Diastereoselective alkylation of β-amino esters: Structural and rate studies reveal alkylations of hexameric lithium enolates

Article abstract of DOI:10.1021/ja045144i

Alkylation of β-amino ester enolates proceeds with high diastereoselectivity. Single crystal, powder, and solution X-ray diffraction studies of the enolate show that the racemic enolate forms prismatic hexamers. ⁶Li NMR spectroscopic studies on partially racemic enolates reveal complex mixtures of homo- and heterochiral hexamers. An implicit fit of the aggregate populations to the Boltzmann distribution provides the free energy differences and equilibrium constants for the ensemble. Rate studies show that enolate alkylation occurs directly from the hexamer with participation by THF. A mechanism based on the alkylation of a ladder-like aggregate is proposed.

Full text of DOI:10.1021/ja045144i

Published on Web 11/23/2004
Diastereoselective Alkylation of â-Amino Esters: Structural
and Rate Studies Reveal Alkylations of Hexameric Lithium
Enolates
Anne J. McNeil,† Gilman E. S. Toombes,‡ Sol M. Gruner,‡ Emil Lobkovsky,†
David B. Collum,*,† Sithamalli V. Chandramouli,§ Benoit J. Vanasse,§ and
Timothy A. Ayers§
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853-1301, Physics Department, Clark Hall,
Cornell UniVersity, Ithaca, New York 14853-2501, and Process DeVelopment Chemistry,
AVentis, Bridgewater, New Jersey 08807
Received August 11, 2004; E-mail: dbc6@cornell.edu
Abstract: Alkylation of â-amino ester enolates proceeds with high diastereoselectivity. Single crystal, powder,
and solution X-ray diffraction studies of the enolate show that the racemic enolate forms prismatic hexamers.
6Li NMR spectroscopic studies on partially racemic enolates reveal complex mixtures of homo- and
heterochiral hexamers. An implicit fit of the aggregate populations to the Boltzmann distribution provides
the free energy differences and equilibrium constants for the ensemble. Rate studies show that enolate
alkylation occurs directly from the hexamer with participation by THF. A mechanism based on the alkylation
of a ladder-like aggregate is proposed.
Introduction
on lithium hexamethyldisilazide (LiHMDS)-mediated ketone
enolizations,6 lithium diisopropylamide (LDA)-mediated ester
Otamixaban (1) is a small molecule, direct factor Xa inhibitor
in phase IIb trials at Aventis for the management of acute
coronary syndromes.1 A key step in the preparative scale
synthesis involves the enolization/alkylation sequence in eq 1.2
The use of an unprotected amino group3 and the resulting high
diastereoselectivity (17:1) are notable. Given the prevalence of
â-amino esters as components of pharmaceutical agents4
(including â-lactam-based antibiotics),5 we were prompted to
investigate the mechanistic details of the alkylation step in the
Otamixaban synthesis.
enolizations,7 and lithium ion solvation by protic amines.8,9
A
detailed understanding of the structures and reactivities of
(1) Guertin, K. R.; Gardner, C. J.; Klein, S. I.; Zulli, A. L.; Czekaj, M.; Gong,
Y.; Spada, A. P.; Cheney, D. L.; Maignan, S.; Guilloteau, J.-P.; Brown, K.
D.; Colussi, D. J.; Chu, V.; Heran, C. L.; Morgan, S. R.; Bentley, R. G.;
Dunwiddie, C. T.; Leadley, R. J.; Pauls, H. W. Bioorg. Med. Chem. Lett.
2002, 12, 1671-1674. Czekaj, M.; Klein, S. I.; Guertin, K. R.; Gardner,
C. J.; Zulli, A. L.; Pauls, H. W.; Spada, A. P.; Cheney, D. L.; Brown, K.
D.; Colussi, D. J.; Chu, V.; Leadley, R. J.; Dunwiddie, C. T. Bioorg. Med.
Chem. Lett. 2002, 12, 1667-1670.
(2) Chandramouli, S. V.; O’Brien, M. K.; Powner, T. H. WO Patent 0040547,
2000. Nagula, G.; Huber, V. J.; Lum, C.; Goodman, B. A. Org. Lett. 2000,
2, 3527-3529.
(3) Enolates derived from unprotected R-amino amides have been alkylated:
Myers, A. G.; Schnider, P.; Kwon, S.; Kung, D. W. J. Org. Chem. 1999,
64, 3322-3327. Myers, A. G.; Gleason, J. L.; Yoon, T.; Kung, D. W. J.
Am. Chem. Soc. 1997, 119, 656-673. Myers, A. G.; Gleason, J. L.; Yoon,
T. J. Am. Chem. Soc. 1995, 117, 8488-8489.
(4) (a) EnantioselectiVe Synthesis of â-Amino Acids; Juaristi, E., Ed.; Wiley-
VCH: New York, 1997. (b) For reviews: Sewald, N. Angew. Chem. Int.
Ed. 2003, 42, 5794-5795. Ma, J.-A. Angew. Chem., Int. Ed. 2003, 42,
4290-4299. Liu, M.; Sibi, M. P. Tetrahedron 2002, 58, 7991-8035. Abele,
S.; Seebach, D. Eur. J. Org. Chem. 2000, 1-15. Cardillo, G.; Tomasini,
C. Chem. Soc. ReV. 1996, 117-128. Cole, D. C. Tetrahedron 1994, 32,
9517-9582. Juaristi, E.; Quintana, D.; Escalante, J. Aldrichimica Acta 1994,
27, 3-11.
(5) For leading references on the syntheses of chiral â-lactams, see: Singh,
G. S. Tetrahedron 2003, 59, 7631-7649. Magriotis, P. A. Angew. Chem.
Int. Ed. 2001, 40, 4377-4379. See also: EnantioselectiVe Synthesis of
â-Amino Acids; Juaristi, E., Ed.; Wiley-VCH: New York, 1997; pp 45-
66.
(6) (a) Zhao, P.; Condo, A.; Keresztes, I.; Collum, D. B. J. Am. Chem. Soc.
2004, 126, 3113-3118. (b) Zhao, P.; Lucht, B. L.; Kenkre, S. L.; Collum,
D. B. J. Org. Chem. 2004, 69, 242-249. (c) Zhao, P.; Collum, D. B. J.
Am. Chem. Soc. 2003, 125, 14411-14424. (d) Zhao, P.; Collum, D. B. J.
Am. Chem. Soc. 2003, 125, 4008-4009. (e) Godenschwager, P. F.; Collum,
D. B., unpublished.
Many of the underlying structural and mechanistic questions
dovetail nicely with investigations carried out at Cornell focusing
(7) (a) Sun, X.; Collum, D. B. J. Am. Chem. Soc. 2000, 122, 2452-2458. (b)
Sun, X.; Collum, D. B. J. Am. Chem. Soc. 2000, 122, 2459-2463. (c)
Sun, X.; Kenkre, S. L.; Remenar, J. F.; Gilchrist, J. H.; Collum, D. B. J.
Am. Chem. Soc. 1997, 119, 4765-4766.
† Department of Chemistry and Chemical Biology, Cornell University.
‡ Physics Department, Cornell University.
§ Aventis.
9
10.1021/ja045144i CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 16559-16568
16559

A R T I C L E S
McNeil et al.
lithium enolates, however, has been slow to develop. Although
numerous X-ray crystal structures of enolates show dimers,
tetramers, and hexamers, analogous structural assignments in
solution are both rare and somewhat tentative.10,11 Colligative
properties of enolate solutions shed light on the degree of
aggregation12 but can be technically challenging (especially
freezing-point measurements in THF) and afford results that
are quite sensitive to adventitious impurities.13,14 NMR spec-
troscopy has thus far afforded limited structural details of lithium
enolates. In contrast to N-lithiated and C-lithiated organo-
lithiums, wherein 6Li-15N and 6Li-13C coupling patterns
provide intimate structural details,15 the Li-O linkages of
enolates and related lithium alkoxides are spectroscopically
opaque. Jackman used a combination of 13C spin-lattice
relaxation times and 7Li quadrupole-splitting constants to show
that enolates form predominantly dimers and tetramers in THF
solution;16 the accuracy and generality of this method, however,
have been questioned.17 Streitwieser used singular value de-
composition of UV-vis spectra to detect monomeric, dimeric,
and tetrameric phenone-derived enolates in dilute solutions.18
in HMPA/THF solutions.17 Despite isolated successes, no
general method to assign structures of lithium enolates in
solution exists.
Mechanistic investigations of enolate alkylations have af-
forded an array of hypotheses as diverse as the enolate/solvent
combinations studied. Both Jackman and Seebach suggested that
enolate alkylations can proceed via tetramer-based transition
structures, although the evidence is circumstantial.19,20 Noyori
recently proffered evidence for a dimer-based mechanism for
the alkylation of cyclopentanone lithium enolate based on
detailed rate data.17 Streitwieser concluded that alkylations of
phenone-derived lithium enolates proceed primarily via mono-
mers.21 To complicate the picture further, Zook, Jackman, and
Streitwieser suggested that mixed aggregates derived from
lithium enolates and lithium halides can alkylate directly.22
We describe herein studies of the alkylation of â-amino ester
6 with benzyl bromide (BnBr) shown in eq 2. Structural studies
show that enolates (R)-9 and rac-9 exist as hexamers both in
the solid state and in THF solution.23 Rate studies reveal direct
alkylations of the observable hexamers with participation by
THF. Alkylations of putative hexameric ladders are discussed
in the context of semiempirical computations.
7
Noyori exploited a combination of Li and 31P NMR spec-
troscopies to assign lithium cyclopentanone enolate as a dimer
(8) (a) Aubrecht, K. B.; Lucht, B. L.; Collum, D. B. Organometallics 1999,
18, 2981-2987. (b) Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1996,
118, 2217-2225. (c) Lucht, B. L.; Collum, D. B. J. Am. Chem. Soc. 1996,
118, 3529-3530. (d) Also, see ref 6b.
(9) For leading references, see: Arvidsson, P. I.; Hilmersson, G.; Ahlberg, P.
J. Am. Chem. Soc. 1999, 121, 1883-1887. Henderson, K. W.; Williard, P.
G. Organometallics 1999, 18, 5620-5626. Vedejs, E.; Kruger, A. W.; Lee,
N.; Sakata, S. T.; Stec, M.; Suna, E. J. Am. Chem. Soc. 2000, 122, 4602-
4607. Laube, T.; Dunitz, J. D.; Seebach, D. HelV. Chim. Acta 1985, 68,
1373-1386.
(10) For examples of related enolate crystal structures, see: (a) Pauer, F.; Power,
P. P. In Lithium Chemistry: A Theoretical and Experimental OVerView;
Sapse, A.-M., Schleyer, P. v. R., Eds.; Wiley & Sons: New York, 1995;
pp 295-392. (b) Williard, P. G. ComprehensiVe Organic Synthesis;
Pergamon: New York, 1991; Vol. 1, pp 1-47. (c) Seebach, D. Angew.
Chem., Int. Ed. Engl. 1988, 27, 1624-1654. (d) Jastrzebski, J. T. B. H.;
van Koten, G.; van de Mieroop, W. F. Inorg. Chim. Acta 1988, 142, 169-
171. (e) Williard, P. G.; Tata, J. R.; Schlessinger, R. H.; Adams, A. D.;
Iwanowicz, E. J. J. Am. Chem. Soc. 1988, 110, 7901-7903. (f) Seebach,
D.; Amstutz, R.; Laube, T.; Schweizer, W. B.; Dunitz, J. D. J. Am. Chem.
Soc. 1985, 107, 5403-5409. (g) Jastrzebski, J. T. B. H.; van Koten, G.;
Christophersen, M. J. N.; Stam, C. H. J. Organomet. Chem. 1985, 292,
319-324. (h) Williard, P. G.; Carpenter, G. B. J. Am. Chem. Soc. 1985,
107, 3345-3346.
Results
Detailed spectroscopic and rate studies are described below.
For brevity, most of the graphic and spectral data have been
archived in the Supporting Information.
(18) Wang, D. Z.; Kim, Y.-J.; Streitwieser, A. J. Am. Chem. Soc. 2000, 122,
10754-10760. Wang, D. Z.; Streitwieser, A. J. Am. Chem. Soc. 1999, 121,
6213-6219. Leung, S. S.-W.; Streitwieser, A. J. Org. Chem. 1999, 64,
3390-3391. Facchetti, A.; Streitwieser, A. J. Org. Chem. 1999, 64, 2281-
2286. Gareyev, R.; Ciula, J. C.; Streitwieser, A. J. Org. Chem. 1996, 61,
4589-4593. Abbotto, A.; Streitwieser, A. J. Am. Chem. Soc. 1995, 117,
6358-6359.
(19) (a) Seebach, D.; Amstutz, R.; Dunitz, J. D. HelV. Chim. Acta 1981, 64,
2622-2626. (b) Jackman, L. M.; Lange, B. C. J. Am. Chem. Soc. 1981,
103, 4494-4499. Aggregate-based transition structures have been proposed
for reactions of related lithium phenolates: (c) Jackman, L. M.; Chen, X.
J. Am. Chem. Soc. 1997, 119, 8681-8684. (d) Jackman, L. M.; Petrei, M.
M.; Smith, B. D. J. Am. Chem. Soc. 1991, 113, 3451-3458.
(20) The concept of aggregate-based reactions of enolates has been endorsed
enthusiastically: (a) Streitwieser, A.; Leung, S. S.-W.; Kim, Y.-J. Org.
Lett. 1999, 1, 145-147. (b) Leung, S. S.-W.; Streitwieser, A. J. Am. Chem.
Soc. 1998, 120, 10557-10558. (c) Solladie´-Cavallo, A.; Csaky, A. G.;
Gantz, I.; Suffert, J. J. Org. Chem. 1994, 59, 5343-5346. (d) Wei, Y.;
Bakthavatchalam, R. Tetrahedron 1993, 49, 2373-2390. (e) Wei, Y.;
Bakthavatchalam, R.; Jin, X.-M.; Murphy, C. K.; Davis, F. A. Tetrahedron
Lett. 1993, 34, 3715-3718. (f) Solladie´-Cavallo, A.; Simon-Wermeister,
M.-C.; Schwarz, J. Organometallics 1993, 12, 3743-3747. (g) Pospisil,
P. J.; Wilson, S. R.; Jacobsen, E. N. J. Am. Chem. Soc. 1992, 114, 7585-
7587. (h) Williard, P. G.; Hintze, M. J. J. Am. Chem. Soc. 1987, 109, 5539-
5541. (i) Horner, J. H.; Vera, M.; Grutzner, J. B. J. Org. Chem. 1986, 51,
4212-4220. (j) Heathcock, C. H.; Lampe, J. J. Org. Chem. 1983, 48, 4330-
4337. (k) House, H. O.; Gall, M.; Olmstead, H. D. J. Org. Chem. 1971,
36, 2361-2371.
(21) Wang, D. Z.; Streitwieser, A. J. Org. Chem. 2003, 68, 8936-8942.
Streitwieser, A.; Kim, Y.-J.; Wang, D. Z. Org. Lett. 2001, 3, 2599-2601.
Streitwieser, A.; Juaristi, E.; Kim, Y.-J.; Pugh, J. K. Org. Lett. 2000, 2,
3739-3741. Wang, D. Z.; Streitwieser, A. Can. J. Chem. 1999, 77, 654-
658. Abbotto, A.; Leung, S. S.-W.; Streitwieser, A.; Kilway, K. V. J. Am.
Chem. Soc. 1998, 120, 10807-10813. Abu-Hasanayn, F.; Stratakis, M.;
Streitwieser, A. J. Org. Chem. 1995, 60, 4688-4689. See ref 18.
(11) For related hexameric lithium aminoalkoxide crystal structures, see:
Strohmann, C.; Seibel, T.; Schildbach, D. J. Am. Chem. Soc. 2004, 126,
9876-9877. Armstrong, D. R.; Davies, R. P.; Raithby, P. R.; Snaith, R.;
Wheatley, A. E. H. New J. Chem. 1999, 23, 499-507.
(12) For examples of colligative measurements on lithium enolates, see: (a)
Seebach, D.; Bauer, W. HelV. Chim. Acta 1984, 67, 1972-1988. (b) Arnett,
E. M.; Moe, K. D. J. Am. Chem. Soc. 1991, 113, 7288-7293. (c) Arnett,
E. M.; Fisher, F. J.; Nichols, M. A.; Ribeiro, A. A. J. Am. Chem. Soc.
1990, 112, 801-808.
(13) For examples of colligative measurements on other organolithiums, see:
(a) Davidson, M. G.; Snaith, R.; Stalke, D.; Wright, D. S. J. Org. Chem.
1993, 58, 2810-2816. (b) Kallman, N.; Collum, D. B. J. Am. Chem. Soc.
1987, 109, 7466-7472. (c) Wanat, R. A.; Collum, D. B.; Van Duyne, G.;
Clardy, J.; DePue, R. T. J. Am. Chem. Soc. 1986, 108, 3415-3422. (d)
West, P.; Waack, R. J. Am. Chem. Soc. 1967, 89, 4395-4399.
(14) Galiano-Roth, A. S.; Collum, D. B. J. Am. Chem. Soc. 1989, 111, 6772-
6778. Zook, H. D.; Gumby, W. L. J. Am. Chem. Soc. 1960, 82, 1386-
1389.
(15) Gu¨nther, H. J. Braz. Chem. 1999, 10, 241-262. Gu¨nther, H. In AdVanced
Applications of NMR to Organometallic Chemistry; Gielen, M., Willem,
R., Wrackmeyer, B., Eds.; Wiley & Sons: New York, 1996; pp 247-290.
Collum, D. B. Acc. Chem. Res. 1993, 26, 227-234.
(16) (a) Jackman, L. M.; Bortiatynski, J. AdV. Carbanion Chem. 1992, 1, 45-
87. (b) Jackman, L. M.; Scarmoutzos, L. M.; DeBrosse, C. W. J. Am. Chem.
Soc. 1987, 109, 5355-5361. (c) Jackman, L. M.; Szeverenyi, N. M. J.
Am. Chem. Soc. 1977, 99, 4954-4962. (d) Jackman, L. M.; Lange, B. C.
Tetrahedron 1977, 33, 2737-2769. (e) Jackman, L. M.; Haddon, R. C. J.
Am. Chem. Soc. 1973, 95, 3687-3692. (f) This spectroscopic technique
has been embraced by others, see: Wang, J. S.; Je´roˆme, R.; Warin, R.;
Teyessie´, P. Macromolecules 1993, 26, 1402-1406. Wen, J. Q.; Grutzner,
J. B. J. Org. Chem. 1986, 51, 4220-4224.
(17) Suzuki, M.; Koyama, H.; Noyori, R. Bull. Chem. Soc. Jpn. 2004, 77, 259-
268. Suzuki, M.; Koyama, H.; Noyori, R. Tetrahedron 2004, 60, 1571-
1579.
9
16560 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004

Diastereoselective Alkylation of
â-Amino Esters
A R T I C L E S
Alkylation. Sequential treatment of 2 with 2.0 equiv of
LiHMDS at -20 °C followed by 1.0 equiv of 3-(bromomethyl)-
benzonitrile (3) affords a 17:1 mixture of 4 and 5 along with
5-10% unalkylated ester. Observed byproducts include C,C-
and C,N-dialkylated amino esters 10 and 11, â-lactam 12, and
stilbene 13. Interestingly, the C,C-dialkylated product was shown
to derive primarily from the enolization and alkylation of the
minor syn isomer (5) due to a markedly higher propensity of 5
to enolize when compared with 4 (krel ) 7).24 The N-alkylated
product 11 derives primarily from the direct alkylation of 4 by
the highly reactive electrophile during workup. â-Lactam 12
is formed from the reaction of the C-alkylated products
with LiHMDS as discussed in more detail below. Stilbene
13 was shown to form by a facile reaction of 3 with Li-
HMDS.25
rates, allowing the reaction to be conveniently monitored at 0
°C, and afforded a somewhat different reaction profile. Control
experiments show that LiHMDS and BnBr do not form stilbene,
and BnBr is not measurably consumed by LiHMDS. However,
residual LiHMDS reacts instantaneously with 7 and 8 to afford
â-lactams 14 and 15 (eq 3). In fact, alkylations using excess
LiHMDS provide 14 in a preparatively useful 54% overall yield
from 6 (eq 4). Finally, the alkylation shows a conversion-
dependent deceleration that was traced, at least in part, to
autoinhibition by LiBr.
Enolate Structure: The General Problem. Determining the
aggregation states of enolates in solution is particularly chal-
lenging due to the absence of Li-O coupling and the high
symmetry of the most common structural forms (16-19, Chart
1). In the following sections, we present data from a number of
analytical methods showing that enolate 9 is hexameric both in
the solid state and in THF solution.
Alkylation of enolate 9 derived from either ammonium
tosylate 2 or free base 6 provides similar diastereoselectivities.
However, free base 6 offered considerable advantages for the
structural and mechanistic studies by affording solutions of
enolate 9 that were free of LiOTs.26-29 Treatment of 6 with 1.0
equiv of LiHMDS effected the quantitative conversion of 6
(1740 cm-1) to enolate 9 (1618 cm-1) followed by in situ IR
spectroscopy.30 Also, we used BnBr as the electrophile instead
of 3 to study the alkylation mechanism. BnBr offered attenuated
Chart 1
(22) (a) Miller, J. A.; Zook, H. D. J. Org. Chem. 1977, 42, 2629-2631. (b)
Jackman, L. M.; Dunne, T. S. J. Am. Chem. Soc. 1985, 107, 2805-2806.
(c) Kim, Y.-J.; Streitwieser, A. Org. Lett. 2002, 4, 573-575. (d) Abu-
Hasanayn, F.; Streitwieser, A. J. Am. Chem. Soc. 1996, 118, 8136-8137.
(e) See ref 19b.
(23) Some of the structural studies were recently communicated: McNeil, A.
J.; Toombes, G. E. S.; Chandramouli, S. V.; Vanasse, B. J.; Ayers, T. A.;
O’Brien, M. K.; Lobkovsky, E.; Gruner, S. M.; Marohn, J. A.; Collum, D.
B. J. Am. Chem. Soc. 2004, 126, 5938-5939.
(24) A similar selective enolization of the syn isomer was previously observed:
Seebach, D.; Sifferlen, T.; Bierbaum, D. J.; Rueping, M.; Jaun, B.;
Schweizer, B.; Schaefer, J.; Mehta, A. K.; O’Conner, R. D.; Meier, B. H.;
Ernst, M.; Gla¨ttli, A. HelV. Chim. Acta 2002, 85, 2877-2917.
Single-Crystal X-ray Diffraction. Although we do not rely
directly on X-ray crystal structures to assign aggregates in
solution, X-ray studies proved critical to the development of a
structural model for enolate 9. The single-crystal X-ray structure
of rac-9 revealed a hexamer (Figure 1) with a drumlike core
displaying S6 symmetry.31 The hexamer consists of three
enolates of S configuration encircling one hexagonal face and
three enolates of R configuration encircling the other (see 20).
The absence of coordinated THF attests to the stability of the
NH2-chelates, but we are less confident that the MeO-Li
(25) LDA-mediated stilbene formation with benzyl bromide has been ob-
served: Mali, R. S.; Jagtap, P. G. Synth. Commun. 1991, 21, 841-848.
(26) A putative LiOTs/9 mixed aggregate forms prior to alkylation. The enolate
generated from tosylate salt 2 is approximately 50% slower than the enolate
generated from free base 6 in the alkylation with BnBr.
(27) For a general discussion of salt effects, see: Tchoubar, B.; Loupy, A. Salt
Effects in Organic and Organometallic Chemistry; VCH: New York, 1992;
Chapters 4, 5, and 7 and references therein.
(29) The role of lithium halides in the anionic polymerization of acrylates has
been thoroughly examined: Dhara, M. G.; Baskaran, D.; Sivaram, S.
Macromol. Chem. Phys. 2003, 204, 1567-1575. Vle´ek, P.; Lochmann, L.
Prog. Polym. Sci. 1999, 24, 793-873. Zune, C.; Je´roˆme, R. Prog. Polym.
Sci. 1999, 24, 631-664. Zune, C.; Dubois, P.; Je´roˆme, R.; Kˇriz´, J.; Dybal,
J.; Lochmann, L.; Janata, M.; Vle´ek, P.; Werkhoven, T. M.; Lugtenburg,
J. Macromolecules 1998, 31, 2744-2755. Wang, J.-S.; Zhang, H.; Je´roˆme,
R.; Teyssie´, P. Macromolecules 1995, 28, 1758-1764. Varshney, S. K.;
Gao, Z.; Zhong, X. F.; Eisenberg, A. Macromolecules 1994, 27, 1076-
1082. Janata, M.; Lochmann, L.; Mu¨ller, A. H. E. Makromol. Chem. 1993,
194, 625-636. Varshney, S. K.; Hautekeer, J. P.; Fayt, R.; Je´roˆme, R.;
Teyssie´, P. Macromolecules 1990, 23, 2618-2622. Fayt, R.; Forte, R.;
Jacobs, C.; Je´roˆme, R.; Ouhadi, T.; Teyssie´, P.; Varshney, S. K.
Macromolecules 1987, 20, 1442-1444.
(30) Reported pKa values are highly solvent and substrate dependent: pKa HMDS
) 25.8 (THF): Fraser, R. R.; Mansour, T. S.; Savard, S. J. Org. Chem.
1985, 50, 3232-3234. pKa EtOAc ) 24.4 (DMSO), pKa NH3 ) 41
(DMSO): Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
(31) rac-9 (0.20 M) was crystallized from a 9.0 M THF/toluene solution held
at -20 °C over 24 h (CCDC 231371).
(28) Salt effects on enolate alkylations and aldol condensations have been
documented: Gutie´rrez-Garc´ıa, V. M.; Lo´pez-Ruiz, H.; Reyes-Rangel, G.;
Juaristi, E. Tetrahedron 2001, 57, 6487-6496. Aurell, M. J.; Gil, S.;
Mestres, R.; Parra, M.; Parra, L. Tetrahedron 1998, 54, 4357-4366. Ru¨ck,
K. Angew. Chem., Int. Ed. Engl. 1995, 34, 433-435. Imai, M.; Hagihara,
A.; Kawasaki, H.; Manabe, K.; Koga, K. J. Am. Chem. Soc. 1994, 116,
8829-8830. Seebach, D.; Beck, A. K.; Studer, A. In Modern Synthetic
Methods; Ernst, B., Leumann, C., Eds.; VCH: New York, 1995; pp 1-178.
Juaristi, E.; Beck, A. K.; Hansen, J.; Matt, T.; Mukhopadhyay, T.; Simson,
M.; Seebach, D. Synthesis 1993, 1271-1290. Bunn, B. J.; Simpkins, N. S.
J. Org. Chem. 1993, 58, 533-534. Hasegawa, Y.; Kawasaki, H.; Koga,
K. Tetrahedron Lett. 1993, 34, 1963-1966. Hatanaka, M.; Park, O.-S.;
Ueda, I. Tetrahedron Lett. 1990, 31, 7631-7632. Murakata, M.; Nakajima,
M.; Koga, K. Chem. Commun. 1990, 1657-1658. Estermann, H.; Seebach,
D. HelV. Chim. Acta 1988, 71, 1824-1839. Narasaka, K.; Ukaji, Y.;
Watanabe, K. Chem. Lett. 1986, 1755-1758. Liotta, C. L.; Caruso, T. C.
Tetrahedron Lett. 1985, 26, 1599-1602. Also, see refs 10c and 22b.
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16561

A R T I C L E S
McNeil et al.
At the lowest attainable NMR probe temperatures (-100 °C),
solvent exchanges,35 conformational equilibrations,36 and chelate
isomerizations37,38 (cf., 21-23) can become slow on NMR
spectroscopic time scales, often eliciting considerable spectral
complexity. The spectra typically simplify on warming above
-100 °C due to time averaging of these facile processes. Further
warming of the probe can lead to a particularly odd effect in
6
which intra-aggregate exchanges of Li nuclei become fast,
whereas inter-aggregate exchanges are still slow,39 causing
aggregates that differ by positional isomerism (cf., 24 and 25)
to appear as a single 6Li resonance. Only aggregates that differ
by subunit composition (cf., 4:2 versus 3:3 mixed hexamers 25
and 26) or by virtue of their aggregation numbers (cf., 16-19)
6
appear as separate species by Li NMR spectroscopy. This
combination of rapid intra-aggregate exchange in conjunction
with slow inter-aggregate exchange proves critical to the
structural assignments of 9.
Figure 1. ORTEP of rac-9 revealing a hexamer of S6 symmetry.
interactions (Li-O ) 2.6-2.8 Å) are stabilizing. Unfortunately,
the highly soluble (R)-9 could not be crystallized.
The 6Li NMR spectrum recorded on [6Li](R)-9 in 9.0 M THF/
toluene at -90 °C shows a single resonance (Figure 2A),
consistent with almost any aggregation state of high symmetry.
(The RnS6-n notations used in Figure 2 are described below.)
X-ray Powder Diffraction. We investigated whether the
structure derived from a single crystal accurately described the
dominant form in the bulk solid. Indeed, powder X-ray
diffraction on bulk rac-9 afforded a unit cell with dimensions
similar to those of the crystal structure, indicating that the
hexamer is the predominant solid-state aggregate of rac-9.
Small-Angle X-ray Scattering (SAXS). SAXS data on THF
solutions of (R)-9 or rac-9 (0.5-1.5 M) showed scattering
diameter and density correlations at a distance of 10 ( 4 Å.
This size compares favorably with the 13 Å unit cell of the
hexamer derived from rac-9. The SAXS data also confirmed
the homogeneity of enolate solutions at the molarities used for
the rate studies.32
Solution NMR Spectroscopy. We found that the inherent
symmetry of the enolate aggregates described in detail above
can be disrupted by using combinations of R and S antipodes
to generate an ensemble of mixed aggregates.23 What began as
a problem of inordinate spectral simplicity resulting from high
symmetry became a problem of extraordinary structural and
spectral complexity. To understand these studies, we must
briefly digress by describing dynamic phenomena that are
commonly observed for organolithium aggregates but may seem
surprising to the nonspecialist.
(33) Attenuated total reflectance (ATR) methods allow IR spectra of liquid phases
to be sampled in heterogeneous mixtures. For solubility studies using in
situ IR spectroscopy, see: Fujiwara, M.; Chow, P. S.; Ma, D. L.; Braatz,
R. D. Cryst. Growth Des. 2002, 2, 363-370. Togkalidou, T.; Tung, H.-
H.; Sun, Y.; Andrews, A.; Braatz, R. D. Org. Process Res. DeV. 2002, 6,
317-322. Fe´votte, G. Int. J. Pharm. 2002, 241, 263-278. Togkalidou, T.;
Fujiwara, M.; Patel, S.; Braatz, R. D. J. Cryst. Growth 2001, 231, 534-
543. Dunuwila, D. D.; Berglund, K. A. Org. Process Res. DeV. 1997, 1,
350-354. Dunuwila, D. D.; Berglund, K. A. J. Cryst. Growth 1997, 179,
185-193. Dunuwila, D. D.; Carroll, L. B., II; Berglund, K. A. J. Cryst.
Growth 1994, 137, 561-568.
6
(34) Enolates rac-9 and (R)-9 were shown by in situ IR spectroscopy, Li and
1H NMR spectroscopies, and GC analysis of the subsequent alkylation
products to be stable in THF solutions for >2 h at 0 °C.
(35) For leading references, see: Lucht, B. L.; Collum, D. B. Acc. Chem. Res.
1999, 32, 1035-1042.
(36) Clegg, W.; Liddle, S. T.; Snaith, R.; Wheatley, A. E. H. New J. Chem.
1998, 22, 1323-1326. Boche, G.; Fraenkel, G.; Cabral, J.; Harms, K.; van
Eikema Hommes, N. J. R.; Lohrenz, J.; Marsch, M.; Schleyer, P. v. R. J.
Am. Chem. Soc. 1992, 114, 1562-1565. Remenar, J. F.; Lucht, B. L.;
Kruglyak, D.; Romesberg, F. E.; Gilchrist, J. H.; Collum, D. B. J. Org.
Chem. 1997, 62, 5748-5754.
(37) Reich, H. J.; Goldenberg, W. S.; Sanders, A. W.; Jantzi, K. L.; Tzschucke,
C. C. J. Am. Chem. Soc. 2003, 125, 3509-3521 and references therein.
Fraenkel, G.; Winchester, W. R. J. Am. Chem. Soc. 1989, 111, 3794-
3797.
(38) For a particularly relevant description of chelate isomerization on a
hexameric aggregate, see ref 8a.
(39) Arvidsson, P. I.; Ahlberg, P.; Hilmersson, G. Chem.-Eur. J. 1999, 5, 1348-
1354. Bauer, W. J. Am. Chem. Soc. 1996, 118, 5450-5455. Bauer, W.;
Griesinger, C. J. Am. Chem. Soc. 1993, 115, 10871-10882. DeLong, G.
T.; Pannell, D. K.; Clarke, M. T.; Thomas, R. D. J. Am. Chem. Soc. 1993,
115, 7013-7014. Thomas, R. D.; Clarke, M. T.; Jensen, R. M.; Young, T.
C. Organometallics 1986, 5, 1851-1857. Bates, T. F.; Clarke, M. T.;
Thomas, R. D. J. Am. Chem. Soc. 1988, 110, 5109-5112. Fraenkel, G.;
Hsu, H.; Su, B. M. In Lithium: Current Applications in Science, Medicine,
and Technology; Bach, R. O., Ed.; Wiley: New York, 1985; pp 273-289.
Heinzer, J.; Oth, J. F. M.; Seebach, D. HelV. Chim. Acta 1985, 68, 1848-
1862. Fraenkel, G.; Henrichs, M.; Hewitt, J. M.; Su, B. M.; Geckle, M. J.
J. Am. Chem. Soc. 1980, 102, 3345-3350.
(32) Follow-up studies using in situ IR spectroscopy afforded the solubility of
rac-9.33 The absorbance of solubilized rac-9 during crystallization was
compared with stock solutions of known molarity, showing the solubility
limit to be 0.22 M in 9.0 M THF/toluene at 0 °C. In contrast, solutions of
(R)-9 were stable at >0.75 M in 9.0 M THF/toluene at 0 °C.34
9
16562 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004

Diastereoselective Alkylation of
â-Amino Esters
A R T I C L E S
Figure 2. 6Li NMR spectra in 9.0 M THF/toluene recorded on (A) [6Li](R)-9 (0.20 M), -90 °C; (B) [6Li]rac-9 (0.20 M), -90 °C; (C) [6Li](R)-9 (0.10 M)
and [6Li]rac-9 (0.10 M), -90 °C; (D) [6Li](R)-9 (0.10 M) and [6Li]rac-9 (0.10 M), -50 °C.
The 6Li NMR spectrum recorded on [6Li]rac-9 affords a single
resonance at a markedly different chemical shift than [6Li](R)-
9 (Figure 2B), suggesting the formation of a symmetric
heterochiral aggregate to the exclusion of the homochiral form.
Partially racemic mixtures using combinations of [6Li](R)-9 and
[6Li](S)-9 at -90 °C (Figure 2C) show both resonances along
isomerism within the R1S5/R5S1 and R2S4/R4S2 pairs in the slow
exchange limit that simplified by rapid intra-aggregate exchange
on warming.
A compelling picture emerges from a variant of a Job plot
(Figure 3) in which the intensities of the four resonances are
plotted as a function of the mole fraction of subunit (R)-9, XR.42
The positions of the maxima reflect the stoichiometries of the
aggregates, although the relationship is more complex than that
found in binary Job plots (vide infra). The concentration
dependencies were modeled as follows:43,44
6
with considerable “noise” in the baseline. Additionally, Li
6
spectra recorded on [6Li,15N](R)-9 and [6Li,15N]rac-9 show -
Li-15N coupling (d, JLi-N ) 3.4 and 3.6 Hz, respectively),
confirming chelation as drawn.40,41
Warming the probe temperature from -90 to -50 °C afforded
a single sharp resonance for [6Li](R)-9, which offered no
evidence of latent stereoisomerism, lower symmetry, or related
structural complexity. Conversely, warming samples containing
varying proportions of [6Li](R)-9 and [6Li](S)-9 revealed two
resonances in lieu of the baseline noise, four resonances in total
(Figure 2D). The data are consistent with deep-seated structural
complexities that simplify by rapid intra-aggregate exchange
at elevated temperatures. Furthermore, the relative intensities
are independent of the enolate (0.04-0.40 M) and the THF
(2.0-9.0 M) concentrations, indicating that the four species are
at the same aggregation and solvation state.
N
n × [R S
]
N-n
∑
n
n)0
XR )
N
N × [R S
]
∑
n
N-n
n)0
[RnSN-n
]
Xn )
N
[R S
]
∑
j
N-j
j)0
We considered models based on homochiral aggregates (RN
or SN) and heterochiral aggregates (RnSN-n). RnSN-n/RN-nSn
and RN/SN refer to pairs of spectroscopically indistinguishable
enantiomers. Dimers (R1S1 and R2/S2) and tetramers (R4/S4,
R1S3/R3S1, and R2S2) are expected to afford two and three 6Li
resonances, respectively. An ensemble of homo- and heterochiral
hexamers, R6/S6, R1S5/R5S1, R2S4/R4S2, and R3S3, should afford
four discrete resonances as observed. We inferred that the so-
called noise in the baseline was due to positional and stereo-
nµR + (N - n)µS
N!
[RnSN-n] ) C ×
× φ × exp
(
)
n
kT
n!(N - n)!
where
-gP
(
φn ) φN-n ) exp
)
kT
Xn, the mole fraction of the aggregate, is an implicit function
of XR and φn and may be solved by an iterative parametric
(40) [15N](S)-6 and [15N]rac-6 were synthesized from [15N]alanine via the
Arndt-Eistert homologation: Podlech, J.; Seebach, D. Liebigs Ann. 1995,
1217-1228.
(42) (a) Job, P. Ann. Chim. 1928, 9, 113-203. (b) Gil, V. M. S.; Oliveira, N.
C. J. Chem. Educ. 1990, 67, 473-478.
(41) For calculations of 6Li-15N coupling constants, see: Parisel, O.; Fressigne´,
C.; Maddaluno, J.; Giessner-Prettre, C. J. Org. Chem. 2003, 68, 1290-
1294. Koizumi, T.; Morihashi, K.; Kikuchi, O. Bull. Chem. Soc. Jpn. 1996,
69, 305-309.
(43) Widom, B. Statistical Mechanics: A Concise Introduction for Chemists;
Cambridge University Press: New York, 2002.
(44) Where µR and µS are the chemical potentials of R and S, gP corresponds to
the free energy of assembly for each permutation, and C is a constant.
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16563

A R T I C L E S
McNeil et al.
concentration with toluene (0.6, 3.2, 6.2, and 12.3 M) for
mixtures of [6Li,15N]LiHMDS and [6Li]rac-9 or [6Li](R)-9
revealed that mixed dimer 27 is favored at lower THF
concentrations.
Although excess LiHMDS influences the alkylation rates, it
is not due to the formation of mixed dimer 27 per se, but rather
due to the LiHMDS-mediated formation of the lithium imidates
28 and 29, which resulted from cyclization of 7 and 8.
Spectroscopic studies of enolate 9 and imidate 28 were
unsuccessful because of the marginal solubility of the lithiated
lactams. The rate studies of the alkylation of 9 by BnBr suggest
that LiBr generated during the course of the alkylation inhibits
the reaction. Putative mixed aggregates were evidenced by the
absence of an enolate homoaggregate in 6Li NMR spectra
recorded on [6Li]9/[6Li]LiBr mixtures, yet no detailed structural
insights were available because of peak broadening.47
Figure 3. Job plot of the mole fraction of RnS6-n/R6-nSn (Xn + X6-n
)
aggregates as a function of the mole fraction of the R enantiomer (XR) at
-50 °C. The best fit to the data is also shown: R3S3 (blue); R2S4/R4S2
(green); R1S5/R5S1 (black); R6/S6 (red).45
Table 1. Results of the Least-Squares Fits Illustrated in Figure 3
K
∆
G (kcal/mol)a
Rate Studies: General. The enolate alkylation was per-
formed under pseudo-first-order conditions ([9] g 10 × [BnBr]).
Rates were monitored by quenching aliquots with aqueous acid,
and subsequent gas chromatographic analysis measured the loss
of BnBr relative to decane as an internal standard. Nonlinear
least-squares fits48 afforded the pseudo-first-order rate constants
(kobsd). Although rate studies of organolithium reactions often
present technical hurdles that must be overcome before highly
reproducible and credible data can be collected, the alkylation
of 9 proved especially challenging to optimize. Some of the
necessary control experiments are described below.
K1 ) (1.0 ( 0.1) × 10-3
K2 ) (5.0 ( 0.3) × 10-3
K3 ) (115 ( 3) × 10-3
1.73 ( 0.04
1.82 ( 0.03
0.83 ( 0.01
a The deviations in ∆G beyond those expected for a statistical distribution
were obtained from the relationship ∆G ) -RT ln(K/Kstatistical).
method. It is instructive to present the results in terms of the
following equilibria:
K1
φ0
R3S3 y\
z 1/2R6 + 1/2S6
K1 )
20φ3
As noted previously, excess LiHMDS in the reaction mixture
converts the alkylated â-amino esters 7 and 8 into lithiated
imidates 28 and 29 with a concomitant 20% increase in the
alkylation rate. Alkylations spiked with 28 displayed analogous
accelerations. In addition, excess LiHMDS beyond that needed
to form the lithiated imidates also marginally accelerated the
enolization. Fortunately, these spurious effects could be elimi-
nated by using a small (typically 20%) excess of â-amino ester
6 relative to LiHMDS.
Autoinhibition attributed to LiBr-derived mixed aggregates
is evidenced by rate constants that are measurably (and
reproducibly) different at 5% and 10% conversion (Figure 5).
The effect is visible in both (R)-9 and rac-9 and is magnified
at low THF concentrations. Exogenous LiBr added at the outset
also produces the rate inhibition. This reaction may be acutely
sensitive to autoinhibition because a mere 5% conversion can,
at least in principle, afford enough LiBr to occlude up to 25%
additional enolate in a mixed hexamer. Regardless of the origins
of the conversion-dependent deceleration, the measured rate laws
were indistinguishable when the rates were monitored to 5%
or 10% conversion (vide infra).
K2
3φ1
R3S3 y\
z 1/2R5S1 + 1/2R1S5
K2 )
10φ3
K3
3φ2
4φ3
R3S3 y\
z 1/2R4S2 + 1/2R2S4
K3 )
If the aggregate distribution is purely statistical (φ0 ) φ1 )
... ) φ6), then K1 ) 0.05, K2 ) 0.30, and K3 ) 0.75. Least-
squares fits illustrated in Figure 3 yield substantially different
values (Table 1). The heterochiral R3S3 hexamer is markedly
more stable than the alternative homo- and heterochiral com-
binations.
Enolate-LiX Mixed Aggregates: Solution NMR Spec-
troscopy. Given the possibility of mixed-aggregate intervention
during the alkylation of 9, we examined a number of 9/LiX
mixtures using 6Li and 15N NMR spectroscopies. Enolates (R)-9
and rac-9 both form mixed dimers (27) with excess LiHMDS.46
The Li NMR spectrum of a 5:1 mixture of [6Li,15N]LiHMDS
and [6Li](R)-9 showed that both lithium resonances of 27 were
6
6
split into a doublet (Figure 4A). The Li NMR spectrum of a
5:1 mixture of [6Li]LiHMDS and [6Li,15N](S)-9 shows splitting
of only one lithium resonance (Figure 4B). Varying the THF
Rate Laws. Plots of kobsd versus [THF] for the alkylation of
(R)-9 and rac-9 reveal approximate first-order dependencies on
(45) For the case in which n ) 3, only X3 is plotted (rather than 2 × X3).
(46) For examples of lithium enolate-lithium amide mixed aggregates, see:
Williard, P. G.; Hintze, M. J. J. Am. Chem. Soc. 1990, 112, 8602-8604.
Romesberg, F. E.; Collum, D. B. J. Am. Chem. Soc. 1994, 116, 9198-
9202. Hall, P. L.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum,
D. B. J. Am. Chem. Soc. 1991, 113, 9575-9585. Also, see refs 7b, 20h,
and 22c.
(47) For examples of lithium enolate-lithium halide mixed aggregates, see: (a)
Henderson, K. W.; Dorigo, A. E.; Liu, Q.-Y.; Williard, P. G.; Schleyer, P.
v. R.; Bernstein, P. R. J. Am. Chem. Soc. 1996, 118, 1339-1347. (b) See
refs 16c, 16e, and 19b.
(48) A fit of the relative area ratios, areaBnBr/areadecane (A), versus time to [A] )
[A]0e(-kt) affords the pseudo-first-order rate constants (kobsd).
9
16564 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004

Diastereoselective Alkylation of
â-Amino Esters
A R T I C L E S
Figure 4. 6Li NMR spectra in 3.1 M THF/cyclopentane recorded at -90 °C on (A) [6Li,15N]LiHMDS (0.14 M) and [6Li](R)-9 (0.03 M), δ 0.00 (d, JLi-N
) 4.9 Hz), 0.62 (d, JLi-N ) 4.0 Hz), 0.71 (d, JLi-N ) 2.3 Hz), 1.00 (t, JLi-N ) 3.7 Hz); (B) [6Li]LiHMDS (0.14 M) and [6Li,15N](S)-9 (0.03 M), δ 0.00 (s),
0.63 (s), 0.71 (d, JLi-N ) 3.2 Hz), 1.00 (s).
Figure 5. Relative area ratios of BnBr (areaBnBr)/decane (areadecane) versus
time for the alkylation of (R)-9 (0.40 M) with BnBr (0.04 M) in 3.0 M
THF/toluene at 0 °C. The curves depict an unweighted least-squares fit to
[A] ) [A]oe(-kt), where kobsd ) (16.3 ( 0.3) × 10-4 for 0-5% conversion
(b), and kobsd ) (12.9 ( 0.1) × 10-4 for 5-10% conversion (O).
Figure 6. Plot of kobsd versus [THF] with toluene for the alkylation of
(R)-9 (O, 0.10 M) and rac-9 (b, 0.10 M), respectively, with BnBr (0.01
M) at 0 °C. The curves depict an unweighted least-squares fit to kobsd
)
a[THF]n, where a ) (1.1 ( 0.1) × 10-4, n ) 1.27 ( 0.04 for (R)-9, and
a ) (6 ( 1) × 10-5, n ) 1.3 ( 0.1 for rac-9.
THF concentration (n ) 1.27 ( 0.04, and n ) 1.3 ( 0.1,
respectively; Figure 6). Plots of kobsd versus [(R)-9] and [rac-9]
reveal approximate first-order dependencies on enolate concen-
tration (n ) 1.05 ( 0.04, and n ) 0.85 ( 0.06, respectively;
Figure 7). Despite deviations from 1.0, we are confident that
the enolate orders of 0.33 or 0.16 anticipated for dimer- and
monomer-based alkylations, respectively, can be excluded. The
idealized rate law (eq 5)49 implicates direct hexamer alkylation
without intervening deaggregation. The approximate first orders
in THF suggest the participation by one (possibly two) THF
ligands.50 The overall mechanism is depicted generically in eqs
6 and 7. Curiously, given the relative energy differences of the
various RnS6-n, the rates and selectivities observed for the
alkylation of 9 are surprisingly insensitive to optical purity
(Table 2).
Discussion
The alkylation of â-amino ester 2 used in the preparative-
scale synthesis of Otamixaban (1) is notable due to the efficiency
of using an unprotected amine. Both the importance and the
residual idiosyncrasies of the protocol prompted us to investigate
the underlying organolithium chemistry in greater detail. As is
often the case, the mechanistic study demanded several com-
promises. First, it is most convenient for the preparative-scale
synthesis to form the free base of ammonium tosylate 2 in situ
with an additional equivalent of LiHMDS. Markedly lower
alkylation rates, however, suggest that LiOTs forms mixed
aggregates with the enolate.26 Conversely, free base 6 is a
superior precursor for the structural and rate studies because it
affords salt-free solutions. Second, the alkylation of enolate 9
with 3 is too fast to monitor at temperatures affording
homogeneous solutions,32 so the less reactive BnBr was used
to achieve tractable rates at 0 °C (eq 2). The lower reactivity of
BnBr also precluded several minor side reactions, including
d[enolate]
dt
(enolate)6 + mTHF y
-
) k′[enolate]1[THF]m[BnBr]1
z (enolate)6(THF)m
(5)
(6)
(m)1-2)
(49) We define the idealized rate law as that obtained by rounding the observed
reaction orders to the nearest rational order.
(enolate)6(THF)m + BnBr f
[(enolate)6(THF)m(BnBr)]q f 7 + 8 + LiBr (7)
(50) An unweighted least-squares fit of the data in Figure 6 to kobsd ) a[THF]n
+ b containing an additional adjustable parameter (b) affords orders in
THF of 1.6 ( 0.1 and 2.2 ( 0.2 for (R)-9 and rac-9, respectively.
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16565

A R T I C L E S
McNeil et al.
tetramers, hexamers, as well as any other oligomers of high
symmetry (Chart 1).
A single-crystal X-ray structure and powder diffraction of
rac-9 revealed a hexameric aggregate containing equal propor-
tions of (R)-9 and (S)-9 antipodes (denoted as R3S3) to be the
dominant crystalline form (Figure 1). The hexamer structure
was unexpected. More to the point, crystal structures do not
offer direct insights into solution structures. For example, in a
recent study of lithium acetylide/lithium aminoalkoxide mixed
aggregates, three species were characterized in solution, two
species were characterized crystallographically, and only one
form was common to both phases.51 Curiously, the crystalline
form that was observed by X-ray, but not observed in solution,
was hexameric. Despite a lingering bias favoring dimers in
solution, the hexamer model for the solution structure of rac-9
was not easily dismissed because the structurally characterized
R3S3 aggregate offered a possible model to explain why rac-9
forms heterochiral aggregates to the exclusion of homochiral
aggregates in THF.
Figure 7. Plot of kobsd versus [9] for the alkylation of (R)-9 (O, 3.0 M
THF/toluene) and rac-9 (b, 8.0 M THF/toluene) with BnBr (0.01 M) at 0
°C. The curves depict an unweighted least-squares fit to kobsd ) a[9]n, where
a ) (4.3 ( 0.2) × 10-3, n ) 1.05 ( 0.04 for (R)-9, and a ) (5.7 ( 0.4)
× 10-3, n ) 0.85 ( 0.06 for rac-9.
A break came during 6Li NMR spectroscopic studies of
partially racemic solutions of 9. Along with single resonances
for both rac-9 and (R)-9, we observed “noise” in the baseline
at -90 °C that coalesced into two new resonances (four total)
on warming to -50 °C. We reasoned that the four resonances
derived from an ensemble of hexamers (RnS6-n). Each aggregate
could be (but is not necessarily) structurally complex due to
chelate stereo and positional isomerism. The key premise was
that warming the samples causes intra-aggregate exchanges,
including site-site exchanges of lithium ions,39 that would be
fast on NMR spectroscopic time scales under conditions in
which inter-aggregate subunit and lithium exchanges remain
slow. Using the simple dimers illustratively, we would expect
Table 2. Relative Alkylation Rates (krel) and Diastereoselectivities
(7:8) versus Optical Purity of Enolate 9a
% ee
krel
7:8
0
50
100
1
1.5
1.8
12:1
12:1
13:1
a Reaction conditions: 9.0 M THF/toluene, 0.10 M LiHMDS, 0.12 M
6, 0.01 M BnBr at 0 °C.
dialkylation and LiHMDS-mediated stilbene formation. By
precluding stilbene formation, the LiHMDS-mediated cycliza-
tion to form â-lactams became significant (eq 3). Although the
one-pot synthesis of â-lactams might be useful (eq 4), the
resulting imidates 28 and 29 elicited a minor autocatalysis that
proved disruptive during the rate studies. The problem was
avoided by using a slight deficiency of LiHMDS for the in situ
enolization.
6
two Li resonances corresponding to R2/S2 and R1S1. (R2 and
S2 would appear as a single resonance because they are
enantiomers.) In contrast, a tetramer would display three
resonances corresponding to R4/S4, R1S3/R3S1, and R2S2. Most
important, the hexamer is predicted to afford four resonances
corresponding to R6/S6, R1S5/R5S1, R2S4/R4S2, and R3S3.
6
The four Li resonances are consistent with an ensemble of
hexamers. Similar ensembles of hexameric sec-butyllithium were
noted by Fraenkel.52 Nonetheless, a more rigorous analysis was
warranted. Monitoring the relative integration of the four
resonances versus the mole fraction of the R enantiomer (optical
purity) provided the results shown in Figure 3. An implicit fit
to the populations using the Boltzmann distribution provides
compelling support for the hexamer assignments as well as the
relative energies of the aggregates. We discuss these energies
further in light of the computational studies and a mechanistic
hypothesis described below. It is instructive, however, to further
explore implications and applications of these unusual Job plots
as follows.
Enolate Structure. At the outset, we confronted one of the
most challenging problems in organolithium chemistry: deter-
mining the solution structures of lithium enolates. In the absence
of Li-O coupling, previous investigators had turned either to
specialized methods or to colligative measurements.12,16-18
Colligative measurements offer a potentially general solution
to the problem by providing solution molalities from which
aggregation numbers can be inferred.12,13 Although we had used
them previously,13b,c it is our opinion that the high sensitivity
to undetectable impurities casts significant doubts on the veracity
of the method. Consequently, we explored fundamentally
different strategies to study the aggregation of enolate 9.
(1) One of the most common applications of simple Job plots
in which a single mixed complex is formed (e.g., AB or AB2)
is that the position of the maximum along the x-axis (mole
fraction) provides the stoichiometry of the complex.42 Figure 8
6
Standard low-temperature spectroscopic studies using Li-
labeled 9 showed that optically pure and racemic enolates, (R)-9
and rac-9, each afforded spectroscopically distinct resonances
consistent with homo- and heterochiral aggregates. 15N-labeled
6
enolate showed Li-15N coupling consistent with the chelates
(51) (a) Thompson, A.; Corley, E. G.; Huntington, M. F.; Grabowski, E. J. J.;
Remenar, J. F.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 2028-2038.
(b) Xu, F.; Reamer, R. A.; Tillyer, R.; Cummins, J. M.; Grabowski, E. J.
J.; Reider, P. J.; Collum, D. B.; Huffman, J. C. J. Am. Chem. Soc. 2000,
122, 11212-11218.
(52) Fraenkel, G.; Henrichs, M.; Hewitt, M.; Su, B. M. J. Am. Chem. Soc. 1984,
106, 255-256.
as drawn. We must confess that we believed enolate 9 was
forming homo- and heterochiral dimers. However, such faith-
based structural assignments have obvious limitations. In fact,
the spectroscopic data were fully consistent with dimers,
9
16566 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004

Diastereoselective Alkylation of
â-Amino Esters
A R T I C L E S
Figure 9. Anti approach of BnBr to the calculated R3S3 hexamer.
results for a statistical distribution of the aggregates, A2, AB,
A1B3, A2B2, A3B1, and B4, and the equations for fitting
experimental data can be found in the Supporting Information.
Mechanism of Alkylation. Rate studies performed under
pseudo-first-order conditions revealed a first-order dependence
of the rate on the enolate concentration and an approximate first-
order dependence of the rate on THF concentration for both
(R)-9 and rac-9. In conjunction with the structural assignments,
the rate studies establish the stoichiometry of the transition
structures as [(enolate)6(THF)1-2(BnBr)]q. Although direct
hexamer alkylations contrast with previous studies implicating
monomer- and dimer-based mechanisms, the chelating amino
group distinguishes 9 from simpler enolates.17,21
We considered mechanisms involving alkylation of the intact
hexameric drum. Although we cannot exclude such models, they
pose several problems. For example, anti approach of the benzyl
bromide to the structurally characterized R3S3 aggregate would
be forced to occur from the congested perimeter (Figure 9). It
is also difficult to imagine what role the THF would play.
Consequently, we proffer an alternative mechanism based on
ladders shown in eq 8. By rearranging a single chelate with the
assistance of one or two THFs, the hexagonal drum can be
cleaved to afford a six-rung ladder (30). The spirocyclic bis-
chelated lithium is a known structural motif and is accessible
to alkylation.54 Although there is no direct evidence of ladder-
based reactions reported to date, organolithium ladders are well
documented in the structural literature largely due to the efforts
of a consortium of European chemists.55
Figure 8. Job plot of the mole fraction of RnS6-n/R6-nSn (Xn + X6-n
)
aggregates as a function of the mole fraction of the R enantiomer (XR) for
the statistical case.
illustrates a plot anticipated for a purely statistical distribution
of hexamers. The R3S3 aggregate exhibits the anticipated
maximum at XR ) 0.5. Although maxima would also appear
for R2S4 and R4S2 at XR ) 0.33 and 0.67, respectively (dotted
green curve), only one maximum is observed for the combined
R2S4/R4S2 curve (solid green curve). Furthermore, when the
maxima are evident in nonstatistical cases (such as in Figure
3), their position on the x-axis depends on the relative energies
within the ensemble. (The same holds true for the R1S5/R5S1
curve.)
(2) An insightful referee for our preliminary communication23
noted that an ensemble of heptamers, R7/S7, R1S6/R6S1, R2S5/
R5S2, and R3S4/R4S3, would also afford four 6Li resonances. It
was tempting to dismiss this proposal on the basis of both the
absence of precedence and even a structural model for such
heptamers. However, it poses the more fundamental question:
Can the fitting protocol rigorously distinguish two ensembles
that differ in aggregation number? The short answer is that the
fit of the experimental data to the heptamer model is as good
as the fit to the hexamer model. This issue may become more
germane in efforts to distinguish dimers (R2/S2 and R1S1) from
trimers (R3/S3 and R1S2/R2S1), both of which are plausible.
(3) The strategy for determining enolate structures by
monitoring an ensemble of mixed aggregates may prove
generally useful. This ensemble could be achieved with pairs
of antipodes via RnSN-n mixtures or via any combination of
two distinct lithium salts. For example, opposite antipodes of
two different enolates, R and S′, would afford the seven
hexamers of RnS′6-n, given that RnS′6-n and R6-nS′n would
no longer be spectroscopically indistinguishable. Preliminary
studies using combinations of different â-amino ester enolates
support this hypothesis.
(4) Simple enolates in coordinating solvents may form
tetramers rather than hexamers.10 Therefore, we explored
qualitatively what the curves would look like for an ensemble
of mixed tetramers, A4, A3B1, A2B2, A1B3, and B4. The requisite
equations and the plot depicting the statistical distribution of
tetramers are archived in the Supporting Information. A similar
distribution of tetramers has been described.53
The hypothesis in eq 8 is highly speculative. Is the ladder
model plausible? Does it explain the stereochemistry of alky-
lation? We turned to semiempirical (MNDO) computational
(53) Desjardins, S.; Flinois, K.; Oulyadi, H.; Davoust, D.; Giessner-Prettre, C.;
Parisel, O.; Maddaluno, J. Organometallics 2003, 22, 4090-4097. For an
application of a Job plot to identify an equilibrium mixture of titanium
alkoxy and amino acid complexes, see: Weingarten, H.; Van Wazer, J. R.
J. Am. Chem. Soc. 1965, 87, 724-730.
(5) We examined scenarios in which two fundamentally
different aggregation states, dimers and tetramers, coexist. The
9
J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004 16567

A R T I C L E S
McNeil et al.
unsolvated ladder was predicted to be thermoneutral. Probing
the more subtle issues pertaining to the stereochemistry of
alkylation,59 however, revealed daunting permutations arising
from conformational isomerism within the chelates, the syn
versus anti approach of the electrophile, the positional isomerism
of the various stoichiometries, and the multiple sites available
for alkylation. More to the point, the energetic differences for
some of these issues appeared well within the accuracy of the
computational methods.
Conclusion
Increasingly complex drug candidates demand state-of-the-
art synthetic methods that are often supported by little or no
mechanistic information. It is in this context that the process
chemistry at Aventis and the mechanistic organolithium chem-
istry at Cornell find considerable synergy. The alkylation of an
unprotected â-amino ester used in the preparative-scale synthesis
of Otamixaban presented challenging structural and mechanistic
problems. Enolates (R)-9 and rac-9 were characterized as an
ensemble of homo- and heterochiral hexamers using a strategy
based on Job plots. Although â-amino ester enolates were
studied for their pharmaceutical importance, this strategy may
solve the long-standing problem of characterizing enolates in
solution.
The formation of enolate hexamers (rather than dimers or
tetramers) was unexpected. Even more surprising, rate studies
revealed that the hexamers alkylate without intervening de-
aggregation, suggested to proceed via ladderlike intermediates.
Aggregation is often accompanied by mixed aggregation effects.
Indeed, enolate 9 appeared to form mixed aggregates with a
number of lithium salts causing minor autocatalysis and
autoinhibition. Nonetheless, such salt effects probably influence
the majority of organolithium reactions,26-29 and the scarcity
of mechanistic insights is conspicuous.60
Figure 10. Transition structure for anti alkylation of the R3S3 ladder with
CH3Br.
studies56 and obtained mixed results. Minimized structures for
the reactants afforded hexagonal drums exemplified by the R3S3
crystal form. The calculations supported experimental evidence
that the chelate orientations affording drums with S6 (R3S3) and
C3 (R6) symmetry are more stable than the C3 (R3S3) and D3
(R6) forms (respectively). Moreover, the experimentally ob-
served preference for the R3S3 over the R6/S6 form was
supported. It is notable, however, that dimers R2/S2 and R1S1
(unsolvated) are predicted to be considerably more stable than
hexamers.57
Preliminary investigations of the alkylation were also promis-
ing. For example, transition structures for the alkylations of the
hexamer drum could not be found, whereas transition structures
stemming from anti alkylation at either end of ladder 30 were
viable (Figure 10).58 Furthermore, opening the drum to an
Acknowledgment. We thank Prof. Cora Lind (University of
Toledo) for the powder X-ray diffraction. A.J.M. and D.B.C.
thank the National Institutes of Health for direct support of this
work. G.E.S.T. and S.M.G. thank the DOE for support (DE-
FG02-97ER62443).
(54) Hilmersson, G.; Arvidsson, P. I.; Davidsson, O.; Håkansson, M. J. Am.
Chem. Soc. 1998, 120, 8143-8149. Hilmersson, G.; Arvidsson, P. I.;
Davidsson, O.; Håkansson, M. Organometallics 1997, 16, 3352-3362.
Hilmersson, G.; Ahlberg, P.; Davidsson, O. J. Am. Chem. Soc. 1996, 118,
3539-3540. Veith, M. Angew. Chem., Int. Ed. Engl. 1987, 26, 1-14. See
also ref 13c.
(55) For leading references to organolithium ladders, see: Mulvey, R. E. Chem.
Soc. ReV. 1998, 27, 339-346. Beswick, M. A.; Wright, D. S. In
ComprehensiVe Organometallic Chemistry II; Abels, E. W., Stone, F. G.
A., Wilkinson, G., Eds.; Pergamon: New York, 1995; Vol. 1, Chapter 1.
Gregory, K.; Schleyer, P. v. R.; Snaith, R. AdV. Inorg. Chem. 1991, 37,
47-142. Mulvey, R. E. Chem. Soc. ReV. 1991, 20, 167-209.
(56) Lithium enolates and their mixed aggregates have been examined compu-
tationally: Pratt, L. M.; Streitwieser, A. J. Org. Chem. 2003, 68, 2830-
2838. Pratt, L. M.; Newman, A.; Cyr, J. S.; Johnson, H.; Miles, B.; Lattier,
A.; Austin, E.; Henderson, S.; Hershey, B.; Lin, M.; Balamraju, Y.;
Sammonds, L.; Cheramie, J.; Karnes, J.; Hymel, E.; Woodford, B.; Carter,
C. J. Org. Chem. 2003, 68, 6387-6391. Abbotto, A.; Streitwieser, A.;
Schleyer, P. v. R. J. Am. Chem. Soc. 1997, 119, 11255-11268. Weiss, H.;
Yakimansky, A. V.; Mu¨ller, A. H. E. J. Am. Chem. Soc. 1996, 118, 8897-
8903. Pratt, L. M.; Khan, I. M. J. Comput. Chem. 1995, 16, 1067-1080.
Dybal, J.; Krˇiz´, J. Collect. Czech. Chem. Commun. 1994, 59, 1699-1708.
Romesberg, F. E.; Collum, D. B. J. Am. Chem. Soc. 1994, 116, 9187-
9197. Rosi, M.; Sgamellotti, A.; Floriani, C. J. Mol. Struct. (THEOCHEM)
1998, 431, 33-46. Also, see ref 47a.
Supporting Information Available: Experimental prepara-
tions, control experiments, rate data, powder X-ray diffraction
data, small-angle X-ray scattering data, NMR spectroscopic data,
computations, mathematical derivations, and X-ray crystal-
lographic data (PDF, CIF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA045144I
(59) For discussions of the stereocontrol in the alkylation of lactam enolates,
see: Tomoda, S.; Ikuta, Y. Org. Lett. 2004, 6, 189-192. Ikuta, Y.; Tomoda,
S. Tetrahedron Lett. 2003, 44, 5931-5934. Groaning, M. D.; Meyers, A.
I. Tetrahedron 2000, 56, 9843-9873. Ando, K.; Green, N. S.; Li, Y.; Houk,
K. N. J. Am. Chem. Soc. 1999, 121, 5334-5335. Meyers, A. I.; Seefeld,
M. A.; Lefker, B. A.; Blake, J. F.; Williard, P. G. J. Am. Chem. Soc. 1998,
120, 7429-7438. Meyers, A. I.; Seefeld, M. A.; Lefker, B. A.; Blake, J.
F. J. Am. Chem. Soc. 1997, 119, 4565-4566. Seebach, D.; Maetzke, T.;
Petter, W.; Klo¨tzer, B.; Plattner, D. A. J. Am. Chem. Soc. 1991, 113, 1781-
1786.
(57) Dimers R2/S2 and R1S1 were predicted to be 30 and 25 kcal/mol more
stable than hexamers R6/S6 and R3S3, respectively.
(58) The absolute activation energies obtained for the ladder-based alkylation
with CH3Br were very large (∼40-43 kcal/mol). However, analogous
alkylations via the monomer and the doubly chelated dimer afforded similar
activation energies (∼44-51 kcal/mol).
(60) For a compelling investigation of autoinhibition in the 1,2-addition of lithium
acetylides to ketones, see ref 51a.
9
16568 J. AM. CHEM. SOC. VOL. 126, NO. 50, 2004