Catalytic Enantioselective Synthesis of ,-Disubstituted Amino Acids

Chapter 7

Downloaded by UNIV OF ARIZONA on December 25, 2012 | http://pubs.acs.org Publication Date: June 14, 2009 | doi: 10.1021/bk-2009-1009.ch007

Catalytic Enantioselective Synthesis of α,α-Disubstituted Amino Acids: The Strecker Reaction of Ketimines Using Chiral Poly-Gadolinium Complexes Masakatsu Shibasaki and Motomu Kanai Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan

Enantiomerically-enriched α,α-disubstituted amino acids are an important group of unnatural amino acids that are useful for pharmaceutical synthesis. The Strecker reaction of ketimines is a direct method for synthesizing α,α-disubstituted amino acids. This chapter summarizes the development, mechanistic insights, and application of the catalytic enantioselective Strecker reaction developed in our group.

Introduction The catalytic enantioselective Strecker reaction is among the most straightforward methods for synthesizing enantiomerically enriched a-amino acids and their derivatives (/). More than a dozen examples have been reported using aldimines as substrates, which can produce a-monosubstituted amino acid derivatives (/). The catalytic enantioselective Strecker reaction of ketimines, however, is not as well established due to the lower reactivity and small difference (sterically and electronically) between the two substituents on the prochiral carbon of ketimines compared to aldimines. To achieve an efficient enantioselective Strecker reaction of ketimines, an asymmetric catalyst must be highly active and strictly enantioselective. The development of a chiral ureacatalyzed reaction by Jaconsen's laboratory in 2000 overcame this issue for the first time (2,5). In 2003 and 2004, we reported the most general and practical 102

© 2009 American Chemical Society

In Asymmetric Synthesis and Application of -Amino Acids; Soloshonok, V., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2009.

103 catalytic enantioselective Strecker reaction of ketimines to date (4). In this chapter, we review the development and synthetic application of our reaction. Progress is actively ongoing in this field (5). The development of enantioselective catalysts that promote tetrasubstituted carbon-forming reactions is a new frontier in organic synthesis (6).


Development of Catalytic Enantioselective Strecker Reaction of Ketimines Using a Chiral Gadolinium Complex In 2001, we developed a catalytic enantioselective cyanosilylation of ketones using a gadolinium (Gd) complex of D-glucose-derived ligand 1 (Scheme 1) (7,8). The optimized catalyst was produced from Gd(0'Pr) and 1 mixed in a 1:2 ratio. Based on the results of mechanistic studies, including labeling experiments and kinetic measurements, we proposed model 4 for the enantio-differentiating cyanation step: (1) the active catalyst is a TMS-containing 2:3 complex of Gd and 1; (2) the active nucleophile is a gadolinium isocyanide generated through transmetalation from T M S C N ; (3) the enantioselective cyanation proceeds through intramolecular transfer of cyanide to an activated ketone coordinating to the Lewis acidic Gd. The positions of both the activated nucleophile and the electrophile are determined by the asymmetric bimetallic complex, thereby affording high enantioselectivity. 3

Gd(0'Pr) -1(1:2) (1-5 mol %)

J

3

O X

,

+

TMSCN

m

Q

V •

R "^R 1

2

CH3CH2CN

66%-97 ee chiral ligand = !

Ph

o ^

x

^-0'

/

N

c

^~^

6

0

TMS

3:X = CI

HO

^

X

4

Scheme 1. Catalytic enantioselective cyanosilylation of ketones

We extended this catalysis to enantioselective Strecker reaction of ketimines. Initial optimization using acetophenone-derived ketimines allowed us to identify two important parameters for reaction efficiency: (1) N-phosphinoyl imines



104 produced high reactivity and enantioselectivity; (2) an electronically tuned catalyst generated from ligand 2 was more active and enantioselective than the catalyst derived from 1. Thus, using 2.5 mol % catalyst, the product from acetophenone-derived ketimine was obtained in 94% yield with 95% ee (24 h) (4a). Heteroaromatic and aliphatic ketimines, however, afforded less satisfactory results. Studies toward expanding substrate generality led us to identify beneficial effects of protic additives. Thus, using 1.5 equiv of T M S C N and 1 equiv of 2,6dimethylphenol (conditions A), excellent enantioselectivity was produced from a wide range of ketimines, including aromatic, heteroaromatic, cyclic, and aliphatic ketimines (Table 1) (4b). The protic additive significantly improved the catalyst activity as well as enantioselectivity. To gain insight into the additive effect, ESI-MS studies of the catalyst were performed in the presence of 2,6-dimethylphenol. A catalyst prepared from a 1:2 ratio of Gd(0'Pr) and 2, in the presence of T M S C N (15 equiv to the catalyst) in acetonitrile, mainly afforded a peak corresponding to the 0-silylated 2:3 complex (Scheme 2, 5). When 2,6-dimethylphenol (10 equiv) was added to the catalyst solution, the peak corresponding to 5 disappeared and a new peak corresponding to an O-protonated 2:3 complex 6 appeared (Scheme 2, eq. 1). Therefore, this protonated complex was a highly active enantioselective catalyst in this asymmetric Strecker reaction of ketimines. 3

+ HCN

Gd(0'Pr) + 2 + 2 TMSCN 3

i Gd-V

O-Gd Co' TMS

NCCN

cJ

5

TMS

W-Gd Gd-.^ CO' NC OJ A 6 HCN

2TMSCN

Scheme 2. Generation of proton-containing asymmetric catalyst

Based on this finding, we expected that the same active catalyst 6 would be generated through the reaction of H C N with 5 (Scheme 2, eq. 2). Protonolysis with H C N produces 2 equiv of T M S C N . Thus, only a catalytic amount of T M S C N is required i f H C N is used as the proton source and stoichiometric


105 cyanide source. This would lead to a further advanced catalytic enantioselective Strecker reaction that might be applicable to industrial scale synthesis of a,ctdisubstituted amino acids.

Table 1. Catalytic Enantioselective Strecker Reaction of Ketimines Q O H

2

N'


r 1

entry

A

Condition A or B

a

NC^N-PPh

-

CH CH2CN, -40 C *

r 2

R

3

substrate

( c a t

. SSJSliol%)

t j m e (h)

1

R yi6

'

2

2

d( % )

e e ( % )

( Q ) p h 2

1

Jj)

2

[Py^ N

A(1) M e

B

.P(0)Ph

30 1

(°- )

A

1

A(1)

4 f r ^ M e N

^P(0)Ph

P(0)Ph N

CH (CH ) 3

2

92 9 0

3 1

9 7

9

5

21

93

93

B(1)

3

99

99

A(1)

22

92

92

A(1)

43

73

90

A (2.5)

2.5

91

80

A(1)

38

93

96

2

c6 K r

94 9 7

2

"l^Me

5

19

2

(\/|

4

e

N

P(0)Ph

2

Me .P(0)Ph N M

2

P n - ^ M e /

" Condition A : catalyst = Gd(0 Pr)

3

(x mol %) + 2 (2x mol %); T M S C N (1.5 equiv), 2,6-

dimethylphenol (1 equiv). Condition B: catalyst = Gd(0'Pr) (x mol %) + 2 (2x mol %); 3

T M S C N (2-5 mol %), H C N (150 mol %).

As expected, the reaction proceeded smoothly in the presence of a catalytic amount of T M S C N and a stoichiometric amount of H C N (Table 1, condition B) (4c). The reaction time was significantly shortened compared to using 2,6dimethylphenol as an additive (condition A). The higher reactivity allowed us to


106 reduce the catalyst amount to as low as 0.1 mol %, while maintaining excellent enantioselectivity. The difference in reactivity when using H C N or 2,6dimethylphenol as the proton source might be partly due to the acidity of the additive. In the presence of H C N (pK = 9.2), the concentration of the active catalyst 6 should be higher than in the presence of 2,6-dimethylphenol (pK = 10). a

a

Gd(0'Pr) + 2 Downloaded by UNIV OF ARIZONA on December 25, 2012 | http://pubs.acs.org Publication Date: June 14, 2009 | doi: 10.1021/bk-2009-1009.ch007

3

Scheme 3. Catalytic cycle of enantioselective Strecker reaction of ketimines

We proposed the catalytic cycle under condition B as shown in Scheme 3. To the active catalyst 6, the substrate ketimine is incorporated and activated. A n intramolecular cyanide transfer from the gadolinium isocyanide to the activated substrate defines the enantioselectivity (8), and the zwitter ionic intermediate 9 is generated. In this step, the characteristics of phosphinoyl ketimines—i.e., the cis/trans isomers equilibrate very rapidly at the reaction temperature (-40 °C) (4a)—should work advantageously in determining the reactive imine geometry. The intermediate 9 should collapse through intramolecular proton transfer, and the product was liberated with regeneration of gadolinium alkoxide complex 7.


107 From 7, successive reactions with T M S C N and H C N reproduced the active catalyst 6. The reaction did not proceed at all in the absence of a catalytic amount of T M S C N at -40 °C. This finding suggests that the active catalyst 6 is generated only through the silylated pre-catalyst 5, and direct generation of 6 from the alkoxide complex 7 and H C N does not occur at -40 °C.


Synthetic Application Due to the versatility of a,ct-disubstituted amino acids in biologically active compounds, there are many potential synthetic targets for our reaction. We achieved catalytic asymmetric synthesis of sorbinil (10) (4b) and lactacystin (11) (9) using the catalytic enantioselective Strecker reaction of ketimines.

O

3

n r

PPh

condition A Gd(0'Pr) -2(1:2)(1 mol %) TMSCN (1.1 equiv) 2,6-dimethylphenol (1 equiv) CH CH CN, -40 °C, 83 h

2

3

O 12 10g- scale

N-PPh

2

or condition B Gd(0'Pr) -2 (1:2) (0.5 mol %) TMSCN (5 mol %) HCN (1.5 equiv) CH CH CN, -40 °C, 1.3 h

13 100%, 98% ee recryst. from 93%, >99% ee J CHCI -hexane (condition A) 96%, 93% ee (condition B)

3

3

2

3

2

1) HCI aq.90 °C 2) KCNO, H 0, 100 °C 2

3) HCI, H 0, 100 °C 67% (3 steps) 2

O sorbinil (10)

Scheme 4. Catalytic asymmetric synthesis of sorbinil

Sorbinil (10) is a potent inhibitor of aldose reductase developed by Pfizer, and could be a therapeutic agent for chronic complications of diabetes mellitus (such as neuropathy) (10). Sorbinil contains a chiral spirohydantoin structure, and its biologic activity resides in the (5)-enantiomer. A straightforward synthesis of sorbinil was achieved using the catalytic enantioselective Strecker reaction of ketimine 12 as a key step (Scheme 4). The reaction proceeded using 1 mol % of catalyst under condition A or 0.5 mol % of catalyst under condition


108 B, affording 13 with excellent enantioselectivity. Enantiomerically pure 13 was obtained by recrystallization of a crude mixture. The reaction was performed in 10 g-scale. Acid hydrolysis followed by hydantoin formation produced 10 in high overall yield. No silica gel chromatography was necessary from 12 to 10.

Gd{N(SiMe ) } -2 3

O

P>

P n

2

3

(1:1.5) (2.5 mol %)

TMSCN (2 equiv) 2,6-dimethylphenol Q HN-PPh (1 equiv) 'Pr CH CH CN -40 °C, 2 days


N

'Pr

3

14

2

15

>99%, 98% ee 1) 0 : Me S 2) NaCI0 3

C0 Et 2

3) H S0 EtOH 81% (3 steps) 2

1) cone. HCI aq. Et0 C NHBoc 2) SOCI , EtOH . Pr 3) Boc 0 NaHC0 16 68% (3 steps) 2

2

1

2

3

'PrMgBr; recryst.

2

2

2

0

2

B

2

3

93% (2 steps)

53% (d.r. = 10:1)

4

1)Ac 0, py. 2) Boc Q, Et N^ 2

C

1)Et NPh SiLi 2

1) LHMDS PhSeBr 2) H 0 2

,

^NBoc

E t

^ J^-C0 Et 2

2

92% (2 steps)

Acd

2

2

NBoc C0 Et

Z n

2) mCPBA, KHF

2

2>

58% (d.r. >15:1) (94% b.r.s.m.)

57% (2 steps) (>30: 1)

19

LDA; Mel HMPA, THF

Donohoe's intermediate C0 H 2

known HO 21

(+)-lactacystin (11)

Scheme 5. Catalytic asymmetric total synthesis of lactacystin

(+)-Lactacystin (11) is a potent and selective proteasome inhibitor, isolated from the Streptomyces sp. by Omura et al ( / / ) . Due to its potent biological activity and complex chemical structure, many synthetic chemists study 11 as a synthetic target (12). We planned to utilize our catalytic enantioselective Strecker reaction for the construction of the tetrasubstituted carbon C-5 of


109 lactacystin. Based on this plan, a-hydroxy ketimines are an obvious starting point. This type of imine, however, is unstable and not isolable in a pure form. Thus, we utilized an enone-derived, stable 14 as a masked a-hydroxy ketimine (Scheme 5). Imine 14, containing a bulky isopropyl group at the cc-position, was barely reactive under the Strecker reaction conditions, and optimization of the reaction conditions was necessary. It was found that the catalyst generated from Gd{N(SiMe ) } and 2 in a 2:3 ratio produced higher activity and enantioselectivity than the catalyst prepared from Gd(0'Pr) . The modified catalyst preparation method afforded the active 2:3 complex in a pure form, which was detected by ESI-MS (13). The catalyst activity difference depending on the preparation methods was attributed to the purity of the active 2:3 complex. Thus, the Strecker reaction of 14 completed using 2.5 mol % catalyst in 2 days, and product 15 was obtained in quantitative yield with 98% ee. Amidonitrile 15 was converted to the protected amino acid derivative 16, which was further transformed to y-lactam 17 through ozonolysis, oxidation, and cyclization. Stereoselective reduction of C-9 ketone using 'PrMgBr (14) via a presumed cyclic transition state produced a diastereomixture of C-9 secondary alcohols with the desired a-alcohol 18 as the major isomer (d.r. = 10:1). Diastereomerically and enantiomerically pure 18 was obtained by recrystallization of the crude mixture from a toluenehexane mixed solvent. After protection of the secondary alcohol and the lactam nitrogen atom with acetyl and Boc groups, respectively, selenenylation and oxidation produced the a,P-unsaturated lactam 19 in excellent yield. The P-hydroxyl group of C-6 was introduced via a stereoselective conjugate addition of the E t N P h S i group from the less hindered side of the enone, followed by Tamao-Fleming oxidation of the silicon with retension of configuration, producing enantiomerically pure known intermediate 20. Methylation at C-7 of 20 under the conditions developed by Donohoe (15) produced the desired stereoisomer 21, which was further converted to lactacystin following the procedures reported by Corey (16) and Donohoe. 3

2

3


3

2

2

Structural Studies of the Poly Gadolinium Asymmetric Catalysts To elucidate the enantio-differentiation mechanism of the Gd catalyst, we studied crystallization of the catalytic species by varying the ligand structure and metal. Colorless, air-stable prisms were obtained from a propionitrile-hexane (2:1) solution of the complex prepared from a 2:3 ratio of Gd(0'Pr) and ligand 3 (crystal A: 80% yield). X-ray crystallographic analysis revealed that crystal A 3


110


was a 4:5 complex of Gd and 3 with a ji-oxo atom surrounded by four gadolinium atoms (Figure l ) ( / 7 ) . The tetranuclear structure was maintained in a solution state, and the sole peak observed by ESI-MS corresponded to crystal A .

Figure I. (a) Crystal structure of 4:5+oxo complex (crystal A) and (b) its chemical structure depiction.

Previously, the M S peak corresponding to crystal A was observed, concomitant with the 2:3 complex (active catalyst), in a solution prepared from a 1:2 ratio of Gd(0'Pr) and 1 or 2. The high yield of crystal A suggested that there is a conversion process from the 2:3 complex to the 4:5+oxo complex (crystal A ) during crystallization. Time-dependent conversion of the 2:3 complex to the 4:5+oxo complex in a solution state was observed by ESI-MS. This conversion was an auto-catalytic process. Thus, seeding the 4:5+oxo complex to a solution containing pure 2:3 complex (prepared from Gd{N(SiMe ) }3) markedly accelerated the formation of the 4:5+oxo complex. The 4:5+oxo complex is thermodynamically more stable than the active kinetically-formed catalyst, the 2:3 complex. The function of crystal A as an enantioselective catalyst was evaluated by Strecker reaction of ketimines. To our surprise, enantioselectivity was completely reversed when crystal A was used as a catalyst, compared to the catalyst prepared in situ (Table 2) (77). This dramatic reversal in enantioselectivity was attributed to the change in the higher-order structure of the chiral polymetallic catalyst. The reaction rate was approximately 5-50 times slower than that using the catalyst prepared in situ. Due to the marked difference in catalyst activity, the major enantiomer obtained using the catalyst prepared from Gd(0'Pr) was (5), whereas the catalyst solution contained a mixture of the 2:3 complex and the 4:5+oxo complex. 3

3

2

3


Ill

Efforts to elucidate the structure of the active 2:3 complex continued. Colorless, air-stable prisms (crystal B) were obtained from a T H F solution of La(0'Pr) and ligand 1 mixed in a 2:3 ratio (47% yield). The x-ray diffraction study revealed the structure to be a pseudo C -symmetric 6:8 complex of La and 1 [Figure 2(a)] (17). Six La atoms were arranged in line as a backbone, and eight chiral ligands were arranged around the backbone. This structure was maintained in solution, based on the fact that the parent peak corresponding to crystal B was observed by ESI-MS. In addition, two major fragment peaks corresponding to metakligand = 2:3 and 4:5+oxo complexes were observed. 3


2

Table 2. Dramatic Reversal of Enantioselectivity Depending on the HigherOrder Structure of the Polymetallic Catalyst 0 /

P

P

h

N 1 ^

R

entry

N

s

5 6 7 8 9 10

N

3

P(0)Ph

P(0)Ph

Ph N ( ° > P

N

N

P h

P(0)Ph

P(0)Ph

catalyst (2.5-10 mol %) TMSCN (1.5 equiv) 2,6-dimethylphenol (1 equiv) CH CH CN,-40 °C

2

substrate

1 2

3 4

R

2

2

2

*

2

2

2

o H ll 2

K

(S)

catalyst (Gd mol %)

time/ h

yield/ %

ee/%

config.

Gd-3 (5) crystal A (7)

0.5 2

99 99

98 91

(S)


0.5 1

99 93

88 96

(S)


2 2

99 95

87 87

Gd-2 (2.5) crystal A (7)

2.5 12

91 99

80 82

(S)


2 14

92 99

74 98

(S)

(R)

(R)

(S) (R)

(R)

(R)

The ESI-MS fragment patterns of crystal B provide insight into the relevance of crystal B to the optimized Gd catalyst (2:3 complex) in solution. The observed MS-fragments (2:3 and 4:5+oxo) can be viewed as subunits of the entire polymetallic assembly. The x-ray structure of crystal B supports this consideration: the distances between the second and the third La atoms from both ends of the 6:8 complex was significantly longer than those between the first and the second La atoms [Figure 2(b)]. In addition, the coordination number



112

Figure 2. (a) Crystal structure of 6:8 complex (crystal B) and (b) its chemical depiction.

of the third L a was smaller than those of the first and the second La, which might induce Lewis base coordination to the third L a and dissociation of the terminal 2:3 subunits. Thus, the 6:8 complex (crystal B) is likely to be constructed via assembly of the 2:3 subunit and the 4:5+oxo subunit. The terminal 2:3 subunits of crystal B might correspond to the optimized active catalyst. This hypothesis is supported by the fact that crystal B (8 mol % La) promoted the catalytic enantioselective Strecker reaction of acetylthiophenederived ketimine, giving the product with 26% ee. Configuration of the major product was the same (S) as when using a catalyst prepared in situ. Crystal B produced a similar level of enantioselectivity as when using a catalyst prepared in situ from La(0'Pr) and 1 (16% ee). Therefore, crystal B represents one of the structures of the lanthanum catalyst in solution. Although the three-dimensional structure of the actual catalyst (2:3 complex) has yet to be clarified, these results demonstrated that the higher-order structure of an artificial asymmetric polymetallic catalyst is a determining factor of the function (enantioselectivity and activity). This finding led to new chiral ligand design (22) (18). Whereas the G d catalyst derived from 22 produced less 3


113


satisfactory results in the Strecker reaction of ketimines compared to 2, it exhibited markedly higher catalyst activity and enantioselectivity in the asymmetric aziridine-opening reaction with T M S C N (Scheme 6). Despite the fact that ligands 2 and 22 contained the same chirality, enantiomer products were produced, which was totally unexpected. The enantio-switching was, however, attributed to the dramatic change in higher-order structure of the polymetallic complexes depending on the ligands. Studies are currently ongoing to identify novel functions of the catalyst derived from 22.

/

Gd(0 'Pr) -2(1:2) (10-20 mol %) TFA (2.5-5 mol %) 3

1

* \

TMSCN (3 equiv) / 2,6-dimethylphenol (1 equiv) / C H C H C N , rt-60 °C

O II

3

J>^S\r

2

N

A

R

C

N

4

P

h

y QT V

80-93% ee

P

h

^ I

I J

H O ^ Y Gd(0'Pr) -22 (1:1.5) _ | %) 3

6

r

v

\

Ar = p - N 0 - C H \ 2

H -^ - | Y 1 Q

1

x

( 1

2 m o

TMSCN (3 equiv) 2,6-dimethylphenol (1 equiv) THF, rt-reflux

. H R v^N 1

Y Y

Ar

22

HO

;

R1 "

Catalytic Enantioselective Synthesis of ,-Disubstituted Amino Acids

Recommend Documents