Mechanistic dichotomy in the asymmetric allylation of aldehydes with allyltrichlorosilanes catalyzed by chiral pyridine N-oxides

FULL PAPER

DOI: 10.1002/chem.201203817

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes with

Allyltrichlorosilanes Catalyzed by Chiral Pyridine N-Oxides

Andrei V. Malkov,*^{[a, b]}Sigitas Stoncˇius,*^{[a, c]}Mark Bell,^{[a, d]}Fabiomassimo Castelluzzo,^{[a, e]}

Pedro Ramꢀrez-Lꢁpez,^{[a, f]}Lada Biedermannovꢂ,^{[a, g]}Vratislav Langer,^[h]

Lubomꢀr Rulꢀsˇek,*^[i]and Pavel Kocˇovsky*^{[a, j]}

´

Dedicated to Professor Reinhard W. Hoffman on the occasion of his 80th birthday

Abstract: Detailed kinetic and compu-

tational investigation of the enantio-

and diastereoselective allylation of al-

dehydes 1 with allyltrichlorosilanes 5,

employing the pyridine N-oxides

METHOX (9) and QUINOX (10) as

chiral organocatalysts, indicate that the

reaction can proceed through a disso-

ciative (cationic) or associative (neu-

tral) mechanism: METHOX apparent-

ly favors a pentacoordinate cationic

transition state, while the less sterically

demanding QUINOX is likely to oper-

ate via a hexacoordinate neutral com-

plex. In both pathways, only one mole-

cule of the catalyst is involved in the

rate- and selectivity-determining step,

which is supported by both experimen-

tal and computational data.

Keywords: allylation · allylsilanes ·

calculations · organocatalysis · pyri-

dine N-oxides

Introduction

blocks in the synthesis of numerous natural products and

drug candidates and for the construction of other functional

molecules.^[2]Allylboranes and boronates (2, M=B) have

been shown to react with aldehydes via the cyclic transition

state 4A, and, as a consequence, the products are formed in

high diastereoisomeric purity (3b and c), reflecting the con-

figuration of the allylic moiety of the reagent (2b and c).

Allylation of aldehydes 1 with organometallics 2^[1,2](M=

B,^[3]Si,^[4]Sn,^[5]Cr,^[6]and Ti,^[7]etc.^[1]) is a powerful strategic

reaction that creates up to two chiral centers (Scheme 1).^[1]

The resulting homoallylic alcohols 3 are the cornerstones of

organic synthesis and, as such, have been utilized as building

ˇ

[a] Prof. Dr. A. V. Malkov, Dr. S. Stoncius, Dr. M. Bell,

[f] Dr. P. Ramꢀrez-Lꢁpez

Dr. F. Castelluzzo, Dr. P. Ramꢀrez-Lꢁpez, Dr. L. Biedermannovꢂ,

Present address: Department of Organic Chemistry

University of Sevilla

ˇ

´

Prof. Dr. P. Kocovsky

Department of Chemistry, WestChem

University of Glasgow, Glasgow G12 8QQ (UK)

Fax : (+44)141-330-4888

Garcꢀa Gonzꢂlez 1, 41012 Sevilla (Spain)

[g] Dr. L. Biedermannovꢂ

Present address: Institute of Biotechnology

Academy of Sciences of the Czech Republic

E-mail: pavelk@chem.gla.ac.uk

ˇ

[b] Prof. Dr. A. V. Malkov

Vꢀdenskꢂ 1083, 14220 Prague 4 (Czech Republic)

Present address: Department of Chemistry

Loughborough University

Loughborough, Leics LE11 3TU (UK)

Fax : (+44)1509-22-3925

[h] Prof. Dr. V. Langer

Environmental Inorganic Chemistry

Department of Chemical and Biological Engineering

Chalmers University of Technology

412 96 Gothenburg (Sweden)

E-mail: a.malkov@lboro.ac.uk

ˇ

[c] Dr. S. Stoncius

ˇ

[i] Dr. L. Rulꢀsek

Present address: Department of Organic Chemistry

Vilnius University

Naugarduko 24, 03225 Vilnius (Lithuania)

E-mail: sigitas.stoncius@chf.vu.lt

Institute of Organic Chemistry and Biochemistry

and Gilead Sciences Research Center

Academy of Sciences of the Czech Republic

ˇ

Flemingovo nꢂmestꢀ 2, 16610 Prague 6 (Czech Republic)

[d] Dr. M. Bell

E-mail: lubos@uochb.cas.cz

Present address: Drug Discovery Unit

College of Life Sciences, University of Dundee

Dundee, DD1 5EH (UK)

ˇ

´

[j] Prof. Dr. P. Kocovsky

Sabbatical address: Department of Organic Chemistry

Arrhenius Laboratory, Stockholm University

SE 10691 Stockholm (Sweden)

[e] Dr. F. Castelluzzo

Present address: ISF

General Medicine Buisness Unit, Bayer S.p.A.

Via Orazio Console, n. 103, 000128 Rome (Italy)

Supporting information for this article, including data for the allylation

products, NMR spectra of new compounds, all kinetic data and equili-

brium structures, and energies of all the computationally studied spe-

cies, is available on the WWW under http://dx.doi.org/10.1002/

chem.201203817.

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

1

&

ÞÞ

These are not the final page numbers!

Scheme 2. Allylation of aldehydes with allylic trichlorosilanes. For 1a–k,

R_E, and R_Z(when different from H or Me), see Tables 1–3.

workup with NaHCO₃, which generates innocuous, readily

separable inorganic by-products (SiO₂and NaCl), leaving

the desired organic product essentially clean. Therefore, it

comes as no surprise that the enantioselective version of this

reaction became a playground for the development of new

chiral Lewis basic catalysts.^[11]Denmark reported on the

first enantioselective allylation that was catalyzed by chiral

phosphoramides.^{[1d,14,18–21]}Nakajima then demonstrated a

similar catalytic effect for the axially chiral biquinoline

N,N’-dioxide,^[22]which inspired Hayashi,^[23]Kotora,^[24,25]

Kwong,^[26]and others to develop a diverse range of bipyri-

dine N,N’-dioxides^[23–25,27]and N,N’,N’’-trioxides,^[26]the enan-

tioselectivities of which were thoroughly studied in the past

few years. In parallel, we have developed novel pyridine-de-

Scheme 1. Allylation of aldehydes (for M, see text).

The enantioselective version of the latter reaction^[3]requires

a chiral (stoichiometric) auxiliary L*, typically derived from

a monoterpene or tartaric acid. In this case however, the

isolation of the desired product and its separation from the

stoichiometric reagents is not trivial on a large scale and

may present serious difficulties.^[8]Allyltrimethylsilanes (2,

M=Si)^[1e,4]represent a less expensive alternative and if the

reaction is catalyzed by a chiral Lewis acid (LA) that coor-

dinates to the carbonyl, preferential formation of one enan-

tiomer can be expected. However, this allylation apparently

proceeds via the open transition state 4B or 4C,^[4,9]allowing

more flexibility so that reduced levels of stereocontrol are

often encountered, with the major product being syn-config-

ured (3c), irrespective of the configuration of the allylic

silane (2b and c). Furthermore, the Me₃SiX that is gradually

released during the reaction can itself act as a Lewis acid

and compete with the chiral Lewis acidic catalyst, which in-

evitably has a negative effect on the enantioselectivity.^[10]

Allylation with allyltrichlorosilanes 5 (Scheme 2) requires

catalysis by a Lewis base,^[11,12]such as DMF,^[13]DMSO,^[13,14]

or HMPA,^[14,15]and is free of the latter drawbacks.^[16,17]

Here, the cyclic transition state 4A (Scheme 1) with the

Lewis base coordinated to the Lewis acidic silicon atom is

believed to be a prerequisite for the reaction to occur; con-

sequently, high levels of stereocontrol can be attained.^[11,12]

Another positive feature of this allylation is the aqueous

rived N-mono

enantio- and diastereoselectivity in the allylation and cro

AHCTUNGTREGaNNNU tion of various aldehydes. Hoveyda then reported on the

ACHTUNGTRENNUNG

first “non-pyridine-type” N-monooxide derived from pro-

line,^[32]and Govender made use of the N-oxides derived

from tetrahydroisoquinolines.^[27d]Chiral, Lewis basic sulfox-

ides,^[33]sul

N

amides,^[34]phosphine oxides, such as BINAPO

ACHTUNGTRENNUNG

AHCTUNGTRENNUNG

ide),^[28e,35]and dinitrones^[36]have also been shown to catalyze

or mediate the allylation reaction.^[37]

Our pyridine-type N-monooxides 9 (METHOX)^[29]and 10

(QUINOX)^[30](Figure 1) are among the most successful

cataACHTNUGRTNEUNG

lysts reported to date for the allylation of aromatic^[29a]

and a,b-unsaturated^[29b]aldehydes with allylic trichlorosi-

lanes, giving ꢀ98% ee and ꢀ99:1 d.r. at 1–5 mol% catalyst

loading. Therefore, this new methodology can be viewed as

a robust alternative to the existing and well established

metal-mediated protocols, which require stoichiometric aux-

iliaries or metal-based Lewis acids. Interestingly, these two

closely related N-oxides exhibit intriguing differences in the

allylation and crotylation of substituted benzaldehydes 1a–

j,^[29,30]suggesting that two fundamentally different mecha-

Abstract in Czech: Enantio- a diastereoselektivnꢀ allylace al-

˚

ˇ

˚

dehydu 1 allyltrichlorsilany 5 s pouzitꢀm derivꢁtu pyridin-N-

oxidu METHOX (9) a QUINOX (10) jako chirꢁlnꢀch orga-

˚

ˇ

´

nokatalyzꢁtoru byla detailne studovꢁna pomocꢀ kinetickych a

nisms ACHTUGNTERNoUNG perate in these reactions. Herein, we present a de-

´

ˇ

˚ ˇ

vypocetnꢀch metod. Reakce muze probꢀhat disociativnꢀm (ka-

tailed investigation into this mechanistic dichotomy, discuss

the effects that govern the reactivity and selectivity of the

pyridine-type N-oxide catalysts, and present a full account

´

tiontovym) nebo asociativnꢀm (neutrꢁlnꢀm) mechanismem: v

ˇ

´

prꢀpade METHOXu je pravdepodobny pentakoordinovany

´

ˇ

ˇ ´

kationtovy tranzitnꢀ stav, zatꢀmco mꢂne stericky nꢁrocny

of the cataACTHNGUTERNNUlG yst design and the reaction scope of METHOX

ˇ

QUINOX zrejme dꢁvꢁ prednost neutrꢁlnꢀmu hexakoordino-

(9).

ˇ

vanꢂmu komplexu. Experimentꢁlnꢀ i vypoctenꢁ data ukazujꢀ,

ˇ

ze se kroku urcujꢀcꢀho rychlost a selektivitu reakce ﬄcastnꢀ v

ˇ

obou prꢀpadech pouze jedna molekula katalyzꢁtoru.

&

2

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

Scheme 3. Krçhnke annulation. a: Ar=Ph; b: Ar=2-MeOC₆H₄; c: Ar=

4-MeOC₆H₄; d: Ar=2-FC₆H₄; e: Ar=2,6-(MeO)₂C₆H₃; f: Ar=2,4,6-

(MeO)₃C₆H₂; g: Ar=3,4,5-(MeO)₃C₆H₂; h: Ar=3,5-(CF₃)₂C₆H₃; i: Ar=

2-MeO-1-naphthyl; j: Ar=3,4,5-(MeO)₃C₆H₂.

the preparation of the Krçhnke salts 20a–i by iodination of

the corresponding methyl aryl ketones 19a–i in pyridine,

where the a-iodo ketone intermediate N-alkylates the sol-

vent. Pinocarvone (ꢁ)-21, obtained in a quantitative yield

by the ene reaction of (+)-a-pinene with singlet oxygen in

the presence of tetraphenylporphine as photosensitizer and

acetic anhydride,^[28i,39]was then heated with the respective

Krçhnke salts 20a–g and AcONH₄to produce the pyridine

derivatives 22a–g in good yields.^[40]Methylation of the latter

products at the benzylic position, mediated by lithium diiso-

propylamide (LDA),^[28i,41]afforded 23a–g with excellent dia-

stereoselectivity. The bistrifluoromethyl and naphthyl deriv-

atives 23h and i were obtained by Krçhnke annulation of

the corresponding salts 20h and i to the Michael acceptor

26, which in turn was obtained in three steps from (+)-isopi-

nocampheol (+)-24 by using the procedure described by

Kwong.^[42]In this Krçhnke annulation, enone 26 (which

Figure 1. Pyridine-type N-oxides as catalysts for allylation reactions.

Results and Discussion

Catalyst design and synthesis: The design of METHOX

(9)^[29]and QUINOX (10)^[30]evolved from the successful

series of our bipyridine N-monooxide catalysts 11–14,^[28a–d]

which performed better than their N,N’-dioxide ana-

1

A

exists as a ꢂ10:1 epimeric mixture, as revealed by H NMR

of the annulated terpene moiety, whereas their congeners 12

and 13 combine the effects of both central and axial chirali-

ty, because the rotation about the bond connecting the two

pyridine moieties is restricted by the two methyl groups and

spectroscopy) was converted into the pyridine derivatives

23h and i in which the configuration of the methyl group

was inverted under the reaction conditions.^[42]Oxidation of

the pyridine nitrogen in 23a–i with m-chloroperoxybenzoic

acid provided the required N-oxides 9 and 15–18.^[43]The last

representative of this series, the naphthyl-derived N-oxide

18, turned out to be a ꢂ1:1 mixture of atropoisomers

(ꢁ)-18a and (ꢁ)-18b, owing to the restricted rotation about

the aryl–aryl bond. The two atropoisomers were separated

by flash chromatography and proved to be configurationally

stable; however, their axial configuration has not been es-

tablished.^[44]

ꢁ

the N O group. METHOX (9) and its desmethoxy ana-

logues 15 retained the central chirality of the terpene unit

but were stripped of the axial chirality, though restricted ro-

tation about the arene–arene bond forced the two aromatic

rings out of plane in the case of 9 and 15e.

The relatively free rotation about the bond linking the

aro

ACHTUNGTRENNUNG

rich tri

N

ACHTUNGTRENNUNG

rivative 17, its electron-poor counterpart. To complete the

mosaic, it was also desirable to synthesize the atropoisomers

18a and b that can be viewed as hybrids of 9 and 10.

Asymmetric allylation of aromatic aldehydes with allyltri-

chlorosilane: Allylation of benzaldehyde (1a) with allyltri-

chlorosilane 5a, catalyzed by PINDOX (11), gave the corre-

sponding homoallylic alcohol (S)-6a in 90% ee (Table 1,

The synthesis of the N-oxides 9 and 15–18 relied on

Krçhnke annulation^[28,29,38](Scheme 3) and commenced with

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

3

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

Table 1. The allylation of benzaldehyde (1a) with allyltrichlorosilane

(5a) catalyzed by Lewis bases (Scheme 2).^[a]

dinated species, the pyridine–pyridine bond becomes a

chiral axis, and its configuration, (R_a) in this instance (as de-

duced from the behavior of 12 and 13), is dictated by the

chirality of the terpene moiety.^[28a]

Entry

Catalyst

[mol%]

Solvent

T

t

Yield

[%]^[b]

ee

A

[^oC]

[h]

[%]^[c,d]

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

(+)-11 (10)

(+)-12 (10)

(ꢁ)-13 (10)

(ꢁ)-14 (10)

(ꢁ)-15a (10)

(+)-15b (10)

(ꢁ)-15c (10)

(ꢁ)-15d (10)

(ꢁ)-15e (10)

(+)-9 (10)

(+)-9 (5)

(+)-9 (1)

(+)-9 (5)

(+)-16 (5)

(ꢁ)-17 (10)

(+)-10 (10)

(ꢁ)-18a (10)

(ꢁ)-18b (10)

CH₂Cl₂

CH₃CN

CH₂Cl₂

CHCl₃

CH₂Cl₂

CH₃CN

CHCl₃

ꢁ60

ꢁ40

ꢁ60

ꢁ40

ꢁ20

0

24

12

24

18

4

78

72^[e]

67

75

66

90 (S)^[f]

98 (S)^[f,g]

82 (R)^[f,g]

96 (S)^[g]

41 (S)^[h,i]

68 (S)^[h,i]

58 (S)^[h,i]

16 (S)^[h,i]

90 (S)^[j]

96 (S)^[h]

95 (S)^[h]

93 (S)^[h]

91 (S)^[h]

88 (S)^[h]

72 (S)^[i]

–

While the latter results seem to suggest a chelation of the

ꢁ

silicon atom between the N O group and the nitrogen atom

of the other pyridine nucleus,^[45]further experiments demon-

strated that the nitrogen coordination does not constitute a

prerequisite for the asymmetric induction: thus, the N-oxide

15a, in which the pyridine ring and the adjacent terpene

unit in the eastern part are replaced by an unsubstituted

phenyl, did catalyze the reaction with a non-negligible selec-

tivity (41% ee, entry 5).^[28b,29a]Enhanced asymmetric induc-

tion was observed for the electronically enriched o- and p-

methoxy derivatives 15b and c (68 and 58% ee, entries 6

and 7).^[28b,29a]By contrast, the electronically poor o-fluoro

derivative 15d reacted sluggishly and with a considerably

lower enantioselectivity (16% ee, entry 8),^[28b,29a]suggesting

that the presence of an electron-rich aromatic system, as in

15b and c, is beneficial. Indeed, the 2,6-dimethoxy deriva-

tive (ꢁ)-15e exhibited good reactivity and provided a re-

markably high level of enantioselectivity (90% ee, entry 9).

A real improvement was attained with the 2,4,6-trimethoxy

derivative METHOX ((+)-9, 96% ee, entry 10),^[29a]especial-

ly for the reactions run in CH₃CN in which the catalyst load-

ing can be substantially reduced with no adverse effect on

enantioselectivity (entries 11–13).^[29a]Furthermore, when

CH₃CN was employed as solvent, the elevated temperature

caused only a slight erosion of enantioselectivity (entries 14–

16), and good stereocontrol was attained even at room tem-

perature (88% ee, entry 16). The 3,4,5-isomer 16 was found

to offer lower selectivity than 9 (72% ee, entry 17) but still

higher than that attained with 15a–c. By contrast, the elec-

tron-deficient 3,5-bis(trifluoromethyl) analogue 17 turned

out to be practically inert (entry 18). This finding, together

with the behavior of the monofluoro derivative 15d

(entry 8), further demonstrates the importance of an elec-

tron-rich aromatic system in this part of the catalyst.^[28b,29]

As reported earlier, the biaryl catalyst QUINOX (10)

55

35^[e]

15^[e]

60

74

ꢃ95

68

ꢃ95

86

RT

ꢁ20

ꢁ40

ꢁ60

18

12

18

87

ꢀ5^[e]

CH₂Cl₂

CH₃CN

CH₂Cl₂

68

87 (R)^[k]

72 (R)^[k]

25 (S)

60

5^[e]

15^[e]

62 (R)

[a] The reaction was carried out on a 0.4 mmol scale with respect to 1a

by using 5a (1.2 equiv), the catalyst (2–10 mol%), and (iPr)₂NEt

(1.5 equiv) as base. [b] Isolated yield (product pure by ¹H NMR spec-

ACHTUNGTRENNUNGtroscopy). [c] Determined by chiral HPLC or GC. [d] All products (6a)

were of S-(ꢁ)-configuration (unless otherwise stated), as revealed by the

comparison of their optical rotations (measured in CHCl₃) and their GC

and HPLC retention times with the literature data and with the behavior

of authentic samples (refs. [22,23,28,29,30]). [e] Incomplete conversion.

[f] Ref. [28a]. [g] Ref. [28c]. [h] Refs. [28b,29a]. [i] Catalyst was of

ꢂ87% ee. [j] The previously reported 82% ee (ref. [29a]) was obtained at

ꢁ608C with a less enantiopure catalyst. [k] Ref. [30].

entry 1).^[28a]The dimethyl derivative 12, in which the rota-

tion about the bond connecting the two pyridine nuclei is re-

stricted, exhibited even higher enantioselectivity (98% ee,

entry 2).^[28a,c]Its atropoisomer 13 catalyzed the formation of

the opposite enantiomer but with lower enantioselectivity

(82% ee, entry 3),^[28a,c]which clearly shows that this combi-

nation of the chiral centers and chiral axis represents the

mismatched case and that the axial chirality overrides the

effect of central chirality residing in the terpenic part of the

scaffold. However, compounds 12 and 13 are not configura-

tionally stable (the equilibration is complete in solution

within several days) so that, although 12 exhibits the highest

enantioselectivity reported to date,^[11]it could hardly

become an off-shelf catalyst. The redesigned, configuration-

ally stable iso-PINDOX (14), lacking the atropoisomerism,

showed the same enantioselectivity as 12 (96% ee,

entry 4)^[28c]but due to the considerably increased steric hin-

AHCTUNGTREGeNNNU xhibited good reactivity and high enantioselectivity

(ꢀ87% ee) in CH₂Cl₂(entry 19), while it was slightly less ef-

fective in CH₃CN (72% ee; entry 20).^[30]

The atropoisomeric naphthalene derivatives 18a and b,

which can be regarded as crossbreeds between the terpene

and biaryl catalysts (9 vs. 10), reacted sluggishly and exhibit-

ed low enantioselectivity (in CH₂Cl₂). Interestingly, 18a and

b induced the formation of the opposite enantiomers of 6a

(entries 21 and 22), which is consistent with the behavior of

axially-chiral terpene-derived bipyridine N-oxides 12 and 13,

confirming that, in these instances, the product configuration

is controlled mainly by the chiral axis.^[28a]

ꢁ

drance in the vicinity of the N O moiety, the reaction was

The electronic and steric effects in the aldehyde substrate

were elucidated with the aid of substituted benzaldehydes

1b–j and catalysts 9, 10, 15e, and 16 (Table 2). When the

2,4,6-trimethoxy derivative METHOX ((+)-9) was em-

ployed as catalyst, benzaldehyde (1a) and its o-, m-, and p-

substituted congeners 1b–h were found to react with a simi-

rather slow and, therefore, of limited practical value. Never-

theless, this experiment demonstrated that if the rotation

about the pyridine–pyridine bond is relatively free, the cata-

lyst can coordinate to the silicon atom of 5a in a way that

maximizes the enantioselectivity of the reaction. In the coor-

&

4

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

Table 2. The allylation of aldehydes 1a–k with allyltrichlorosilane (5a) catalyzed by Lewis bases (Scheme 2)^[a]

lowered enantioselectivity can

be attributed to the relatively

unhindered rotation about the

Entry Aldehyde Ar

Catalyst [mol%] Solvent

T [^oC] t [h] Yield [%]^[b]ee [%]^[c,d]

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

1a

1b

1c

1d

1e

1 f

1g

1h

1i

Ph

(+)-9 (5)

CH₃CN ꢁ40

18

2

ꢃ95

86

ꢃ95

87

ꢃ95

80

81

75

0

96 (S)

93 (S)

96 (S)

95 (S)

89 (S)

94 (S)

97 (S)

92 (S)

–

ꢁ

4-CF₃-C₆H₄

4-MeO-C₆H₄

3-MeO-C₆H₄

2-MeO-C₆H₄

4-Cl-C₆H₄

3-Cl-C₆H₄

2-Cl-C₆H₄

3,5-Me₂-C₆H₄(+)-9 (5)

2,6-Me₂-C₆H₄(+)-9 (5)

Ph

4-CF₃-C₆H₄

4-MeO-C₆H₄

2-MeO-C₆H₄

4-Cl-C₆H₄

3-Cl-C₆H₄

2-Cl-C₆H₄

CH₃CN ꢁ40

C C bond connecting the pyri-

dine and benzene nuclei in the

catalyst and to less steric con-

gestion in the vicinity of the re-

action center. It is worth noting

that the latter rings in 9 and

15e are forced to be approxi-

mately perpendicular (the tor-

sion angle in the crystal of 9

was found to be 1158; see

below), whereas 16 can be

more flexible, imposing less re-

striction to the approach and

orientation of the substrate. As

a further result, which stands in

stark contrast to the behavior

of 9 and 15e, catalyst 16 did

promote the allylation of 3,5-di-

methylbenzaldehyde (1i), al-

though with a rather low level

1j

0

–

1a

1b

1c

1e

1 f

1g

1h

1i

1a

1k

1c

1i

1a

1b

1c

1d

1e

1i

(ꢁ)-15e (10)

CHCl₃

CH₃CN ꢁ20

ꢁ40

60

34

25

53

63

54

75

0

87

58

72

73

68

85

70

75

82

73

40

68

0

90 (S)

85 (S)

91 (S)

75 (S)

88 (S)

89 (S)

86 (S)

–

(ꢁ)-15e (10)

3,5-Me₂-C₆H₄(ꢁ)-15e (10)

Ph

(+)-16 (5)

72 (S)^[e]

70 (S)^[e]

62 (S)^[e]

87 (R)^[f]

96 (R)^[f]

16 (S)^[f,g]

72 (R)

45 (R)

80 (R)^[f]

37 (R)^[f]

81 (R)^[h]

–

4-F-C₆H₄

4-MeO-C₆H₄

3,5-Me₂-C₆H₄(+)-16 (5)

Ph

(+)-10 (5)

(ꢁ)-10 (5)

(+)-10 (5)

CH₂Cl₂

ꢁ40

ꢁ20

0

ꢁ40

2

4-CF3-C6H4

4-MeO-C₆H₄

18

12

16

18

4-MeO-C₆H₄

3-MeO-C₆H₄

2-MeO-C₆H₄

3,5-Me₂-C₆H₄(+)-10 (5)

2,6-Me₂-C₆H₄(+)-10 (5)

of

asymmetric

induction

(62% ee, entry 22), apparently

as a result of the less congested

chiral pocket.

1j

In contrast to the terpene-de-

rived catalysts 9, 15e, and 16,

which gave high enantioselec-

tivities across the spectrum

from electron-deficient to elec-

tron-rich benzaldehydes (see

above), QUINOX (10) exhibit-

ed strong electronic depend-

ence. Thus, while benzaldehyde

[a] The reaction was carried out on a 0.4 mmol scale with respect to aldehyde 1 by using 5a (1.2 equiv), the

catalyst (5–10 mol%), and (iPr)₂NEt (1.5 equiv) as base. [b] Isolated yield (product pure by ¹H NMR spec-

ACHTUNGTRENNUNGtroscopy). [c] Determined by chiral HPLC or GC. [d] The configuration of the products (6) was established by

comparison of their optical rotations (measured in CHCl₃) and their GC and HPLC retention times with the

literature data and with the behavior of authentic samples (refs. [22,23,28,29,30]). [e] The catalyst was of

87% ee. [f] Ref. [30]. [g] Note that the opposite enantiomer of the catalyst was used. [h] The catalyst was of

85% ee.

lar level of asymmetric induction (89–97% ee) and at com-

parable rates (Table 2, entries 1–8), irrespective of their

electronics or the steric hindrance exercised by the o-sub-

stituent. By contrast, no reaction was observed for 3,5-di-

(1a) and its electron-poor p-trifluoromethyl analogue 1b re-

acted with high enantioselectivity and enhanced reaction

rates (87 and 96% ee, Table 2, entries 23 and 24), a nearly

racemic product was obtained from the electron-rich p-me-

thoxybenzaldehyde (1c, 16% ee, entry 25); in the last in-

stance, considerable deceleration was also observed.^[30]Fur-

thermore, o-methoxybenzaldehyde (1e) shared the low

enantioselectivity (37% ee, entry 29), whereas its m-isomer

1d (in which the MeO group has a slight electron-withdraw-

ing effect) resembled the behavior of 1a and b (80% ee,

ACHTUNGTRENNUNGmethACHTUNGTRENNUNGylbenzaldehyde (1i, entry 9), presumably owing to the

steric bulk of the substrate that cannot enter the chiral

cavity created in the transition state. Its 2,6-isomer 1j was

equally unreactive as a result of the general steric hindrance

to the addition to the carbonyl group (entry 10).

The 2,6-dimethoxy catalyst (ꢁ)-15e followed the suite,

though with slightly lower enantioselectivities (75–90% ee)

and reduced reaction rates, reflected in lower conversions

(entries 11–17): again, little difference was observed be-

tween the electron-rich and electron-poor aldehydes, and

the sterically bulky 3,5-dimethylbenzaldehyde (1i) also

proved to be inert (entry 18), which is consistent with the

observation made for METHOX (entry 9).

entry 28). 3,5-DiACTHNUGTRENNUGmethACHUTTGNRENyNUGN lbenzaldehyde (1i) reacted readily

(81% ee, entry 30), which again is in a stark contrast to the

behavior of 9 and 15e. Nevertheless, the congested 2,6-

isomer 1j proved inert, as expected (entry 31).

The huge difference in enantioselectivity observed for 1b

and c (96 vs. 16% ee) in the reaction catalyzed by QUINOX

(10) is striking and, to our knowledge, represents the largest

difference in enantioselectivity as a function of electronic ef-

fects reported to date for any enantioselective reaction.

However, the asymmetric induction in the case of 1c was

only low at ꢁ408C (16% ee, entry 25); at ꢁ208C, it reached

The 3,4,5-trimethoxy catalyst (+)-16 exhibited good reac-

tivity but lower enantioselectivity than 9 and 15e. Again,

little difference was observed for benzaldehyde and its p-

substituted congeners 1a–c (70–72% ee, entries 19–21). The

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

5

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

a much higher level (72% ee, entry 26),^[30b]indicating a large

entropy effect (or a change of mechanism), which does not

operate in the case of METHOX (9) and its congeners. No-

tably, further increase of the reaction temperature to 08C

resulted in the “normal” decrease of enantioselectivity to

45% ee (entry 27).^[30b,c]

The difference in the reactivities of catalysts 9 and 10 ex-

tends to asymmetric crotylation (Scheme 2). With the major-

ity of the known catalysts,^[11]the anti/syn ratio of the prod-

ucts reflects the trans/cis ratio of the starting crotylsilane.

However, METHOX (9) was found to exhibit a strong ki-

netic preference towards the trans isomer 5b.^[29a,46]Thus,

levels but the stereochemistry copied the scenario observed

for trans-5b (entries 6–12). The vinyl bromide 5h proved to

be inert (entry 13) and required different catalysts.^[28e]The

remaining terpene-derived N-oxide catalysts 11 (PINDOX)

and 14 (iso-PINDOX) mimicked the reactivity of

METHOX (entries 14–18), although the kinetic preference

for the trans isomer 5b was less accentuated.

In contrast to METHOX, the crotylation catalyzed by

QUINOX (10) proceeded slightly faster with cis-5c (com-

pare entries 19–21 and 22–24), and the reaction reproduced

the original E/Z ratio with good fidelity (see also discussion

in ref. [30b]). Again, the enantioselectivities were consider-

crotylation of benz

G

ACHTUNGTRENaNUGN bly affected by the electronics of the aldehyde, that is, 1b

equivalents of crotyltrichlorosilanes 5b (E/Z 87:13)^[46,47]in

the presence of 9 as catalyst (5 mol%) in CH₃CN as the op-

timized solvent, afforded the pure anti-configured product

(ꢁ)-7a (anti/syn ꢃ99:1) of high enantiopurity (97% ee,

Table 3, entry 1). By contrast, the reaction of 1a with pure

5c (Z/E ꢃ98:2)^[48]turned out to be sluggish (26% conver-

sion), affording the syn product (ꢁ)-8a in low stereochemi-

cal purity (6:1 d.r., 26% ee, entry 2).^[29a,77]p-Substituted

versus 1c (compare the ee values in entry 20 vs. 21 and 23

vs. 24), in the same way as in the allylation (Table 2,

entry 24 vs. 25).^[30]The reactivity with bulky reagents 5g and

h (see Table 3 for the R_Esubstituent) mirrored that of

METHOX (entries 25 and 26).

Kinetic experiments: As shown above, the two key catalysts,

METHOX (9) and QUINOX (10), differ dramatically in

their behavior, which indicates fundamental differences in

their mode of action. As an initial step on the way to under-

stand this divergence, we investigated their kinetics. Previ-

ously, we have reported that in the case of QUINOX (10),

benzACHTUNGTRENNUNGaldehydes 1c, k, and l behaved in the same way as 1a

(entries 3–5). Bulky, trans-configured crotyl-type reagents

5d–g (see Table 3 for the R_Esubstituent) required a higher

catalyst loading to maintain the reaction rate at practical

Table 3. Reaction of benzaldehydes 1 with (E)- and (Z)-crotyltrichlorosilanes 5b and c (Scheme 2).^[a]

Entry Aldehyde Ar

Catalyst

[mol%]

Silane R_E

R_ZE/Z

Solvent

T

t

Yield anti/syn^[c]Product 7/8

G

[^oC]

[%]^[b]

ee [%]^[d,e]

1

1a

1c

1k

1l

1a

1b

1d

1k

1a

1b

1c

1a

1b

1c

1b

Ph

(+)-9 (5)

(E)-5b Me

H

87:13 CH₃CN ꢁ40 18h 90

ꢃ99:1

1:6

97 (1S,2S)^[l]

26 (1S,2R)

ꢃ96 (1S,2S)

98 (1S,2S)

2^[f]

(Z)-5c

H

Me ꢀ2:98 CH₃CN ꢁ40 24h 26

3^[g]

4-MeO-C₆H₄(+)-9 (5)

4-NO₂-C₆H₄(+)-9 (5)

(E)-5b Me

(E)-5d Pr

(E)-5e PhCH₂

(E)-5 f BnOCH₂

(E)-5g Me₃SiCH₂

H

Br

H

87:13 CH₃CN ꢁ40 24h 87

87:13 CH₃CN ꢁ40 24h 82

87:13 CH₃CN ꢁ40 24h 90

ꢃ98:2 CH₃CN ꢁ40 24h 80

ꢃ98:2 CH₃CN ꢁ40 24h 85

ꢃ25:1

ꢃ99:1

n/a

93:7

96:4

93:7

98:2

10:90

95:5

96:4

83:17

1:99

4^[g]

5^[g]

4-Me-C₆H₄

Ph

4-CF₃-C₆H₄

3-MeO-C₆H₄(+)-9 (15)

4-NO₂-C₆H₄(+)-9 (20)

Ph

(+)-9 (5)

6^[g]

(+)-9 (20)

(+)-9 (15)

97 (1S,2S)

7^[g]

8^[g]

97 (1S,2R)^[m]

93 (1S,2R)^[m]

90 (1S,2R)^[m]

96 (1S,2R)^[m]

94 (1S,2R)^[m]

n/a

9^[h]

ꢃ99:1 CH₃CN ꢁ35 7d

ꢃ99:1 CH₃CN ꢁ20 24h

52

55

41

46

0

10^[h]

11^[h]

12^[h]

13^[i]

14^[j]

15^[j]

16^[i]

17^[j]

18^[j]

19^[k]

20^[k]

21^[k]

22^[k]

23^[k]

24^[k]

25^[h]

26^[i]

(+)-9 (10)

(Z)-5h

H

(+)-11 (10) (E)-5b Me

(ꢁ)-14 (10) (E)-5b Me

87:13 CH₂Cl₂ꢁ60 24h 54

87:13 CH₂Cl₂ꢁ60 15h 44

87:13 CH₂Cl₂ꢁ60 18h 27

87:13 CH₃CN ꢁ40 18h 88

87 (1S,2S)

ꢃ99 (1S,2S)

Ph

98 (1S,2S)

(ꢁ)-14 (10) (Z)-5c

H

Me ꢀ2:98 CH₃CN ꢁ40 18h 37

87 (1S,2R)

(ꢁ)-10 (5)

(E)-5b Me

H

ꢃ98:2 CH₂Cl₂ꢁ40 24h 65

ꢃ98:2 CH₂Cl₂ꢁ40 24h 75

ꢃ98:2 CH₂Cl₂ꢁ40 24h 40

66 (1S,2S), 77 (1S,2R)

92 (1S,2S), 90 (1S,2R)

25 (1S,2S), 12 (1R,2S)

69 (1S,2S), 79 (1S,2R)

n.d., 94 (1S,2R)

40 (1R,2R), 60 (1S,2R)

88 (1S,2R)^[m]

n/a

4-CF₃-C₆H₄

4-MeO-C₆H₄(ꢁ)-10 (5)

Ph

(ꢁ)-10 (5)

(Z)-5c

H

Me ꢀ2:98 CH₂Cl₂ꢁ40 24h 78

Me ꢀ2:98 CH₂Cl₂ꢁ40 24h 85

Me ꢀ2:98 CH₂Cl₂ꢁ40 24h 50

4-CF₃-C₆H₄

4-MeO-C₆H₄(ꢁ)-10 (5)

4:96

4-CF₃-C₆H₄

(ꢁ)-10 (20) (E)-5g Me₃SiCH₂

H

Br

ꢃ98:2 CH₂Cl₂ꢁ35 9d

50

0

ꢃ99:1

(ꢁ)-10 (10) (Z)-5h

H

ꢃ99:1 CH₂Cl₂ꢁ20 24h

n/a

[a] The reaction was carried out on a 1.0 mmol scale with respect to 1 (entries 1–18) and/or on a 0.4 mmol scale (entries 19–26) by using 5 (1.2 or

1.5 equiv), the catalyst (5–10 mol%), and (iPr)₂NEt (1.5 equiv, entries 1–13, 15, and 17–26) or Bu₄NI (1 equiv, entries 14 and 16). The enantiopurity of

the catalysts was as follows: 9 (ꢃ99%), 11 (ꢃ96%), 14 (ꢃ96%), and 10 (98%). Note that here the S-(ꢁ)-enantiomer of 10 was employed (in contrast

to Tables 1 and 2). [b] Isolated yield. [c] The anti/syn ratio was determined by GC. [d] The enantiopurity was determined by chiral GC. [e] The absolute

configuration was inferred from the optical rotation (measured in CHCl₃) and from the GC retention times by comparison with the literature data

(refs. [22,23,24,25,26]) and with the behavior of authentic samples (refs. [28,29,30]). [f] Ref. [29a]. [g] Ref. [46]. [h] Ref. [29c]. [i] Ref. [28e].

[j] Ref. [28c]. [k] Ref. [30]. [l] The slightly lower ee value (95% ee) reported earlier (ref. [29a]) was due to the less pure catalyst. [m] Note the change in

priority of the substituents according to the Cahn–Ingold–Prelog convention. n.d. =not determined; n/a=not applicable.

&

6

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

the reaction followed first-order kinetics in all components:

aldehyde, allyltrichlorosilane, and the catalyst. To determine

the kinetic parameters for the METHOX-catalyzed reac-

tion, we employed the same approach, namely the method

of initial rates in which aliquots were taken at certain inter-

vals and analyzed by GC to monitor the product formation.

Good reproducibilities were achieved up to 20% conver-

sion. Again, the reaction followed first-order kinetics in al-

dehyde 1b (1.09, in the range of 1.0–2.0 equiv), allyltri-

ACHTUNGTRENNUNGchloroACHTUNGTRENNUNGsilane 5a (1.01, within the 0.5–2.0 equiv range), and

the catalyst (0.98, with 0.05–0.10 equiv).

Next, the secondary kinetic isotope effect (SKIE)^[49]was

measured by comparing the relative reactivities of ArCD=O

versus ArCH=O. For this purpose, the allylation was carried

out by using 9 (2 mol%), 5a (1 equiv), and deuterated and

standard substrate aromatic aldehydes (10 equiv each). In

the case of benzaldehyde (1a) and its p-methoxy derivative

1c, a SKIE value of 0.73 was found, which indicates a late

transition state, closely resembling the sp³-hybridized prod-

uct. For p-trifluoromethylbenzaldehyde (1b), a decrease in

the SKIE value to 0.80 signals a partial increase of the sp²

character of the TS. With QUINOX (10), a similar trend

was observed, the respective SKIE values for 1a, b, and c

being 0.81, 0.86, and 0.70.^[30b]

A set of equilibria operating in the allylation reaction is

shown in Equations (1)–(4). First, the catalyst is assumed to

reversibly bind to the allylsilane, generating the pentacoor-

dinate donor–acceptor complex 27 [Eq. (1)]. An extracoor-

dination has been known to dramatically enhance the Lewis

acidity of the silicon center,^[12f,45a,50]which should facilitate

the coordination of the weakly Lewis basic aldehyde to pro-

duce the ternary complex 28 [Eq. (2)]. The reaction would

then take place in the coordination sphere of 28 to release

the product [Eq. (3)]. It is reasonable to assume that the

rate of equilibration in Equation (1) is much higher, while in

Equation (2) it is comparable to the rate of allylation in

Equation (3); then the reaction rate v can be expressed by

Quantum Chemistry Calculations

To better understand the mechanistic aspects of the allyla-

tion reactions and provide energetic and structural data for

some of the above arguments, a thorough computational

analysis was carried out. Since the existence of the complex

of the catalyst with allyltrichlorosilane was established (see

above), the following mechanisms can be envisioned: 1) as-

sociative, 2) dissociative, 3) second sphere, and 4) a nonco-

valent interaction that may also be considered as an addi-

tional scenario. The first option is characterized by the hexa-

coordinate reactant complex, whereas in the second mecha-

nism the substitution (dissociation) of a chloride anion by

the aldehyde species is assumed (implying the existence of a

transient ion pair). In the third mechanism, the aldehyde

moiety can, in principle, react with the “activated allyl” in

the second sphere, though this mechanism seems to be quite

unlikely. Finally, the non-covalent alternative, recently sug-

gested by Roithovꢂ and Kotora,^[25]does not involve a direct

bonding between Si and the catalyst but rather the interac-

Equation (4), known in enzymatic catalysis for rapid equi

ACHTUNGTRENNUNG

ACHTUNGTNERlNUNG ib-

rium-ordered bisubstrate systems.^[51]The first order kinetic

ꢁ

in aldehyde and silane suggests that under the catalytic con-

ditions, in the range of the tested concentrations, K_iAK_Bis

much larger than the sum of the two other factors in the de-

nominator in Equation (4). When the concentration of the

aldehyde is increased 20-fold, as in the case of the SKIE de-

termination, the system seems to approach saturation in al-

dehyde and thus, gradually moves towards zero order in

substrates. Under these conditions, the carbon–carbon bond

formation becomes the rate-limiting step (SKIE 0.73). Natu-

rally, saturation is achieved sooner for the more Lewis basic

1c, followed by 1a and b. From the kinetic point of view,

the only difference between the catalysis by 9 and 10 is that

the saturation is reached at different concentrations of the

aldehyde. The similarity in the kinetic pattern observed for

both 9 and 10 suggests that the differences in the mechanism

of their catalytic behavior is likely to reside in the structure

of the transient species 28; therefore, we turned to the com-

putational analysis of the respective transition states.

tion of the N O group with various C H groups of the reac-

tant(s), in reminiscence of enzymatic catalysis. However,

since this mechanism was originally proposed for bipyridine-

type N,N’-dioxide catalysts and a non-polar medium, it was

not investigated in this work, which deals with the reactivity

of mono-N-oxide METHOX in CH₃CN.

Non-catalyzed reaction: We first investigated the non-cata-

lyzed reaction of allyltrichlorosilane with benzaldehyde in

CH₃CN. The reaction profile, including the structures of the

reactant complex (RC), the transition state (TS), and the

product are depicted in the Supplementary Information

(Figure S1). The calculated barrier at the RI-PBE+D/def2-

TZVP level of theory (including solvation, zero-point vibra-

tion energies, and thermal corrections to the Gibbs energies)

is DG^°=28.1 kcalmol^ꢁ1; the calculated thermodynamics of

the process predict that it should be exergonic by DG=

ꢁ10.3 kcalmol^ꢁ1, whereas the formation of the reactant

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

7

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

complex (RC) is endergonic by DG_RC=20.7 kcalmol^ꢁ1.

Structurally, the TS is characterized by the formation of a

ꢁ

C C bond and breaking of a Si C bond. At the TS, they

attain values of r_CꢁC=1.97 and r_SiꢁC=2.11 ꢄ, which can be

compared with r_CꢁC=3.09 and r_SiꢁC=1.93 ꢄ in the reactant

complex. Because the uncatalyzed reaction is experimentally

very slow, the reaction barrier of DG^°_noncatalyzed=28.1 kcal

mol^ꢁ1can be regarded as the upper limit for any plausible

reaction mechanism of the catalyzed reaction.

The six-membered ring in the TS complex adopts a chair

conformation (see 4A, Scheme 1), which mirrors the reac-

tion catalyzed by QUINOX (in which the preference of the

chair conformation over the twisted boat conformation for

both R- and S-channels was calculated to be 1.5–2.0 kcal

mol^ꢁ1).^[30b]

To further ascertain the accuracy of the computed values,

calculations of the reaction barrier in this model system

were carried out by using the high level CCSD(T) method

with the large basis set aug-cc-pVDZ. The value of DE^°=

15.3 kcalmol^ꢁ1, obtained this way, is in excellent agreement

with the DE^°=15.6 kcalmol^ꢁ1calculated by using the RI-

PBE(+D)/def2-TZVP method, which is used throughout

this study. This result supports the reliability of the data pre-

sented herein, not only from the qualitative but also from

the quantitative perspective.

Figure 2. The equilibrium geometries of the reactants, the reactant com-

plex (RC), the transition state (TS), and the products along the reaction

pathway of the allylation of benzaldehyde (1a) with 5a catalyzed by pyri-

dine N-oxide. All distances are in [ꢄ]. The calculated values were ob-

tained at the RI-PBE(+D)/def2-TZVP//RI-PBE(+D)/6-31G(d) level.

The free energy profile of the reaction is depicted as well.

tion of the complex has been calculated at DG₁=25.0 kcal

mol^ꢁ1, that is, this complex is less stable than the three sol-

vated reactant molecules. The dominant term in the calcu-

lated DG₁is an (anticipated) unfavorable entropy change

that accompanies the formation of a trimolecular complex.

Enthalpically, the reactant complex is DH₁=ꢁ14.3 kcal

mol^ꢁ1more stable than the reactants. As for the three other

structural variants described above, it can be observed that

they are almost isoenergetic (within 2.9 kcalmol^ꢁ1) with the

minimum-energy conformation described above.

Pyridine N-oxide associative mechanism: Since both

METHOX (9) and QUINOX (10) contain a pyridine N-

oxide functional group that is crucial for their catalytic

activi

ty, we first studied the simplest catalyzed reaction bear-

Following the latter pathway, a TS was found that con-

nects the reactant complex with a product structure that is

only by 2.5 kcalmol^ꢁ1(measured by free energy change) less

stable than the reactant complex. Therefore, the reaction

barrier for the reaction of benzaldehyde and allyltrichlorosi-

lane catalyzed by pyridine N-oxide is DG^°=27.5 kcalmol^ꢁ1,

which is 0.6 kcalmol^ꢁ1lower than the reaction barrier calcu-

lated for the uncatalyzed reaction. This corresponds to an

increase of a reaction-rate constant by approximately a

factor of three. Therefore, pyridine N-oxide should only

function as a rather inefficient catalyst, which is qualitatively

supported by the experimentally observed very slow reac-

tion.

ing the important structural features of the allylation reac-

tion. This is a minimal system that allowed us to study the

three plausible mechanisms: associative, dissociative, and

second sphere. It also allowed to partially discriminate be-

tween the electronic and steric effects in the catalysis by

more complex catalysts.

The overall thermodynamic of the reaction (after dissocia-

tion of the catalyst) is, of course, identical to the previous

value of DG=ꢁ10.3 kcalmol^ꢁ1, but the reaction coordinate

was found to be more complex. First, the existence of the re-

actant complex for the associative and dissociative reaction

mechanism can be postulated. Here, benzaldehyde (1a) and

the allyl moiety of 5a must occupy the neighboring positions

in the coordination polyhedron (octahedron for the associa-

tive and trigonal bipyramid for the dissociative mechanism).

This leaves us with four positions of the pyridine N-oxide in

the associative mechanism (and three for the dissociative al-

ternative), which were all systematically investigated. The

structure of the reactant complex then determines the TS

through which it leads to the product complex. Below, the

results are discussed separately for the individual reaction

mechanisms.

Structurally, the TS is characterized by the two bond

lengths of r_CꢁC=2.21 and r_SiꢁC=2.05 ꢄ that can be com-

pared to those of 2.80 and 1.98 ꢄ in the reactant complex

(Figure 2). The apparent structural and energetic similarity

between the reactant complex and the transition state has

prompted us to also calculate the TS structure for the least

stable of the reactant complexes, namely, that with the pyri-

dine N-oxide in a trans position (axial) with respect to benz-

AHCTUNGTRENNUNG

aldehyde. Indeed, it was found to be 0.8 kcalmol^ꢁ1less

stable than the TS corresponding to the lowest (free) energy

pathway. This result constitutes supporting evidence for the

hypothesis that the most stable reactant complex(es) should

lead to the lowest (free) energy TS(s). With this phenome-

We first investigated the associative mechanism

(Figure 2). The most stable reactant complex contains the

pyridine N-oxide in trans position (axial) with respect to the

allyl group and in cis position (equatorial) with respect to

benzaldehyde. The free energy associated with the forma-

AHCTUNGTRENNUNG

&

8

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

to the reactant complexes (which only requires simple

energy minimization) and avoid tedious TS searches for all

possible conformers.

The next step in the reaction is the formation of the prod-

uct complex, which is DG₂=ꢁ17.9 kcalmol^ꢁ1more stable

than the reactant complex. Still, the value is DG₁+DG₂=

7.1 kcalmol^ꢁ1; that is, the product complex is less stable (on

the free-energy scale) than the isolated reactants. The prod-

uct structure is depicted in Figure 2. The last step in the

change and the formation of the reactant complex is DG₁=

23.8 kcalmol^ꢁ1for the most stable structure with pyridine

N-oxide in a trans position (axial) with respect to benzalde-

hyde. This value is by 1.2 kcalmol^ꢁ1lower in energy than

that for the formation of the hexacoordinate associative

complex. In contrast to the associative mechanism, the other

structural variants of the reactant complex differ by ꢂ6 kcal

mol^ꢁ1.

According to this scenario, the reaction should proceed

through a pentacoordinate TS and a product complex with

assumed reinsertion of Cl^ꢁinto the Si coordination shell, re-

turning the system to the associative reaction pathway with

the free-energy changes being identical to those described

above. The calculated free energy barrier is DG^°=31.2 kcal

mol^ꢁ1, which is 3.7 kcalmol^ꢁ1higher than that for the asso-

ciative mechanism. Therefore, this reaction is unlikely to

proceed through a dissociative mechanism.

catalACHTUNGTRENNUNGysis is the pyridine N-oxide dissociation, which is char-

acterized by a value of DG₃=ꢁ17.4 kcalmol^ꢁ1(with respect

to the product complex), thus yielding DG=ꢁ10.3 kcal

mol^ꢁ1for the total reaction, as discussed above.

Pyridine N-oxide dissociative mechanism: With the pyridine

N-oxide model system, the dissociative mechanism as an al-

ternative reaction pathway was also explored. This endeavor

involved the participation of a transient ion pair and the TS

featuring a pentacoordinate complex with two chlorine

atoms, pyridine N-oxide, allyl, and benzaldehyde as ligands.

Two possible computational strategies can be envisaged: 1)

treating the whole system as an ion pair with the chloride

ion bound in the second sphere of the Si complex; and 2) to

consider an ion-separated system. Not being aware of theo-

retical arguments that would strongly favor one of the two

alternatives, we chose the latter, admitting that comparing

the dissociative and associative mechanism is not a trivial

task. In this approach, the first step is the exchange of one

chloride for benzaldehyde and the formation of the penta-

coordinate reactant complex RC (Figure 3) with the chloride

Pyridine N-oxide second-sphere mechanism: Finally, the

ꢁ

variant that the C C bond between benzaldehyde and the

allyl is formed in the second sphere, that is, without a direct

coordination of benzaldehyde to the silicon atom, was also

explored. To this end, a number of second-sphere reactant

complexes were optimized to find the structures in which

benzaldehyde is in a favorable orientation to the allyl

moiety of trichlorosilane. The formation free energy of one

of the most stable second-sphere complexes (Figure S2 in

the Supporting Information) was calculated to be DG₁=

18.5 kcalmol^ꢁ1. The second-sphere complex is by DDG₁=

6.5 kcalmol^ꢁ1more stable than the first-sphere complex,

with the dominant contribution to this energy difference

originating, as expected, in the entropy term D

G

5.5 kcalmol^ꢁ1. However, the reaction coordinate corre-

ꢁ

sponding to the second-sphere C C bond formation (i.e.,

ꢁ

C C relaxed PES scan) shows that this reaction path is un-

likely to be favorable (for details, see the Supporting Infor-

mation). The barrier is higher than 50 kcalmol^ꢁ1(after

adding DE₂=ꢂ27 kcalmol^ꢁ1to DG₁=18.5 kcalmol^ꢁ1and

considering ꢁTDS equal to ꢂ5 kcalmol^ꢁ1), and the last

point on the R_CꢁCcoordinate is not a local minimum. There-

fore, the second-sphere mechanism as the potential reaction

channel can be safely ruled out. This finding is in agreement

with the qualitative considerations related to the observed

stereoselectivity of the reactions catalyzed by METHOX

and QUINOX. In fact, with the second-sphere mechanism,

it would be difficult to find a plausible explanation of the

factors energetically discriminating between the TSs corre-

sponding to the formation of R- and S-configured products.

In fact, many isoenergetic minima do exist for both orienta-

tions of benzaldehyde leading to R and S products in this

mechanism (see below) but none appears to be substantially

favored in terms of energy.

Figure 3. The equilibrium geometries of the reactants, the reactant com-

plex, the transition state, and the product along the dissociative reaction

pathway of the allylation of benzaldehyde catalyzed by pyridine N-oxide.

The calculated values were obtained at the RI-PBE(+D)/def2-TZVP//

RI-PBE(+D)/6-31G(d) level. All distances are in [ꢄ].

ion being presumably in the bulk (solution). Therefore, in

that step, the free energy of chloride solvation in CH₃CN

must be considered, and it was estimated to be DG

(Cl^ꢁ)=

_solvACHTGNUTRENNUNG

ꢁ84.8 kcalmol^ꢁ1by using the same method as for the rest of

the systems and the Sackur-Tetrode equation for the calcula-

tion of the translational entropy. As depicted in Figure 3,

the free-energy change accompanying the PhCHOꢅCl^ꢁex-

METHOX-catalyzed allylation of benzaldehyde (1a) with

allyltrichlorosilane (5a): Finally, we addressed the reaction

mechanism of the allylation catalyzed by METHOX (9).

The results are summarized in Table 4; as in the previous in-

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

9

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

DG^°_dissoc=25.6 and 23.3 kcalmol^ꢁ1for the R and S reaction

channels, respectively (Figure 4), showing that the S channel

should be prefered by 2.3 kcalmol^ꢁ1. The latter difference

corresponds to 96 (at 258C) and 98.5% ee (at ꢁ408C) in

favor of the S enantiomer, which is in excellent agreement

with the experimentally observed 88% ee at room tempera-

ture (Table 1, entry 16) and 96% ee at ꢁ408C (Table 1,

entry 12), respectively. On the energy scale, the difference

Table 4. The calculated thermochemical data (Gibbs free energies, DG)

for the allylation of benzaldehyde (1a) with allyltrichlorosilane (5a) cata-

lyzed by METHOX (9).^[a]

Entry

Mechanism

Config.

Reactant

complex

Transition

state

Product

complex

1

2

3

4

associative

dissociative

R

S

R

S

24.4

–

25.0

17.8

29.3

28.7

25.6

23.3

ꢁ2.0

ꢁ4.8

ꢁ2.0

ꢁ4.8

[b]

[a] All DG values are in kcalmol^ꢁ1. [b] No stable complex in vacuo struc-

between the experimental and theoretical is DDG_R/S

=

0.6 kcalmol^ꢁ1, which is likely to be at the error bar of the

computational method. This is in contrast to both the

QUINOX and pyridine N-oxide catalyzed reactions report-

ed previously by us^[30b]and here. We hypothesize that it is

not the Lewis basicity of the catalyst but solely the steric de-

mands of METHOX that prevent an energetically feasible

formation of the hexacoordinate reactant complex. At the

ture was found.

stances, only the data for the most stable structures in the

associative and dissociative mechanism are presented. The

formation of the reactant complex is characterized by a

value of DG₁=17.8–25.0 kcalmol^ꢁ1(the range applies to

multiple conformers studied here), which is similar to the

corresponding value for QUINOX as catalyst.^[30b]However,

no transition state on the associative pathway with the barri-

er lower than that for an uncatalyzed reaction

(DG^°_noncatalyzed=28.1 kcalmol^ꢁ1) could be identified. Here,

same time, we believe that the steric demands of ACTHNUGTRNEUNGMETHOX

and favorable dispersion interactions between its methoxy-

phenyl ring and benzaldehyde (in the reactant complex and

the transition state) cause the different structural arrange-

ment of the silicon coordination sphere in comparison with

the pyridine N-oxide (which is in the apical position with re-

spect to the benzaldehyde).

Following the reaction pathway, the product complexes

are then by DG₂=ꢁ27.6 and ꢁ28.1 kcalmol^ꢁ1more stable

than the transition states, that is, the overall thermochemis-

try is predicted to be DG₁+DG₂=ꢁ2.0 and ꢁ4.8 kcalmol^ꢁ1

before and DG₁+DG₂=ꢁ10.3 kcalmol^ꢁ1after the catalyst

dissociation (for the R and S reaction channels, respective-

the lowest-energy transition states were found at DG^°

=

assoc

29.3 and 28.7 kcalmol^ꢁ1for the R and S reaction channels,

respectively. Furthermore, the marginal energy difference

between these channels, namely 0.6 kcalmol^ꢁ1in favor of

the S product, would not rationalize the high enantioselec-

tivity observed experimentally. This all rules out the associa-

tive pathway and leaves us with the dissociative mechanism

as the only candidate; here, the calculated barriers are

Figure 4. The equilibrium geometries of the most stable reactant complexes, transition states, and product complexes along the reaction coordinate for

the dissociative pathway of the allylation of benzaldehyde cataACHTNUGRTENUNGlyzed by METHOX (9). The calculated values were obtained at the RI-PBE(+D)/def2-

TZVP//RI-PBE(+D)/6-31G(d) level. All distances are in [ꢄ].

&

10

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

Table 5. Calculated transition state energies (DG^°) for the allylation of

p-substituted benzaldehydes 1a–c with allyltrichlorosilane (5a) catalyzed

by METHOX (9) by using the dissociative mechanism (Scheme 2).^[a]

Entry ArCHO Ar

Product TS

ee [%]

(calc.)^[b]

ee [%]

ACHTUNGTRENNUNG

(exp.)^[c]

U

1

2

3

4

5

6

1a

1b

1c

Ph

R

S

R

S

R

S

25.6

23.3 98.6

27.5

24.3 99.8

28.1

25.5 99.2

–

96^[d]

–

4-CF₃-C₆H₄

4-MeO-C₆H₄

–

93

–

96

–

[a] All DG values are in kcalmol^ꢁ1. [b] Calculated for 233.15 K at 1atm.

[c] Established for ꢁ408C (Table 1). [d] The enantioselectivities at

ꢁ208C, 08C, and room temperature were 93, 91, and 88% ee, respective-

ly (Table 1, entries 14–16).

derivatives 1b and c, were submitted to the same theoACHTUNGTRENNUNGretical

Figure 5. Transition states A and B refer to the allylation of benzalde-

hyde (1a) with allyltrichlorosilane (5a) catalyzed by METHOX (9),

giving rise to (S)-6a as the major product (from A) and its enantiomer

(from B). Allyl carbon atoms are shown in orange, benzaldehyde carbon

atoms in turquoise. Transition states C and D refer to the crotylation of

benzaldehyde (1a) with (E)-crotyltrichlorosilane (5b) and (Z)-crotyltri-

chlorosilane (5c) catalyzed by METHOX (9), giving rise to the anti and

syn products 7 and 8, respectively; the terminal methyl group of the

crotyl group is shown in magenta.

treatment. As shown in the previous paragraph, the prefer-

ence for the S channel over its R counterpart was computed

to be 2.3 kcalmol^ꢁ1for 1a (Table 5, entries 1 and 2). Similar

differences, namely 3.2 and 2.6 kcalmol^ꢁ1, were now found

for 1b and c, respectively, as revealed by comparison of en-

tries 3 and 4 with entries 5 and 6. This result translates to

theoretical values of 98.6, 99.8, and 99.2% ee at ꢁ408C for

1a–c, which qualitatively correlates with 96, 93, and 96% ee

attained experimentally (at the same temperature), allowing

for at least ꢂ1 kcalmol^ꢁ1errors in the calculated activation

energies.

ly); the major S product is thus favored both kinetically and

thermodynamically. This result stands in stark contrast to

QUINOX, in which both R and S product complexes were

almost isoenergetic (for the preferred associative mecha-

nism).^[30b]

The space-filling models for the key allylation of benzal-

dehyde through the dissociative mechanism (A and B in

Figure 5) clearly show the cavity available for the incoming

benzaldehyde. The preferred spatial positioning of the part-

ners in the transition state A, giving rise to the S product,

becomes obvious by comparison with the more encumbered

transition state B that would lead to the R enantiomer (a

very minor product).

METHOX-catalyzed reaction of benzaldehyde (1a) with

trans- and cis-crotyltrichlorosilane (5b and c): The strong ki-

netic preference for trans-crotyltrichlorosilane (5b) ob-

served for the METHOX-catalyzed crotylation (see above),

which stands in contrast to the sluggish reaction of the cis

isomer (5c), raised the question to the origin of this behav-

ior. Therefore, the two reactions were examined by the

same theoretical method as that for the allylation but con-

fined to the dissociative mechanism (Table 6). These calcula-

METHOX-catalyzed reaction of 4-substituted benzalde-

hydes 1b and c with allyltrichlorosilane (5a): Allylations

catalyzed by QUINOX (10), which are likely to operate

through an associative mechanism, have been shown by us

to exhibit strong electronic dependence.^[30]Thus, while benz-

Table 6. Crotylation of benzaldehyde (1a) with (E)- and (Z)-crotyltri-

chlorosilane (5b/c) catalyzed by METHOX (9).^[a]

Entry

Crotylsilane

Product

TS

1

2

3

4

(E)-5b

(Z)-5c

1R

1S

1R

1S

19.3

18.2

22.7

19.8

ACHTUNGTRENNUNG

[a] All DG values are in kcalmol^ꢁ1

.

AHCTUNGTRENNUNG

ACHTUNGTRENNUNG

16% ee (Table 2, entry 25). Our previous calculation re-

vealed relatively minor but visible distortions in the TS for

these reactions, which resulted in a decrease of the differen-

ces in the TS barriers for the R and S channels from 2.0 for

1a to 0.8 kcalmol^ꢁ1for 1c.^[30]Since METHOX (9) did not

exhibit this type of dependence (affording the products in

93–96% ee, Table 2, entries 1–3), it was of interest to exam-

ine the reactivity theoretically.

tions show the following trend: 1) the activation barriers for

the crotylation are lower than those for the allylation, 2) the

trans isomer gives rise to a transition structure that is lower

in energy by 1.6 kcalmol^ꢁ1for the observed 1S product, pre-

dicting a faster reaction (by ca. 1–2 orders of magnitude),

which is in a qualitative agreement with the experiment, and

3) both isomers should preferentially give rise to the 1S-con-

figured product 7 or 8. In the case of the trans isomer, the

calculated difference of 1.1 kcalmol^ꢁ1(Table 6) should trans-

late to 83% ee at ꢁ408C, which is in a good agreement with

Assuming the dissociative mechanism, analogues of benz-

ACHTUNGTRENNUNGaldehyde (1a), namely the p-trifluoromethyl and p-methoxy

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

11

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

the experimentally observed 97% ee at ꢁ408C (Table 3,

entry 1). The corresponding transition states are shown in

the Supporting Information (Figure S3). The space-filling

representation of the transition states for the reaction of the

trans- and cis-crotyltrichlorosilanes (5b and c) with benzal-

dehyde are shown in Figure 5 (C and D), with the terminal

methyl group of the crotyl group highlighted. The transition

state C, giving rise to the formation of the anti-configured

solvation, entropic, and dispersion effects, with the latter

one being slightly more important.

Conclusion

The enantiopure, terpene-derived pyridine N-oxide organo-

catalyst METHOX ((+)-9), synthesized in three steps from

the inexpensive chiral pool (Scheme 3), has been shown to

give a very high level of stereocontrol in the allylation and

crotylation of aromatic and a,b-unsaturated aldehydes 1

product 7 (from 5b) is clearACTHNUTRGNEUGNly less cumbersome than D,

which was found for the formation of the syn product 8

(from 5c).

with allyltrichlorosilanes

5

(ꢀ97% ee and >99:1 d.r.,

The origin of stereoselectivity in the METHOX- and

QUINOX-catalyzed reactions: Let us assume that the reac-

tion proceeds through the mechanisms and transition states

described above, that is, through a dissociative pathway for

9 and an associative route for 10. Since many discussions

have been devoted to the origin of the stereoselectivity of

this class of reactions, mostly based on general argu-

Scheme 2 and Tables 1-3). Acetonitrile as the solvent, a tem-

perature of ꢁ408C, and a catalyst loading of 5 mol% have

been identified as the optimal conditions in terms of reac-

tion rate and selectivity; furthermore, a catalyst loading as

low as 1 mol% has been found to give satisfactory results.

Kinetic and computational investigations indicate that the

METHOX-catalyzed allylation proceeds through a dissocia-

tive pathway involving the cationic, pentacoordinate com-

plex 29 (Scheme 4). This finding stands in stark contrast to

the mechanism proposed by us^[30b]for the related N-oxide

QUINOX (10), which apparently favors an associative path-

ACHTUNGTRENNUNG

ments,^[31a,52]several relevant observations based on the

actual bonding situation in these reactions can be added. In

the preceding discussion, mainly the free energy values,

which are directly related to the experimental thermochemi-

cal parameters, were presented. However, they can be fur-

ther decomposed into the contributions originating from

zero-point energy corrections, entropy, solvation energies,

and dispersion energies of the catalyst to provide a better

understanding of the factors governing the observed stereo-

selectivity (Table 7).

In the case of METHOX, it can be seen that the domi-

nant term is given by the steric requirements, since the DE_gp

term is dominant, reflecting the steric bulk of the catalyst,

which also accounts for the high energy of the TS of the as-

sociative pathway compared to that of the uncatalyzed reac-

tion (see above). The dispersion energy alone favors the less

stable TS, that is, that the one leading to the R enantiomer.

Furthermore, Table 5 highlights a notable difference in the

relative contributions of METHOX and QUINOX to

Scheme 4. Mechanism of the allylation reaction catalyzed by chiral pyri-

dine-N-oxides.

way via the neutral, hexacoordinate complex 30. The differ-

ence can be attributed to the increased steric congestion in

the case of METHOX, which disfavors the sterically more

demanding hexacoordinate complex. Significantly, in both

pathways only one molecule of the catalyst is in-

DACHTUNGTRENNUNG

ACHTUNGTRENNUNG

Table 7. The decomposition of the differences in free energy barriers for METHOX

volved in the rate- and selectivity-determining

step, which contrasts with the behavior of Den-

markꢆs chiral phosphorACTHNUGRTNEUNGamides that have been

(9) and QUINOX (10) into the individual contributions.^[a]

[d]

[e]

[f]

[g]

Entry

Catalyst

Product

DE_gp

DG_solv

D

A

DE_disp

shown to operate through a transition state that

1

2

3

4

METHOX (9)^[b]

QUINOX (10)^[c]

R

S

R

S

0.0

ꢁ3.2

0.0

ꢁ0.3

0.0

ꢁ0.4

0.0

1.9

0.0

1.1

contains two catalyst molecules.^[18d,53]

ꢁ0.6

0.0

0.5

The second-sphere mechanism (i.e., one not in-

volving a direct coordination of the aldehyde to

the Si center) has been ruled out. The noncovalent

mechanism, proposed by Roithovꢂ and Kotora^[25]

for the reaction catalyzed by N,N’-dioxides in non-

polar media (which was not investigated in this

work), appears equally unlikely for METHOX op-

erating in CH₃CN.^[54]

0.7

[a] Contributions originating in molecular energies, zero-point energy corrections, en-

tropy, solvation energies, and dispersion energies. Only the energy differences between

isomers are listed here. All values are in kcalmol^ꢁ1and were calculated at the same

level of theory as the data reported in Tables 4–6 and as described in the computation-

al details. [b] Dissociative mechanism. [c] Associative mechanism. [d] DE_gpis the dif-

ference in the in vacuo energies between the (R) and (S) enantiomers. [e] DG_solvis the

difference in solvation free energies between the (R) and (S) enantiomers.

[f] D

(ꢁTDS)_gpis the difference in the in vacuo entropic terms between the (R) and (S)

The experimental ee values have been repro-

duced by calculations with great fidelity, along

enantiomers. [g] DE_dispis the difference in the dispersion energy stabilizations between

(R) and (S) enantiomers.

&

12

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

6.0 mg, 0.02 mmol, 0.05 equiv), N,N-diisopropylethylamine (106 mL,

0.6 mmol, 1.5 equiv),^[58]and the aromatic aldehyde (0.4 mmol) in CH₂Cl₂

(2 mL) under dry argon at ꢁ408C. The reaction mixture was stirred at

the same temperature for 24 h. The reaction was quenched with an aque-

ous solution of NaHCO₃(10%, 1 mL), and the mixture was diluted with

water and CH₂Cl₂(10 mL). The layers were separated, and the aqueous

layer was extracted with CH₂Cl₂(2ꢅ10 mL). The combined organic ex-

tracts were washed with saturated aqueous NaCl (20 mL), dried over

MgSO₄, and filtered. The solvent was removed in vacuo, and the residue

was purified by column chromatography on silica gel (1ꢅ20 cm) with a

petroleum ether/ethyl acetate mixture (20:1 to 10:1). The d.r. and ee

values are given in Table 3 for the crude products; this table also shows

the isolated yields. The enantiopurity of the resulting alcohols was deter-

mined by chiral GC, using a Supelco b-DEX 120 column or Supelco g-

DEX 120 columns, or by chiral HPLC, using a Chiralcel OD-H column.

The resulting homoallylic alcohols were identical with the authentic

samples.^[28-30,46]

with the catalytic efficiency of METHOX and QUINOX

(together with the inefficiency of the bare pyridine N-

oxide^[55]).^[56]

Experimental Section

General methods and materials: Optical rotations were recorded in

CHCl₃, unless otherwise indicated, with an error of ꢀ ꢄ0.1. The [a]_D

values are given in 10^ꢁ1degcm²g^ꢁ1. The NMR spectra were recorded in

CDCl₃, ¹HNMR at 400 and ¹³CNMR at 100.6 MHz with CDCl₃(¹H: d=

7.26 ppm; ¹³C: d=77.0 ppm) as internal standard, unless otherwise indi-

cated. Various 2D-techniques and DEPT experiments were used to estab-

lish the structures and to assign the signals. The IR spectra were recorded

in KBr disc, unless otherwise indicated. The mass spectra (EI and/or CI)

were measured on a dual sector mass spectrometer by using direct inlet

and the lowest temperature enabling evaporation. All reactions were per-

formed under an atmosphere of dry oxygen-free argon in oven-dried

glassware. Yields are given for isolated products, showing one spot on a

TLC plate and no impurities detectable in the NMR spectrum. The iden-

tity of the products prepared by different methods was checked by com-

parison of their NMR spectra. The enantiomeric excesses were deter-

mined by GC analysis using Supelco b-DEX 120 and/or Supelco g-CD

120 columns (oven: 1008C for 2 min, then 0.58Cmin^ꢁ1to 2008C, 10 min

at that temperature). The chiral GC methods were calibrated with the

corresponding racemates. The absolute configuration of the products was

determined by comparison of their optical rotations (measured in CHCl₃)

and their GC retention times with the literature data and with authentic

samples.^[28-30,46]All solvents and reagents for the reactions were of re-

agent grade and were dried and distilled under argon or nitrogen imme-

diately before use as follows: CH₃CN, CH₂Cl₂, triethylamine, and

Hꢇnigꢆs base from calcium hydride. Petroleum ether refers to the fraction

boiling in the range from 40–608C. Geometrically pure (E)- and (Z)-cro-

tyltrichlorosilanes were synthesized according to the literature meth-

ods.^[47,48]The deuterated aromatic aldehydes [D]-1a–c are known com-

pounds and were synthesized by the reduction of the corresponding ethyl

benzoates with LiAlD₄, followed by oxidation with pyridinium chloro-

chromate;^[57]the procedure for the KIE study is given in the Supporting

Information. (+)-a-Pinene was purchased from Aldrich; according to GC

analysis on a Supelco b-DEX 120 column, the ee was 90%. (ꢁ)-Pinocar-

vone (ꢁ)-21 was obtained on a 10 g scale from (+)-a-pinene through

photochemical oxidation (99%) according to the Mihelich procedure.^[39]

(+)-Isopinocampheol (+)-24, purchased from Aldrich, was of 96% ee

and was converted into enone 26 according to the literature protocol.^[42a]

General procedure for the preparation of the Krçhnke salts: The aryl

ketone (48.0 mmol) was heated in pyridine (20 mL) until a clear solution

was obtained. Crystalline iodine (17.4 g, 57.0 mmol) was added in por-

tions, and the resulting solution was refluxed for 1 h and then cooled to

room temperature. The brownish precipitate was filtered off and succes-

sively washed with absolute pyridine (3ꢅ15 mL) to give the Krçhnke salt

as a pale yellow solid, which could be crystallized from water or used di-

rectly in the Krçhnke annulation. In the course of preparation of the

3,4,5-trimethoxy salt 20g, partial loss of one methyl group was observed,

so that 1-[2-(3’,5’-dimethoxy-4’-hydroxyphenyl)-2-oxo-ethyl]-pyridinium

iodide was obtained, along with 20g.^[40]The mixture was used in the an-

nulation experiment, and the product mixture was methylated to afford

22g.

1-[2-(2’,6’-Dimethoxyphenyl)-2-oxo-ethyl]pyridinium iodide (20e): Yield:

1

50%; m.p. 201–2038C (water); H NMR (400 MHz, [D₆]DMSO): d=3.85

(s, 6H), 6.08 (s, 2H), 6.82 (d, J=8.4 Hz, 2H), 7.51 (t, J=8.4 Hz, 1H),

8.21 (t, J=6.7 Hz, 2H), 8.74 (t, J=7.8 Hz, 1H), 8.94 ppm (d, J=5.5 Hz,

2H); ¹³C NMR (100 MHz, [D₆]DMSO): d=56.65 (CH₃O), 69.4 (CH₂),

104.9 (CH), 114.6 (C), 128 4 (CH), 134.0 (CH), 146.4 (CH), 146.9 (CH),

158.0 (C), 194.0 ppm (CO); HRMS (FAB): m/z calcd for C₁₅H₁₆NO₃:

258.1130; found: 258.1130.

1-[2-(2’,4’,6’-Trimethoxyphenyl)-2-oxo-ethyl]pyridinium iodide (20 f):

Yield: 98%; m.p. 202–2068C (methanol/toluene); ¹H NMR (400 MHz,

[D₆]DMSO): d=3.85 (s, 6H), 3.88 (s, 3H), 6.01 (s, 2H), 6.37 (s, 2H), 8.24

(t, J=7.2 Hz, 2H), 8.69 (t, J=7.6 Hz, 1H), 8.74 ppm (t, J=5.8 Hz, 2H);

¹³C NMR (100 MHz, [D₆]DMSO): d=56.2 (CH₃O), 56.6 (CH₃O), 69.7

(CH₂), 91.6 (CH), 128.2 (CH), 146.4 (CH), 146.6 (CH), 160.8 (C), 164.8

(C), 190.7 ppm (CO); MS (FAB): m/z (%): 288.3 [MꢁI]⁺(69); HRMS

(FAB): m/z calcd for C₁₆H₁₈NO₄: 288.1236; found: 288.1233; in agree-

ment with the literature.^[63]

General procedure for the reaction of aldehydes 1 with allyltrichlorosi-

lanes 5a–h: Allylic trichlorosilane 5 (0.47 mmol) was added to a solution

of the catalyst (0.04 mmol), diisopropylethylamine (0.7 mmol),^[58]and al-

dehyde (0.4 mmol) in CH₃CN (2 mL), CH₂Cl₂(2 mL), or CHCl₃(2 mL)

under nitrogen at the specified temperature (Tables 1–3). The mixture

was stirred at the same temperature for the specified period of time

(Tables 1–3) and then quenched with aqueous saturated NaHCO₃

(1 mL). The aqueous layer was extracted with ethyl acetate (3ꢅ10 mL),

and the combined organic extracts were washed with brine and dried

over Na₂SO₄. The solvent was removed in vacuo, and the residue was pu-

rified by flash chromatography on a silica gel column (15 cmꢅ1 cm) with

a petroleum ether/ethyl acetate mixture (95:5). The d.r. and ee values are

given in Tables 1–3 for the crude products; these tables also show the iso-

lated yields. The enantiopurity of the resulting alcohols was determined

by chiral GC, using a Supelco b-DEX 120 column or Supelco g-DEX 120

columns, or by chiral HPLC, using a Chiralcel OD-H column. The result-

ing homoallylic alcohols were identical with the authentic sam-

ples,^{[28-30,46,59-62]}and their data are given in the Supporting Information.

Cases in which the opposite enantiomer was obtained are also listed

there.

1-[2-(3’,4’,5’-trimethoxyphenyl)-2-oxo-ethyl]pyridinium iodide (20g) and

1-[2-(3’,5’-dimethoxy-4’-hydroxyphenyl)-2-oxo-ethyl]pyridinium

iodide

(20j): Yield: 90% (obtained as a ꢂ1:1 mixture);^{[40] 1}H NMR (400 MHz,

[D₆]DMSO): showed, amongst others, d=3.81 (s, 3H) 3.88 (s, 6H),

3.90 ppm (s, 6H). This mixture was employed in the Krçhnke annulation

without separation.

1-{2-[3’,5’-Bis(trifluoromethyl)phenyl]-2-oxo-ethyl}pyridinium

iodide

(20h): Yield: 73%; m.p. 179–1818C (water); ¹H NMR (400 MHz,

[D₆]DMSO): d=6.60 (s, 2H), 8.32 (t, J=7.6 Hz, 2H), 8.61 (s, 1H), 8.86

(s, 2H), 8.76 (t, J=7.8 Hz, 1H), 8.96 ppm (d, J=5.5 Hz, 2H); ¹³C NMR

(100 MHz, [D₆]DMSO): d=66.8 (CH₂), 121.8 (C), 124.4 (C), 127.9 (CH),

128.4 (CH), 129.2 (CH), 131.3 (q, CF₃), 136.3 (C), 146.6 (CH), 147.1

(CH), 189.6 ppm (CO).

1-{2-[2’-Methoxynaphthalen-1’-yl]-2-oxo-ethyl}pyridinium iodide (20i):

Yield: 30%; ¹H NMR (400 MHz, [D₆]DMSO): d=4.15 (s, 3H), 6.30 (s,

2H), 7.48 (t, J=7.8 Hz, 1H), 7.58 (t, J=7.8 Hz, 1H), 7.68 (d, J=9.2 Hz,

1H), 7.99–8.02 (m, 2H), 8.26–8.33 (m, 3H), 8.79 (t, J=7.8 Hz, 1H),

9.11 ppm (d, J=5.5 Hz, 2H); ¹³C NMR (100 MHz, [D₆]DMSO): d=57.4

(CH₃), 70.1 (CH₂), 113.9 (CH), 118.3 (C), 124.2 (C), 124.9 (C), 128.2

(CH), 128.3 (CH), 128.8 (CH), 128.9 (C), 130.8 (C), 135.4 (CH), 146.7

(CH), 146.8 (CH), 157.7 (C), 189.6 ppm (CO).

General procedure for the asymmetric crotylation of aldehydes 1 with

(E)-crotyltrichlorosilane (5b): (E)-Crotyltrichlorosilane (E/Z 87:13;

92 mL, 0.60 mmol) was added dropwise to a solution of METHOX (9,

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

13

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

General procedure for the Krçhnke annulation by using (ꢁ)-pinocarvone

(ꢁ)-21 as Michael acceptor: Anhydrous ammonium acetate (110.0 g) was

heated in acetic acid (100 mL) at 1108C until it dissolved. Krçhnke salt

20 (47.0 mmol) was then added, and the mixture was left at 1108C until

the Krçhnke salt dissolved. Pinocarvone (ꢁ)-21 (43.0 mmol) was added,

and the solution was stirred at 1108C for 48 h. Water (20 mL) was then

added, and the mixture was extracted with ethyl acetate (3ꢅ50 mL). The

combined organic layers were washed successively with water (3ꢅ50 mL)

and brine (30 mL) and dried (MgSO₄). The solvent was removed in

vacuo to afford an oil that was purified by flash chromatography on silica

gel (25 g) using petroleum ether, followed by a petroleum ether/ethyl

acetate mixture (9:1) to give the pure product as a yellow oil or solid. In

the preparation of 22g, a mixture of 20g and the monodemethyl Krçhnke

salt (see above)^[40]was employed to produce the corresponding mixture

(d, J=2.8 Hz, 2H), 3.92 (s, 3H), 3.97 (s, 6H), 7.20 (s, 2H), 7.25 (d, J=

7.8 Hz, 1H), 7.35 ppm (d, J=7.8 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃):

d=21.7 (CH₃), 23.0 (CH₃), 32.4 (CH₂), 37.1 (CH₂), 39.9 (CH), 40.6 (C),

46.6 (CH), 56.6 (OCH₃), 61.3 (OCH₃), 104.4 (CH), 117.7 (CH), 134.0

(CH), 136.2 (C), 138.8 (C), 140.8 (C), 153.8 (C), 154.9 (C), 157.1 ppm

(C); MS (EI): m/z (%): 339 [M]⁺(100), 324 [MꢁMe]⁺(42); HRMS

(EI): m/z calcd for C₂₂H₂₇NO₃: 339.1834, found: 339.1834.

(+)-5-(3’,5’-Dimethoxy-4’-hydroxyphenyl)-10,10-dimethyl-6-aza-tricyclo-

AHCTUNGTERNNUNG ]AHCNUTGTREGuNNNU ndeca-2(7),3,5-triene) (+)-(22j): Yield: 54%; m.p. 98–1008C

[7.1.1.0^2,7

(toluene/ethyl acetate); [a]²_D⁰= +81.0 (c=1.0, CH₂Cl₂); ¹H NMR

(400 MHz, CDCl₃): d=0.71 (s, 3H), 1.32 (d, J=9.5 Hz, 1H), 1.44 (s, 3H),

2.42 (m, 1H), 2.72 (app dt, J=9.4, 5.9 Hz, 1H), 2.79 (app t, J=5.8 Hz,

1H), 3.20 (d, J=2.7 Hz, 2H), 4.00 (s, 6H), 7.24–7.28 (m, 3H), 7.35 ppm

(d, J=7.6 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃): d=21.7 (CH₃), 26.5

(CH₃), 32.4 (CH₂), 37.1 (CH₂), 39.9 (C), 40.7 (CH), 46.7 (CH), 56.8

(OCH₃), 104.1 (CH), 117.1 (CH), 132.0 (C), 134.0 (CH), 135.7 (C), 138.5

(C), 140.4 (C), 147.6 (C) 155.0 (C), 157.1 ppm (C); MS (EI): m/z (%):

325 [M]⁺(93), 310 [MꢁMe]⁺(34); HRMS (EI): m/z calcd for

C₂₁H₂₅NO₃: 339.1834, found: 339.1835.

of 22g and its monodemethyl analogue, which was subsequently meth

ated to obtain 22g.

[7.1.1.0^2,7

(+)-5-(4’-Methoxyphenyl)-10,10-dimethyl-6-aza-tricycloACHTNUGTRENNUGN ]AHCTUNGERTGuNNUN ndeca-

ACHTUNGERTNyNUNG l-

ACHTUNGTRENNUNG

2(7),3,5-triene (+)-(22c): Yield: 35%; m.p. 99–1018C (petroleum ether/

ethyl acetate); [a]²_D⁰= +82.5 (c=1.0, CHCl₃); ¹H NMR (400 MHz,

CDCl₃): d=0.61 (s, 3H), 1.34 (d, J=5.5 Hz, 1H), 1.44 (s, 3H), 2.42 (m,

1H), 2.71 (dt, J=9.5, 5.8 Hz, 1H), 2.79 (t, J=5.6 Hz, 1H), 3.19 (d, J=

2.6 Hz, 2H), 3.89 (s, 3H), 6.94 (d, J=8.9 Hz, 2H), 7.25 (d, J=7.7 Hz,

1H), 7.36 (d, J=7.7 Hz, 1H), 7.94 ppm (d, J=8.9 Hz, 2H); ¹³C NMR

(100 MHz, CDCl₃): d=21.7 (CH₃), 26.5 (CH₃), 32.5 (CH₂), 37.2 (CH₂),

40.7 (CH), 46.6 (CH), 56.7 (OCH₃), 113.9 (CH), 116.8 (CH),128.3 (CH),

133.1 (C), 133.9 (CH), 140.0 (C),154.9 (C), 157 (C), 160.2 ppm (C); MS

(EI): m/z (%): 279 [M]⁺(100), 264 (60), 250 (25), 236 (75), 224 (14), 193

(13), 83 (14), 49 (18); HRMS (EI): m/z calcd for C₁₉H₂₁NO: 279.1623;

found: 279.1623.

General procedure for the methylation of the pyridine derivatives 22a–g:

A solution of n-butyllithium (1.9m in hexane, 9.2 mL, 17.7 mmol) was

added dropwise to a solution of diisopropylamine (2.76 mL, 19.5 mmol)

in THF (30 mL) at ꢁ408C, the mixture was brought to 08C, stirred for

30 min, and cooled to ꢁ408C. A solution of the respective pyridine deriv-

ative 22a–g (4.0 g, 11.8 mmol) in THF (30 mL) was added at ꢁ408C,

turning the solution dark red. The mixture was stirred at that tempera-

ture for 2 h. Methyl iodide (1.12 mL, 17.7 mmol) was then added drop-

wise, and the mixture was stirred at room temperature overnight. Water

(30 mL) was added, and the mixture was extracted with CH₂Cl₂(3ꢅ

30 mL). The organic layers were combined, washed with brine (30 mL),

and dried (MgSO₄), and the solvent was removed under reduced pres-

sure. The crude product was purified by using flash chromatography on

silica gel (80 g) with a petroleum ether/ethyl acetate mixture (9:1) to give

the pure product as a white solid.

(+)-5-(2’,6’-Dimethoxyphenyl)-10,10-dimethyl-6-aza-tricycloACHTNUGRTNEUNG

[7.1.1.0^2,7]-

undeca-2(7),3,5-triene (+)-(22e): Yield: 50%; [a]²_D⁰= +47.1 (c=1.0,

1

CHCl₃); H NMR (400 MHz, CDCl₃): d=0.62 (s, 3H), 1.28 (d, J=9.2 Hz,

1H), 1.33 (s, 3H), 2.30 (m, 1H), 2.58 (dt, J=9.6, 5.7 Hz, 1H), 2.67 (t, J=

5.6 Hz, 1H), 3.09 (d, J=2.6 Hz, 2H), 3.63 (s, 6H), 6.52 (d, J=8.4 Hz,

2H), 6.88 (d, J=7.6 Hz, 1H), 7.16 ppm (m, 2H); ¹³C NMR (100 MHz,

CDCl₃): d=21.8 (CH₃), 26.5 (CH₃), 32.1 (CH₂), 36.7 (CH₂), 39.9 (C), 40.7

(CH), 46.7 (CH), 56.3 (OCH₃), 104.3(CH), 119.6 (C), 122.9 (CH), 129.7

(CH), 133.4 (CH), 139.9 (C), 151.4 (C), 156.3 (C), 158.6 ppm (C); MS

(EI): m/z (%): 309 [M]⁺(60), 294 (45), 266 (15), 83 (100), 47 (21);

HRMS (EI): m/z calcd for C₂₀H₂₃NO₂: 309.1729; found: 309.1729.

(+)-5-(2’,4’,6’-Trimethoxyphenyl)-8,10,10-trimethyl-6-aza-tricyclo-

AHCTUNGTERNNUNG ]AHCNUTGTREGuNNNU ndeca-2(7),3,5-triene (+)-(23 f): Yield: 79%; m.p. 110–1128C

[7.1.1.0^2,7

(toluene/ethyl acetate); [a]²_D⁰= +14.9 (c=1.0, CH₂Cl₂); ¹H NMR

(400 MHz, CDCl₃): d=0.60 (s, 3H), 1.28 (d, J=7.0 Hz, 2H), 1.30 (s, 3H),

1.32 (s, 3H), 2.05 (m, 1H), 2.45 (dt, J=9.8, 5.6 Hz, 1H), 2.79 (t, J=

5.8 Hz, 1H), 3.26 (dq, J=7.1, 2.8 Hz, 1H), 3.61 (s, 6H), 3.74 (s, 3H), 6.11

(s, 2H), 6.85 (d, J=7.2 Hz, 1H), 7.09 ppm (d, J=7.2 Hz, 1H); ¹³C NMR

(100 MHz, CDCl₃): d=19.0 (CH₃), 21.3 (CH₃), 26.7 (CH₃), 28.9 (CH₂),

39.0 (CH), 41.8 (C), 47.3 (CH), 47.4 (CH), 55.7 (OCH₃), 56.5 (OCH₃),

91.9 (CH), 113.7 (C), 123.2 (CH), 132.9 (CH), 138.5 (C), 139.2 (C), 151.4

(C), 159.5 (C), 160.2 (C), 161.4 ppm (C); MS (EI): m/z (%): 353 [M]⁺

(36), 338 [MꢁMe]⁺(100); HRMS (EI): m/z calcd for C₂₂H₂₇NO₃:

353.1991; found: 353.1991.

(+)-5-(2’,4’,6’-Trimethoxyphenyl)-10,10-dimethyl-6-aza-tricycloACHTNUGRTNEUNG

[7.1.1.0^2,7]-

undeca-2(7),3,5-triene (+)-(22 f): Yield: 95%; m.p. 98–1008C (toluene/

ethyl acetate); [a]_D²⁰= +49.5 (c=1.0, CH₂Cl₂); ¹H NMR (400 MHz,

CDCl₃): d=0.62 (s, 3H), 1.31 (d, J=9.4 Hz, 1H), 1.38 (s, 3H), 2.29 (m,

1H), 2.59 (dt, J=9.6, 5.6 Hz, 1H), 2.68 (t, J=5.6 Hz, 1H), 3.07 (d, J=

2.8 Hz, 2H), 3.61 (s, 6H), 3.72 (s, 3H), 6.13 (s, 2H), 7.13 (d, J=7.6 Hz,

1H), 7.20 ppm (d, J=7.6 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃): d=21.8

(CH₃), 26.5 (CH₃), 32.1 (CH₂), 37.0 (CH₂), 39.9 (C), 40.6 (CH), 46.7

(CH), 55.7 (OCH₃), 56.3 (OCH₃), 91.2 (CH), 113.3 (C), 123.1 (CH), 133.1

(CH), 139.5 (C), 151.6 (C), 156.3 (C), 159.3 (C) 161.3 ppm (C); MS (EI):

m/z (%): 339 [M]⁺(95), 324 [MꢁMe]⁺(53); HRMS (EI): m/z calcd for

C₂₁H₂₅NO₃: 339.1834; found: 339.1835.

(+)-5-(3’,4’,5’-Trimethoxyphenyl)-8,10,10-trimethyl-6-aza-tricyclo-

AHCTUNGTERNNUNG ]AHCNUTGTREGuNNNU ndeca-2(7),3,5-triene (+)-(23g): Yield: 38% (yellow foam);

[7.1.1.0^2,7

[a]²_D⁰= +17.5 (c=1.0, CH₂Cl₂); ¹H NMR (400 MHz, CDCl₃): d=0.56 (s,

3H), 1.24 (d, J=9.8 Hz, 1H), 1.31 (s, 3H), 1.34 (d, J=7.2 Hz, 3H), 2.05

(m, 1H), 2.45 (dt, J=9.8, 5.6 Hz, 1H), 2.66 (t, J=5.4 Hz, 1H), 3.26 (dq,

J=7.0, 2.3 Hz, 1H), 3.80 (s, 3H), 3.84 (s, 6H), 7.16–7.10 (m, 3H),

7.24 ppm (d, J=7.5 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃): d=19.0

(CH₃), 21.3 (CH₃), 23.0 (CH₃), 29.1 (CH₂), 39.3 (CH), 41.8 (C), 47.2

(CH), 56.5 (OCH₃), 61.3 (OCH₃), 104.2 (CH), 117.2 (CH), 133.7 (CH),

136.2 (C), 138.8 (C), 140.7 (C), 153.8 (C), 154.5 (C), 160.9 ppm (C); MS

(EI): m/z (%): 353 [M]⁺(90), 338 [MꢁMe]⁺(100); HRMS (EI): m/z

calcd for C₂₂H₂₇NO₃: 353.1991; found: 353.1992.

(+)-5-(3’,4’,5’-Trimethoxyphenyl)-10,10-dimethyl-6-aza-tricycloACHTNUGRTNEUNG

[7.1.1.0^2,7]-

undeca-2(7),3,5-triene (+)-(22g): Potassium carbonate (23.49 g, 0.16 mol)

was added to a solution of (+)-5-(3’,5’-dimethoxy-4’-hydroxyphenyl)-

10,10-dimethyl-6-aza-tricyclo[7.1.1.0^2,7]undeca-2(7),3,5-triene

((+)-22j,

6.92 g, 21 mmol) in acetone (200 mL) and the mixture was stirred for

30 min. Dimethyl sulfate (2.18 mL, 22 mmol) was added, and the mixture

left to stir at room temperature for 1.5 h. Water (100 mL) was then

added, acetone was removed by using a rotary evaporator, and the prod-

uct was extracted with CH₂Cl₂(3ꢅ30 mL). The organic layers were com-

bined, washed with brine (30 mL), and dried (MgSO₄), and the solvent

was removed under reduced pressure. The crude product was purified by

using flash chromatography on silica gel (80 g) with a petroleum ether/

ethyl acetate mixture (1:1) to give pure (+)-22g as a yellow foam. Yield:

91% (6.46 g); [a]²_D⁰= +72.7 (c=1.0, CH₂Cl₂); ¹H NMR (400 MHz,

CDCl₃): d=0.69 (s, 3H), 1.31 (d, J=9.5 Hz, 1H), 1.42 (s, 3H), 2.40 (m,

1H), 2.70 (app dt, J=9.6, 5.8 Hz, 1H), 2.79 (app t, J=5.6 Hz, 1H), 3.18

General procedure for the Krçhnke annulation by using the Michael

AHCTUNGTRENNUNG

acceptor 26:^[42a]The procedure was carried out in the same way as in the

case of (ꢁ)-pinocarvone ((ꢁ)-18, see above) by employing Krçhnke salts

20h and 20i.

(+)-5-[3’,5’-Bis(trifluoromethyl)phenyl]-8,10,10-trimethyl-6-aza-tricyclo-

AHCTUNGTRENNUNG ]ACHTUNGTRENNUNG

[7.1.1.0^2,7undeca-2(7),3,5-triene (+)-(23h): Yield 35%; [a]²_D⁰= +60.2 (c=

1.0, CHCl₃); ¹H NMR (400 MHz, CDCl₃): d=0.61 (s, 3H), 1.28 (s, 3H),

1.31 (d, J=6.3 Hz, 1H), 2.12 (m, 1H), 2.54 (dt, J=10.2, 5.7 Hz, 2H), 2.81

(t, J=6.3 Hz, 1H), 3.25 (qd, J=6.6, 2.8 Hz, 1H), 7.25 (d, J=7.7 Hz, 1H),

&

14

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

7.41 (d, J=7.5 Hz, 1H), 7.78 (s, 1H), 8.39 ppm (s, 2H); ¹³C NMR

(100 MHz, CDCl₃): d=18.6 (CH₃), 21.3 (CH₃), 26.6 (CH₃), 28.9 (CH₂),

39.2 (CH), 41.8 (C), 47.1 (CH), 47.4 (CH), 122.0 (C), 122.5 (CH), 127.0

(CH), 130.2 (CH), 132.1 (q, CF₃), 142.2 (C), 142.6 (C), 151.5 ppm (C);

MS (EI): m/z (%): 399 [M]⁺(2), 356 (12), 85 (100), 83 (100), 47 (100);

HRMS (EI): m/z calcd for C₂₁H₁₉F₆N: 399.1422; found: 399.1422.

ꢁ49.4 (c=1.0, CHCl₃); ¹H NMR (400 MHz, CDCl₃): d=0.61 (s, 3H),

1.37 (s, 3H), 1.40 (s, 1H), 1.42 (d, J=8.4 Hz, 3H), 2.12 (m, 1H), 2.54 (dt,

J=10.2, 5.7 Hz, 1H), 2.77 (t, J=5.6 Hz, 1H), 3.38 (qd, J=6.6, 2.8 Hz,

1H), 6.88 (d, J=7.8 Hz, 1H), 7.41 (d, J=7.8 Hz, 1H), 7.84 (s, 1H),

8.22 ppm (s, 2H); ¹³C NMR (100 MHz, CDCl₃): d=14.9 (CH₃), 21.0

(CH₃), 26.1 (CH₃), 28.6 (CH₂), 35.4 (CH), 41.9 (C), 47.4 (CH), 47.7 (CH),

122.2 (C), 122.9 (CH), 123 (CH), 123.4 (CH), 130.3 (C), 132.2 (q, CF₃),

135.8 (C), 144.9 (C), 146.5 (C), 151.1 ppm (C); MS (EI): m/z (%): 415

[M]⁺(4), 398 (10), 356 (12), 83 (100), 47 (15); HRMS (EI): m/z calcd for

C₂₁H₁₉F₆NO: 415.1371; found: 415.1371.

(+)-5-(2’-Methoxy-1’-naphthyl)-8,10,10-trimethyl-6-aza-tricycloACHTNUGRTNEUNG

[7.1.1.0^2,7]-

undeca-2(7),3,5-triene (+)-(23i): Yield: 20%; [a]²_D⁰= +100.9 (c=1.0,

1

CHCl₃); H NMR (400 MHz, CDCl₃): d=0.67 (s, 3H), 1.34 (d, J=7.0 Hz,

3H), 1.38 (s, 3H), 1.40 (d, J=5.8 Hz, 1H), 2.10–2.13 (m, 1H), 2.53 (dt,

J=9.8, 5.7 Hz, 1H), 2.76 (t, J=5.7 Hz, 1H), 3.21 (m, 1H), 3.75 (s, 3H),

7.04 (d, J=7.5 Hz, 1H), 7.16–7.30 (m, 4H), 7.47 (d, J=9.1 Hz, 1H), 7.69–

7.73 (m, 1H), 7.79 ppm (d, J=9.1 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃):

d=18.9 (CH₃), 21.4 (CH₃), 26.8 (CH₃), 29.0 (CH₂), 39.2 (CH), 41.9 (C),

47.3 (CH), 47.5 (CH), 57.6 (OCH₃), 114.9 (CH), 123.5 (CH), 123.9 (CH),

125.5 (C), 125.6 (CH), 126.8 (CH), 128.2 (CH), 129.8 (C), 130.1 (CH),

133.2 (CH), 134.1 (C), 139.2 (C), 153.1 (C), 154.8 (C), 160.8 ppm (C); MS

(EI): m/z (%): 343 [M]⁺(18), 328 [MꢁOMe]⁺(35); HRMS (EI): m/z

calcd for C₂₄H₂₅NO: 343.1936; found: 343.1936.

(ꢁ)-5-(2’-Methoxy-1’-naphthyl)-8,10,10-trimethyl-6-aza-tricyclo

[7.1.1.0^2,7]-

ACHTUNGTRENNUNG

undeca-2(7),3,5-triene 6-oxide (ꢁ)-(18a): The mixture resulting from the

N-oxidation of 23g (95 mg) was chromatographed on a silica gel column

(10ꢅ1 cm) with a petroleum ether/ethyl acetate mixture (2:1) to afford

(ꢁ)-18a as the less polar amorphous product. Yield: 34%, (35 mg);

[a]²_D⁰=ꢁ163 (c=1.0, CHCl₃); ¹H NMR (400 MHz, CDCl₃): d=0.68 (s,

3H), 1.39 (s, 3H), 1.43 (d, J=6.4 Hz, 3H), 1.49 (d, J=9.8 Hz, 1H), 2.12

(m, 1H), 2.52 (dt, J=9.7, 5.6 Hz, 1H), 2.78 (t, J=5.7 Hz, 1H), 3.39 (dq,

J=5.6, 2.3 Hz, 1H), 3.81 (s, 3H), 6.86 (d, J=7.6 Hz, 1H), 7.06, (d, J=

7.6 Hz, 1H), 7.12 (d, J=8.4 Hz, 1H), 7.26–7.32 (m, 3H), 7.73 (d, J=

7.8 Hz, 1H), 7.85 ppm (d, J=9.2 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃):

d=15.1 (CH₃), 21.0 (CH₃), 26.3 (CH₃), 28.6 (CH₂), 35.3 (CH), 42.0 (C),

47.3 (CH), 47.9 (CH), 56.9 (OCH₃), 114.0 (CH), 122.6 (CH), 124.0 (CH),

124.6 (CH), 126.5 (CH), 127.4 (CH), 128.4 (CH), 129.2 (C), 131.2 (CH),

133.3 (C), 144.8 (C), 150.6 (C), 155.1 ppm (C); MS (EI): m/z (%): 359

[M]⁺(8), 328 [M-OMe]⁺(100); HRMS (EI): m/z calcd for C₂₄H₂₅NO₂:

359.1884; found: 359.1885.

N-Oxidation of the pyridine derivatives 23a–i: m-Chloroperoxybenzoic

acid (MCPBA; 70%, 1.86 g, 9.38 mmol) was added in portions to a solu-

tion of the respective pyridine derivative (8.52 mmol) in CH₂Cl₂(60 mL)

at 08C, and the mixture was stirred at room temperature for 2 h. The re-

action mixture was then diluted with diethyl ether (60 mL) and washed

successively with saturated NaHCO₃(3ꢅ30 mL) and brine (30 mL). The

ethereal solution was dried over Na₂SO₄, and the solvent was evaporated

in vacuo. The resulting crude product was purified by column chroma-

(ꢁ)-5-(2’-Methoxy-1’-naphthyl)-8,10,10-trimethyl-6-aza-tricyclo

[7.1.1.0^2,7]-

ACHTUNGTRENNUNG

ACHTUNGTRENNUNGtography on silica gel (10ꢅ2.5 cm) with a petroleum ether/ethyl acetate

mixture (10:1) to elute the unreacted starting material, followed by ethyl

acetate, to obtain the N-oxide as a white solid.

undeca-2(7),3,5-triene 6-oxide (ꢁ)-(18b): Continued elution with a petro-

leum ether/ethyl acetate mixture (2:1), after isolation of (ꢁ)-18a, afford-

ed (ꢁ)-18b^[44]as the more polar amorphous product. Yield: 27%

(28 mg,); [a]_D²⁰=ꢁ21.5 (c=1.0, CHCl₃);^{[44] 1}H NMR (400 MHz, CDCl₃):

d=0.65 (s, 3H), 1.39 (s, 3H), 1.43 (d, J=6.4 Hz, 3H), 1.49 (d, J=9.8 Hz,

1H), 2.11 (m, 1H), 2.54 (dt, J=9.7, 5.6 Hz, 1H), 2.78 (t, J=5.7 Hz, 1H),

3.39 (dq, J=5.6, 2.3 Hz, 1H), 3.80 (s, 3H), 6.85 (d, J=7.6 Hz, 1H), 7.04

(d, J=7.6 Hz, 1H), 7.23–7.32 (m, 4H), 7.74 (d, J=7.8 Hz, 1H), 7.88 ppm

(d, J=9.2 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃): d=14.5 (CH₃), 21.1

(CH₃), 26.3 (CH₃), 28.6 (CH₂), 35.3 (CH), 42.0 (C), 47.4 (CH), 47.9 (CH),

57.1 (OCH₃), 113.7 (CH), 122.4 (CH), 124.0 (CH), 124.6 (CH), 126.5

(CH), 127.4 (CH), 128.6 (CH), 129.2 (C), 131.2 (CH), 133.3 (C), 144.8

(C), 150.6 (C), 155.1 ppm (C); MS (EI): m/z (%): 359 [M]⁺(8), 328

[MꢁOMe]⁺(100); HRMS (EI): m/z calcd for C₂₄H₂₅NO₂: 359.1884;

found: 359.1885.

(+)-5-(2’,4’,6’-Trimethoxyphenyl)-8,10,10-trimethyl-6-aza-tricyclo-

ACHTUNGTRENNUNG ]ACHTUNGTRENNUNGundeca-2,4,6-triene 6-oxide (+)-(9): When the N-oxidation of

[7.1.1.0^2,7

23 f was complete, the reaction mixture was diluted with diethyl ether,

and the resulting solution was washed with aqueous NaOH (5ꢅ30 mL)

to remove both the m-chlorobenzoic acid and the phenolic by-prod-

ucts,^[64]and with brine (1ꢅ30 mL). The ethereal solution was dried over

Na₂SO₄, and the solvent was evaporated in vacuo. The product obtained

by chromatography (see the general procedure in the previous para-

graph) was crystallized from an ethyl acetate/n-hexane mixture to afford

the highly pure compound (+)-9. Yield: 67% (1.08 g); m.p. 135–1368C;

[a]²_D⁰= +8.6 (c=1.0, CH₂Cl₂);^{[29a] 1}H NMR (400 MHz, CDCl₃): d=0.67 (s,

3H), 1.41 (m, 4H), 2.1 (m, 1H), 2.48 (dt, J=10.2, 6.6 Hz, 1H), 2.70 (t,

J=6.6 Hz, 1H), 3.34 (dq, J=6.6 Hz, 2.8 Hz, 1H), 3.76 (d, J=5.4 Hz,

6H), 6.55 (t, J=7.6 Hz, 2H), 6.8 (d, J=7.8 Hz, 1H), 6.9 (d, J=7.7 Hz,

1H), 7.25 ppm (t, J=8.4 Hz, 1H); ¹³C NMR (100 MHz, CDCl₃): d=15.1

(CH₃), 21.0 (CH₃), 26.3 (CH₃), 28.5 (CH₂), 35.2 (CH), 42.0 (C), 47.3

(CH), 47.9 (CH), 55.8 (OCH₃), 56.3 (OCH₃), 56.5 (OCH₃), 91.2 (CH),

91.3 (CH), 105.4 (C), 122.3 (CH), 126.3 (CH), 143.7(C), 144.0 (C), 149.9

(C), 159.2 (C), 159.9 (C), 162.5 ppm (C); MS (EI): m/z (%): 369 [M]⁺

(10), 338 [MꢁOMe]⁺(100); HRMS (EI): m/z calcd for C₂₂H₂₇NO₄:

369.1940; found: 369.1940. Crystals for X-ray analysis were obtained by

vapor diffusion of petroleum ether into a solution of METHOX in ethyl

acetate.

Computational details: Density functional theory (DFT) calculations

were carried out by using the program Turbomole 6.1.^[65]The Perdew–

Burke–Ernzerhof (PBE) functional^[66]has been used throughout. The cal-

culations were expedited by expanding the Coulomb integrals in an aux-

AHCTUNGTRENNUNG

iliary basis set, the resolution-of-identity (RI-J) approximation.^[67]All

geomACHTNUTRGNEUNGetry optimizations were carried out by using the 6–31G(d) basis

set,^[68]whereas the single point energies were recomputed by using the

def2-TZVP basis set.^[69]To account for the description of dispersion

forces (which are not described by the standard DFT functionals), the

empirical dispersion parameters were added to both the energy and gra-

dients during geometry optimization and the calculations of the single

point energies (DFT(+D) method).^[70]Here, default parameters^[70b]were

used, except for the C₆parameters for Si, in which the values for P were

used, and Cl, in which we used the values of Neumann and Perrin.^[71]For

the non-catalyzed model reaction of allyltrichlorosilane with benzalde-

hyde, DFT+D has been shown to yield the activation energies within

ꢂ1 kcalmol^ꢁ1accuracy in comparison with the reference CCSD(T)

values (as shown in the Results and Discussion).

(+)-5-(3’,4’,5’-Trimethoxyphenyl)-8,10,10-trimethyl-6-aza-tricyclo-

ACHTUNGTRENNUNG ]ACHTUNGTRENNUNGundeca-2,4,6-triene 6-oxide (+)-(16): Yield: 44% (pink foam);

[7.1.1.0^2,7

[a]²_D⁰= +1.7 (c=1.0, CH₂Cl₂); ¹H NMR (400 MHz, CDCl₃): d=0.62 (s,

3H), 1.36 (s, 3H), 1.38 (d, J=11.5 Hz, 1H), 1.53 (d, J=7.0 Hz, 3H), 2.1

(m, 1H), 2.50 (app dt, J=10.0, 5.7 Hz, 1H), 2.73 (app t, J=5.6 Hz, 1H),

3.39 (app dq, J=6.6 Hz, 2.8 Hz, 1H), 3.81 (s, 3H), 3.83 (s, 6H), 6.83 (d,

J=7.5 Hz, 1H), 6.95 (s, 2H), 7.10 ppm (d, J=7.6 Hz, 1H); ¹³C NMR

(100 MHz, CDCl₃): d=15.1 (CH₃), 21.0 (CH₃), 26.3 (CH₃), 28.6 (CH₂),

35.4 (CH), 42.0 (C), 47.3 (CH), 47.8 (CH), 56.7 (OCH₃), 61.2 (OCH₃),

107.6 (CH), 123.4 (CH), 124.4 (CH), 129.3 (C), 139.1 (C), 144.6 (C),

147.5 (C), 150.5 (C), 153.3 ppm (C); MS (EI): m/z (%): 369 [M]⁺(48),

354 [MꢁMe]⁺, 56), 338 [MꢁOMe]⁺(61); HRMS (EI): m/z calcd for

C₂₂H₂₇NO₄: 369.1940; found: 369.1942.

To account for solvation effects, the conductor-like screening model

(COSMO)^[72]was employed with the dielectric constant corresponding to

CH₃CN (e_r=36.6). Gibbs free energy was then calculated as the sum of

these contributions [Eq. (5)]:

(ꢁ)-5-[3’,5’-Bis(trifluoromethyl)phenyl]-8,10,10-trimethyl-6-aza-tricyclo-

[7.1.1.0^2,7

]

G

=

G ¼ E_elðþDÞþG_solvþ E_ZPEꢁRT lnðq_transq_rotq_vib

Þ

ð5Þ

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

15

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

in which DE_el(+D)is the in vacuo energy of the system (at the RI-

PBE(+D)/def2-TZVP level, with the geometry optimized at the RI-

PBE(+D)/6–31G(d) level; see above), DG_solvis the solvation free energy

S. A. Snyder, Classics in Total Synthesis II, Wiley-VCH, Weinheim,

2003; d) L. Kꢇrti, B. Czakꢁ, Strategic Applications of Named Reac-

tions in Organic Synthesis, Elsevier, Amsterdam, 2005; e) T. Hud-

(at the RI-PBE(+D)/6–31G(d) level) and the E_ZPEꢁRT ln(q_transq_rotq_vib

)

´

licky, J. W. Reed, The Way of Synthesis, Wiley-VCH, Weinheim,

term is the zero-point energy, thermal corrections to the Gibbs free

energy, and entropic term (obtained from a frequency calculation with

the same method and software as for the geometry optimizations at the

RI-PBE/6–31G(d) level, 298 K and 1 atm pressure, by using an ideal-gas

approximation).^[73]

ˇ

´

ˇ

2007; f) P. Kocovsky, F. Turecek, J. Hꢂjꢀcek, Synthesis of Natural

Products: Problems of Stereoselectivity, Vol. 1 and 2, CRC Press,

Boca Raton, 1986.

[3] a) W. R. Roush, in Comprehensive Organic Synthesis (Eds.: B. M.

Trost, I. Fleming, C. H. Heathcock), Pergamon, Vol. 2, Oxford, 1991,

p. 1; b) D. G. Hall, H. Lachance, Org. React. 2008, 73, 1; for a selec-

tion of the initial seminal papers, see: c) R. W. Hoffmann, T.

Herold, Chem. Ber. 1981 , 114 , 375; d) T. Herold, U. Schrott, R. W.

Hoffmann, G. Schnelle, W. Ladner, K. Steinbach, Chem. Ber. 1981 ,

114 , 359; e) W. R. Roush, A. E. Walts, L. K. Hoong, J. Am. Chem.

Soc. 1985 , 107 , 8186; f) W. R. Roush, R. L. Halterman, J. Am.

Chem. Soc. 1986 , 108 , 294; g) W. R. Roush, M. A. Adam, A. E.

Walts, D. J. Harris, J. Am. Chem. Soc. 1986 , 108 , 3422; h) W. R.

Roush, L. K. Hoong, M. A. J. Palmer, J. C. Park, J. Org. Chem. 1990 ,

55 , 4109; i) W. R. Roush, L. K. Hoong, M. A. J. Palmer, J. A. Straub,

A. D. Palkowitz, J. Org. Chem. 1990 , 55 , 4117; j) B. W. Gung, X. W.

Xue, W. R. Roush, J. Am. Chem. Soc. 2002 , 124 , 10692 (correction:

B. W. Gung, X. W. Xue, W. R. Roush, J. Am. Chem. Soc. 2003 , 125 ,

3668 ); for a selection of recent studies, see: k) M. Chen, W. R.

Roush, Org. Lett. 2010 , 12 , 2706; l) S. M. Winbush, W. R. Roush,

Org. Lett. 2010 , 12 , 4344; m) M. Chen, D. H. Ess, W. R. Roush, J.

Am. Chem. Soc. 2010 , 132 , 7881.

[4] a) H. Sakurai, Pure Appl. Chem. 1982 , 54 , 1; b) A. Hosomi, Acc.

Chem. Res. 1988 , 21 , 200; c) I. Fleming, J. Dunoguꢈs, R. Smithers,

Org. React. 1989, 37, 57; d) I. Fleming, A. Barbero, D. Walter,

Chem. Rev. 1997 , 97 , 2063.

[5] a) W. R. Roush, M. S. VanNieuwenhze, J. Am. Chem. Soc. 1994 , 116 ,

8536; b) B. W. Gung, Org. React. 2004, 64, 1.

[6] a) A. Fꢇrstner, Chem. Rev. 1999 , 99 , 991; b) G. C. Hargaden, P. I.

Guiry, Adv. Synth. Catal. 2007 , 349 , 2407.

[7] R. O. Duthaler, A. Hafner, Chem. Rev. 1992 , 92 , 807.

By using this definition, all values calculated in this work refer to a stand-

ard state corresponding to the 1 atm concentration of the species in-

volved. A correction of (1.9 Dn) kcalmol^ꢁ1(corresponding to the differ-

ence between the concentration of the ideal gas at 298 K and 1 atm and

its 1 molL^ꢁ1concentration; Dn is the change in number of moles in the

reaction) needs to be applied to refer to 1 molL^ꢁ1standard state (not

done in this work to be able to compare the results with our previous

work).

Transition state (TS) optimizations were performed as follows: first, the

TS initial geometry was obtained by the restricted geometry optimization

ꢁ

on the interpolated reaction coordinate, defined by the Si(1) C(2) and

ꢁ

C(4) C(5) distances, which were assumed to attain the discrete values

between the reactant and product with all the remaining internal coordi-

nates optimized. The maximum on this approximate reaction coordinate

was then taken as the initial guess for the saddle-point optimization

along the eigenvector corresponding to the vibration with the imaginary

frequency. The approximate method was used to obtain the TSs in solu-

tion: 1) the in vacuo optimization of transition state according to the

ꢁ

above procedure; 2) constraining the Si(1) C(2) and C(4) C(5) distances

(corresponding to the distance highlighted in Figure 4) to the values cor-

responding to the TS acquired in step 1 and optimizing all other degrees

of freedom by the DFT(+D)/COSMO method. The difference between

the DFT(+D)/COSMO energy at the in vacuo TS geometry and the re-

laxed geometry is denoted as the solvent relaxation term and is regarded

to represent an accurate approximation to the true transition state in sol-

ution, calculated with the DFT(+D) method.

Finally, for benchmarking purposes on the non-catalyzed model reaction,

we have used the CCSD(T) method^[74]and the aug-cc-pVDZ basis set.^[75]

These calculations were carried out by using the MOLPRO program.^[76]

[8] a) J. Prunet, 2011, personal communication; for details, see: L. Gri-

maud, Thesis, Ecole Polytechnique, Palaiseau (France), 1999;

a

promising recent protocol, using the pinacol-derived allylboronate

and a chiral Brønsted acid as catalyst, may serve as a remedy: b) P.

Jain, J. C. Antilla, J. Am. Chem. Soc. 2010 , 132 , 11884; for a recent

analogous allylation of aminals (as equivalents of imines), see: c) Y.-

Y. Huang, A. Chakrabarti, N. Morita, U. Schneider, S. Kobayashi,

Angew. Chem. 2011 , 123 , 11317; Angew. Chem. Int. Ed. 2011 , 50 ,

11121; for a chiral Brønsted acid catalyzed allylation with allyltri-

chlorosilane, see: d) K. Cheng, T. Fan, J. Sun, Chinese J. Chem.

2011 , 29 , 1669.

Acknowledgements

We thank the EPSRC for grant No. GR/T27051/01, the Ministry of Edu-

cation and Science of Spain for a fellowship to P.R.-L., the University of

Glasgow for a studentship to M.B., the University of Rome for a fellow-

[9] a) S. E. Denmark, N. G. Almstead, J. Org. Chem. 1994 , 59 , 5130; for

an open transition state operating in the case of the analogous allyl-

stannanes, see: b) S. E. Denmark, S. Hosoi, J. Org. Chem. 1994 , 59 ,

5133; for a comprehensive experimental and computational study of

the origin of syn/anti diastereoselectivity in the aldehyde and ketone

crotylation, see: c) L. F. Tietze, T. Kinzel, S. Schmatz, J. Am. Chem.

Soc. 2006 , 128 , 11483.

ˇ

ship to F.C., the Ministry of Education of the Czech Republic (MSMT

ˇ

CR) for grant No. LC512, the European Socrates-Erasmus exchange pro-

gram for a scholarship to L.B., the European Social Fund under the

ˇ

Global Grant measure (VP1-3.1-SMM-07-K-01-030), the Wenner–Gren

Foundation for a sabbatical scholarship to PK, and Dr. Alfred Bader for

additional support. L.B. was an exchange graduate student from the Insti-

ˇ

tute of Organic Chemistry and Biochemistry, AVCR. COST Office sup-

[10] a) T. K. Hollis, B. Bosnich, J. Am. Chem. Soc. 1995 , 117 , 4570; for a

recent contribution, see: b) M. Mahlau, P. Garcꢀa-Garcꢀa, B. List,

Chem. Eur. J. 2012 , 18 , 16283.

port via ORCA action is also appreciated.

ˇ

´

[11] P. Kocovsky, A. V. Malkov, in Enantioselective Organocatalysis (Ed.:

[1] For leading reviews, see: a) R. W. Hoffmann, Angew. Chem. 1987 ,

99 , 503; Angew. Chem. Int. Ed. Engl. 1987 , 26 , 489; b) Y. Yamamoto,

N. Asao, Chem. Rev. 1993 , 93 , 2207; c) S. E. Denmark, R. A. Sta-

venger, Acc. Chem. Res. 2000 , 33 , 432; d) S. E. Denmark, J. Fu,

Chem. Commun. 2003 , 167; e) S. E. Denmark, J. Fu, Chem. Rev.

2003 , 103 , 2763; f) J. W. J. Kennedy, D. G. Hall, Angew. Chem. 2003 ,

115 , 4880; Angew. Chem. Int. Ed. 2003 , 42 , 4732; g) S. E. Denmark,

N. G. Almstead, J. Mex. Chem. Soc. 2009, 53, 174; h) M. Yus, J. C.

Gonzꢂlez-Gꢁmez, F. Foubelo, Chem. Rev. 2011 , 111 , 7774.

P. I. Dalko), Wiley-VCH, Weinheim, 2007, p. 255.

[12] For reviews on Lewis basic catalysis, including discussions of hyper-

valent silicon, see: a) S. Rendler, M. Oestreich, Synthesis 2005, 1727;

b) Y. Orito, M. Nakajima, Synthesis 2006, 1391; c) A. V. Malkov, P.

ˇ

´

ˇ

´

Kocovsky, Eur. J. Org. Chem. 2007 , 29; d) P. Kocovsky, A. V.

Malkov, Pure Appl. Chem. 2008, 80, 953; e) M. Benaglia, S. Guizzet-

ti, L. Pignataro, Coord. Chem. Rev. 2008 , 252 , 492; f) S. E. Denmark,

G. L. Beutner, Angew. Chem. 2008 , 120 , 1584; Angew. Chem. Int.

Ed. 2008 , 47 , 1560.

[2] a) E. J. Corey, X.-M. Cheng, The Logic of Chemical Synthesis,

Wiley, New York, 1989; b) K. C. Nicolaou, E. J. Sorensen, Classics

in Total Synthesis, Wiley-VCH, Weinheim, 1996; c) K. C. Nicolaou,

[13] a) S. Kobayashi, K. Nishio, Tetrahedron Lett. 1993 , 34 , 3453; b) S.

Kobayashi, K. Nishio, J. Org. Chem. 1994 , 59 , 6620; c) S. Kobayashi,

&

16

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

K. Nishio, Synthesis 1994 , 457; d) J. D. Short, S. Attenoux, D. J. Ber-

risford, Tetrahedron Lett. 1997 , 38 , 2351.

[14] S. E. Denmark, D. M. Coe, N. E. Pratt, B. D. Griedel, J. Org. Chem.

1994 , 59 , 6161.

[15] Z. Wang, G. Xu, D. Wang, M. E. Pierce, P. N. Confalone, Tetrahe-

dron Lett. 2000 , 41 , 4523.

[16] For other substances possessing a strongly Lewis basic oxygen, such

as various formamides, urea derivatives, and catecholates, that have

been shown to catalyze or mediate the allylation, see ref. [13] and

the following: a) K. Iseki, S. Mizuno, Y. Kuroki, Y. Kobayashi, Tetra-

hedron Lett. 1998 , 39 , 2767; b) K. Iseki, S. Mizuno, Y. Kuroki, Y. Ko-

bayashi, Tetrahedron 1999 , 55 , 977; c) C. Ogawa, M. Sugiura, S. Ko-

bayashi, Chem. Commun. 2003 , 192; d) I. Chataigner, U. Piarulli, C.

Piarulli, T e trahedron Lett. 1999 , 40 , 3633; e) M. Kira, K. Sato, H. Sa-

kurai, J. Am. Chem. Soc. 1988 , 110 , 4599; f) M. Kira, L. C. Zhang, C.

Kabuto, H. Sakurai, Organometallics 1998 , 17 , 887; g) H. Sakurai,

Synlett 1989, 1.

[27] For the preparation of selected bipyridine N,N’-dioxides, see:

a) S. E. Denmark, F. Yu, Tetrahedron: Asymmetry 2006 , 17 , 687; for

the synthesis of bipyridine N,N’-dioxides with planar chirality, see:

b) J. R. Fulton, J. E. Glover, L. Kamara, G. J. Rowlands, Chem.

Commun. 2011 , 47 , 433; for recent examples of allylation catalyzed

by pyridine-type N-oxides, see: c) D. R. Boyd, N. D. Sharma, L. Sbir-

cea, D. Murphy, J. F. Malone, S. L. James, C. C. R. Allen, J. T. G.

Hamilton, Org. Biomol. Chem. 2010 , 8 , 1081; d) T. Naicker, P. I. Ar-

vidsson, H. G. Kruger, G. E. M. Maguire, T. Govender, Eur. J. Org.

Chem. 2011 , 6923; e) G. Chelucci, G. Murineddu, G. A. Pinna, Tetra-

hedron: Asymmetry 2004 , 15 , 1373; f) B. Bai, L. Shen, J. Ren, H. J.

Zhu, Adv. Synth. Catal. 2012 , 354 , 354; for an analogous use of bis-

AHCTUNGTREGiNNUN midazole N-oxides in the allylation, see: g) P. Kwiatkowski, P.

´

Mucha, G. Mloston, J. Jurczak, Synlett 2009, 1757.

[28] a) A. V. Malkov, M. Orsini, D. Pernazza, K. W. Muir, V. Langer, P.

ˇ

´

Meghani, P. Kocovsky, Org. Lett. 2002 , 4 , 1047; b) A. V. Malkov, M.

ˇ

´

Bell, M. Vassieu, V. Bugatti, P. Kocovsky, J. Mol. Catal. A 2003, 196,

[17] For catalysis by fluoride, see refs. [16e–g] and the following: a) S. R.

Chemler, W. R. Roush, J. Org. Chem. 2003 , 68 , 1319; b) N. Aoyama,

T. Hamada, K. Manabe, S. Kobayashi, Chem. Commun. 2003 , 676;

c) N. Aoyama, T. Hamada, K. Manabe, S. Kobayashi, J. Org. Chem.

2003 , 68 , 7329; d) S. Yamasaki, K. Fuji, R. Wada, M. Kanai, M. Shi-

basaki, J. Am. Chem. Soc. 2002 , 124 , 6536.

179; c) A. V. Malkov, M. Bell, M. Orsini, D. Pernazza, A. Massa, P.

ˇ

´

Herrmann, P. Meghani, P. Kocovsky, J. Org. Chem. 2003 , 68 , 9659;

d) A. V. Malkov, M.-M. Westwater, A. Gutnov, P. Ramꢀrez-Lꢁpez, F.

ˇ

Friscourt, A. Kadlcꢀkovꢂ, J. Hodacovꢂ, Z. Rankovic, M. Kotora, P.

ˇ

´

Kocovsky, Tetrahedron 2008 , 64 , 11335; e) A. V. Malkov, C. Mac-

ˇ

´

Donald, P. Kocovsky, Tetrahedro n : A symmetry 2010 , 21 , 1173; for

ˇ

´

overviews, see ref. [11] and [12d] and the following: f) P. Kocovsky,

[18] a) S. E. Denmark, J. Fu, J. Am. Chem. Soc. 2001 , 123 , 9488; b) S. E.

Denmark, J. Fu, Org. Lett. 2002 , 4 , 1951; c) S. E. Denmark, J. Fu, J.

Am. Chem. Soc. 2003 , 125 , 2208; d) S. E. Denmark, J. Fu, D. M.

Coe, X. P. Su, N. E. Pratt, B. D. Griedel, J. Org. Chem. 2006 , 71 ,

1513; e) S. E. Denmark, J. Fu, M. J. Lawler, J. Org. Chem. 2006 , 71 ,

1523.

ˇ

´

A. V. Malkov, Izv. Akad. Nauk, Ser. Khim. 2004, 1733; P. Kocovsky,

A. V. Malkov, Russian Chem. Bull. Int. Ed. 2004, 53, 1806; for the

synthesis of the bipyridine scaffold, see: g) A. V. Malkov, M. Bella,

ˇ

´

V. Langer, P. Kocovsky, Org. Lett. 2000 , 2 , 3047; h) A. V. Malkov,

I. R. Baxendale, M. Bella, V. Langer, J. Fawcett, D. R. Russell, D. J.

ˇ

´

Mansfield, M. Valko, P. Kocovsky, Organometallics 2001 , 20 , 673;

[19] For an analogous allylation with the corresponding AllylSnR₃re-

agents, see ref. [18c] and the following: S. E. Denmark, T. Wyn, J.

Am. Chem. Soc. 2001 , 123 , 6199.

´

i) A. V. Malkov, D. Pernazza, M. Bell, M. Bella, A. Massa, F. Teply,

ˇ

´

P. Meghani, P. Kocovsky, J. Org. Chem. 2003 , 68 , 4727, and referen-

[20] For an analogous catalysis in the Mukaiyama-aldol reaction, see

ref. [1e].

ces cited therein; for another application of pyridine N-oxides as or-

ˇ

ganocatalysts, see: j) A. V. Malkov, M. R. Gordon, S. Stoncius, J.

ˇ

´

[21] For the synthesis of Denmarkꢆs bases, see: a) S. E. Denmark, X. Su,

Y. Nishigaichi, D. M. Coe, K.-T. Wong, S. B. D. Winter, J. Y. Choi, J.

Org. Chem. 1999 , 64 , 1958; for other phosphoramides, see: b) T.

Nemoto, T. Hitomi, H. Nakamura, L. Jina, K. Hatano, Y. Hamada,

Tetrahedron: Asymmetry 2007 , 18 , 1844.

Hussain, P. Kocovsky, Org. Lett. 2009 , 11 , 5390.

ˇ

´

[29] a) A. V. Malkov, M. Bell, F. Castelluzzo, P. Kocovsky, Org. Lett.

ˇ

2005 , 7 , 3219; b) A. V. Malkov, M. Barłꢁg, Y. Jewkes, J. Mikusek, P.

Kocovsky, J. Org. Chem. 2011 , 76 , 4800; c) A. V. Malkov, O. Kysilka,

M. Edgar, A. Kadlcꢀkovꢂ, M. Kotora, P. Kocovsky, Chem. Eur. J.

ˇ

´

ˇ

´

[22] a) M. Nakajima, M. Saito, M. Shiro, S. Hashimoto, J. Am. Chem.

Soc. 1998 , 120 , 6419; b) M. Nakajima, M. Saito, S. Hashimoto,

Chem. Pharm. Bull. 2000, 48, 306.

2011 , 17 , 7162.

ˇ

´

[30] a) A. V. Malkov, L. Dufkovꢂ, L. Farrugia, P. Kocovsky, Angew.

Chem 2003 , 115 , 3802; Angew. Chem. Int. Ed. 2003 , 42 , 3674;

ˇ

b) A. V. Malkov, P. Ramꢀrez-Lꢁpez, L. Biedermannovꢂ, L. Rulꢀsek,

[23] a) T. Shimada, A. Kina, S. Ikeda, T. Hayashi, Org. Lett. 2002 , 4 ,

2799; b) T. Shimada, A. Kina, T. Hayashi, J. Org. Chem. 2003 , 68 ,

6329; c) A. Kina, T. Shimada, T. Hayashi, Adv. Synth. Catal. 2004 ,

346 , 1169.

ˇ

´

L. Dufkovꢂ, M. Kotora, F. Zhu, P. Kocovsky, J. Am. Chem. Soc.

2008 , 130 , 5341; for a general discussion of the isoinversion effect in

the correlation of enantioselectivity versus temperature, see for ex-

ample: c) P. J. Walsh, M. C. Kozlowski, Fundamentals of Asymmetric

Catalysis, University Science Books, Mill Valley, 2009, p. 9.

ˇ

[24] a) R. Hrdina, I. G. Starꢂ, L. Dufkovꢂ, S. Mitchell, I. Cꢀsarovꢂ, M.

ˇ

Kotora, Tetrahedron 2006 , 62 , 968; b) R. Hrdina, A. Kadlcꢀkovꢂ, I.

ˇ

Valterovꢂ, J. Hodacovꢂ, M. Kotora, T e trahedro n : A symmetry 2006 ,

[31] For other pyridine N-monooxide catalysts, see: a) Q. Chai, C. Song,

Z. Sun, Y. Ma, C. Ma, Y. Dai, M. B. Andrus, Tetrahedron Lett. 2006 ,

47 , 8611; b) L. Pignataro, M. Benaglia, R. Annunziata, M. Cinquini,

F Cozzi, J. Org. Chem. 2006 , 71 , 1458; c) G. Chelucci, S. Baldino,

G. A. Pinna, M. Benaglia, L. Buffa, S. Guizzetti, T e trahedron 2008 ,

64 , 7574; d) J. S. Chen, B. Captain, N. Takenaka, Chem. Eur. J. 2009 ,

15 , 7268; e) J. S. Chen, B. Captain, N. Takenaka, Org. Lett. 2011 , 13 ,

1654; f) M. S.-I. Veitꢀa, M. Joudat, M. Wagner, A. Falguiꢉres, A.

Guy, C. Ferroud, Heterocycles 2011, 83, 2011.

ˇ

17 , 3185; c) R. Hrdina, I. Valterovꢂ, J. Hodacovꢂ, I. Cꢀsarovꢂ, M.

Kotora, Adv. Synth. Catal. 2007 , 349 , 822; d) R. Hrdina, M. Dracꢀn-

sky, I. Valterovꢂ, J. Hodacovꢂ, I. Cꢀsarovꢂ, M. Kotora, Adv. Synth.

Catal. 2008 , 350 , 1449; e) R. Hrdina, T. Boyd, I. Valterovꢂ, J. Hoda-

covꢂ, M. Kotora, Synlett 2008, 3141; f) R. Hrdina, F. Opekar, J. Roi-

thovꢂ, M. Kotora, Chem. Commun. 2009 , 2314; g) A. Kadlcꢀkovꢂ, R.

Hrdina, I. Valterovꢂ, M. Kotora, Adv. Synth. Catal. 2009, 351, 1279;

h) K. Vlasanꢂ, R. Hrdina, I. Valterovꢂ, M. Kotora, Eur. J. Org.

ˇ

´

ˇ

Chem. 2010, 7040; i) A. Kadlcꢀkovꢂ, K. Vlasanꢂ, M. Kotora, Collect.

Czech. Chem. Commun. 2011, 76, 415; for an overview; see: j) M.

Kotora, Pure Appl. Chem. 2010 , 82 , 1813.

[32] a) J. F. Traverse, Y. Zhao, A. H. Hoveyda, M. L. Snapper, Org. Lett.

2005 , 7 , 3151. For a follow up, see: b) V. Simoninia, M. Benaglia, L.

Pignataroa, S. Guizzettia, G. Celentano, Synlett 2008, 1061; for

recent examples of tetrahydroisoquinoline-derived N-oxides, see

ref. [27e].

ˇ

[25] a) A. Kadlcꢀkovꢂ, I. Valterovꢂ, L. Duchꢂckovꢂ, J. Roithovꢂ, M.

ˇ

Kotora, Chem. Eur. J. 2010, 16, 9442; b) L. Duchꢂckovꢂ, A. Kadlcꢀ-

kovꢂ, M. Kotora, J. Roithovꢂ, J. Am. Chem. Soc. 2010, 132, 12660.

[26] a) W.-L. Wong, C.-S. Lee, H. K. Leung, H.-L. Kwong, Org. Biomol.

Chem. 2004 , 2 , 1967; for polymeric poly-N-oxides, see: b) C. A.

Mꢇller, T. Hoffart, M. Holbach, M. Reggelin, Macromolecules 2005 ,

38 , 5375; for an overview, see: c) H.-L. Kwong, H.-L. Yeung, C.-T.

Yeung, W.-S. Lee, C.-S. Lee, W.-L. Wong, Coord. Chem. Rev. 2007 ,

251 , 2188.

ˇ

´

[33] a) A. Massa, A. V. Malkov, P. Kocovsky, A. Scettri, Tetrahedron

Lett. 2003 , 44 , 7179; b) G. J. Rowlands, W. K. Barnes, Chem.

Commun. 2003 , 2712; c) P. Wang, J. M. Chen, L. F. Cun, J. G. Deng,

J. Zhu, J. Liao, Org. Biomol. Chem. 2009 , 7 , 3741; d) A. Massa,

M. R. Acocella, V. De Sio, R. Villano, A. Scettri, Tetrahedro n :

Asymmetry 2009 , 20 , 202; e) V. De Sio, A. Massa, A. Scettri, Org.

Biomol. Chem. 2010 , 8 , 3055; f) V. De Sio, M. R. Acocella, R. Villa-

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

17

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

ˇ

´

no, A. Scettri, Tetrahedro n : A symmetry 2010 , 21 , 1432; g) G.

Monaco, C. Vignes, F. De Piano, A. Bosco, A. Massa, Org. Biomol.

Chem. 2012 , 10 , 9650.

[46] a) A. V. Malkov, M. A. Kabeshov, M. Barłꢁg, P. Kocovsky, Chem.

Eur. J. 2009 , 15 , 1570; b) A. V. Malkov, M. Barłꢁg, L. Miller-

Potuckꢂ, M. A. Kabeshov, L. Farrugia, P. Kocovsky, Chem. Eur. J.

ˇ

´

[34] J. R. Fulton, L. M. Kamara, S. C. Morton, G. J. Rowlands, Tetrahe-

dron 2009 , 65 , 9134.

2012, 18, 6873.

[47] Prepared as a trans/cis mixture (87:13) by the CuCl-catalyzed reac-

tion of crotyl chloride with Cl₃SiH: K. Iseki, Y. Kuroki, M. Takaha-

shi, S. Kishimoto, Y. Kobayashi, Tetrahedron 1997 , 53 , 3513.

[48] Prepared as a practically pure cis isomer by Pd-catalyzed addition of

Cl₃SiH to butadiene: J. Tsuji, M. Hara, K. Ohno, Tetrahedron 1974 ,

30 , 2143.

[49] For recent excellent reviews on isotope effects, see: a) M. Gꢁmez-

Galleo, M. A. Sierra, Chem. Rev. 2011, 111, 4857; b) E. M. Simmons,

J. F. Hartwig, Angew. Chem. 2012 , 124 , 3120; Angew. Chem. Int. Ed.

2012 , 51 , 3066.

[50] a) M. S. Gordon, M. T. Carroll, L. P. Davis, L. W. Burrgraf, J. Phys.

Chem. 1990 , 94 , 8125; b) M. Kira, K. Sato, H. Sakurai, M. Hada, M.

Izawa, J. Ushio, Chem. Lett. 1991 , 387; c) M. A. Brook, Silicon in

Organic, Organometallic and Polymeric Chemistry, Wiley, New

York, 2000, pp. 133; d) H. Fleischer, Eur. J. Inorg. Chem. 2001 , 393;

for a more detailed investigation, see: e) K. Sakata, H. Fujimoto,

Organometallics 2010 , 29 , 1004.

[35] a) M. Nakajima, S. Kotani, T. Ishizuka, S. Hashimoto, T e trahedron

Lett. 2005 , 46 , 157; b) S. Kotani, S. Hashimoto, M. Nakajima, Tetra-

hedron 2007 , 63 , 3122; c) V. Simonini, M. Benaglia, T. Benincori,

Adv. Synth. Catal. 2008 , 350 , 561; d) D. R. Boyd, M. Bell, K. S.

Dunne, B. Kelly, P. Stevenson, J. F. Malone, C. C. R. Allen, Org.

Biomol. Chem. 2012 , 10 , 1388.

[36] Y. S. Oh, S. Kotani, M. Sugiura, M. Nakajima, Tetrahedron: Asym-

metry 2010 , 21 , 1833.

[37] For the non-catalytic modification of allyltrichlorosilane with a stoi-

chiometric chiral auxiliary and its application in asymmetric allyla-

tion, see: a) K. Kubota, J. L. Leighton, Angew. Chem. 2003 , 115 ,

976; Angew. Chem. Int. Ed. 2003 , 42 , 946; b) H. Kim, S. Ho, J. L.

Leighton, J. Am. Chem. Soc. 2011 , 133 , 6517.

[38] For a review on Krçhnke annulation, see: a) F. Krçhnke, Synthesis

1976, 1; for recent overviews of the Krçhnke application in the syn-

thesis of terpenoid pyridine derivatives, see refs. [26c], [28c], and

the following: b) U. Knof, A. von Zelewsky, Angew. Chem. 1999 ,

111 , 312; Angew. Chem. Int. Ed. 1999 , 38 , 302; c) G. Chelucci, R. P.

Thummel, Chem. Rev. 2002 , 102 , 3129; d) N. C. Fletcher, J. Chem.

[51] V. Leskovac, Comprehensive Enzyme Kinetics, Kluwer Academic

Publishers, New York, 2003.

[52] M. O. Sinnokrot, C. D. Sherrill, J. Am. Chem. Soc. 2004 , 126 , 7690.

[53] a) S. E. Denmark, J. Fu, J. Am. Chem. Soc. 2000 , 122 , 12021; b) S. E.

Denmark, Proc. Natl. Acad. Sci. USA 2004 , 101 , 5439; c) S. E. Den-

mark, T. Bui, J. Org. Chem. 2005 , 70 , 10393; d) S. E. Denmark, S. M.

Pham, R. A. Stavenger, X. Su, K.-T. Wong, Y. Nishigaichi, J. Org.

Chem. 2006 , 71 , 3904.

ˇ

´

Soc. Perkin Trans. 1 2002 , 1831; e) A. V. Malkov, P. Kocovsky, Curr.

Org. Chem. 2003, 7, 1737.

[39] E. D. Mihelich, D. J. Eickhoff, J. Org. Chem. 1983 , 48 , 4135.

[40] The synthesis of 22g was complicated by a partial loss of the methyl

group of the 4-OMe group in the preparation of the Krçhnke salt

20g from 19g. Here, as a result of the S_N2 substitution by the strong-

ly nucleophilic I^ꢁ, the desmethyl analogue, that is, 3,5-(MeO)₂-4-

[54] Note that Kotora and Roithovꢂ^[25]have found the non-covalent tran-

sition states for the biaryl-type N,N’-dioxides to be of considerably

lower in energy than those corresponding to either of the covalent

(OH)C₆H₂-COCH₂ACHTUNGTRENNUNG

(N⁺C₅H₅)I^ꢁ(20j), was obtained along with the

expected salt 20g (in ꢂ1:1 ratio). This mixture of Krçhnke salts was

then condensed with (ꢁ)-21 and AcONH₄under the standard condi-

tions to afford mainly the monodesmethyl analogue of 22j (appa-

rently, further loss of the methyl group had occurred). Methylation

of the 4-OH group in the latter product with Me₂SO₄in the pres-

ence of K₂CO₃gave rise to the desired 3,4,5-trimethoxy derivative

22g.

mechanisms (in a non-polar environment). However, the latter

mechanism would give rise to the opposite enantiomer of the prod-

uct than found experimentally. Furthermore, these authors have also

revealed a dramatic, experimentally observed switch from one enan-

tiomer to the opposite one by changing the solvent from nonpolar

to polar,^[24f]which indicates a dramatic change in the reaction

mechaACTHNUTRGNEUGnN ism.

[41] For the method, see: a) D. Lçtscher, S. Rupprecht, H. Stoeckli-

Evans, A. von Zelewsky, Tetrahedron: Asymmetry 2000 , 11 , 4341;

b) B. Kolp, D. Abeln, H. Stoeckli-Evans, A. von Zelewsky, Eur. J.

Inorg. Chem. 2001 , 1207; c) S. Lçtscher, S. Rupprecht, P. Collomb, P.

Belser, H. Viebrock, A. von Zelewsky, P. Burger, Inorg. Chem. 2001 ,

40 , 5675.

[42] a) H.-L. Kwong, W.-L. Wong, W.-S. Lee, L.-S. Cheng, W.-T. Wong,

Tetrahedron: Asymmetry 2001 , 12 , 2683; b) G. Chelucci, A. Saba, F.

Soccolini, Tetrahedron 2001 , 57 , 9989; c) Y. Bessiꢉre-Chrꢈtien, C.

Grison, Bull. Soc. Chim. Fr. 1970, 3103.

[43] The N-oxides 9 and 15–17 were synthesized from (+)-a-pinene of

96% ee (via pinocarvone 21); therefore, they should be of the same

enantiopurity (at least). In fact, METHOX (9) was obtained enan-

tiopure (ꢃ99% ee), because it could be easily purified by crystalli-

zation. (+)-Isopinocampheol (+)-24, which can be prepared by hy-

droboration of (ꢁ)-a-pinene, was purchased from Aldrich (declared

[a]²_D⁰= +35.1; c=20, EtOH) and employed for the synthesis of 18.

[44] While 18a was obtained pure, we were not able to fully purify 18b,

which contained ꢂ17% of 18a, as shown by the ¹H NMR spectrum

(see the Supporting Information). Therefore, the observed optical

rotation for 18b is not accurate.

[55] That pyridine-N-oxide itself is rather inefficient as catalyst has also

been noted experimentally by Denmark.^[14]Thus, the more complex

molecules METHOX and QUINOX apparently offer additional fea-

tures, such as arene–arene interactions, that lower the activation bar-

rier.

[56] For a recent computational study of propargylation of benzaldehyde

with CH₂=C=CHSiCl₃, catalyzed by helical bipyridine N-monooxide,

see: a) T. Lu, R. Zhu, Y. An, S. E. Wheeler, J. Am. Chem. Soc. 2012 ,

134 , 3095; for a computational study and comparison of propargyla-

tion with CH₂=C=CHSiCl₃and allylation with CH₂=CHCH₂SiCl_3,

catalyzed by bipyridine N-monooxides, see: b) T. Lu, M. A. Porter-

field, S. E. Wheeler, Org. Lett. 2012 , 14 , 5310.

[57] a) G. Vassilikogiannakis, M. Hatzimarinaki, M. Orfanopoulos, J.

Org. Chem. 2000 , 65 , 8180; b) D. OꢆHagan, R. J. M. Goss, A. Med-

dour, J. Courtieu, J. Am. Chem. Soc. 2003, 125, 379.

[58] For aged samples of allyltrichlorosilane, which may contain HCl, re-

sulting from a slow partial decomposition, the base was used in

larger excess (2 or up to 5 equiv); no effect on enantioselectivity

was observed.

[59] H. Doucet, M. Santelli, Tetrahedro n : A symmetry 2000 , 11 , 4163.

[60] T. P. Loh, J. R. Zhou, Tetrahedron Lett. 2000 , 41 , 5261.

[61] S. Kobayashi, S. Nagayama, T. Busujima, Chem. Lett. 1997, 26, 959.

[62] H. Yamataka, K. Nishikawa, T. Hanafusa, Chem. Lett. 1990, 19,

1711.

[63] a) J. L. Hartwell, S. R. L. Kornberg, J. Am. Chem. Soc. 1946 , 68 , 868;

b) L. C. King, M. McWhirter, R. L. Rowland, J. Am. Chem. Soc.

1948 , 70 , 239; c) L. C. King, J. Am. Chem. Soc. 1948 , 70 , 242; d) J.

Alvarez-Builla, E. Gꢂlvez, A. M. Cuadro, F. Florencio, S. Garcia-

Blanco, J. Heterocycl. Chem. 1987 , 24 , 917.

[45] For coordination of pyridine to trichlorosilane, see: a) G. W. Fester,

J. Wagler, E. Brendler, U. Bçhme, D. Gerlach, E. Kroke, J. Am.

Chem. Soc. 2009 , 131 , 6855; b) G. W. Fester, J. Wagler, E. Brendler,

E. Kroke, Eur. J. Inorg. Chem. 2008 , 5020; c) A. V. Malkov, A. J. P.

ˇ

Stewart Liddon, P. Ramꢀrez-Lꢁpez, L. Bendovꢂ, D. Haigh, P. Kocov-

´

sky, Angew. Chem. 2006 , 118 , 1460; Angew. Chem. Int. Ed. 2006 , 45 ,

ˇ

´

1432; d) A. V. Malkov, K. Vrankovꢂ, S. Stoncius, P. Kocovsky, J.

Org. Chem. 2009 , 74 , 5839.

&

18

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Mechanistic Dichotomy in the Asymmetric Allylation of Aldehydes

FULL PAPER

[64] Compared to the original protocol,^[29a]this procedure represents an

improvement: apparently, in addition to the desired N-oxidation,

the electron-rich 2,4,6-trimethoxyphenyl group is slowly oxidized by

m-CPBA to generate a phenolic by-product, which was previously

separated by chromatography. Now, washing the solution of the

crude product with aqueous NaOH removed the latter by-product

prior to chromatography. The isolated yield of (+)-9 was thus im-

proved, since the phenolic material interfered neither with the chro-

matography nor with the final crystallization.

[71] M. A. Neumann, M. A. Perrin, J. Phys. Chem. B 2005 , 109 , 15531.

[72] a) A. Klamt, G. Schuurmann, J. Chem. Soc. Perkin Trans. 2 1993 ,

799; b) A. Schꢊfer, A. Klamt, D. Sattel, J. C. W. Lohrenz, F. Eckert,

Phys. Chem. Chem. Phys. 2000 , 2 , 2187.

[73] F. Jensen, Introduction to Computational Chemistry, Wiley, New

York, 1999.

ˇ

[74] a) J. Cꢀzek, J. Chem. Phys. 1966, 45, 4256; b) J. Cꢀzek, Adv. Chem.

Phys. 1969, 14, 35; c) K. Raghavachari, G. W. Trucks, J. A. Pople, M.

Head-Gordon, Chem. Phys. Lett. 1989 , 157 , 479.

[65] R. Ahlrichs, M. Bꢊr, M. Hꢊser, H. Horn, C. Kçlmel, Chem. Phys.

Lett. 1989 , 162 , 165.

[66] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996 , 77 , 3865.

[67] a) K. Eichkorn, O. Treutler, H. ꢋhm, M. Hꢊser, R. Ahlrichs, Chem.

Phys. Lett. 1995 , 240 , 283; b) K. Eichkorn, F. Weigen, O. Treutler, R.

Ahlrichs, Theor. Chim. Acta 1997, 97, 119.

[75] D. E. Woon, T. H. Dunning, J. Chem. Phys. 1993 , 98 , 1358.

[76] R. D. Amos, J. S. Andrews, N. C. Handy, P. J. Knowles, Chem. Phys.

Lett. 1991 , 185 , 256.

[77] Note added in proof: The difference between the reactivity of (E)-

and (Z)-crotylsilanes 5b and c in the allylation of aldehydes cata-

lyzed by METHOX (9; Scheme 2, Table 3, and Figure 5) parallels

the behavior of (E)- and (Z)-crotylboronates (R. W. Hoffmann, H.-

J. Zeiß, J. Org. Chem. 1981 , 46 , 1309).

[68] W. J. Hehre, L. Radom, P. I. R. von Schleyer, J. A. Pople, Ab Initio

Molecular Orbital Theory, Wiley, New York, 1986.

[69] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005 , 7 , 3297.

ˇ

[70] a) S. Grimme, J. Comput. Chem. 2004 , 25 , 1463; b) P. Jurecka, J.

Received: October 25, 2012

Published online: && &&, 0000

ˇ

´

Cerny, P. Hobza, D. Salahub, J. Comput. Chem. 2007, 28, 555.

Chem. Eur. J. 2013, 00, 0 – 0

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

&

19

&

ÞÞ

These are not the final page numbers!

ˇ

´

A. V. Malkov, S. Stoncius, L. Rulꢀsek, P. Kocovsky et al.

Kinetic and computational investiga-

tions of the enantio- and diastereose-

lective allylation of aldehydes with

allyltrichlorosilanes, employing

METHOX as chiral organocatalyst,

indicate that the reaction proceeds

through a dissociative (cationic) mech-

anism with a pentacoordinate cationic

transition state (see scheme). The less

sterically demanding QUINOX cata-

lyst operates via a hexacoordinate neu-

tral complex. In both pathways, only

one molecule of the catalyst is

Asymmetric Allylation

ˇ

A. V. Malkov,* S. Stoncius,* M. Bell,

F. Castelluzzo, P. Ramꢀrez-López,

L. Biedermannovꢁ, V. Langer,

ˇ

´

L. Rulꢀsek,* P. Kocovsky* . . &&&& — &&&&

Mechanistic Dichotomy in the Asym-

metric Allylation of Aldehydes with

Allyltrichlorosilanes Catalyzed by

Chiral Pyridine N-Oxides

involved in the rate- and selectivity-

determining step.

&

20

&

www.chemeurj.org

ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 0000, 00, 0 – 0

ÝÝ

These are not the final page numbers!

Article Doi

Mechanistic dichotomy in the asymmetric allylation of aldehydes with allyltrichlorosilanes catalyzed by chiral pyridine N-oxides

DOI: 10.1002/chem.201203817

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.201203817

Full text of DOI:10.1002/chem.201203817

Products guided by the article

R&D Labs maybe for 1314113-34-2

Relevant to this article

Hot Product