Synthesis of novel quaternary amino acids using molybdenum-catalyzed asymmetric allylic alkylation

Article abstract of DOI:10.1021/ja020290e

The Mo-catalyzed asymmetric allylic alkylation using azlactones provides extraordinary levels of selectivity. Thus, a wide range of cinnamyl-type substrates react with 2-methyl and 2-benzyl azlactones to give only the product resulting from attack at the more substituted carbon. Using other alkyl substituents such as 2-methylthioethyl, isobutyl, allyl, and isopropyl provides products that still retain excellent regioselectivity but small quantities of the linear product are also observed. In all cases, excellent diastereo- and enantioselectivity of the branched alkylated product are observed. This new asymmetric reaction provides ready access to unusual quarternary amino acids, important building blocks for biological applications. The reactions complements the Pd AAA wherein the cinnamyl substrate leads to only the product of attack at the primary terminus of the allyl moiety. Copyright

Full text of DOI:10.1021/ja020290e

Published on Web 05/29/2002
Synthesis of Novel Quaternary Amino Acids Using Molybdenum-Catalyzed
Asymmetric Allylic Alkylation
Barry M. Trost* and Kalindi Dogra
Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080
Received February 25, 2002
The growing importance of unusual amino acids in the design
of drugs stimulates the development of new synthetic methods.
Among the most challenging for asymmetric synthesis are quater-
nary amino acids since they cannot be made by asymmetric catalytic
hydrogenation. We previously disclosed the ability to synthesize
such systems using the palladium asymmetric allylic alkylation1
(AAA) reaction (eq 1, path A) with azlactones with good enantio-
selectivity did not bode well for an even more sterically demanding
nucleophile that would be required to make quaternary amino acids.
Thus, despite our misgivings regarding reactivity issues, we chose
to explore the use of azlactones as reported herein (eq 1, path b).
Initial studies examined the reaction shown in eq 3 with Ar )
selectivity. An even more challenging task is shown in eq 1, path
B. For such regioselectivity, Mo-2,3 and W-catalyzed2,4 allylic
alkylations are preferred. However, the Mo AAA reaction did not
seem promising since the use of more stabilized anions such as
those derived from 1,3-diketones or â-ketoesters reacted signifi-
cantly more poorly compared to malonate anion which did not bode
well for a rather highly stabilized anion like that derived from an
azlactone. As a result, we turned to the examination of a glycine
ester 1 as in eq 2. Indeed, the lithium enolate of 1 reacted rapidly
Ph and R ) CH3. Because the cleanliness of the generation of the
enolate of 7 with sodium hydride was poor, we examined the use
of the hexamethyldisilamide base. Whereas the use of sodium or
potassium hexamethyldisilamide gave some linear product as well
as the desired branched product 9 (Ar ) Ph, R ) CH3), employment
of the lithium base gave only the branched product in 83% yield
and excellent diastereoselectivity (dr ) 96:4) using the tert-
butylcarbonate leaving group. Switching to the diethyl phosphate
leaving group, the product was formed as a single diastereomer
(dr > 98:2) also in 80% yield. The azlactone was solvolyzed in
basic methanol to give the aminoester 10 (Ar ) Ph, R ) CH3)
quantitatively. Analysis by chiral HPLC showed the ee in both cases
to be 99%. An improved procedure avoids the isolation of the
alkylated azlactone. Directly adding basic methanol to the initial
reaction mixture using 8 (Ar ) Ph, X ) OCO2Boc) produced a
90% yield of 10 instead of 83% for the two step protocol. Using
the one pot protocol with 8 (Ar ) Ph, X ) OCO2CH3) gave 92%
yield of 10 having a 97:3 dr wherein the major diastereomer had a
99% ee.
Adopting the one-pot protocol as our standard, variation of the
Ar and R substituents, as in eq 3, was examined using the methyl
carbonate leaving group. Table 1 summarizes the results. In nearly
all cases, only a single diastereomer was detected (entries 3-12).
Using the methyl- and benzyl-substituted azlactones, only the
branched regioisomer was formed using allyl substrates such as 8
(entries 1-4, 6-8). Employing the chiral racemic substrate as the
allylating partner as in entry 5 did see some deterioration in the
regio- and enantioselectivity as we noted in our earlier work.2 In
this case, two diastereomeric complexes must form when the
reactions are run to 100% conversion. Apparently, their rate of
with cinnamyl phosphate 2 in the presence of the chiral Mo catalyst
derived from ligand 3 and precatalyst 4 to give a 2:1 ratio of the
branched (5) and linear (6) products.5 Gratifyingly, the dr (20:1)
and ee (98%) of 5 was excellent. However, the modest regio-
* To whom correspondence should be addressed. E-mail: bmtrost@stanford.edu.
9
7256
J. AM. CHEM. SOC. 2002, 124, 7256-7257
10.1021/ja020290e CCC: $22.00 © 2002 American Chemical Society

C O M M U N I C A T I O N S
Table 1. Catalyzed AAA to Quaternary Amino Acids
Figure 2.
that resulting from A. This correlation suggests that the product
derives from a late transition state which reflects the relative
thermodynamic stability of the two possible diastereomers.
The Mo-catalyzed AAA reaction leads to an unprecendented high
level of regio-, diastero-, and enantioselectivity in the reactions of
azlactones as nucleophiles. Thus, access to highly unusual quater-
nary amino acids result. The products may also serve as sources of
additional novel amino acids. For example, using ring-closing
metathesis as in eq 4 (product of entry 11), a cyclic amino ester of
a In all cases, X ) OCO2CH3. b N.D. ) not detected. c Substrate was
high diastereo- and enantioselectivity is generated. This represents
the first examples of control of stereochemistry at a nucleophile in
a Mo-catalyzed AAA.
The ability of these ligands to provide such exquisite control
even at 65 °C is noteworthy. Providing an understanding must await
more details of the structure of the catalytically active species.
Nevertheless, at present, they show increasing promise as an
important tool for asymmetric synthesis of important building
blocks. This process complements the Pd AAA which provides only
the linear product with cinnamyl substrates.
Figure 1.
equilibration, although reasonably fast, is still somewhat competitive
with the rate of nucleophilic attack. Curiously, using longer alkyl
substituents than methyl (entries 9 and 11) or branched alkyl groups
(entries 10 and 12) also led to some linear regioisomers. Neverthe-
less, in every case, the branched product was isolated in excellent
yields and stereoselectivities.
The impact of catalyst loading was examined in the reaction of
entry 6. Dropping the Mo to 5 mol % and 7.5 mol % ligand gave
quite comparable results after 7 hs86% isolated yield of product
as a single diastereomer of 95% eesalbeit with 6% of the linear
product observed. Further reduction to 2 mol % Mo and 3.5 mol
% ligand gave only a 52% isolated yield of a 4:1 branched:linear
product wherein the branched product still was only one dia-
stereocomer of 93% ee.
The relative and absolute stereochemistry was established by
X-ray crystallographic analysis of the product of entry 5 as shown
in Figure 1. The stereochemistry of all the remaining examples are
then assumed by analogy. The facial selectivity with respect to the
allyl is the same as with malonate.2 Figure 2 depicts the two
different diastereomeric transition states that minimize steric
hindrance. The crystal structure shows the approach in A to be
favored. Simple MM2 calculation on the products resulting from
A and B gave energy differences of 2-10 kcal/mol in favor of
Acknowledgment. We thank the National Science Foundation
and the National Institutes of Health for their generous support of
our programs. We thank Dr. Adam Cole for obtaining the X-ray
structure. Mass spectra were provided by the Mass Spectrometry
Regional Center of the University of California-San Francisco,
supported by the NIH Division of Research Resources.
Supporting Information Available: Experimental procedure and
characterization data for compound in Table 1 (PDF). X-ray crystal-
lographic file in CIF format. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Trost, B. M.; Ariza, X. Angew. Chem., Int. Ed. Engl. 1997, 36, 2635:
Trost, B. M.; Lee, C. B. J. Am. Chem. Soc. 1998, 120, 6818.
(2) Trost, B. M.; Hachiya, I. J. Am. Chem. Soc. 1998, 120, 1104; Trost, B.
M.; Hildbrand, S.; Dogra, K. J. Am. Chem. Soc. 1999, 121, 10416.
(3) Also see: Glorious, F.; Neuberger, M.; Pfaltz, A. HelV. Chem. Acta 2001,
84, 3178; Kaiser, N.-F.; Bremberg, U.; Larhed, M.; Moberg, C.; Hallberg,
A. Angew. Chem., Int. Ed. 2000, 39, 3596; Belda, O.; Kaiser, N.-F.;
Bremberg, U.; Larhed, M.; Hallberg, A.; Moberg, C. J. Org. Chem. 2000,
65, 5868; Janosen, U. P.; Helmchen, G. Tetrahedron Lett. 1997, 38, 8025.
(4) Lloyd-Jones, G. C.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1995, 34,
462.
(5) Trost, B. M.; Bouix-Peter, C. Unpublished work in these laboratories.
JA020290E
9
J. AM. CHEM. SOC. VOL. 124, NO. 25, 2002 7257