Reversal of Regioselection in the Sharpless Asymmetric Aminohydroxylation of Aryl Ester Substrates

Article abstract of DOI:10.1021/ol9903032

(Matrix Presented) The asymmetric synthesis of β-hydroxy-α-amino acids is reported which relies on the use of α,β-unsaturated aryl ester substrates and the dihydroquinyl alkaloid ligand system (DHQ)₂-AQN to control the regio- and enantioselectivity of the asymmetric aminohydroxylation (AA) process. α,β-Unsaturated ester substrates of type 1 have a significant effect on the substrate - ligand recognition event which results in a reversal of regioselectivity in the AA reaction.

Full text of DOI:10.1021/ol9903032

ORGANIC
LETTERS
1999
Vol. 1, No. 12
1949-1952
Reversal of Regioselection in the
Sharpless Asymmetric
Aminohydroxylation of Aryl Ester
Substrates
Adam J. Morgan,† Craig E. Masse,‡ and James S. Panek*
Department of Chemistry, Metcalf Center for Science and Engineering,
Boston UniVersity, Boston, Massachusetts 02215
panek@chem.bu.edu
Received September 28, 1999
ABSTRACT
The asymmetric synthesis of â-hydroxy-r-amino acids is reported which relies on the use of r,â-unsaturated aryl ester substrates and the
dihydroquinyl alkaloid ligand system (DHQ)2-AQN to control the regio- and enantioselectivity of the asymmetric aminohydroxylation (AA)
process. r,â-Unsaturated ester substrates of type 1 have a significant effect on the substrate−ligand recognition event which results in a
reversal of regioselectivity in the AA reaction.
â-Hydroxy-R-amino acid derivatives frequently occur as
constituents of biologically active peptides, precursors to
â-lactam antibiotics, and synthons for the preparation of
several neurologically active natural products. For instance,
3-hydroxyleucine and 3-hydroxylysine derivatives have at-
tracted considerable attention as unusual amino acid com-
ponents of numerous peptide antibiotics such as azinothricin,1
telomycin,2 lysobacin,3 and the protein kinase C inhibitor
(-)-balanol (Figure 1).4 More recently, 3-hydroxyleucine has
been sought after as a key synthon in the synthesis of (+)-
lactacystin and its analogues (Figure 1).5
In conjunction with our efforts toward the total synthesis
of (+)-lactacystin,5a we began to explore the asymmetric
aminohydroxylation (AA) reaction as an efficient method
to access the hydroxyleucine synthon. Numerous approaches
to the 3-hydroxyleucine have been reported. However, they
lack the flexibility to prepare stereoisomers, require the
preparation of a chiral catalyst system, or are prohibitively
lengthy and not practical for large scale preparations.6 In
our report on the synthesis of (+)-lactacystin, we discovered
that using the AA reaction on a p-bromophenyl ester olefinic
substrate provided the hydroxyleucine synthon with useful
† Recipient of a PREPARE fellowship sponsored by Pfizer Pharmaceu-
ticals.
(5) (a) Panek, J. S.; Masse, C. E. Angew. Chem. Int. Ed. 1999, 38, 1093-
1095. (b) Corey, E. J.; Li, W.; Reichard, G. A. J. Am. Chem Soc. 1998,
120, 2330-2336. (c) Corey, E. J.; Li, W.; Nagamitsu, T. Angew. Chem.
Int. Ed. 1998, 37, 1676-1679. (d) Nagamitsu, T.; Sunazuka, T.; Omura,
S.; Sprengler, P. A.; Smith, A. B. III. J. Am. Chem. Soc. 1996, 118, 3584-
3590.
(6) (a) Horikawa, H.; Petersen, J. B.; Corey, E. J. Tetrahedron Lett. 1999,
3843-3846. (b) Sunazuka, T.; Nagamitsu, T.; Tanaka, H.; Omura, S.;
Sprengler, P. A.; Smith, A. B. III. Tetrahedron Lett. 1993, 28, 4447-4448.
(c) Corey, E. J.; Lee, D.-H.; Choi, S. Tetrahedron Lett. 1992, 33, 6735-
6738. (d) Caldwell, C. G.; Bundy, S. S. Synthesis 1990, 34-36. (e) Jung,
M. E.; Jung, Y. H. Tetrahedron Lett. 1989, 30, 6636-6640. (f) Evans, D.
A.; Sjogren, E. B.; Weber, A. E.; Conn, R. E. Tetrahedron Lett. 1987, 28,
39-43.
‡ Recipient of a Graduate Fellowship from the Medicinal Chemistry
Division of the American Chemical Society 1998-1999, sponsored by
Hoechst Marion Roussel.
(1) Maehr, H.; Liu, C.-M.; Palleroni, N. J.; Smallheer, J.; Todan, L.;
Williams, T. H.; Blount, J. F. J. Antibiot. 1986, 39, 17-21.
(2) Sheehan, J. C.; Maeda, K.; Sen, A. K.; Stock, J. A. J. Am. Chem.
Soc. 1962, 84, 1303-1305.
(3) Tymiak, A. A.; McCormick, T. J.; Unger, S. E. J. Org. Chem. 1989,
54, 1149-1157.
(4) Kulanthaivel, P.; Hallock, Y. F.; Boros, C.; Hamilton, S. M.; Janzen,
W. P.; Ballas, L. M.; Loomis, C. R.; Jiang, J. B. J. Am. Chem. Soc. 1995,
115, 6452-6453.
10.1021/ol9903032 CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/10/1999

Our initial experiments with aliphatic substrates utilized
alkyl esters (1, Scheme 1) with the (DHQ)2-AQN ligand.
Scheme 1
However, this reaction afforded the undesired R-hydroxy-
â-amino ester as the sole regioisomeric product with high
levels of enantioselection (83% ee), consistent with literature
precedent for substrates of this type with the (DHQ)2-PHAL
ligand.7
With the presumption that the interactions of the aromatic
group of the cinnamates and the (DHQ)2-AQN alkaloid
ligand used in the Sharpless investigation contributed to the
regiochemical course of the reaction, a series of p-substituted
aryl esters of type 2 were screened in the AA using the
(DHQ)2-AQN ligand (Table 1).10 These substrates proved
to affect the regioselectivity pattern of the AA reaction to
provide the â-hydroxy-R-amino ester as the major regio-
Figure 1. Amino acid components of (+)-lactacystin and (-)-
balanol.
levels of regio- and enantioselection. This successful ap-
proach constituted the first example of a reversal in the
regioselection of the AA reaction with noncinnamate ester
substrates. Herein, we describe an efficient approach to the
asymmetric synthesis of certain â-hydroxy-R-amino acid
derivatives including a further investigation of the hydroxy-
leucine substrate. This process employs commercially avail-
able materials utilizing the Sharpless catalytic asymmetric
aminohydroxylation (AA) of aryl ester acrylate substrates.7
Sharpless and co-workers have previously shown that cin-
namates provided the desired â-hydroxy-R-amino ester when
the (DHQ)2-AQN ligand is utilized.8 At the time of our initial
investigation, the regiochemical outcome of the AA process
for aliphatic olefins of type 1 was unknown with the recently
available (DHQ)2-AQN ligand. The recent work of Janda
and co-workers9 with allylic alcohol substrates using the
(DHQD)2-PHAL ligand has shown that the steric bulk and
the electronic properties of the substituents on either side of
the olefinic substrate can be used to control the regiochem-
istry of the AA reaction. Despite the recent discovery of a
method for obtaining â-hydroxy-R-aminocinnamate esters,
the development of routes to enantiomerically pure â-alkyl-
substituted â-hydroxy-R-amino derivatives using the AA
process remains an active area of research. We decided to
investigate if this trend of the reversal of regiochemistry was
applicable to â-substituted aliphatic acrylate esters in an effort
to expand the scope to â-alkyl-substituted â-hydroxy-R-
amino derivatives and to further understand the nature of
these electronic effects.
Table 1. Reversal of Regioselection with Aryl Ester Substrates
(7) Li, G.; Chang, H.-T.; Sharpless, K. B. Angew. Chem., Int. Ed. Engl.
1996, 35, 451-454.
(8) Tao, B.; Schlingloff, G.; Sharpless, K. B. Tetrahedron Lett. 1998,
a Ratios of products were determined by 1H NMR (400 MHz) operating
at S/N of >200:1. b Determined by chiral HPLC analysis using a Chiracel
OD-H column. c After 2× recrystallization from EtOH/H2O (1:1). c Starting
olefin was recovered with varying amounts of saponification.
39, 2507-2510.
(9) Han, H.; Cho, C. W.; Janda, K. D. Chem. Eur. J. 1999, 5, 1565-
1569.
1950
Org. Lett., Vol. 1, No. 12, 1999

isomeric product (Table 1, entries 2-7).11 The turnover of
the regioselection for these aryl esters appears to rely on
subtle changes in the substrate-ligand recognition event. To
more fully investigate the electronic contributions of the aryl
esters on the aminohydroxylation reaction, a series of
p-substituted aryl esters were surveyed for use in a Hammett-
type analysis.12
preference for penetration of the olefinic substrate into the
active site of the catalyst.
The reversal of regioselection may arise from a confor-
mational change induced by the aryl ester functionality. To
determine the conformation of the aryl ester substrates in
solution, a series of NOE experiments were carried out with
substrates 2d and 1. These NOE experiments were performed
in d4-methanol to mimic the alcoholic solvent used in the
AA reaction. Substrate 2d exhibited a 6% NOE between the
â-olefinic proton and the ortho protons of the aryl ester (H4
and H1, respectively, see Figure 3). The observation of this
The ready accessibility of the p-substituted aryl ester
substrates allows for a survey of the electronic characteristics
of the AA using a Hammett-type analysis on the data
presented in Table 1 (Figure 2). These data clearly suggested
Figure 3. Solution conformation and 3-D minimized view of
substrate 2d.
NOE is indicative of a s-trans type conformer which
positions the ortho protons of the aryl ester (H1) in close
proximity to the â-olefinic proton (H4).13 Not surprisingly,
NOE experiments on the ethyl ester substrate (1) showed
no observable NOE’s upon irradiation of either the â-olefinic
proton or the methylene protons of the ethyl ester in accord
with the literature precedent of Wiberg and co-workers on
the conformation of such alkyl esters.14 Wiberg has shown
that the alkyl ester substrates prefer a s-cis conformer with
a cis orientation of the alkyl group relative to the carbonyl
moiety. This conformer minimizes dipole-dipole interac-
tions.14 The altered conformation of the aryl ester substrates
might allow for the assembly of a productive π-complex
between the osmium bound to the alkaloid ligand, the aryl
moiety of the ester, and the anthroquinone ring of the
Figure 2. Hammett-type analysis for the aminohydroxylation
process.
an empirical correlation between the electronic nature of the
aryl ester and the enantioselection of the aminohydroxylation
reaction. In general, electron-donating groups provided higher
levels of enantioselection than electron-withdrawing groups,
which if sufficiently electron withdrawing (e.g., p-NO2, entry
8) rendered the substrate unreactive. Interestingly, the p-halo-
substituted esters were found to have a unique effect on the
AA reaction, providing higher levels of regio- and enantio-
selection compared to those of the other substituents. Indeed,
the halogens defined a separate linear correlation distinct
from that of the other functional groups. This may indicate
a subtle change in the substrate-catalyst recognition event.
The negative slope associated with both the halo-substituted
esters and the other substituents surveyed indicated that
electron-donating groups stabilize the developing partial
positive charge on the olefinic carbons in the transition state,
thus accelerating the reaction rates. The type of halogen atom
also seemed to play a key role in the observed regioselection,
with the bromine-substituted ester (entry 4) being the optimal
substrate for the cases examined. The iodine-substituted ester
(entry 9) proved unreactive even after prolonged exposure
(48 h) to the reaction conditions. This may indicate a steric
(10) For the preparation of olefins 2a-i and 4a-d, see the Supporting
Information.
(11) To determine the effect of the (DHQ)2-AQN ligand on the
regiochemical outcome of the AA, we subjected substrate 2e to the AA
protocol using (DHQ)2-PHAL as the ligand. The AA reaction afforded
the R-hydroxy-â-amino ester as the major regiosiomeric product (ratio )
2.5:1).
(12) Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96-99. For a similar
type of analysis, see: Palucki, M.; Finney, N. S.; Pospisil, P. J.; Gu¨ler, M.
L.; Ishida, T.; Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120, 948-954.
(13) Molecular modeling studies with substrate 2d (Chem 3D Pro, MM2
minimization) indicate an internuclear distance between H1 and H4 to be
ca. 3.0 Å.
(14) Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1987, 109, 5935-
5943.
Org. Lett., Vol. 1, No. 12, 1999
1951

ligand.15 This orientation of the substrate-catalyst complex
may account for the turnover in regioselectivity of the AA
and the formation of the â-hydroxy-R-amino ester.
In an effort to expand the scope of the AA, additional
p-bromo-substituted aryl esters were prepared (Table 2).
the standard conditions of the AA reaction.7 The degree of
saponification can be attenuated by careful monitoring of
the reaction; however, yields could be significantly improved
by limiting the competing hydrolysis pathway. Attempts to
reduce saponification by using the analogous p-bromobenzyl
esters were less successful as these substrates proved less
reactive and afforded generally lower levels of regio- and
enantioselection. However, saponification of the p-bromo-
phenyl esters was greatly reduced by utilizing an aqueous
solution of OsO4 (4 wt % in H2O) in the reaction. This
allowed the use of a reduced amount of NaOH (1.5 equiv)
as it was no longer necessary to in situ generate the OsO4
from the potassium osmate dihydrate. This procedure was
performed using substrate 2d on a 40 mmol scale, providing
3d in 60% yield. The regioselectivity and enantioselectivity
of the product (3d) remained unchanged; hence this protocol
is readily scalable to multigram quantities using this proce-
dure.
Table 2. Aminohydroxylation of p-Bromophenyl Ester
Substrates
yield of A regioselection enantioselection
entry
R1
iPr, 2d
Me, 4a
H, 4b
Cl(CH2)3,4c
Cl(CH2)3, 4d
(%)a
(A:B)b
(% ee of A)c
1
2
3
4
60
40
52
53
50
7:1 (>20:1)d
1:1
4:1
87 (>99),d 3d
2, 5a
69, 5b
In summary, the Os(VIII)/(DHQ)2-AQN promoted AA of
certain R,â-unsaturated aryl esters provides a useful route
to â-alkyl-â-hydroxy-R-amino acid derivatives using a
convenient procedure and readily available catalyst system.
Presently, the AA process provides efficient access to the
important amino acids, 3-hydroxyleucine and 3-hydroxy-
lysine. Further, this study clearly indicates a trend for the
reversal of AA regioselection by simple variation of the ester
moiety which apparently alters the substrate-ligand recogni-
tion event. In that context, this study constitutes a useful
observation that subtle alterations of the substrate can be an
effective alternative to catalyst modification. It has also been
shown that the level of enantioselection is controlled, at least
in part, by the electronic properties of the aryl ester
functionality on the olefinic substrate. We are currently
undertaking further investigations in our laboratories to
enhance the scope of this process.
>20:1
4:1
82,e 5c
5f
90,e 5d
a Yields of isolated product after chromatography on silica gel. b Ratios
of products were determined by 1H NMR (400 MHz) operating at a S/N of
>200:1. c Determined by chiral HPLC analysis using a Chiracel OD-H
column. d After 2× recrystallization from EtOH/ H2O (1:1). e Optical purity
1
determined by H NMR of the mandelate ester. f Ar ) p-fluorophenyl.
These included crotonate and acrylate olefins (entries 2 and
3) which, in principle, provide access to the threonine and
serine synthons, respectively. The â-substituted acrylate
systems (entries 4 and 5) were also included to provide access
to unnatural amino acid derivatives as well as a hydroxy-
lysine derivative.
The acrylate ester (entry 3) reacted, as expected, to favor
the desired â-hydroxy-R-amino ester with moderate levels
of regio- and enantioselection. Unexplainably, the crotonate
derivative (entry 2) exhibited no regio- or enantioselection!
This surprising result indicates that even subtle variations
in the nature of the alkyl groups on the olefin can have a
dramatic effect on the selectivity of the AA reaction.
Interestingly, the p-fluorophenyl ester derivative (entry 5)
gave the highest ee (90%) of these substrates but showed a
slight decrease in regioselectivity.
Acknowledgment. This work has been financially sup-
ported by the National Institutes of Health (RO1GM55740).
The authors are grateful to Mr. Scott Schaus (Harvard
University), Professor John Porco, and Professor John K.
Snyder for helpful discussions.
Supporting Information Available: General experimen-
tal procedures and full characterization of compounds 1-5d.
This material is available free of charge via the Internet at
http://pubs.acs.org.
The high levels of enantioselection with these aryl ester
substrates offset their susceptibility to saponification under
(15) For precedent relative to the formation of such aryl-osmium
π-complexes, see: Wallis, J. M.; Kochi, J. K. J. Org. Chem. 1988, 53,
1679-1686.
OL9903032
1952
Org. Lett., Vol. 1, No. 12, 1999