
success. Diez and co-workers9 reported similar difficulties in preparing chiral a-methylene
2-piperidones. In an effort to avoid these problems, the enolate anions of both lactams 3a and
3b were treated with diethyl phosphorochloridite10 followed by immediate air oxidation of the
resultant reaction products (Scheme 2). In both cases (lactams 7a and 7b), the 31P NMR spectra
of the crude reaction mixtures revealed a 2.5:1 ratio of CꢀP to OꢀP bond formation products—
the CꢀP products being a mixture of exo–endo diastereomers (3:1). This crude mixture was
treated with NaH followed by paraformaldehyde to provide a 43% yield (55% yield based on
recovered starting material) of the desired a-methylene bicyclic lactams 4a and 4b over the two
steps.
Initial experiments to introduce vicinal hydroxyls were performed on lactam 4a, which was
treated with OsO4–NMO to furnish a disappointing 3:1 (NMR) mixture of diastereomers (Table
1). In order to increase the selectivity of the dihydroxylation, quinuclidine was added as a ligand
since it was anticipated that the ligand–OsO4 complex11 would provide increased facial selectiv-
ity due to its greater steric bulk as compared to OsO4 alone. The selectivity did increase,
providing a 7:1 mixture of diastereomers at the newly formed tertiary alcohol center (diol 5a).
Attention was now turned to bicyclic lactam 4b. Treatment of 4b with OsO4–quinuclidine,
however, gave a poor mixture of diastereomers (3:1) as compared to the model system (7:1).
This result suggested that the dihydroxylation might be occurring preferentially from the
exo-face; the decrease in exo-facial selectivity with lactam 4b perhaps arising from the larger
angular benzyloxymethyl substituent shielding the b-face of the lactam. It was felt that the
addition of a ligand bulkier than quinuclidine might improve the facial selectivity of the
dihydroxylation. Lactam 4b was subsequently treated with AD-mix-a and AD-mix-b,
respectively11 providing a greater than 15:1 (NMR) mixture of diastereomers. It is interesting to
note that in the reactions using either AD-mix-a or AD-mix-b, the same ratio of diastereomers
was observed (15:1) with the same sense of stereochemistry. This strongly suggests that the

9379
Scheme 2.
Table 1
Lactam
Ligand
de
3:1
7:1
3:1
Yield (%)
4a
4a
4b
4b
4b
–
89
96
93
98
100
Quinuclidine
Quinuclidine
AD-mix-a
AD-mix-b
15:1
15:1
increase in the observed facial selectivity is not due to the chirality of these ligands, but is due
solely to their steric bulk. The stereochemistry of the dihydroxylation products 5, were
tentatively assigned, at this point, as taking place from the exo-face.
The two hydroxyl groups were transformed to their benzyl ethers, 8, and the CꢀO cleavage
step was next addressed to prepare pyrrolidine 9. From earlier results in our laboratory,12 it was
anticipated that this reaction would proceed with retention of configuration of the angular
substituent. Thus, treatment of 8 with DIBAL-H furnished an 88% yield of pyrrolidine 9; the
reductive cleavage of the ring CꢀO bond occurring with very high diastereoselectivity (no trace
of C-2 diastereomers being observed). The 15:1 mixture of diastereomers obtained from the
dihydroxylation could also be separated by column chromatography at this point, providing
compound 9 as a single diastereomer. The stereochemistry at C-2 was tentatively assigned as

9380
having occurred with retention of configuration. Catalytic hydrogenation (H2, 4 atm, Pd(OH)2) in
the presence of Boc2O afforded branched pyrrolidine 6, which was immediately acetylated to
provide 10 in 68% yield. It is interesting to note that if Boc2O is not added to the hydrogenation,
only N-benzyl deprotection is observed and the O-benzyl groups remain intact.13 Pyrrolidine 10
provides a convenient entry into branched pyrrolidines, because by simply treating 10 with TFA
in CH2Cl2, the Boc protecting group can be removed (pyrrolidine 11), leaving the nitrogen free to
react with any number of electrophiles. The ease at which 10 is transformed into N-benzylated
pyrrolidine 2 illustrates the usefulness of this intermediate.
With an efficient, enantioselective synthesis of pyrrolidine 10 in hand (ten steps from
2,3-dihydrofuran, 12% overall yield), it was necessary to confirm the absolute configuration. Com-
mercially available (S)-(+)-5-(hydroxymethyl)-2-pyrrolidinone 12 was dibenzylated to afford
pyrrolidinone 13 in 80% yield (Scheme 3). The latter was olefinated [LDA/(EtO)2PCl, air, NaH/
(CH2O)n) to the a-methylene lactam 14 in 65% yield. Dihydroxylation of 14 employing OsO4–
NMO gave a 2.5:1 (NMR) mixture of diastereomers (diol 15). Similar results were obtained
using quinuclidine. Surprisingly, addition of AD-mix-a or AD-mix-b decreases the diastereoselec-
tivity, affording a 1:1 ratio of diastereomers. It was assumed that the major product in the OsO4–
NMO dihydroxylation arose from a-facial selectivity; the trans product being favored due to
shielding of the b-face by the benzyloxymethyl C-5 substituent. Dibenzylation of the diol moiety
followed by a DIBAL-H reduction furnished the tetrabenzyl pyrrolidine 17 in high yield.
Hydrogenolyses of the N-benzyl and O-benzyl groups in the presence of Boc2O afforded 18, which
was immediately acetylated to provide pyrrolidine 19 as a 3:1 mixture of diastereomers. When
1
1
the H NMR spectrum of pyrrolidine 19 was compared to the H NMR spectrum of pyrrolidine
10 prepared via the bicyclic lactam route, it became apparent that the minor diastereomer of 19
was identical to 10. With the absolute configuration of 19 known, this sequence proved the relative
stereochemistry of 10, with the acetoxymethyl substituents lying anti to each other. Since (+)-2
had previously been prepared in the literature,5b to confirm the absolute configuration of 10, it
only became necessary to determine the stereochemistry of 2 prepared from 10 above ([h]D lit.=
Scheme 3.

9381
+30.0,14 [h]D of 2=−44.2). The negative optical rotation observed for 2 proved that the
stereochemical assignment was correct, with the dihydroxylation proceeding from the exo-face
and the CꢀO ring reduction occurring with retention of configuration to afford (−)-2.
In conclusion, an efficient, enantioselective route to branched, N-substituted pyrrolidines has
been accomplished using chiral bicyclic lactams. A convenient entry into branched pyrrolidines
is now available using intermediate 10, which may be elaborated in a variety of ways.
Acknowledgements
The authors thank Merck for a post-doctoral fellowship (to M.F.M.) and the National
Institutes of Health for financial support.
References
1. (a) Baxter, E. W.; Reitz, A. B. J. Org. Chem. 1994, 59, 3175. (b) Sinnott, L. M. Chem. Rev. 1990, 90, 1171.
2. Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetrahedron: Asymmetry 2000, 11, 1645.
3. (a) Bernaki, R. J.; Korytnyk, W. Cancer Metastasis Rev. 1985, 4, 81. (b) Tsukamoro, K. et al. Clin. Res. 1989,
37A, 722.
4. (a) Fleet, G. W. J. et al; FEBS Lett. 1998, 237. (b) Gruters, R. A.; Niefies, J. J.; Tersmette, M. I.; DeGroede, R.
E.; Tulp, A.; Heisman, H. G.; Miedema, F.; Ploegh, H. L. Nature 1987, 330, 74.
5. (a) Ichikawa, M.; Igarashi, Y.; Ichikawa, Y. Tetrahedron Lett. 1995, 36, 1767. (b) Bennis, K.; Gelas, J.;
Thomassigny, C. Carbohydr. Res. 1995, 279, 307.
6. Meyers, A. I.; Andres, C. J.; Resek, J. E.; Woodall, C. C.; McLaughlin, M. A.; Lee, P. H.; Price, D. A.
Tetrahedron 1999, 55, 8931.
7. Romo, D.; Meyers, A. I. Tetrahedron 1991, 47, 9503.
8. Grieco, P. A. Synthesis 1975, 67 and references cited therein.
9. Forns, P.; Fernndez, M. M.; Diez, A.; Rubiralta, M.; Cherrier, M. P.; Bonin, M.; Quirion, J.-C. Synthesis 1999,
258.
10. Lee, K.; Wiemer, D. F. J. Org. Chem. 1991, 56, 5556.
11. Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. Rev. 1994, 94, 2483.
12. Meyers, A. I.; Burgess, L. E. J. Org. Chem. 1991, 56, 2294.
13. Czech, B. P.; Bartsch, R. A. J. Org. Chem. 1984, 49, 4076.
14. The [h]D value of our sample, −44.2, does not compare well to the literature value, +30.0, because their
compound was contaminated with trace amounts of another undetermined isomer—see Ref. 5b.
.