Highly enantioselective 1,2-addition of lithium acetylide-ephedrate complexes: Spectroscopic evidence for reaction proceeding via a 2:2 tetramer, and X-ray characterization of related complexes

Article abstract of DOI:10.1021/ja0022728

The key step in the manufacturing process for the HIV reverse transcriptase inhibitor efavirenz (Sustiva) involves addition of the 2:2 tetrameric complex 6 [formed from lithium cyclopropylacetylide (5) and lithium (1R,2S)-N-pyrrolidinylnorephedrate (4)] to ketone 2, to give 3 in 95% yield and 98% enantioselectivity. Studies of acetylide-alkoxide complexes in solution by NMR spectroscopy and in the solid state by X-ray crystallography are described. Studies of the asymmetric addition reaction involving 2:2 tetramer 6 using low-temperature NMR spectroscopy provide conclusive evidence for formation of 2:1:1 tetramer 9 containing the product alkoxide 3. Observation of this reaction intermediate strongly supports the proposed reaction mechanism involving the tetramer 6 in the stereo-determining step.

Full text of DOI:10.1021/ja0022728

11212
J. Am. Chem. Soc. 2000, 122, 11212-11218
Highly Enantioselective 1,2-Addition of Lithium Acetylide-Ephedrate
Complexes: Spectroscopic Evidence for Reaction Proceeding via a
2:2 Tetramer, and X-ray Characterization of Related Complexes
Feng Xu,* Robert A. Reamer,* Richard Tillyer, Jordan M. Cummins,
Edward J. J. Grabowski, Paul J. Reider, David B. Collum,† and John C. Huffman‡
Contribution from the Department of Process Research, Merck Research Laboratories, P.O. Box 2000,
Rahway, New Jersey 07065, Department of Chemistry, Cornell UniVersity, Ithaca, New York 14853, and
Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405
ReceiVed June 23, 2000. ReVised Manuscript ReceiVed September 22, 2000
Abstract: The key step in the manufacturing process for the HIV reverse transcriptase inhibitor efavirenz
(Sustiva) involves addition of the 2:2 tetrameric complex 6 [formed from lithium cyclopropylacetylide (5) and
lithium (1R,2S)-N-pyrrolidinylnorephedrate (4)] to ketone 2, to give 3 in 95% yield and 98% enantioselectivity.
Studies of acetylide-alkoxide complexes in solution by NMR spectroscopy and in the solid state by X-ray
crystallography are described. Studies of the asymmetric addition reaction involving 2:2 tetramer 6 using low-
temperature NMR spectroscopy provide conclusive evidence for formation of 2:1:1 tetramer 9 containing the
product alkoxide 3. Observation of this reaction intermediate strongly supports the proposed reaction mechanism
involving the tetramer 6 in the stereo-determining step.
Introduction
Scheme 1. Highly Enantioselective 1,2-Addition Reaction
Recent efforts within the Merck Research Laboratories to
discover new compounds for the treatment of HIV infection
have resulted in indinavir (Crixivan), a protease inhibitor, as
well as efavirenz (1)1-3 (Sustiva),4 a nonnucleoside reverse
transcriptase inhibitor. Efavirenz has shown excellent potency
against a variety of HIV-1 mutants when used in combination
with Crixivan or with other reverse transcriptase inhibitors, and
was recently approved for use by the FDA. A practical
asymmetric synthesis5 of efavirenz has been implemented for
large scale manufacture.6 The key step in this process (Scheme
1) involves the 1,2-addition of lithium cyclopropylacetylide (5)
to trifluoromethyl ketoaniline 2 using stoichiometric amounts
of lithium (1R,2S)-N-pyrrolidinylnorephedrate (4) as chiral
mediator.7
† Cornell University.
‡ Indiana University.
(1) Young, S. D.; Britcher, S. F.; Tran, L. O.; Payne, L. S.; Lumma, W.
C.; Lyle, T. A.; Huff, J. R.; Anderson, P. S.; Olsen, D. B.; Carrol, S. S.;
Pettibone, D. J.; O’Brien, J. A.; Ball, R. G.; Balani, S. K.; Lin, J. H.; Chen,
I.-W.; Schleif, W. A.; Sardana, V. V.; Long, W. J.; Byrnes, V. W.; Emini,
E. A. Antimicrob. Agents Chemother. 1995, 39, 2602.
(2) Romero, D. L. Annu. Rep. Med. Chem. 1994, 29, 123. Tucker, T. J.;
Lyle, T. A.; Wiscount, C. M.; Britcher, S. F.; Young, S. D.; Sanders, W.
M.; Lumma, W. C.; Goldman, M. E.; O’Brien, J. A.; Ball, R. G.; Homnick,
C. F.; Schlief, W. A.; Emini, E. A.; Huff, J. R.; Anderson, P. S. J. Med.
Chem. 1994, 37, 2437. Quinn, T. C. Lancet 1996, 348, 99. De Clercq, E.
J. Med. Chem. 1995, 38, 2491.
Optimal enantioselectivity (98% ee) and full conversion (95%
isolated yield)6 are obtained using THF as solvent and require
(3) Mayers, D.; Riddler, S.; Bach, M.; Stein, D.; Havlir, M. D.; Kahn,
J.; Ruiz, N.; Labriola, D. F. Interscience Conference on Antimicrobial Agents
and Chemotherapy 37th Congress (ICAAC); Toronto, Ontario, Sept. 28-
Oct. 1, 1997. Mayers, D. Interscience Conference on Antimicrobial Agents
and Chemotherapy 36th Congress (ICAAC); New Orleans, LA, September
15-18, 1996. Ruiz, N.; Riddle, S.; Mayers, D.; Wagner, K.; Bach, M.;
Stein, D.; Kahn, J.; Labriola, D. RetroVirus and Opportunistic Infections
4th Annual Conference; Washington, D. C., January 22-26, 1997.
(4) Efavirenz is sold under the trademark STOCRIN in certain countries.
(5) Thompson, A. S.; Corley, E. G.; Huntington, M. F.; Grabowski, E.
J. J. Tetrahedron Lett. 1995, 36, 8937.
(6) Pierce, M. E.; Parsons, R. L., Jr.; Radesca, L. A.; Lo, Y. S.; Silverman,
S.; Moore, J. R.; Islam, Q.; Choudhury, A.; Fortunak, J. M. D.; Nguyen,
D.; Luo, C.; Morgan, S. J.; Davis, W. P.; Confalone P. N.; Chen, C.; Tillyer,
R. D.; Frey, L.; Tan, L.; Xu, F.; Zhao, D.; Thompson, A. S.; Corley, E. G.;
Grabowski, E. J. J.; Reamer, R.; Reider, P. J. J. Org. Chem. 1998, 63,
8536.
(7) For examples for asymmetric 1,2-addition of acetylide mediated by
chiral additives, see: Mukaiyama, T.; Suzuki, K.; Soai, K.; Sato, T. Chem.
Lett. 1979, 447. Mukaiyama, T.; Suzuki, K. Chem. Lett. 1980, 255. Niwa,
S.; Soai, K. J. Chem. Soc., Perkin Trans. 1 1990, 937. Ramos Tombo, G.
M.; Didier, E.; Loubinoux, B. Synlett 1990, 547. Corey, E. J.; Cimprich,
K. A. J. Am. Chem. Soc. 1994, 116, 3151. Ye, M.; Logaraj, S.; Jackman,
L. M.; Hillegass, K.; Hirsh, K.; Bollinger, A. M.; Grosz, A. L.; Mani, V.
Tetrahederon 1994, 50, 6109. Huffman, M. A.; Yasuda, N.; DeCamp, A.
E.; Grabowski, E. J. J. J. Org. Chem. 1995, 60, 1590. For 1,2-additions
mediated by chiral lithium alkoxides, see: Mukaiyama, T.; Soai, K.; Sato,
T.; Shimizu, H.; Suzuki, K. J. Am. Chem. Soc. 1979, 101, 1455. Alberts,
A. H.; Wynberg, H. J. Am. Chem. Soc. 1989, 111, 7265. Alberts, A. H.;
Wynberg, H. J. Chem. Soc., Chem. Commun. 1990, 453. Gallagher, D. J.;
Wu, S.; Nikolic, N. A.; Beak, P. J. Org. Chem. 1995, 60, 8148. Hashimoto,
K.; Kitaguchi, J.; Mizuno, Y.; Kobayashi, T.; Shirahama, H. Tetrahedron
Lett. 1996, 37, 2275.
10.1021/ja0022728 CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/28/2000

Lithium Acetylide-Ephedrate Complexes
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11213
Scheme 2. Proposed Reaction Pathways
Figure 1. Tetramer structures.
the use of 2.0 equiv of 5 and 2.0 equiv of alkoxide 4, which
must be allowed to equilibrate at temperatures > -40 °C prior
to reaction with keto aniline 2 at < -50 °C. Using 1.0 equiv
each of equilibrated 4 and 5 provided only 50% conversion (98%
ee) below -50 °C; subsequent warming of the reaction mixture
to 0 °C provided 90% conversion at the expense of selectivity
(82% ee). Generation of the alkoxide-acetylide mixture at low
temperature without aggregate equilibration and subsequent
reaction with keto-aniline 2 provided 3 in only 85% ee.8
The requirement for 2.0 equiv of 4 and 5 for full conversion
(or 1.0 equiv of tetramer) was explained in terms of the
formation of an unreactive 2:1:1 complex 9 containing product
alkoxide, acetylide 5, and 2 equiv of ephedrate 4. This
rationalization was consistent with the known low reactivity of
the closely related 3:1 alkoxide:acetylide complex (7 or 8).
Although these studies have enormously increased our
understanding of lithium acetylide-alkoxide complexes, some
aspects of this interesting puzzle warranted further investigation.
First, although there was ample evidence for generation of cubic
tetrameric alkoxide-acetylide complexes in THF, there was no
reported evidence for the formation of these species in other
solvents or in the solid state. Second, although equimolar
amounts of pyrrolidinylnorephedrine lithium alkoxide 4 and
acetylide 5 rapidly equilibrate to a single tetrameric complex 6
above -40 °C, this behavior had not been confirmed for other
N-substituted 1,2-aminoalkoxide-acetylide mixtures. These
points are particularly pertinent, given the observed solvent and
aminoalkoxide structural effects on the enantioselectivity of the
1,2-addition.5 Finally, there was no direct spectroscopic evidence
supporting the proposed tetramer-based mechanism for the 1,2-
addition reaction.
6
Initial mechanistic studies on this reaction employing Li,
13C, and 15N NMR spectroscopy and MNDO calculations have
shed much light on these experimental observations.8 6Li NMR
spectroscopic studies on THF solutions containing lithium
ephedrate 4 and lithium acetylide 5 in various proportions
confirmed a slow aggregate equilibration at low temperature,
which was rapid above -40 °C. It was determined that a 1:1
mixture of 4 and 5 equilibrates to a single C2-symmetrical 2:2
cubic tetramer 6 (Figure 1), which was fully characterized by
6Li NMR spectroscopy. It was also shown that a 3:1 mixture
of 4 and 5 equilibrates to a single cubic tetramer, assigned as 7
6
6
or 8 based on Li-13C and Li-15N couplings. The structure
of 7 was supported by MNDO model calculations (Figure 1).
It was concluded that reaction of the ketoaniline 2 with the
equilibrated 2:2 mixture of alkoxide and acetylide proceeds via
the 2:2 tetramer 6. A mechanism accounting for the observed
stereochemical outcome was proposed on the basis of MNDO
calculations (Scheme 2)8,9 and was supported by several key
observations: (1) the 2:2 tetramer 6 is the only detectable species
present after equilibration of the equimolar mixture of alkoxide
4 and acetylide 5 and prior to addition of ketone 2; (2) the 1,2-
addition is fast at low temperature, as compared to the slow
aggregate equilibration; (3) the reaction displays asymmetric
amplification [50% ee N-pyrollidinylnorephedrine (4a) provides
77%ee product].
The studies described herein were carried out to further
explore the structures of acetylide-alkoxide complexes, to
account for the experimental observations in the 1,2-addition,
and to further substantiate the reaction mechanism proposed in
the initial work.
Results and Discussions
Lithium Acetylide-Alkoxide Complexes: Solvent Effects.
In the initial work on the asymmetric addition reaction, THF
was found to give the best enantioselectivity (98%).5 Compa-
rable selectivity was observed in ethylene glycol dimethyl ether
(DME) (92% ee), but lower ee’s were obtained in diethyl ether
(68% ee) or tert-butyl methyl ether (MTBE) (26% ee), and close
to zero selectivity was observed in toluene. For reactions in
THF, the maximum ee was obtained using solutions that had
fully equilibrated to give a single 2:2 cubic tetramer 6 containing
(8) 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.
(9) Reaction of the ketoaniline 2 with the tetramer 6 possibly occurs via
formation of a precomplex (replacement of THF in 6 with the carbonyl of
2). The sense of facial selectivity in this reaction as predicted by MNDO
calculations is readily explained in terms of steric effects. In the preferred
transition state, the bulky aryl portion of 2 points away from the
norephedrine substituents, as indicated in Scheme 3.

11214 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Xu et al.
Table 1. Ring Size Effects on Reaction Enantioselectivity
a 6Li NMR was recorded in THF/hexane (1:1) at -60 °C. b 6Li NMR
was recorded in THF/hexane (1:1) at -40 °C.
and 11) were selected for this study to probe the effects of
N-cycloalkyl ring size.10 Thus, the complexes were prepared
using equimolar quantities of lithium acetylide 5 and norephe-
drine alkoxides 4, 10, and 11. In each case, the formation of
fully equilibrated tetrameric complexes was found to be
complete at -40 °C to give two-line spectra as expected (Table
1). The 1,2-addition reactions were then performed at -70 °C
to determine ee% (Table 1). These data clearly show that the
size of the N-cycloalkyl substituent is critical in this process
and confirms the initial observation that (1R,2S)-N-pyrrolidi-
nylnorephedrine (4a) is optimal.9
Lithium Acetylide-Alkoxide Complexes: X-ray Crystal-
lography. The tendency of lithium acetylides to form dimers,
tetramers, and higher oligomers in the solid state is well-known.
Lithium acetylides have been crystallized and characterized as
tetramers by X-ray analysis,11 but structural studies on mixed
lithium acetylide-lithium alkoxide complexes have not been
carried out.12 Accordingly, we were quite interested in preparing
X-ray quality crystals of these complexes to obtain structural
information in the solid state.
X-ray quality crystals were readily obtained from a 3:1
mixture of alkoxide 4 and acetylide 5. A solution of 3.0 equiv
(1R,2S)-N-pyrrolidinylnorephedrine (4a) and 1.0 equiv cyclo-
propylacetylene (5a) was treated with 4 equiv n-BuLi in THF/
Figure 2. 6Li NMR spectra of equilibrated samples containing 4 and
5 (1:1) (a) in THF/hexane/toluene (5:5:1) at -40 °C, (b) in DME/
hexane/toluene (5:5:1) at -60 °C, (c) in Et2O/hexane/toluene (5:5:1)
at -40 °C, and (d) in t-BuOMe/hexane/toluene (5:5:1) at -40 °C.
2 mol of THF. This was the first piece of (circumstantial)
evidence implicating the cubic tetramer itself in the addition
reaction. It was suspected that it might be possible to correlate
the observed solvent effects with the presence or absence of a
single tetrameric structure (related to 6), as determined from
6Li NMR spectra of the alkoxide-acetylide mixtures. Thus, the
6Li NMR spectra shown in Figure 2 were obtained for solutions
containing equimolar amounts of acetylide 5 and alkoxide 4,
which had been allowed to equilibrate at 22 °C. In DME, a
spectrum similar to that in THF was observed, which supports
formation of a tetrameric complex similar to 6 (possibly with
an η1-bound DME in place of the THF). In diethyl ether, the
6Li NMR spectrum showed two singlets, as in THF and DME,
but there were other species formed. In MTBE, a complex
mixture of species was formed. The high enantioselectivity is
observed only in solvents which allow for formation of a single
tetrameric complex at equilibrium prior to reaction with
ketoaniline 2.
(10) Zhao, D.; Xu, F.; Chen, C.; Tillyer, R.; Tan, L.; Pierce, M.; Moore,
J. Org. Synth. Vol. 77, 12.
(11) Lithium acetylide X-ray structures: (a) Goldfuss, B.; Schleyer, P.
v. R.; Hampel, F. J. Am. Chem. Soc. 1997, 119, 1072. (b) Geissler, M.;
Kopf, J.; Schubert, B.; Weiss, E.; Neugebauer, W.; Schleyer, P. v. R.;
Angew. Chem., Int. Ed. Engl. 1987, 26, 587. (c) Schubert, B.; Weiss, E.
Angew. Chem., Int. Ed. Engl. 1983, 22, 496. (d) Schubert, B.; Weiss, E.
Chem. Ber. 1983, 116, 3212.
(12) Theoretical and spectroscopic studies of lithium acetylides: Buhl,
M.; van Eikema Hommes, N. J. R.; Schleyer, P. v. R.; Fleischer, U.;
Kutzelinigg, W. J. Am. Chem. Soc. 1979, 101, 2848. Ritchie, J. P.
Tetrahedron Lett. 1982, 23, 4999. Fraenkel G.; Pramanik, P. J. Chem. Soc.,
Chem. Commun. 1983, 1527. Bachrach, S. M.; Streitwieser, A., Jr. J. Am.
Chem. Soc. 1984, 106, 2283. Buhl, M.; van Eikema Hommes, N. J. R.;
Schleyer, P. v. R.; Fleischer, U.; Kutzelnigg, W. J. Am. Chem. Soc. 1991,
113, 2459. Dorigo, A. E.; van Eikema Hommes, N. J. R.; Krogh-Jespersen,
K.; Schleyer, P. v. R. Angew. Chem., Int. Ed. Engl. 1992, 31, 1602. For
theoretical investigations of RLi/R′OLi mixed aggregates, see: Morey, J.;
Costa, A.; Deya, P. M.; Suner, G.; Saa, J. M. J. Org. Chem. 1990, 55,
3902. Sorger, K.; Schleyer, P. v. R.; Fleischer, R.; Stalke, D. J. Am. Chem.
Soc. 1996, 118, 6924. For investigations of RLi/R′OLi mixed aggregation,
see: McGarrity, J. F.; Ogle, C. A. J. Am. Chem. Soc. 1985, 107, 1805.
McGarrity, J. F.; Ogle, C. A.; Brich, Z.; Loosli, H. J. Am. Chem. Soc. 1985,
107, 1810. Marsch, M.; Harms, K.; Lochmann, L.; Boche, G. Angew. Chem.,
Int. Ed. Engl. 1990, 29, 308. Saa, J. M.; Martorelli, G.; Frontera, A. J.
Org. Chem. 1996, 61, 5194. Goldfuss, B.; Schleyer, P. v. R.; Hampel, F. J.
Am. Chem. Soc. 1996, 118, 12183. For recent investigations of RLi
aggregation, see: Hoffmann, D.; Collum, D. B. J. Am. Chem. Soc. 1998,
120, 5810. Reich, H. J.; Green, D. P.; Medina, M. A.; Goldenberg, W. S.;
Gudmundsson, B. O.; Dykstra, R. R.; Phillips, N. H. J. Am. Chem. Soc.
1998, 120, 7201. Hilmersson, G.; Arvidsson, P. I.; Davidsson, O.;
Hakansson, M. J. Am. Chem. Soc. 1998, 120, 8143. Sugasawa, K.; Shindo,
M.; Noguchi, H.; Koga, K. Tetrahedron Lett. 1996, 37, 7377. Hoppe, I.;
Marsch, M.; Harms, K.; Boche, G.; Hoppe, D. Angew. Chem., Int. Ed. Engl.
1995, 34, 2158.
Lithium Acetylide-Alkoxide Complexes: Ephedrine Sub-
stituents. In the initial studies, N-pyrrolidinylnorephedrine
lithium alkoxide 4 was identified as the optimal chiral mediator
on the basis of studying a variety of different chiral aminoalkox-
ides in the addition reaction.5 Much of these data were generated
before the realization of slow aggregate equilibration and under
conditions which did not guarantee complete equilibration of
complexes (-40 °C). Thus, a study of the aggregate equilibra-
6
tions of various acetylide 5-alkoxide mixtures using Li NMR
was carried out. Three closely related aminoalkoxides (4, 10,

Lithium Acetylide-Ephedrate Complexes
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11215
Figure 3. Perspective ORTEP drawings of 3:1 tetramer 8. Hydrogen
atoms have been omitted for clarity.
Table 2. Selected Bond Lengths (Å) and Angles (deg) for 3:1
Tetramer 8
Bond Distances
C(5)-Li(1)
C(5)-Li(3)
C(5)-Li(4)
O(17)-Li(2)
O(17)-Li(3)
2.13(2)
2.29(3)
2.27(3)
2.00(2)
1.97(2)
O(17)-Li(4)
O(55)-Li(2)
N(20)-Li(3)
N(35)-Li(4)
N(50)-Li(1)
1.86(2)
1.97(3)
2.04(2)
2.07(1)
2.07(3)
Bond Angles
95.5(12) C(16)-O(17)-Li(3) 97.0(7)
O(32)-Li(1)-O(47)
O(32)-Li(1)-C(5)
O(47)-Li(1)-N(50)
O(47)-Li(1)-C(5)
O(17)-Li(2)-O(32)
O(17)-Li(2)-O(47)
O(32)-Li(2)-O(47)
104.8(10) Li(2)-O(17)-Li(3)
78.9(8)
82.5(8)
87.1(9)
71.0(8)
70.3(9)
Figure 4. 4:2 hexamer 12 structure and its perspective ORTEP
drawings. Hydrogen atoms have been omitted for clarity.
90.2(6)
101.6(6)
94.2(8)
96.6(7)
95.8(12)
Li(2)-O(17)-Li(4)
Li(3)-O(17)-Li(4)
Li(1)-C(5)-Li(3)
Li(1)-C(5)-Li(4)
alternative stereoisomer 7 (Figure 1) was predicted to be more
stable on the basis of MNDO model calculations. If the solution
and solid-state structures of the 3:1 tetramer are indeed different,
the behavior of the 3:1 alkoxide:acetylide mixture, which was
prepared by dissolving the crystalline material in THF, would
require the facile conversion of 8 to 7 in solution.
hexane at -10 to 0 oC under nitrogen. The mixture was
concentrated in vacuo, and the solvent ratio was adjusted to
approximately 40% THF in hexane at a concentration of about
0.15 M (for the 3:1 tetramer). X-ray quality single crystals were
obtained from this mixture at room temperature under an
atmosphere of N2. The 1H, 6Li, and 13C NMR spectra measured
on a solution prepared by dissolving the crystals in THF-d8 at
room temperature were identical to the spectra previously
obtained from an equilibrated solution of the 3:1 aggregate.
Treatment of this solution with ketoaniline 2 at -40 °C provided
the addition product 3 in 99% ee and 60% yield, as expected
for a reaction involving the 3:1 tetramer.8
The structure of the crystalline material, as determined by
X-ray diffraction analysis at -165 °C, was shown to be the 3:1
tetramer 8 (Figure 1). The ORTEP drawing is displayed in
Figure 3, and the selected bond lengths and angles are
summarized in Table 2. It is interesting that the three Li-C
bonds in 3:1 tetramer 8 are slightly longer than the Li-O bonds,
which distorts the cubic structure somewhat. Surprisingly, the
THF O-Li bond [O(55)-Li(2)] is almost the same length as
other O-Li bonds and shorter than the N-Li bonds, which
indicates a strong binding of THF to lithium in the tetrameric
structure. Interestingly, each of the three internal O-Li-N
angles (inside the five membered ring) are right angles.
The structure of the 3:1 cubic tetramer 8 in the solid state is
identical to one of the two solution structures that are assigned
on the basis of the earlier NMR studies.8 However, the
Encouraged by our success in obtaining X-ray quality crystals
from the 3:1 complex, we attempted to prepare crystalline
material from a solution of the 2:2 alkoxide 4:acetylide 5
mixture. A crystalline solid was obtained from this mixture only
with difficulty, and this required removal of most of the THF.
Thus, a THF-hexane solution of a 1:1 mixture of cyclopropyl-
acetylene (5a) and (1R,2S)-N-pyrrolidinylnorephedrine (4a) was
treated with 2.0 equiv n-BuLi at -10 to 0 °C. The solution
was concentrated in vacuo and more hexane was added to adjust
the solvent ratio to 1-2% THF in hexane (concentration of 2:2
tetramer, approximately 0.2 M). X-ray quality single crystals
were obtained from this mixture at room temperature under a
nitrogen atmosphere. X-ray diffraction analysis carried out at
-174 °C under an atmosphere of N2 revealed hexamer 12
(Figure 4). The ORTEP drawing is displayed in Figure 4, and
selected bond lengths and angles are summarized in Table 3.
The preferential crystallization of the 4:2 hexamer 12 was
an interesting, yet unexpected, result, and all subsequent attempts
to obtain crystals of the 2:2 tetramer 6 have failed.13 Some
interesting features of this 4:2 complex were noted. The C2-
symmetrical hexamer 12 consists of two stacked chairlike six-
membered rings (displaced approximately 60° relative to each

11216 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Xu et al.
Table 3. Selected Bond Lengths (Å) and Angles (deg) for 4:2
Hexamer 12
Bond Distances
O(16)-Li(1)
O(16)-Li(2)
O(31)-Li(3)
O(16)-Li(3)′
1.906(5)
1.850(5)
2.005(5)
1.940(10)
N(19)-Li(1)
C(4)-Li(1)
C(4)-Li(1)′
C(4)-Li(3)
2.121(10)
2.215(11)
2.262(11)
2.121(5)
Bond Angles
Li(1)-O(16)-Li(2)
Li(1)-O(16)-Li(3)
Li(1)-C(4)-Li(1′)
Li(1)-C(4)-Li(3)
113.9(4)
84.1(3)
77.9(4)
103.4(3)
O(16)-Li(1)-N(19)
O(16)-Li(1)-C(4)′
O(16)-Li(1)-C(4)
O(16)-Li(2)-O(31)′
87.8(4)
125.6(3)
99.7(4)
99.7(4)
Figure 5. 6Li NMR spectrum at -100 °C after treatment of the 2:2
tetramer 6 with 0.5 equiv of ketone 2 in THF/pentane (1:1) at -100
°C.
other) that contain two alkoxides and one acetylide. Two lithium
atoms (those directly opposite to the acetylide carbon atoms)
are three-coordinate and are not bonded to THF. Additionally,
each of the four internal O-Li-N angles (within the five-
membered ring) are close to right angles, which is similar to
that observed in 3:1 tetramer 8.
Dissolving the crystalline 4:2 hexamer 12 in THF at room
temperature, followed by cooling to -70 °C and reaction with
ketone 2, provided the addition product 3 in 99% ee. Identical
results were obtained by using an equilibrated solution of 4 equiv
alkoxide 4 and 2 equiv acetylide 5, which suggests the formation
of solutions having similar composition in both cases. To better
6
understand these results, Li NMR spectra were recorded for
both solutions in THF-d8. It was immediately apparent that
dissolving the hexamer 12 in THF and a 4:2 mixture of alkoxide
4 and acetylide 5 generated in situ provide mixtures containing
predominantly the 2:2 tetramer 6 and 3:1 tetramer 7 or 8. This
is consistent with the highly enantioselective reactions that are
observed in both cases with ketone 2 at -70 °C.5,8 The
conclusion is that hexamer 12 is less stable than either tetrameric
structure in THF, that equilibration is fast at room temp, and
that 12 is not important in the addition reaction pathway.
Acetylide-Transfer Mechanism: 6Li NMR Studies. Ac-
cording to the proposed mechanism outlined in Scheme 2,
reaction of cubic tetramer 6 with the ketone 2 at low temperature
results in the rapid transfer of one acetylide to give a less-
reactive product complex 9 containing two ephedrates, one
product alkoxide, and one acetylide. Strong support for this
Figure 6. 6Li NMR spectra at -100 °C after treatment the 2:2 tetramer
6 with ketone 2 in THF/pentane (1:1) at -100 °C (a) using [1-13C]-
labeled acetylene 5a and (b) using [1-13C]-labeled acetylene 5a with
waltz-16 13C decoupling.
Table 4. NMR Spectroscopic Data for 2:1:1 Tetramer 9
6
mechanism was obtained by following the reaction using Li
NMR spectroscopy. A solution of 1.0 equiv mixed 2:2 tetramer
6 in THF and hexane, generated as usual, was cooled to -100
°C.14 To this solution was added, slowly and with good mixing,
a solution of approximately 0.5 equiv ketone 2 in THF. The
6Li NMR spectrum at -100 °C showed the appearance of four
new singlets, with equal intensities (1.13, 0.95, 0.77, and 0.60
ppm) (Figure 5). Further conversion of the 2:2 tetramer to this
four-line intermediate was achieved upon further addition of
keto aniline 2.15
The four-line spectrum is fully consistent with the formation
of a mixed cubic tetramer 9 containing the product alkoxide
and provides the first strong evidence in favor of the 1,2-addition
reaction proceeding via the 2:2 tetramer 6. Employing [1-13C]-
labeled acetylene 5a revealed that three of four 6Li resonances
were split into doublets due to spin-spin coupling to 13C. This
Li atom
LiA, LiB
ppm
13C-6Li coupling
15N-6Li coupling
1.13
0.95
0.77
0.60
d (J ) 6.8 Hz)
d (J ) 6.8 Hz)
d (J ) 4.0 Hz)
none
d (J ) 2.9 Hz)
d (J ) 2.9 Hz)
none
LiC
LiD
none
6
coupling was verified by a Li NMR spectrum (spectrum B,
Figure 6) with 13C decoupling. A broad signal at 116.2 ppm in
the 13C NMR spectrum at -100 °C is consistent with an alkyne
bound to several Li nuclei, and a signal of equal intensity at
6
(13) Preferential crystallization of the 4:2 hexamer from the 2:2 mixture
was found to occur reproducibly, regardless of solvent composition or
crystallization temperature.
(14) The best quality spectral data were obtained at -100 °C. Interest-
ingly, subsequent warming to -70 °C did not afford any new species, but
caused broadening of the signal assigned as LiC (Figure 5), which sharpened
upon re-cooling.
75.45 ppm is consistent with the alkyne carbon in the product
alkoxide itself. Using 15N-labeled ephedrine 4a,10,16 revealed
that the 6Li NMR spectrum showed splitting of signals at 1.13
and 0.95 ppm into doublets due to a coupling with 15N (Table
4). Using [15N]ketoaniline 2,6,17 the 6Li NMR spectrum obtained
at -100 °C showed no connection between the nitrogen atom
(15) Reaction of the sample in the NMR tube with ketoaniline 2 provided
the product 3 in 99% ee.

Lithium Acetylide-Ephedrate Complexes
J. Am. Chem. Soc., Vol. 122, No. 45, 2000 11217
Icon Services Inc. n-Bu6Li was prepared21 in pentane from commercial
6Li and n-butyl chloride and was filtered to remove solids before use.
15N-labeled (1R,2S)-N-pyrrolidineylnorephedrine 4a was prepared from
15N-labeled (1R,2S)-norephedrine, which was synthesized from pro-
piophenone via a literature procedure using 15NH2OH‚HCl.10,16 15N-
labeled ketoaniline 2 was prepared from [15N]-4-chloroaniline using
the existing procedures.6 [15N]-4-Chloroaniline was prepared from
4-chlorobenzoic acid via Hofmann rearrangement using 15NH3 as the
source of the label.17
in addition product 3 and any lithium atom in the product
alkoxide complex. This is surprising because the 3:1 tetramer
8 features bonding of the nitrogen atom in each ephedrate to a
lithium atom in the tetramer.18 On the basis of these data, the
product complex is assigned as the tetramer 9,19 the logical
product of asymmetric 1,2-addition of the tetramer 6 to the
ketoaniline 2.
Summary
Instrument Descriptions. NMR Spectroscopic Analyses: For 1H,
spectra were obtained on a Bruker AMX-400 operating at 399.87 MHz.
6Li and 13C spectra were recorded at 58.85 and 100.55 MHz,
Continued studies of lithium 1,2-aminoalkoxide-lithium
acetylide mixtures have substantially increased our understand-
ing of these complexes in solution and in the solid state. In
particular, it was confirmed that single tetrameric aggregates
are formed at equilibrium in THF and DME, while complex
mixtures are present in diethyl ether and MTBE. In addition, it
was shown that equimolar mixtures of lithium acetylide 5 and
N-cycloalkyl lithium aminoalkoxides equilibrate to single tet-
rameric species at approximately the same rate in THF and
confirmed that there is a substantial ring-size effect on the
enantioselectivity of the 1,2-addition reaction. X-ray crystal-
lographic studies revealed a symmetrical cubic tetrameric
structure 8 for the 3:1 aggregate which is in close agreement
with a structure proposed on the basis of solution NMR data.
A new hexameric complex 12 was prepared by crystallization
from a 2:2 mixture of alkoxide:acetylide and was characterized
by X-ray crystallography. This structure is not stable in the
solution and provides mixtures containing predominantly the
2:1 and 3:1 tetramer aggregates.
6
respectively. In the Li spectrum of tetramer 6 at -40 °C, the low-
6
field signal was referenced to δ ) 1.19 ppm.8 Subsequent Li spectra
were referenced to the 2H lock frequency. 13C spectra were referenced
to the high-field signal from THF-d8 (δ ) 25.4 ppm). Low-temperature
calibration was determined using 4% CH3OH in CD3OD and the Bruker
variable temperature calibration curve. 13C-decoupled 6Li spectra were
obtained using standard Waltz-16 decoupling on an “X” tuneable 13C/
1H triple-resonance probe.
X-ray Crystallography: All crystallographic studies were performed
on a Picker four-circle goniostat using a locally designed interface and
nitrogen gas flow cooling system of local design. In both studies,
crystals were affixed to the end of glass fibers using silicone grease
and transferred to the system where they were cooled for characteriza-
tion and data collection. Details of the diffractometer, low-temperature
facilities, and computational procedures employed by the Molecular
Structure Center are available elsewhere.22 The structures were solved
by direct methods (SHELXTL) and Fourier methods and refined on
F2 using full-matrix least-squares techniques.
[1-13C]Cyclopropylacetylene (5a).8,23 To a solution of PPh3 (15.8
g, 60.3 mmol) in 40 mL of dry methylene chloride at -40 °C was
added a solution of 13CBr4 (10.0 g, 30.1 mmol) in 40 mL of dry THF
at such a rate as to keep the temperature below -20 °C. After aging
15 min at -20 °C, the mixture was cooled to -70 °C, and Et3N (5.0
mL, 36.2 mmol) was added dropwise. A solution of cyclopropane
carboxaldehyde (4.23 g, 60.3 mmol) in 5 mL of dichloromethane was
added at such a rate as to keep the temperature below -55 °C. The
reaction mixture was aged at -40 °C for 1.5 h (reaction completion
determined by GC analysis) and was then quenched in 50 mL of water.
The organic layer was separated, and the aqueous layer was extracted
using 15 mL of dichloromethane. The combined organic solution was
dried over Na2SO4 and concentrated under reduced pressure to give a
white slurry. 100 mL of hexane was added, and the solid was removed
by filtration through a thin pad of Celite. Concentration of the filtrate
provided 5.5 g of [1-13C]-1,1-dibromocyclopropylethylene as a colorless
liquid (81% yield). The spectral data for this material exactly matched
those reported in the literature.23d
Further studies on the mechanism of the 1,2-addition reaction
have resulted in the spectroscopic characterization of the
tetrameric reaction intermediate 9, which leads us to conclude
that the reaction proceeds via the cubic tetramer 6 and that
pathways involving dimeric intermediates are not relevant in
this process.20
Experimental Section
General Considerations. Unless otherwise noted, all of the reactions
and manipulations were carried out using dry glassware under a nitrogen
atmosphere. All solvents were dried over 4-Å molecular sieves and
checked for water content by Karl-Fisher titration. 6Li (95%, enriched),
15NH3 (98%, enriched), and 15NH2OH‚HCl (99%, enriched) were
purchased from Aldrich. 13CBr4 (99%, enriched) was purchased from
(16) Oppolzer, W.; Tamura, O.; Sundarababu, G.; Signer, M. J. Am.
Chem. Soc. 1992, 114, 5900. Soai, K.; Yokoyama, S.; Hayasaka, T. J. Org.
Chem. 1991, 56, 4264.
(17) Huang, X.; Keillor, J. W. Tetrahedron Lett. 1997, 38, 313.
(18) It was pointed out correctly by a referee that the 3:1 tetramer 8
may not predict reactivity of the intermediate 9 due to differences in Li-N
bonding in these structures.
(19) The NMR data are also consistent with the diastereoisomers 13 and
14. The structure 9 is proposed because it derived immediately from the
1,2-addition reaction, and it is plausibe that internal reorganization is slow
at low temperature.
To a solution of 5.1 g (22.6 mol) of [1-13C]-1,1-dibromocyclopro-
pylethylene in 8 mL of dry toluene and 11 mL of dry THF was slowly
added 5.0 mL of n-BuLi (50 mmol, 10 N in hexane) while keeping the
temperature below -10 °C. The reaction mixture was aged between
-10 and 25 °C for 1 h (reaction completion determined GC analysis),
and then 10 mL of 10% NH4Cl was added while keeping the
temperature below -15 °C. The organic phase was separated at 0 to
-5 °C and was dried over Na2SO4 at 0 °C. Short-path distillation of
the resulting solution under an atmosphere of nitrogen gave 14.0 g of
a solution of 13C-labeled cyclopropylacetylene solution in THF-hexane
(7.2 wt % solution, 68% yield) and 55% overall yield.
General Procedure for the Preparation of 2:2 Tetramer 6. To a
solution of (1R,2S)-N-pyrrolidinylnorephedrine (4a) (10.0 g, 45.79
mmol) and Ph3CH (10 mg) in 80 mL of dry THF was charged n-BuLi
(1.6 M in hexane) while maintaining the temperature below -10 °C.
(21) Amonooneiger, E. H.; Shaw, R. A.; Skovlin, D. O.; Smith, B. C.
Inorg. Synth. 1966, 8, 19.
(22) Chisholm, M. H.; Folting, K.; Huffman, J. C.; Kirkpatrick, C. C.
Inorg. Chem. 1984, 23, 1021.
(23) (a) Ramirez, F.; Desai, N. B.; McKelvie, N. J. Am. Chem Soc. 1962,
84, 1745. (b) Corey, E. J.; Fuchs, P. L. Tetrahedron Lett. 1972, 13, 3769.
(c) Grandjean, D.; Pale, P.; Chuche, J. Tetrahedron Lett. 1994, 35, 3529.
(d) Baldwin, J. E.; Villarica, K. A. J. Org. Chem. 1995, 60, 186.
(20) Dissociation of the tetramer 6 to its corresponding dimer followed
by reaction with ketone 2 and recombination to a single product 9 is
considered unlikely.

11218 J. Am. Chem. Soc., Vol. 122, No. 45, 2000
Xu et al.
The amount of n-BuLi added at the red end point was noted. The same
amount of n-BuLi was then added (total amount of n-BuLi, 91.58
mmol), and then a solution of cyclopropylacetylene (5a) in THF-
heptane (45.8 mmol) was added below 0 °C until the red color
disappeared. The mixture was aged at 0-5 °C for 30 min to complete
the formation of tetramer 6.
addition of three-fourths of the theoretical amount of n-BuLi. Cyclo-
propylacetylene (a solution either in THF-heptane or neat) was added
at -10 to 0 °C until the reaction solution became colorless. The solution
was aged at room temperature for 30 min and then concentrated in
vacuo to a volume of approximately 10 mL. A portion (15mL) of THF
and 70 mL of pentane were added. The concentration of 3:1 tetramer
was approximately 0.15 M in approximately 30% THF:70% hydro-
carbon. The solution was allowed to stand at room temperature under
a nitrogen atmosphere for 6 h, which resulted in the slow growth of
the 3:1 tetramer crystals 8.
6
Generation of Li-Labeled Tetramer 6 and Observation of the
6
Addition Reaction by Li NMR Spectroscopy. The 2:2 tetramer 6
was prepared as described in the general procedure, except that the
deprotonation of both ephedrine 4a and acetylene 5a was carried out
using n-Bu6Li (1.4 M in pentane) below -70 °C. Immediately after
the addition of acetylene 5a, the solution was transferred into a 5-mm
NMR tube while taking care to maintain low temperature. Low-
temperature 6Li NMR spectra were recorded on this solution as
described in the general considerations section. The 6Li NMR spectrum
at -70 °C was quite complicated (see spectrum in Supporting
Information). Upon gradual warming of the solution, growth of 2:2
Crystallization of 4:2 Hexamer. A solution of the 2:2 tetramer 6
in THF-hexane was generated as in the general procedure starting
from 5.0 g (27.47 mmol) norephedrine 4a. The solution was concen-
trated in vacuo to a volume of approximately 7 mL, and 60 mL of
pentane was added. The concentration of 2:2 tetramer 6 was ap-
proximately 0.2 M in approximately 2% THF-hydrocarbon. The
solution was allowed to stand at room temperature under a nitrogen
atmosphere for several days, which resulted in the slow growth of the
4:2 hexamer crystals 12.
6
tetramer peaks at 1.19 and 0.47 ppm was observed in the Li NMR
spectrum. As soon as the temperature reached to -40 °C, the formation
of 2:2 tetramer 6 was complete (concentration, approximately 0.13 M).
Solutions of the 2:2 tetramer 6 solution in THF-hexane were
generated according to the general procedure using n-Bu6Li in pentane
and labeled or nonlabeled norephedrine 4a and acetylene 5a. A portion
(0.4 mL) of the 2:2 tetramer solution (0.13 M, 0.052 mmol) was
transferred into a 5-mm NMR tube and cooled to -100 to -105 °C.
A solution of labeled or nonlabeled ketone 2 (0.17 mL, 0.233 M, 0.040
mmol) in THF-d8 was slowly added with good mixing. NMR spectra
were recorded starting at -100 °C (Table 4), as described in the text.
Spectra are also included as Supporting Information.
Supporting Information Available: 6Li NMR spectra for
the tetramers in Table 1 and the 3:1 tetramer. 6Li NMR spectra
after treatment of the 2:2 tetramer 6 with ketone 2 at -100 °C
using 15N-labeled ketone 2 and using 15N-labeled ephedrine 4a.
6Li NMR spectra at -40 °C (2:1 ratio of 4 and 5 generated in
situ and equilibrated at room temperature, hexamer 12 crystals
dissolved in THF-d8, and lithium ephedrate 4). X-ray structural
data for 8 and 12, including complete atomic coordinates and
thermal parameters, bond distances, and angles. This material
is available free of charge via the Internet at http:/pubs.acs.org.
Crystallization of 3:1 Tetramer. To a solution of (1R,2S)-N-
pyrrolydinylnorephedrine (4a) (5.0 g, 27.47 mmol) and Ph3CH (5 mg)
in 40 mL of dry THF at -10 °C was added n-BuLi (1.6 M, 25.2 mL,
40.18 mmol). The color of the reaction solution became red after the
JA0022728