Asymmetric catalysis with silicon-based cuprates: Enantio- and regioselective allylic substitution of linear precursors

Article abstract of DOI:10.1002/chem.201501371

An enantio- and regioselective allylic silylation of linear allylic phosphates that makes use of catalytically generated cuprate-type silicon nucleophiles is reported. The method relies on soft bis(triorganosilyl) zincs as silicon pronucleophiles that are prepared in situ from the corresponding hard lithium reagents by transmetalation with ZnCl₂. With a preformed chiral N-heterocyclic carbene-copper(I) complex as catalyst, exceedingly high enantiomeric excesses are achieved. The new method is superior to existing ones using a silicon-boron reagent as the source of the silicon nucleophile. Zinc about it: Cuprates catalytically generated from bis(triorganosilyl)zinc compounds and a preformed chiral N-heterocyclic carbene-copper(I) complex allow for the highly enantio- and regioselective allylic silylation of various allylic phosphates. The method is broadly applicable with regard to the silicon pronucleophile and the allylic acceptor. It is the first example of asymmetric catalysis with a silicon-based cuprate.

Full text of DOI:10.1002/chem.201501371

DOI: 10.1002/chem.201501371
Communication
&
Allylic Substitution
Asymmetric Catalysis with Silicon-Based Cuprates: Enantio- and
Regioselective Allylic Substitution of Linear Precursors
Alexander Hensel and Martin Oestreich*[a]
Dedicated to Professor Reinhard Brückner on the occasion of his 60th birthday
asymmetric poses a formidable challenge. For example, enan-
tioselective conjugate addition had initially failed, mainly be-
cause of the excess LiCl that comes with the silicon pronucleo-
phile.[9] This hypothesis was verified by the use of
(Me2PhSi)2Zn·4KCl that is not burdened with overstoichiometric
amounts of the lithium Lewis acid, and promising but still low
Abstract: An enantio- and regioselective allylic silylation
of linear allylic phosphates that makes use of catalytically
generated cuprate-type silicon nucleophiles is reported.
The method relies on soft bis(triorganosilyl) zincs as silicon
pronucleophiles that are prepared in situ from the corre-
sponding hard lithium reagents by transmetalation with
ZnCl2. With a preformed chiral N-heterocyclic carbene–
copper(I) complex as catalyst, exceedingly high enantio-
meric excesses are achieved. The new method is superior
to existing ones using a silicon–boron reagent as the
source of the silicon nucleophile.
Silicon-based cuprates such as Fleming’s (Me2PhSi)2CuLi·LiX or
Me2PhSiCu·LiX (X=I or CN) are, despite being stoichiometric in
copper(I), established reagents for the formation of carbon–sili-
con bonds.[1] Catalytic variants had been elusive for decades[2]
until our laboratory introduced (Me2PhSi)2Zn·4LiCl[3] as a pronu-
cleophile in copper(I) catalysis.[4] Combinations of
(Me2PhSi)2Zn·4LiCl and CuX are broadly applicable to common
acceptors[5–8] but rendering these copper(I)-catalyzed reactions
Scheme 2. Enantioselective g-selective copper(I)-catalyzed allylic silylation.
Table 1. Survey of leaving groups.
Entry Allylic precursor Leaving group g/a[a]
Yield [%][b] ee [%][c]
Scheme 1. Racemic copper(I)-catalyzed allylic substitution with zinc- and
boron-based silicon pronucleophiles. pin=pinacolato.
1
2
3
4
5
6
E-1a
E-3a
E-4a
E-5a
E-6a
E-7a
Cl
>95:5
>95:5
>95:5
>95:5
–
80
80
44
98
>99
17
60
–
OP(O)(OEt)2
Br
OC(O)OEt
OC(O)NHPh
OC(O)Me
traces[d]
[e]
–
–
[e]
[a] A. Hensel, Prof. Dr. M. Oestreich
–
–
Institut für Chemie, Technische Universität Berlin
Strasse des 17. Juni 115, 10623 Berlin (Germany)
E-mail: martin.oestreich@tu-berlin.de
[a] Determined by GLC and 1H NMR analysis; [b] combined yield of analyt-
ically pure regioisomers after purification by flash chromatography on
silica gel; [c] determined by HPLC analysis using a chiral stationary phase;
[d]not isolated in analytically pure form; [e] no conversion. LG=leaving
group.
Homepage: http://www.organometallics.tu-berlin.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501371.
Chem. Eur. J. 2015, 21, 9062 – 9065
9062
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
enantioinduction was indeed obtained.[9] Asymmetric allylic
substitution was, however, not achieved with this catalytic
system.[10] These problems were recently overcome by a novel
approach to the catalytic generation of cuprate-type silicon re-
agents, namely the transmetalation of the silicon–boron bond
in Me2PhSiBpin[11] by copper(I) alkoxides.[12] With chiral biden-
tate phosphines or monodentate N-heterocyclic carbenes
(NHCs) as ligands, essentially all typical carbon–silicon bond-
forming reactions were accomplished in enantioselective fash-
ion with excellent enantioselectivities.[13–16]
vantage of this procedure is that it introduces two solvents to
a transformation that is known to be highly sensitive toward
the solvent, particularly with regard to the regioselectivity.[6c,14d]
We nevertheless tested the zinc method with L1·CuCl in that
solvent system and were delighted to find that both allylic
chloride E-1a and phosphate E-3a were converted into allylic
silane g-(R)-2a with 92% ee and 91% ee, respectively (not
shown). Moreover, the g/a ratios were surprisingly high in
comparison to the a-selectivity typically seen with complex
L1·CuCl in THF.[14d,19] A solvent change to Et2O was achieved by
Both the zinc[6c] and the
boron[17] methods are perfectly
applicable to branched-selective
allylic substitution of linear allylic
Table 2. Enantio- and regioselective copper(I)-catalyzed branched-selective allylic silylation of allylic phos-
chlorides (Scheme 1). The g/
ratios are generally higher
phates.
a
with the latter and good regio-
selectivity is also obtained with
allylic
phosphates.
Our
group,[14b,d] as well as Hayashi
and co-workers,[14c] were recently
able to render the boron
method enantioselective by
using copper(I) complexes with
different chiral NHC ligands. By
this method, allylic phosphates
with and without substitution at
the b and g positions were trans-
formed into a-chiral allylic si-
lanes with high enantiocontrol
Entry
1
Phosphate
Allylic silane
g/a[a]
Yield [%][b]
ee [%][c]
>99 (R)
ee [%][d]
E-3a
>95:5
80
96[14b] (R)
2
3
4
E-3b
E-3c
E-3d
>95:5
>99:1
95:5
80
97 (R)
98 (R)
–
74
95[14b] (R)
and
excellent
g-selectivity
(Scheme 2, top).[14b,d] This was
achieved by using preformed
94[e]
>99 (R)
93[14b] (R)
NHC–copper(I)
complexes
L1·CuCl and L2·CuCl, introduced
by McQuade and co-workers
(Scheme 2, bottom).[18] We dis-
close herein that McQuade’s cat-
alysts are also applicable to the
zinc method. The new protocol
is superior to the boron meth-
od[14b,d] not only in terms of
enantio- and regiocontrol but
also with regard to the leaving
group of the allylic acceptor. The
present work is, to our knowl-
edge, the first example of
5
6
7
8
9
E-3e
E-3 f
E-3g
E-3h
E-3i
>99:1
>99:1
97:3
76
96[f] (S)
>99[f] (R)
94 (S)
88[14b] (S)
95[14b] (R)
97[14b] (S)
78[14b] (R)
93[14d] (R)
70
70
92:8
65[g]
70[g] (R)
94[h] (R)
a
highly
enantioselective
carbon–silicon bond-forming re-
action with a catalytically gener-
94:6
75
ated
silicon-based
cuprate
reagent.
1
[a] Determined by GLC and H NMR analysis; [b] combined yield of analytically pure regioisomers after purifica-
tion by flash chromatography on silica gel; [c] determined by HPLC analysis using chiral stationary phases;
[d] enantiomeric excesses obtained with the boron method in CH2Cl2 using precatalyst L1·CuCl (cf. Sche-
me 2);[14b,d] [e] contaminated with disilane; [f] no full baseline separation of the enantiomers achieved by HPLC
analysis; [g] not stable on silica gel, and impurities seen in the HPLC analysis; [h] value likely to be higher as
minor enantiomer co-elutes with impurity.
We began with a few control
experiments. (Me2PhSi)2Zn·4LiCl
is made in situ by mixing
Me2PhSiLi·LiCl in THF (2m) and
ZnCl2 in Et2O (1m). The disad-
Chem. Eur. J. 2015, 21, 9062 – 9065
www.chemeurj.org
9063
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
evaporation under reduced pressure after the salt metathesis
and dissolution of the dark brown residue in Et2O. The result-
ing suspension was filtered to remove undissolved LiCl.[6c] Re-
peating the previous allylic silylations with this ethereal solu-
tion of the zinc reagent boosted the enantioselectivity. Allylic
chloride E-1a afforded g-(R)-2a with 98% ee and allylic phos-
phate E-3a yielded g-(R)-2a with a superb >99% ee (Table 1,
entries 1 and 2). Unexpectedly,[6c] allylic bromide E-4a reacted
with little enantioinduction in moderate yield (Table 1, entry 3).
Other acceptors such as E-5–E-7a, with oxygen leaving groups
that are displaced a-selectively in the racemic series,[6c] showed
poor conversion or did not react at all (Table 1, entries 4–6). It
is also worth noting that L1·CuCl (>99% ee) was again superi-
or to L2·CuCl (70% ee, not shown) in the reaction of the allylic
phosphate, in agreement with earlier findings.[14d]
to afford the distillable and storable SiÀB compound.[11] How-
ever, applications of the boron method have to date been
largely limited to Me2PhSiBpin,[12,16a] despite the availability of
the common substitution patterns at the silicon atom.[11] The
ease of the preparation of the zinc reagent allowed us to test
three other silicon nucleophiles in the allylic silylation (Table 3).
Table 3. Survey of silyl groups.
Entry
R3ÀnPhnSi
Allylic silane
g/a[a]
Yield [%][b]
ee [%][c]
Owing to the excellent enantiomeric excess obtained with E-
3a, we decided to continue with allylic phosphates as sub-
strates (Table 2). For comparison, we have included the enan-
tiomeric excesses achieved with the boron method (Table 2,
column 7).[14b,d] As a general observation with linear acceptors
E-3a–f, the enantioselectivities are substantially higher with
the zinc method at similar levels of regiocontrol (Table 2, en-
tries 1–6); E-3g is an exception with a slightly lower enantio-
meric excess (Table 2, entry 7). Acceptors with substitution at
either the b or g position, such as E-3h and E-3i, afforded dra-
matically improved g/a ratios with comparable enantioselectiv-
ities. The regioselectivity for b-methyl-substituted E-3h in-
creased from 65:35[14b] to 92:8 (Table 2, entry 8) and for g-
methyl-substituted E-3i from 73:27[14d] to 94:6 (Table 2,
entry 9), again demonstrating the superiority of the zinc over
the boron method.
1
2
3
4
Me2PhSi
MePh2Si
Ph3Si
g-(R)-2a
g-(R)-8a
g-(R)-9a
g-(R)-10a
>95:5
>99:1
>99:1
88:12
80
78
60
60
>99
96
68
tBuPh2Si
44
[a] Determined by GLC and 1H NMR analysis; [b] combined yield of analyt-
ically pure regioisomers after purification by flash chromatography on
silica gel; [c] determined by HPLC analysis using chiral stationary phases.
The key finding is that both the enantioselectivity and, to
a lesser extent, the regioselectivity decrease with increasing
steric bulk at the silicon atom (e.g., MePh2Si vs. tBuPh2Si;
Table 3, entries 2 and 4). With the Ph3Si group the g/a ratio re-
mained unaffected while the enantiomeric excess collapsed
(Table 3, entry 3).
Another feature of the boron method is its stereoconver-
gence with E-3a and Z-3a furnishing identical regioselectivities
and enantioselectivities in the same order of magnitude.[14b]
That stereoconvergence was also found for the zinc method
yet with significantly deteriorated levels of regio- and enantio-
control (Scheme 3). This comparison underscores once more
that the reagent cocktails of the boron and zinc methods
indeed lead to different product distributions.
In summary, we have reported herein the first application of
silicon-based cuprates in asymmetric catalysis. Exceedingly
high enantioselectivities were achieved in branched-selective
allylic silylations of linear allylic phosphates, and the method
was also shown to work with chloride as leaving group.[20]
Unlike previously reported methods that relied on a silicon–
boron reagent as the silicon pronucleophile,[14] the new proce-
dure makes use of soft bis(triorganosilyl) zincs that are formed
in situ from the corresponding hard lithium reagents by trans-
metalation with ZnCl2. The zinc method is superior in terms of
regio- and particularly enantiocontrol. Although both the es-
tablished boron method and the new zinc method are likely to
involve similar silicon–copper complexes, the individual salt ad-
ditives lead to different results. NaOMe brings in Lewis basic
methoxide, whereas LiCl is Lewis acidic.
Scheme 3. Stereoconvergence of the zinc versus boron method.
Acknowledgements
A convenience of the zinc over the boron method is the
straightforward reagent preparation. Reductive metalation of
an aryl-substituted chlorosilane, typically R3ÀnPhnSiCl with n=
1–3, with sodium-rich lithium metal followed by salt metathe-
sis with ZnCl2 produces the ready-to-use zinc reagent,[4] possi-
bly requiring a solvent change. The same reductive metalation
is utilized in the preparation of the boron reagent but the in-
termediate anion then reacts with a pinacolborane derivative
M.O. is indebted to the Einstein Foundation (Berlin) for an en-
dowed professorship.
Keywords: allylic substitution · asymmetric catalysis · copper ·
silicon · zinc
Chem. Eur. J. 2015, 21, 9062 – 9065
www.chemeurj.org
9064
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[1] a) R. K. Dieter in Modern Organocopper Chemistry (Ed.: N. Krause), Wiley-
VCH, Weinheim, 2002, pp. 79–144; b) I. Fleming in Organocopper Re-
agents. A Practical Approach (Ed.: R. J. K. Taylor), Oxford Academic Press,
New York, 1994, pp. 257–292.
[2] For a noteworthy exception using zincate (Me2PhSi)ZnMe2Li as the sili-
con pronucleophile together with the carbon-based cuprate
Me2CuLi·LiCN as the catalyst, see: B. H. Lipshutz, J. A. Sclafani, T. Takana-
mi, J. Am. Chem. Soc. 1998, 120, 4021–4022.
d) L. B. Delvos, A. Hensel, M. Oestreich, Synthesis 2014, 2957–2964;
e) L. B. Delvos, M. Oestreich, Synthesis 2015, 924–933.
[15] Aldehyde addition: V. Cirriez, C. Rasson, T. Hermant, J. Petrignet, J. Díaz
lvarez, K. Robeyns, O. Riant, Angew. Chem. Int. Ed. 2013, 52, 1785–
1788; Angew. Chem. 2013, 125, 1829–1832.
[16] Imine addition: a) A. Hensel, K. Nagura, L. B. Delvos, M. Oestreich,
Angew. Chem. Int. Ed. 2014, 53, 4964–4967; Angew. Chem. 2014, 126,
5064–5067; b) T. Mita, M. Sugawara, K. Saito, Y. Sato, Org. Lett. 2014, 16,
3028–3031; c) C. Zhao, C. Jiang, J. Wang, C. Wu, Q.-W. Zhang, W. He,
Asian J. Org. Chem. 2014, 3, 851–855.
[3] Y. Morizawa, H. Oda, K. Oshima, H. Nozaki, Tetrahedron Lett. 1984, 25,
1163–1166.
[4] For a review, see: A. Weickgenannt, M. Oestreich, Chem. Eur. J. 2010, 16,
402–412.
[17] D. J. Vyas, M. Oestreich, Angew. Chem. Int. Ed. 2010, 49, 8513–8515;
Angew. Chem. 2010, 122, 8692–8694.
[5] Reaction with a,b-unsaturated acceptors: M. Oestreich, B. Weiner, Syn-
lett 2004, 2139–2142.
[18] a) J. K. Park, D. T. McQuade, Synthesis 2012, 1485–1490; McQuade and
co-workers successfully applied L1·CuCl and L2·CuCl in enantioselective
carbon–boron bond-forming reactions involving the related boron–
boron bond activation: b) J. K. Park, H. H. Lackey, M. D. Rexford, K.
Kovnir, M. Shatruk, D. T. McQuade, Org. Lett. 2010, 12, 5008–5011 (con-
jugate addition); c) J. K. Park, H. H. Lackey, B. A. Ondrusek, D. T.
McQuade, J. Am. Chem. Soc. 2011, 133, 2410–2413; d) J. K. Park, D. T.
McQuade, Angew. Chem. Int. Ed. 2012, 51, 2717–2721; Angew. Chem.
2012, 124, 2771–2775 (allylic substitution).
[6] Reactions with allylic acceptors: a) M. Oestreich, G. Auer, Adv. Synth.
Catal. 2005, 347, 637–640; b) E. S. Schmidtmann, M. Oestreich, Chem.
Commun. 2006, 3643–3645; c) D. J. Vyas, M. Oestreich, Chem. Commun.
2010, 46, 568–570 and cited references; for the related method with
(Bu3Sn)2Zn·4LiCl as tin pronucleophile, see: d) E. S. Schmidtmann, M.
Oestreich, Angew. Chem. Int. Ed. 2009, 48, 4634–4638; Angew. Chem.
2009, 121, 4705–4709.
[7] Reaction with propargylic acceptors: C. K. Hazra, M. Oestreich, Org. Lett.
2012, 14, 4010–4013.
[19] The solvent change to pure THF had little effect in the substitution of
the allylic phosphate [E-3a!g-(R)-2a]: 93% ee, g/a>99:1, 62%.
[20] For alternative a-chiral allylic silane syntheses by copper(I)-catalyzed
SN2’ allylic displacement with carbon nucleophiles, see: a) M. A. Kacpr-
zynski, T. L. May, S. A. Kazane, A. H. Hoveyda, Angew. Chem. Int. Ed.
2007, 46, 4554–4558; Angew. Chem. 2007, 119, 4638–4642 (aryl and
alkyl zinc reagents); b) Y. Lee, K. Akiyama, D. G. Gillingham, M. K. Brown,
A. H. Hoveyda, J. Am. Chem. Soc. 2008, 130, 446–447 (alkenyl aluminum
reagents); c) F. Gao, Y. Lee, K. Mandai, A. H. Hoveyda, Angew. Chem. Int.
Ed. 2010, 49, 8370–8374; Angew. Chem. 2010, 122, 8548–8552 (aryl
and heteroaryl lithium reagents with Et2AlCl); d) R. Shintani, K. Takatsu,
M. Takeda, T. Hayashi, Angew. Chem. Int. Ed. 2011, 50, 8656–8659;
Angew. Chem. 2011, 123, 8815–8818 (aryl and alkenyl boron reagents);
e) J. A. Dabrowski, F. Gao, A. H. Hoveyda, J. Am. Chem. Soc. 2011, 133,
4778–4781 (alkynyl aluminum reagents); f) B. Jung, A. H. Hoveyda, J.
Am. Chem. Soc. 2012, 134, 1490–1493 (allenyl boron reagents); g) Y.
Shido, M. Yoshida, M. Tanabe, H. Ohmiya, M. Sawamura, J. Am. Chem.
Soc. 2012, 134, 18573–18576 (alkyl boron reagents).
[8] Reactions with CÀC multiple bonds: G. Auer, M. Oestreich, Chem.
Commun. 2006, 311 –313.
[9] G. Auer, B. Weiner, M. Oestreich, Synthesis 2006, 2113–2116.
[10] D. J. Vyas, dissertation, Westfälische Wilhelms-Universität Münster, 2011.
[11] M. Suginome, T. Matsuda, Y. Ito, Organometallics 2000, 19, 4647–4649.
[12] For reviews of SiÀB chemistry, see: a) M. Oestreich, E. Hartmann, M.
Mewald, Chem. Rev. 2013, 113, 402–441; b) E. Hartmann, M. Oestreich,
Chim. Oggi 2011, 29, 34–37.
[13] Conjugate addition: a) E. Hartmann, D. J. Vyas, M. Oestreich, Chem.
Commun. 2011, 47, 7917–7932 (review); b) K.-s. Lee, A. H. Hoveyda, J.
Am. Chem. Soc. 2010, 132, 2898–2900; c) H. Y. Harb, K. D. Collins, J. V.
Garcia Altur, S. Bowker, L. Campbell, D. J. Procter, Org. Lett. 2010, 12,
5446–5449; d) K.-s. Lee, H. Wu, F. Haeffner, A. H. Hoveyda, Organometal-
lics 2012, 31, 7823–7826; e) V. Pace, J. P. Rae, H. Y. Harb, D. J. Procter,
Chem. Commun. 2013, 49, 5150–5152; f) V. Pace, J. P. Rae, D. J. Procter,
Org. Lett. 2014, 16, 476–479.
[14] Allylic substitution: a) L. B. Delvos, M. Oestreich, Chim. Oggi 2013, 31,
74–76 (review); b) L. B. Delvos, D. J. Vyas, M. Oestreich, Angew. Chem.
Int. Ed. 2013, 52, 4650–4653; Angew. Chem. 2013, 125, 4748–4751;
c) M. Takeda, R. Shintani, T. Hayashi, J. Org. Chem. 2013, 78, 5007–5017;
Received: April 8, 2015
Published online on May 12, 2015
Chem. Eur. J. 2015, 21, 9062 – 9065
www.chemeurj.org
9065
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim