A memory of chirality approach to the stereoselective synthesis of 4-hydroxy-α-methylprolines

Article abstract of DOI:10.1021/ol071023m

To extend the memory of chirality (MOC) methodology to structurally more diverse compounds, the synthesis of 4-hydroxy-a-methylprolines was undertaken. Yield and selectivity were very good, with an unexpected reversal in selectivity observed for the cycli

Full text of DOI:10.1021/ol071023m

ORGANIC
LETTERS
2007
Vol. 9, No. 16
3029-3032
A Memory of Chirality Approach to the
Stereoselective Synthesis of
4-Hydroxy-r-methylprolines
Lawrence Kolaczkowski* and David M. Barnes
Abbott Global Pharmaceutical Research and DeVelopment, Process Chemistry,
1401 Sheridan Road, North Chicago, Illinois 60064-6285
larry.kolaczkowski@abbott.com
Received May 11, 2007
ABSTRACT
To extend the memory of chirality (MOC) methodology to structurally more diverse compounds, the synthesis of 4-hydroxy-r-methylprolines
was undertaken. Yield and selectivity were very good, with an unexpected reversal in selectivity observed for the cyclization of one adduct
with an unprotected hydroxyl.
Incorporation of R-alkyl-R-amino acids into synthetic pep-
tides and proteins can have a significant impact on structure.1
These conformationally restricted building blocks play an
important role as peptide conformation modifiers and as
enzyme inhibitors. The constraining effects exerted by these
nonproteinogenic building blocks are even more pronounced
when placed in a cyclic system such as R-methylproline.2
The diastereomers of 4-hydroxy-R-methylproline (2 and 3,
and their enantiomers) are structurally interesting variants
that have been underutilized due in part to the challenge of
stereoselective synthesis. While some of the four diastereo-
mers have been synthesized, most reported approaches3-7
to these targets rely on alkylation of a protected 4-hydroxy-
proline substrate followed by separation of the isomeric
products (eq 1).
One stereoselective synthesis has been reported; Seebach
found that pivaldehyde N,O-acetal alkylation (eq 2) cleanly
afforded trans-isomer 6 (ds >95%) after hydrolysis. How-
ever, this route is limited by the availability of the trans-4-
hydroxyproline,8-10 and does not allow access to the cis
isomers.
(1) Mendel, D.; Ellman, J.; Schultz, P. G. J. Am. Chem. Soc. 1993, 115,
4359-4360.
(2) Thaisvongs, S.; Pals, D. T.; Lawson, J. A.; Turner, S. R.; Harris, D.
W. J. Med. Chem. 1987, 30, 536-541.
(3) Noe, C. R.; Knollmu¨eller, M.; Vo¨llenkle, H.; Noe-Letschnig, M.;
Weigand, A.; Mu¨ehl, J. Pharmazie 1996, 15, 800-804.
(4) Huck, B. R.; Gellman, S. H. J. Org. Chem. 2005, 70, 3353-
3362.
(5) Ohtake, H.; Imada, Y.; Murahashi, S. Bull. Chem. Soc. Jpn. 1999,
72, 2737-2754.
10.1021/ol071023m CCC: $37.00
© 2007 American Chemical Society
Published on Web 07/12/2007

Memory of chirality (MOC)11-13 is a potentially powerful
strategy for the stereospecific construction of quaternary
centers. Kawabata and co-workers recently described an
MOC route to cyclic R-substituted amino acids (see Figure
1),14 a novel and operationally uncomplicated approach to
(eq 4) of the diastereomeric products. Inspection of models
suggests that crowding in the transition state may be
occurring, disfavoring the expected (2R,4R) product.
Figure 1. Kawabata’s MOC approach to cyclic amino acids with
a quaternary stereocenter.
this synthetic problem. In an effort to extend this methodol-
ogy to functionally more complex cyclic R-alkyl amino acids,
we undertook a study targeting the stereoselective synthesis
of 4-hydroxy-R-methylproline diastereomers.
Our approach to the synthesis of 4-hydroxy-R-methylpro-
line builds on the observation that epichlorohydrin reacts
stereospecifically with carboxy-protected alanine to give the
corresponding chiral chlorohydrin.15 Installation of protecting
groups and cyclization under standard conditions should lead
to the desired 4-hydroxy-R-methylproline. By judicious
choice of commercially available optically active starting
materials it should be possible to construct any of the four
stereoisomers. To demonstrate proof of principle we chose
to react the naturally occurring (S)-alanine benzyl ester with
(R)- and (S)-epichlorohydrin (Scheme 1).
We reasoned that a smaller alcohol protecting group might
decrease steric congestion and tilt the balance in favor of
the expected diastereomer. To test this theory we made the
corresponding MOM ether and carried out the cyclization
under identical conditions (eq 5). We were gratified to see
a small but noticeable increase in the selective formation of
the (2R,4R) diastereomer with the sterically less demanding
protecting group, supporting our hypothesis.
As expected, when optically active starting materials were
used, a single chlorohydrin diastereomer was formed. We
protected the nitrogen as the tert-butylcarbamate14,16 and the
hydroxyl as the TES (triethylsilyl) ether in preparation for
cyclization. Reaction of the (2S,4R)-chlorohydrin 9 under
standard conditions14 followed by deprotection of the hy-
droxyl led to the expected (2R,4S) isomer as the major
product with excellent diastereoselectivity17 (eq 3).
The smallest C4 group would of course be the unprotected
alcohol. However, reaction of this substrate could be
complicated by cyclization to the epoxide (Figure 2). The
On the other hand, we were surprised to discover in the
(S)-chlorohydrin series that cyclization gave a 1:1 mixture
Scheme 1. Proposed MOC approach to (2R) diastereomers of 4-hydroxy-R-methylproline
3030
Org. Lett., Vol. 9, No. 16, 2007

coordinating solvent (DMF) only the R-methylproline prod-
ucts were observed (Scheme 3). And while the TES protected
Scheme 3. Cyclization of Unprotected Hydroxy Substrates
with LiHMDS
Figure 2. Possible cyclizations of the unprotected chlorohydrin.
resulting epoxyenolate (X) could then undergo either exo
closure to the azetidine (path a) or endo to the pyrrolidine
(path b).18
When the two unprotected chlorohydrins were exposed
to excess KHMDS in toluene at -60 °C, the epoxides were
the only identifiable products (Scheme 2). The proton and
Scheme 2. KHMDS Treatment of Unprotected Hydroxy
Compounds
(2S,4S) diastereomer 12 gave a 1:1 mixture of products 16
and 17 (eq 4), the unprotected hydroxyl compound 11 gave
almost exclusively the expected (2R,4R) product 16. Surpris-
ingly, when the (2S,4R) isomer 8 was cyclized, the major
product was not the expected (2R,4S) isomer, but rather the
(2S,4S), where the carboxylate bearing stereocenter has been
inverted.
During the course of the reaction we observed (by HPLC)
what appeared to be the epoxide, but it is not clear if this is
a true intermediate, or an artifact of our sample preparation
and analysis. When authentic epoxide was treated with
LiHMDS in THF, even in the presence of excess LiCl, no
reaction was observed.
As suggested by Kawabata,14 we believe that where
retention of configuration is observed (Scheme 4, path A),
deprotonation and cyclization occur from the shown con-
carbon NMR spectra of compounds 19 and 20 are virtually
indistinguishable, suggesting that scrambling occurred at
C2.
Interestingly, when the base was switched to LiHMDS (1
M in THF) and the chlorohydrin substrate dissolved in a
(6) Nagumo, S.; Mizukami, M.; Akutsu, N.; Nishida, A.; Kawahara, N.
Tetrahedron Lett. 1999, 40, 3209-3212.
(7) Sato, T.; Kawasaki, S.; Oda, N.; Yagi, S.; Bialy, S.; Uenishi, J.;
Yamauchi, M.; Ikeda, M. J. Chem. Soc., Perkin Trans. 1 2001, 2623-
2631.
Scheme 4. Possible Mechanism for Stereoselective
Cyclizations
(8) Seebach, D.; Weber, T. HelV. Chim. Acta 1985, 68, 155-161.
(9) Remuzon, P.; Bouzard, D.; Guiol, C.; Jaquet, J.-P. J. Med. Chem.
1992, 35, 2898-2909.
(10) Remuzon, P.; Massoudi, M.; Bouzard, D.; Jaquet, J.-P. Heterocycles
1992, 34, 679-684.
(11) Fuji, K.; Kawabata, T. Chem. Eur. J. 1998, 4, 373-376.
(12) Kawabata, T.; Fuji, K. Top. Stereochem. 2003, 23, 175-205.
(13) Zhao, H.; Hsu, D. C.; Carlier, P. R. Synthesis 2005, 1-16.
(14) Kawabata, T.; Kawakami, S.; Majumdar, S. J. Am. Chem. Soc. 2003,
125, 13012-13013.
(15) (a) Bergeron, R.; Xia, M.; Phanstiel, O. J. Org. Chem. 1993, 58,
6804-6806. (b) Morie, T.; Kato, S.; Harada, H.; Matsumoto, J. Heterocycles
1994, 38, 1033-1040.
(16) For a discussion of the critical nature of the nitrogen protecting
group on preservation of chirality during enolate formation and cyclization
see: (a) Kawabata, T.; Wirth, T.; Yahiro, K.; Suzuki, H.; Fuji, K. J. Am.
Chem. Soc. 1994, 116, 10809-10810. (b) Kawabata, T.; Suzuki, H.; Nagae,
Y.; Fuji, K. Angew. Chem., Int. Ed. 2000, 39, 2155-2157. (c) Kawabata,
T.; Chen, J.; Suzuki, H.; Nagae, Y.; Kinoshita, T.; Chancharunee, S.; Fuji,
K. Org. Lett. 2000, 2, 3883-3885.
Org. Lett., Vol. 9, No. 16, 2007
3031

to the proton. Additional base then deprotonates to give the
enolate, which cyclizes to the cis product.19
Scheme 5. Summary of Cyclizations
In conclusion, we have demonstrated a straightforward
method for the synthesis of the four diastereomers of
4-hydroxy-R-methylproline. Using memory of chirality
methodology, cyclization of the protected (2S,4R) compound
9 gave the trans product 14 with excellent selectivity.
Omission of the hydroxy protecting group for the (2S,4S)
compound 11 led to the technically more challenging cis
diastereomer 16. Unexpectedly, cyclization of the unprotected
(2S,4R) isomer 8 gave predominately the cis diastereomer
15, via inversion of the carboxylate stereocenter. The (2S,4S)
and (2S,4R) 4-hydroxy-R-methylproline diastereomers are
accessible starting with (R)-alanine.
Acknowledgment. Many thanks to our colleagues in
Abbott Structural Chemistry for all their help: Rodger Henry
for the X-ray analysis on 21, Ian Marsden for help interpret-
ing the NMR data, and Paul West for HRMS analysis.
Supporting Information Available: Experimental pro-
cedures and analytical data for compounds 8, 9, 11, 12, 14,
15, and 18-21. This material is available free of charge via
the Internet at http://pubs.acs.org.
OL071023M
former A, with base approaching opposite to the sterically
more demanding Boc group. Cyclization is rapid relative to
racemization of the enolate. However, in the case of the
unprotected alcohol 8 we believe that the alkoxide forms a
complex (C) that locks the Boc group into a syn relationship
(18) Metal-halogen exchange induced cyclization of ω-iodopentyl
epoxides led predominately to the cyclobutylmethanol. Addition of certain
Lewis acids reverses selectivity to the cyclopentanol. See: Cooke, M. P.;
Houpis, I. N. Tetrahedron Lett. 1985, 26, 3643-3646. Reaction of
ω-bromopentyl epoxides with activated copper in the presence of trialkyl
or triaryl phosphines leads predominately to the cyclopentanols. See: Wu,
T.-C.; Rieke, R. D. Tetrahedron Lett. 1988, 29, 6753-6756. See also: Stork,
G.; Cohen, J. F. J. Am. Chem. Soc. 1974, 96, 5270-5272.
(17) The NMR of the major product compares favorably with that
previously reported.4 The trans stereochemistry of the major product (14)
was confirmed by removal of the benzyl ester and single-crystal X-ray
analysis of the resultant N-boc-4-hydroxy-R-methylproline (21).
(19) During the review process an alternate mechanism was proposed
in which the asymmetric alkylation results from kinetic resolution of racemic
enolates by the alkoxide derived from the chiral chlorohydrin. We thank
the referee for this suggestion.
3032
Org. Lett., Vol. 9, No. 16, 2007