Remarkable effect of silyl groups on asymmetric induction in a conjugate addition reaction with O-methylnorephedrine-based N-silylamidocuprates

Article abstract of DOI:10.1021/ja0455805

Scalemic N-silylamidocuprates, prepared from O-methylnorephedrine in a two-flask procedure, are exceedingly reactive and give excellent enantiomeric excesses with chalcone, a classic substrate for studies of organocuprate conjugate addition. For example, the butyl N-diphenylmethylsilylamidocuprate (T = Bu) affords a 99.2% ee (99% yield) of (S)-3-butyl-1,3-diphenyl-1-propanone. The remarkably high yields with this and other silyl ligands are attributed to the extraordinary activating effect of a β-silicon atom on organocuprate reactions. The high ee is a consequence of the larger size of the silyl ligands compared to that of the unsubstituted or methyl-substituted analogues. Copyright

Full text of DOI:10.1021/ja0455805

Published on Web 01/13/2005
Remarkable Effect of Silyl Groups on Asymmetric Induction in a Conjugate
Addition Reaction with O-Methylnorephedrine-Based N-Silylamidocuprates1
Steven H. Bertz,*,† Craig A. Ogle,* and Abhinav Rastogi
Department of Chemistry, UniVersity of North Carolina-Charlotte, UNCC Station, Charlotte, North Carolina 28223,
and Complexity Study Center, Mendham, New Jersey 07945
Received July 22, 2004; E-mail: cogle@email.uncc.edu; sbertz@complexitystudycenter.org.
Scheme 1
Asymmetric induction in the conjugate addition reaction of
organocuprates has received a great deal of attention over the past
two decades, and steady progress has been made, as documented
in reviews by Rossiter,2 Krause,3 and Alexakis.4 Scalemic amido-
cuprates were an attractive approach,5-7 owing to the ready
availability of the corresponding amines from the “chiral pool”;
however, they have been hamstrung by low chemical or optical
yieldssor often both. We report herein “proof of principle” that
the â-Si effect8,9 can be used to achieve excellent chemical and
optical yields from a typical amidocuprate reaction, viz. conjugate
addition of scalemic N-silylamidocuprates based on O-methyl-
norephedrine to chalcone.
Commercially available (1R,2S)-(-)-norephedrine was converted
to our starting material, (1R,2S)-O-methylnorephedrine 1a by an
improved method, as detailed in the Supporting Information. Pyne
used 1a as a chiral auxiliary attached to the substrate for conjugate
additions of organolithium and organocopper reagents and obtained
high (g90%) optical yields in favorable cases.10 In previous work
by one of us,5 a heterocuprate prepared from norephedrine gave
poor results, which made this ligand a good candidate for
improvement.
The new silyl ligands were prepared in situ from 1a and
chlorosilanes 2a-e, as shown in Scheme 1. Thus, solutions of 1a
in ether, dimethyl sulfide (DMS), or tetrahydrofuran (THF) were
treated with 1 equiv of BuLi at -78 °C to afford N-lithio-O-
methylnorephedrine, and 2 was added to yield N-silylamines 3a-
e, which were not isolated. They were likewise lithiated to give
N-lithio-N-silyl-O-methylnorephedrines 4a-e, which were trans-
ferred via syringe or cannula to a -78-°C suspension of MeCu,
prepared from MeLi and CuI in the same solvent. The resulting
methyl N-silylamidocuprates 5a-e were annealed at 0 °C for 0.1
h,11,12 and then they were cooled to -78 °C for further reaction.
Analogously, addition of 4a-e to BuCu in the final stage gave
butyl N-silylamidocuprates 6a-e, respectively. This procedure has
been referred to as “Method B”.13
Chalcone 7 proved to be the ideal yardstick by which to measure
The new N-silylamidocuprates 5a-d and 6a-d gave very high
our results, as it has been studied by many investigators using a wide
chemical and optical yields of 1,4-addition products in ether. The
variety of reagents and ligands.14-22 Optical yields ranging from 0
best examples are butyl cuprates 6b and 6c, based on dimethylphe-
nylsilyl (DMPS) and diphenylmethylsilyl (DPMS) groups, respec-
tively, which gave >99% ee in 99% yield (Table 1, entries 2, 3).
The corresponding methyl cuprates 5b and 5c gave g99% ee and
to 97% ee (82% chemical yield)21 have been reported for organo-
copper reagents. Chemical yields range from 0 to 99% (90% ee).16
Upon aqueous workup, the initial product, enolate 8a or 8b, was
converted to the final one, (S)-3-methyl-1,3-diphenyl-1-propanone
g90% yields (entries 16, 17). The chemical and optical yields with
9a or (S)-3-butyl-1,3-diphenyl-1-propanone 9b, respectively. Chemi-
the trimethylsilyl (TMS) and triphenylsilyl (TPS) ligands were lower
for both T ) Bu (entries 1, 4) and T ) Me (entries 15, 18);
nevertheless, they were still very good. The pattern of eesDMPS,
DPMS > TMS > TPSsis the same for both the butyl and methyl
cal yields were measured by GLC and optical yields (ee) by HPLC
with a Daicel ChiralPak AD column. All 1,4-adducts were
predominantly S, based on Tomioka’s correlation of elution order
with optical rotation.16 While the products are drawn as the S
enantiomers in Scheme 1, minor amounts of the R isomer are
present in all cases.23
series.
Since DMS is an excellent solvent for organocuprate conjugate
addition reactions,24 we had hoped that the optical yields with the
TMS and TPS ligands might be improved in it, but in fact they
† Complexity Study Center.
9
1372
J. AM. CHEM. SOC. 2005, 127, 1372-1373
10.1021/ja0455805 CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Table 1. 1,4-Addition of N-Silylamidocuprates to Chalcone
common substrate. On the basis of the superb results with
derivatives of (1R,2S)-(-)-norephedrine, we believe the â-Si effect
deserves to be tested broadly with organocupratessand other
reagents as well.27
entry (cuprate)
T
solvent
Cu(I) salt
Si group
yd,a
%
ee,b
%
1 (6a)
2 (6b)
3 (6c)
4 (6d)
5 (6e)
6 (6a)
7 (6a)
8 (6a)
9 (6a)
10 (6c)
11 (6d)
12 (6a)
13 (6c)
14 (6d)
15 (5a)
16 (5b)
17 (5c)
18 (5d)
19 (10a)
20 (10b)
Bu
ether
CuI
Me3Si
90c
97c
99.6
99.2
96
10
37
96
10
93
96
Me2PhSi
Ph2MeSi
Ph3Si
tBuMe2Si
Me3Si
99
99
90
45
62
87
46
88
92
90
1
Acknowledgment. Partial support of this work by NSF Grant
0353061 is gratefully noted. S.H.B. thanks Daicel Corporation for
the gift of HPLC columns and Prof. M. Suzuki for helpful advice.
Bu3PCuI
Me2SCuBr Me3Si
CuCN
Supporting Information Available: Typical experimental proce-
dures, including the preparation of 1a. This material is available free
of charge via the Internet at http://pubs.acs.org.
Me3Si
Me3Si
Ph2MeSi
Ph3Si
Me3Si
Ph2MeSi
Ph3Si
Me3Si
Me2PhSi
Ph2MeSi
Ph3Si
DMS CuI
THF
94
NMd
33
References
9
2
(1) (a) New Copper Chemistry, part 31. For part 30 see (b) Bertz, S. H.; Carlin,
C. M.; Deadwyler, D. A.; Murphy, M. D.; Ogle, C. A.; Seagle, P. H. J.
Am. Chem. Soc. 2002, 124, 13650-13651.
(2) Rossiter, B. E.; Swingle, N. M. Chem. ReV. 1992, 92, 771-806.
(3) Krause, N. Angew. Chem., Int. Ed. Engl. 1997, 36, 187-204.
(4) Alexakis, A.; Benhaim, C. Eur. J. Org. Chem. 2002, 3221-3236.
(5) Bertz, S. H.; Dabbagh, G.; Sundararajan, G. J. Org. Chem. 1986, 51,
4953-4959.
NMd
97
Me ether
CuI
88
90
92
88
25
14
99.0
99.6
95
10
85
Bu
none
none
(6) Dieter, R. K.; Tokles, M. J. Am. Chem. Soc. 1987, 109, 2040-2046.
(7) Only one ligand, (S)-N-methyl-1-phenyl-2-(1-piperidinyl)ethanamine, gave
examples of enantiomeric excesses (ee) >90%; see Miao, G.; Rossiter,
B. E. J. Org. Chem. 1995, 60, 8424-8427.
a All reactions were run at -78 °C for 20 h. b %S - %R. c After 2 h at
-78 °C, the yield was 85%, and the ee was 96%. d Not measured, owing
to low yield.
(8) Bertz, S. H.; Miao, G.; Rossiter, B. E.; Snyder, J. P. J. Am. Chem. Soc.
1995, 117, 11023-11024.
(9) Bertz, S. H.; Eriksson, M.; Miao, G.; Snyder, J. P. J. Am. Chem. Soc.
1996, 118, 10906-10907.
were slightly lower in all cases (entries 9-11). In THF the yields
were very low (entries 12-14). The yields of 1,4-adduct from Bu2-
CuLi‚LiI were also significantly lower in THF (16%) than in ether
(52%) or DMS (49%).
(10) Pyne, S. G. J. Org. Chem. 1986, 51, 81-87.
(11) Bertz, S. H.; Gibson, C. P.; Dabbagh, G. Tetrahedron Lett. 1987, 28,
4251-4254.
(12) Tanaka, K.; Matsui, J.; Suzuki, H.; Watanabe, A. J. Chem. Soc., Perkin
Trans. 1 1992, 1193-1194.
The use of other Cu(I) compounds as precursors was also
investigated (entries 6-8). Whereas in some applications Bu3PCuI
was superior to CuI,25 our results with it were much inferior.
Substitution of CuBr in the preferred form of its DMS complex26
gave results that were comparable to those from CuI. The use of
CuCN under the same conditions gave very poor results.
To better understand the interplay between electronic and steric
factors, we prepared unsubstituted amidocuprate 10a and methyl-
substituted 10b from 1a and 1b, respectively. Cuprate 10a gave a
low yield and very low ee (entry 19). Upon going from H to methyl
(i.e., 10a to 10b), the ee improved dramatically from 10% to 85%,
while the yield fell from 25% to 14% (entry 20). Both effects can
be attributed to the larger size of methyl vs H. Upon going from
methyl to trimethylsilyl (i.e., 10b to 6a), substituent size again
increases, and so too did the ee, from 85% to 97% (entry 1).
Incredibly, the yield rebounded from 14% to 90%.
Thus, by incorporating the promethean TMS group (i.e., going
from 10a to 6a), the ee was improved from 10% to 97%, and the
yield, from 25% to 90%. The extraordinary increases in yield and
ee can be attributed to the powerful activating effect of â-Si on
Cu.8,9,27 It appears that the tert-butyldimethylsilyl group is too bulky
for a favorable fit of 6e with our substrate; nevertheless, the yield
from this cuprate (45%) is significantly higher than from 10a or
10b (25% and 14%, respectively), which have much smaller
substituents.
(13) Method A involves adding a lithium reagent to an amidocopper(I) species,5
Method B involves adding a lithium amide to an organocopper(I)
compound,5 and Method C involves adding a mixture of alkyllithium and
lithium amide (from the amine and 2 equiv of lithium reagent) to the
Cu(I) salt; see Eriksson, J.; Arvidsson, P. I.; Davidsson, O¨ . J. Am. Chem.
Soc. 2000, 122, 9310-9311. In selected experiments with method A, the
results were virtually the same. Method C gave essentially the same ee
but significantly lower yields.
(14) Huche´, M.; Berlan, J.; Pourcelot, G.; Cresson, P. Tetrahedron Lett. 1981,
22, 1329-1332.
(15) Leyendecker, F.; Laucher, D. NouV. J. Chim. 1985, 9, 13-19.
(16) Kanai, M.; Nakagawa, Y.; Tomioka, K. Tetrahedron 1999, 55, 3831-
3842.
(17) Morimoto, T.; Yamaguchi, Y.; Suzuki, M.; Saitoh, A. Tetrahedron Lett.
2000, 41, 10025-10029.
(18) Alexakis, A.; Burton, J.; Vastra, J.; Benhaim, C.; Fournioux, X.; van den
Heuvel, A.; Leveˆque, J.-M.; Maze´, F.; Rosset, S. Eur. J. Org. Chem. 2000,
4011-4027.
(19) Martorell, A.; Naasz, R.; Feringa, B. L.; Pringle, P. G. Tetrahedron:
Asymmetry 2001, 12, 2497-2499.
(20) Shintani, R.; Fu, G. C. Org. Lett. 2002, 4, 3699-3702.
(21) Hu, Y.; Liang, X.; Wang, J.; Zheng, Z.; Hu, X. J. Org. Chem. 2003, 68,
4542-4545.
(22) Ogle, C. A.; Human, J. B. Tetrahedron: Asymmetry 2003, 14, 3281-
3283.
(23) To obtain the best results, solutions of 7 were cooled with dry ice and
added slowly down the walls of the reaction vessel, so that upon entering
the rapidly stirred reaction mixture, they were as close as possible to -78
°C. When the reaction of 6a and 7 was run at -50 °C, instead of -78
°C, the ee dropped from 97% to 77%, which was not unexpected.5
(24) Bertz, S. H.; Dabbagh, G. Tetrahedron 1989, 45, 425-434.
(25) Nicewicz, D. A. MS Thesis, University of North Carolina-Charlotte, 2001.
(26) House, H. O.; Chu, C.-Y.; Wilkins, J. M.; Umen, M. J. J. Org. Chem.
1975, 40, 1460-1469.
(27) The activating effect of â-Si has been extended to Zn reagents: Jones, P.;
Reddy, C. K.; Knochel, P. Tetrahedron 1998, 54, 1471-1490.
(28) Berrisford, D. J.; Bolm, C.; Sharpless, K. B. Angew. Chem., Int. Ed. Engl.
1995, 34, 1059-1070.
(29) When an additional 1 equiv of BuLi was added to 6a before 7 in ether,
the yield of 1,4-adduct was 27% (32% ee), the yield of 1,2-adduct was
3.5%, and the recovery of 7 was 53% after 2 h at -78 °C (cf. Table 1,
note c). With 1 equiv of BuLi alone, the yield of 1,4-adduct was 25%
and the yield of 1,2-adduct was 67%.
(30) The reaction was run in the usual way, but when it was complete, another
1 equiv of BuLi was added, and the reaction mixture was annealed at 0
°C for 0.1 h, as usual. Upon cooling to -78 °C, an additional 1 equiv of
7 was added, and the reaction mixture was stirred for 2 h at -78 °C to
complete cycle 2. With n cycles the final mol % of catalyst is 100/n. For
example, with n ) 5 the effective catalyst loading is 20 mol %, and the
yield was 73% (80% ee).
Sharpless has described “ligand-accelerated catalysis” in which
a catalytic amount of ligand reacts with a reagent to form a (small
amount of) a new reagent that has significantly higher reactivity
toward the substrate.28 As lithium reagents are highly reactive
toward R-enones, we did not expect to be able to use our new
cuprates for such classical catalysis.29 Nevertheless, following
Tanaka et al.,12 we were able to achieve encouraging results with
a “batch-catalytic” process,30 e.g., after three cycles, the yield was
99% and the ee was 84%.
In conclusion, the activating effect of â-Si on organocuprate
reactivity has been demonstrated to extend to the enantiomeric
excesses from typical silyl derivatives of an amidocuprate with a
JA0455805
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1373