Mechanism of enantioselective Ti-catalyzed strecker reaction: Peptide-based metal complexes as bifunctional catalysts

Article abstract of DOI:10.1021/ja011875w

Kinetic, structural, and stereochemical data regarding the mechanism of Ti-catalyzed addition of cyanide to imines in the presence of Schiff base peptide ligands are disclosed. The reaction is first order in the Ti·ligand complex; kinetic studies reveal Δ

Full text of DOI:10.1021/ja011875w

11594
J. Am. Chem. Soc. 2001, 123, 11594-11599
Mechanism of Enantioselective Ti-Catalyzed Strecker Reaction:
Peptide-Based Metal Complexes as Bifunctional Catalysts
Nathan S. Josephsohn, Kevin W. Kuntz, Marc L. Snapper,* and Amir H. Hoveyda*
Contribution from the Department of Chemistry, Merkert Chemistry Center, Boston College,
Chestnut Hill, Massachusetts 02467
ReceiVed August 2, 2001
Abstract: Kinetic, structural, and stereochemical data regarding the mechanism of Ti-catalyzed addition of
cyanide to imines in the presence of Schiff base peptide ligands are disclosed. The reaction is first order in the
q
-1
-1
Ti‚ligand complex; kinetic studies reveal ∆S ) -45.6 ( 4.1 cal K mol , indicating a highly organized
transition structure for the turnover-limiting step of the catalytic cycle. A mechanistic model consistent with
the kinetic and stereochemical data is presented, where the Ti center is coordinated to the Schiff base unit of
the ligand and the AA2 moiety of the peptidic segment of the chiral ligand associates and deliVers HNC to the
activated bound substrate. Thus, these studies illustrate that these non-C2-symmetric catalysts likely operate in
a bifunctional fashion.
Introduction
Chart 1
From the standpoint of catalytic asymmetric reaction devel-
opment, peptidic Schiff bases represented by I-III (Chart 1)
are attractive for several reasons: (i) their structural modularity
1
allows for facile preparation of catalyst libraries, (ii) their non-
2
C2-symmetric structure readily lends itself to diversification,
and (iii) the availability of multiple Lewis acid binding sites
3
raises the possibility of multifunctional catalysis. These chiral
dipeptides have thus emerged as highly versatile ligands for
enantioselective catalysis: Ligands I effect an assortment of
asymmetric transformations that include Ti-catalyzed addition
4
5
6
of cyanide to aldehydes, epoxides, and imines, Al-catalyzed
alkylzincs. Such advances and the related high levels of
reactivity and enantioselectivity render mechanistic information
regarding the inner workings of this class of chiral catalysts
valuable, especially in connection with future efforts aimed
toward the development of new catalytic asymmetric processes.
Herein, we disclose kinetic and structural data that, for the
first time, shed light on one of the key transformations promoted
by this versatile class of chiral ligands. The Ti-catalyzed addition
7
addition of cyanide to ketones, and Zr-catalyzed addition of
8
dialkylzincs to imines. Moreover, dipeptides II and III have
been employed to promote facile and enantioselective Cu-
9
10
catalyzed conjugate additions and allylic substitutions with
(
1) Shimizu, K. D.; Snapper, M. L.; Hoveyda, A. H. Chem.sEur. J.
998, 4, 1885-1889.
2) Pfaltz, A. In Stimulating Topics in Organic Chemistry; Shibasaki,
M., Stoddard, J. F., Vogtle, F., Eds.; VCH-Wiley: Weinheim, 2000, pp
1
(
6
of cyanide to imines is a C-C bond forming process that is
8
9-103.
(
3) (a) Steinhagen, H.; Helmchen, G. Angew. Chem., Int. Ed. Engl. 1996,
promoted by ligands I and affords nonracemic amino nitriles
(e.g., 3, Scheme 1) which may then be converted to optically
3
5, 2339-2342. (b) Shibasaki, M.; Sasai, H.; Arai, T. Angew. Chem., Int.
Ed. Engl. 1997, 36, 1236-1256. (c) Rowlands, G. J. Tetrahedron 2001,
11
pure R-amino acids (e.g., 4, Scheme 1). The results presented
5
7, 1865-1882.
4) Mori, A.; Abe, H.; Inoue, S. Appl. Organomet. Chem. 1995, 9, 189-
97 and references therein.
5) (a) Cole, B. M.; Shimizu, K. D.; Krueger, C. A.; Harrity, J. P.;
Snapper, M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1996, 35,
668-1671. (b) Shimizu, K. D.; Cole, B. M.; Krueger, C. A.; Kuntz, K.
W.; Snapper, M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1997,
6, 1704-1707.
6) (a) Krueger, C. A.; Kuntz, K. W.; Dzierba, C. D.; Wirschun, W. G.;
Gleason, J. D.; Snapper, M. L.; Hoveyda, A. H. J. Am. Chem. Soc. 1999,
21, 4284-4285. (b) Porter, J. R.; Wirschun, W. G.; Kuntz, K. W.; Snapper,
below show that in effecting the Ti-catalyzed addition, these
(
1
non-C2 symmetric catalysts likely operate in a bifunctional
(
(9) Degrado, S. J.; Mizutani, H.; Hoveyda, A. H. J. Am. Chem. Soc.
2001, 123, 755-756.
(10) Luchaco-Cullis, C. A.; Mizutani, H.; Murphy, K. E.; Hoveyda, A.
H. Angew. Chem., Int. Ed. Engl. 2001, 40, 1456-1460.
1
3
(
(11) For related studies on catalytic asymmetric cyanide addition to
imines, see: (a) Iyer, M. S.; Gigstad, K. M.; Namdev, N. D.; Lipton, M. J.
Am. Chem. Soc. 1996, 118, 4910-4911. (b) Sigman, M. S.; Jacobsen, E.
N. J. Am. Chem. Soc. 1998, 120, 4901-4902. (c) Sigman, M. S.; Jacobsen,
E. N. J. Am. Chem. Soc. 1998, 120, 5315-5316. (d) Ishitani, H.; Komiyama,
S.; Kobayashi, S. Angew. Chem., Int. Ed. Engl. 1998, 37, 3186-3188. (e)
Corey, E. J.; Grogan, M. Org. Lett. 1999, 1, 157-160. (f) Ishitani, H.;
Komiyama, S.; Hasegawa, Y.; Kobayashi, S. J. Am. Chem. Soc. 2000, 122,
762-766. (g) Takamura, M.; Hamashima, Y.; Usuda, H.; Kanai, M.;
Shibasaki, M. Angew. Chem., Int. Ed. Engl. 2000, 39, 1650-1652. (h)
Sigman, M. S.; Vachal, P.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl.
2000, 39, 1279-1281. (i) Nogami, H.; Matsunaga, S.; Kanai, M.; Shibasaki,
M. Tetrahedron Lett. 2001, 42, 279-283.
1
M. L.; Hoveyda, A. H. J. Am. Chem. Soc. 2000, 122, 2657-2658. (c) For
a review on catalytic asymmetric additions to imines, see: Kobayashi, S.;
Ishitani, H. Chem. ReV. 1999, 99, 1069-1094. (d) Enders, D.; Reinhold,
U. Tetrahedron: Asymmetry 1997, 8, 1895-1946.
(7) Deng, H.; Isler, M. P.; Snapper, M. L.; Hoveyda, A. H. Manuscript
in preparation.
(8) (a) Porter, J. R.; Traverse, J. F.; Hoveyda, A. H.; Snapper, M. L. J.
Am. Chem. Soc. 2001, 123, 984-985. (b) Porter, J. R.; Traverse, J. F.;
Hoveyda, A. H.; Snapper, M. L. J. Am. Chem. Soc. 2001, 123, 10409-
1
0410.
1
0.1021/ja011875w CCC: $20.00 © 2001 American Chemical Society
Published on Web 11/01/2001

EnantioselectiVe Ti-Catalyzed Strecker Reaction
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11595
Scheme 1
Figure 1. Plot of % ee product vs % ee ligand.
2
the reaction is first order in the catalyst (R ) 0.9973). The
necessary data were obtained by monitoring reaction progress
(disappearance of substrate CdN absorption) with a ReactIR
1
000 instrument. Therefore, the cyanide addition likely involves
manner. The Ti‚Schiff base (SB) coordinates with the substrate,
while an amide moiety within the peptide segment associates
and delivers cyanide to the activated imine. Experimental
evidence and subsequent modeling studies indicate that proper
disposition of different Lewis basic sites within the ligand
structure allow the SB and amide carbonyls in AA1 and AA2
to provide complementary functions, giving rise to high yields
and enantioselectivities.
a monomeric Ti-based complex. A similar protocol was used
to establish the overall rate expression and order for each of
the reaction parameters. In these studies, substrate 5 and ligand
7
were employed (toluene, 18 °C), and i-PrOH was introduced
at a constant rate. The order in 5 was established by determi-
nation of the kinetic rate constants at various concentrations of
the substrate with 10 mol % Ti(Oi-Pr)4 and ligand 7. A plot of
ln[5]0 versus ln k gave a straight horizontal line, indicating a
zero-order dependence on the concentration of the substrate.
When data were collected under the same conditions, but with
varying amounts of TMSCN, a zero-order dependence on
TMSCN was observed. Similar studies with i-PrOH indicated
zero rate order. Catalytic additions carried out in the presence
of TMSCN, i-PrOD, and 7 (pretreated with i-PrOD to ensure
complete exchange of acidic protons) gave k /k ) 1.0 ( 0.08,
Results and Discussion
A key observation regarding the Ti-catalyzed process is that
slow addition of an alcohol (e.g., i-PrOH) to the reaction mixture
6a,b
is required to achieve appreciable levels of conversion (>98%).
We thus reasoned that HCN (or HNC), not TMSCN, may be
participating in the transformation. That is, slow generation of
HCN, from reaction of the alcohol additive and TMSCN, might
ensure minimal levels of uncatalyzed and nonselective C-C
bond formation (HCN can add to imines in the absence of a
catalyst). We subsequently established that slow addition of
cyanide (without added i-PrOH, under identical conditions) does
indeed lead to similar levels of reactivity and asymmetric
H
D
suggesting that the turnover-limiting step of the catalytic cycle
does not involve cleavage of a CH, NH, or an OH bond.
Various structural features of the chiral peptide ligand were
systematically altered, and relative rates and enantioselectivities
were measured. These studies, summarized in Scheme 2, led
to several key observations. (1) Not only is the presence of the
AA2 moiety critical to reactivity and enantioselectivity (compare
6
a
induction as observed with TMSCN in the presence of
i-PrOH.12
7
with 8), its stereochemical identity is of notable significance
Because a number of Ti-catalyzed reactions have been shown
to proceed through pathways involving dimeric or oligomeric
complexes,5 we determined that there is a linear correlation
(compare 7 with 9 and 10). (2) The presence of a more Lewis
basic amide carbonyl (vs a carboxylic ester) influences the rate
of asymmetric cyanide addition. For instance, the initial rate of
reaction (90 min) with ligand 7 is 2.3 times faster than that for
the derived methyl ester 11. The importance of the presence of
AA2 and the significant influence of the related local chirality
suggest that the peptide segment of the ligand might be actiVely
participating in the asymmetric reaction, rather than simply
providing steric differentiation that leads to absolute π-facial
selectivity.
b,13
14
between catalyst and product ee (Figure 1). Subsequent kinetic
-
5
-1
studies (kobs ) 5.35 ( 0.50 × 10
s
at 22 °C) indicate that
(12) An alternative pathway would involve direct participation of
TMSCN (vs HCN), where the effect of added i-PrOH is not to generate
HCN. Rather, the AA2 carbonyl may first react with TMSCN to form a
zwitterionic cyanide complex such as i. After the addition of CN to the
substrate, i-PrOH would react with the silylated amide to generate
i-PrOTMS, releasing the catalytic active ligand. Such a pathway is unlikely,
because treatment of 50 mol % of 7 (cf., Scheme 2) and Ti(Oi-Pr)4 (complex
preformed by removal of i-PrOH in vacuo) with imine 5 and TMSCN leads
to only ∼10% conversion (4 °C, 18 h, toluene; no i-PrOH).
The proposed participation of the AA2 unit suggests that the
turnover-limiting step involves a highly ordered transition state.
Accordingly, activation parameters were measured for the
asymmetric Ti-catalyzed cyanide addition to 5 through kobs
values measured at 5 °C intervals between -15 and +20 °C.
Calculations from an Eyring plot (Figure 2) give ∆H ) 8.98
(
q
-
1
q
-1
-1
1.3 kcal mol and ∆S ) -45.6 ( 4.1 cal K mol .
The large and negative entropy of activation, together with
the aforementioned data, are consistent with a rapid and
(
13) For reviews of nonlinear effects in asymmetric catalysis, see: (a)
Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. Engl. 1998, 37, 2922-
959. (b) Blackmond, D. G. Acc. Chem. Res. 2000, 33, 402-411.
14) See the Supporting Information for all plots and details of kinetic
measurements and spectroscopic studies.
reversible H-bond formation between HNC and AA2 carbonyl
15
(
cf., ester terminus less effective than AA2 amide) and a
2
turnover-limiting cyanide delivery to the Ti-bound substrate via
(
1
6-18
the rigid transition structure in Figure 3.
Subsequent and

11596 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Josephsohn et al.
Scheme 2
peptide complex, the AA2 carbonyl cannot promote C-C bond
formation readily, because the bulky AA2 substituent would
be projected toward the catalyst structure during cyanide delivery
(steric hindrance). The imine substrate, in its lower energy trans
configuration, coordinates in a manner that minimizes steric
interactions with the chiral Ti-peptide complex. It is in such a
substrate-metal ensemble that the amide carbonyl of AA2 can
readily reach and deliver HNC to the CdN bond to produce
the major product enantiomer.
This proposal is consistent with the available kinetic and
stereochemical data, including the importance of the stereo-
chemical identity of the AA2 moiety (cf., Scheme 2). The model
in Figure 3 provides a plausible rationale as to why the absence
of such a substituent leads to diminution in reactivity and
selectivity (compare 7 and 8 in Scheme 2). The stereochemistry
of the AA2 moiety is critical to catalyst preorganization,
rendering H-bonding between HNC and AA2 amide carbonyl
and the formation of the C-C bond entropically more favorable.
A Ti‚peptide complex that bears an AA2 with the alternative
stereochemistry is a less effective chiral catalyst (compare 7
and 10 in Scheme 2), as cyanide delivery would engender costly
steric interactions involving the amino acid substituent and the
O-i-Pr ligands of the transition metal. Finally, substrate coor-
dination to the available equatorial site situates the imine in
such a manner as to give rise to unfavorable steric repulsions
between the phenyl moieties of the N-protecting group and the
AA2 residue, which must remain within a proper distance to
be able to deliver HNC.
a
Enantioselectivities measured by chiral HPLC (chiralpak AD);
1
conversion measured by H NMR.
Extensive spectroscopic studies of Ti‚ligand complexes have
been carried out. Treatment of ligand 2 with 1 equiv of Ti(Oi-
Pr)4 at 22 °C leads to the formation of multiple species (400
1
MHz H NMR; toluene-d8); the resulting solution can be used
to promote cyanide additions to imines efficiently and selec-
tively. When the sample is heated to 100 °C, the spectrum
simplifies significantly and remains so when it is cooled to 22
1
°
C. H NMR analysis indicates the presence of two complexes
1
(
9:1), the major component of which was analyzed by H and
C NMR, gCOSY, and HETCOR spectral analyses, suggest-
1
3
13
2
1
ing that 12 is the major entity. As the representative data in
Scheme 3 indicate, significant downfield shifts in the 13C NMR,
especially of the two amide carbons, imply association of both
the AA1 and AA2 sites to the transition metal center. Disap-
13
Figure 2. Eyring plot for Ti-catalyzed cyanide addition to imine 5.
(
16) 13C NMR spectroscopic studies involving TMS CN do not indicate
the presence of a Ti-CN complex. Similar conclusions are reached after
the results of IR studies are compared with reported literature values. See:
Belokon, Y. N.; Green, B.; Ikonnikov, N. S.; Larichev, V. S.; Lokshin, B.
V.; Moscalenko, M. A.; North, M.; Orizu, C.; Peregudov, A. S.; Timofeeva,
G. I. Eur. J. Org. Chem. 2000, 2655-2661.
(
17) For a related mechanism involving a bimolecular delivery of HCN
to aldehydes, see: Corey, E. J.; Wang, Z. Tetrahedron Lett. 1993, 34, 4001-
004.
18) It is unlikely that the active complex is one where the AA2 carbonyl
is bound to the Ti center (no internal delivery). If this were the case (simple
cyanide addition to the bound imine), the value for entropy of activation
would likely be significantly smaller, because a notably lesser degree of
rigidity would take place within the transition structure. Moreover, in the
absence of a base, it is unlikely that the AA2 carbonyl can displace an
O-i-Pr ligand from the Ti center.
(19) Alternatively, it may be suggested that a rapid and reversible
association of HCN with the AA2 amide bond precedes cyanide addition.
This scenario is less favored on the basis of the substantial pKa difference
between HCN and a protiated amide (∼9 and <0, respectively); see:
Handbook of Chemistry and Physics, 73rd ed.; Lide, D. R., Ed.; CRC
Press: Ann Arbor, MI, 1993.
4
(
Figure 3.
facile loss of proton leads to the formation of the desired amino
nitrile (cf., kH/kD ) 1.0 ( 0.08).19
20
Molecular modelings and calculations involving imine 5
and peptidic Schiff base 7 suggest that, as shown in Figure 3,
association of substrate to one face of the distorted octahedral
Ti-Schiff base complex is energetically more favorable (top
vs bottom face in Figure 3). More importantly, if metal-
substrate association occurs from the alternative face of the Ti-
(20) Calculations were carried out with Spartan SGI (version 5.1.1) at
(
15) Because a deuterium isotope effect is not detected, it follows that
the semiempirical PM3 (tm) level. See the Supporting Information for
details.
(21) For previous reports regarding the structure of related Ti complexes,
see: (a) Hayashi, M.; Miyamoto, Y.; Inoue, T.; Oguni, N. J. Org. Chem.
1993, 58, 1515-1522. (b) Flores-Lopez, L. Z.; Parra-Hake, M.; Somanathan,
R.; Walsh, P. J. Organometallics 2000, 19, 2153-2160 and references therein.
dissociation of N-H bond does not occur during the turnover-limiting step.
These considerations indicate that H-bonding likely involves the N-H of
HNC, not the C-H of HCN, because in the latter case, cleavage of a C-H
bond would have to occur in the slowest step of the catalytic cycle (as the
C-C bond is being formed).

EnantioselectiVe Ti-Catalyzed Strecker Reaction
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11597
Scheme 3. Representative H and 13C NMR Data for the
1
identity of the AA2 moiety is crucial to the reaction efficiencies
and selectivities. Investigations toward establishing the general-
ity of the principles presented here, and their application to the
development of new catalytic asymmetric C-C bond forming
reactions, are in progress.
Unreactive Ti‚L Complex (12)
Experimental Section
General. ReactIR spectra were recorded on an ASI Inc. ReactIR
-
1
spectrophotometer, νmax in cm . ReactIR and QuantIR software are
products of ASI Inc. Infrared (IR) spectra were recorded on a Perkin-
pearance of the AA1 amide NH and the significant downfield
shifts in the H NMR spectrum further support this proposal
-
1
1
Elmer 781 spectrophotometer, νmax in cm . Bands are characterized
1
as broad (br), strong (s), medium (m), and weak (w). H NMR spectra
(∆δ in parentheses). However, we soon discovered that the
were recorded on a Varian Unity 300 (300 MHz) or Varian Gemini
complex generated upon heating, which is stable after cooling
to 22 °C and exhibits the simpler NMR spectra, is catalytically
inactiVe.
2
000 (400 MHz). Chemical shifts in ppm from tetramethylsilane are
reported with the solvent resonance as the internal standard (CHCl , δ
.26; PhCH , δ 2.34). Data are reported as follows: chemical shift,
3
7
3
Several plausible explanations may be put forward for the
inactivity of 12. (1) According to molecular models, AA2
carbonyl association with the Ti center is geometrically more
constrained than that of the AA1 unit. Furthermore, the
covalently bound oxygen ligand of AA1 moiety in 12 is
expected to be less electron donating than an Oi-Pr. Thus, the
relatively electron poor Ti in 12 would be less prone to convert
to a catalytically active cationic Ti complex by loss of an
alkoxide.22 (2) The more rigidly constrained AA2 unit in 12 is
expected to be less effective in delivering HNC to the
coordinated substrate (see Figure 3). Previous considerations
suggest that a complex such as 13 should be more active,
because the derived cationic Ti would be more accessible. An
Oi-Pr ligand may more readily dissociate from a tri-isopropoxide
Ti complex such as 13, because the metal center is more electron
rich by (i) the more accessible AA1 carbonyl as the coordinating
Lewis base (vs AA2 carbonyl) and (ii) the additional stronger
Lewis basic isopropoxide (vs AA1 amide oxygen).23
integration, multiplicity (s ) singlet, d ) doublet, t ) triplet, q )
quartet, br ) broad, m ) multiplet), coupling constants (Hz), and
assignment. 13C NMR spectra were recorded on a Varian Unity 300
(
75 MHz) or Varian Gemini 2000 (100 MHz) with complete proton
decoupling. Chemical shifts are reported in ppm with the solvent as
the internal reference (CDCl , δ 77.0; PhCH , δ 21.0). Enantiomer ratios
3
3
were determined by high-pressure liquid chromatography (HPLC) using
a Shimadzu chromatograph and a chiralpak AD column (4.6 mm ×
2
50 mm chiral column by Chiral Technologies).
All reactions were conducted in oven (135 °C) and flame-dried
glassware under an inert atmosphere of dry argon or nitrogen. DMF
(99.9%), diphenylmethyl amine, tert-leucine, and all commercially
available aldehydes were purchased from Aldrich and used without
further purification. Trimethylsilylcyanide and Ti(Oi-Pr)
purchased from Aldrich and distilled prior to use. All amino acids,
-(3-dimethylaminopropyl)-3-ethyl-carbodiimide (EDC), N-hydroxy-
4
(99.99%) were
1
benzotriazole (HOBt), and piperidine were purchased from Advanced
Chemtech and used without further purification. Toluene was distilled
from sodium/benzophenone ketal. 2-Propanol was distilled from CaSO
-Methoxy-salicyl-t-Leu-Thr(t-Bu)-NHBu (7). Fmoc-Thr(t-Bu) (20
g, 50 mmol) was dissolved in 200 mL of CH Cl and cooled to 0 °C.
HOBT (7.7 g, 50 mmol), EDC (9.6 g, 50 mmol), and n-butyl amine
11 mL, 110 mmol) were added directly to the reaction mixture in
succession. The solution was allowed to warm to 22 °C and stirred for
2 h. The reaction was quenched upon addition of 100 mL of a 10%
4
.
5
2
2
Conclusions
(
Findings disclosed herein demonstrate that, by the virtue of
their structural and stereochemical identity, peptidic Schiff bases
can serve as bifunctional catalysts to deliver appreciable
reactivity and high enantioselectivity. It must be noted that,
although the above mechanistic hypotheses offer a plausible and
consistent rationale that accounts for numerous features of the
asymmetric Ti-catalyzed cyanide additions, they fail to explain
the more subtle attributes of this class of transformations. One
such example involves variations in the Schiff base (SB) moiety
of the optimal chiral ligand as a function of the identity of
different substrates.6 It is, after all, such subtle energetic
modulations that often make it nearly impossible to predict, a
1
citric acid solution, and the mixture was washed with 2 × 200 mL
portions of citric acid and 2 × 100 mL portions of saturated aqueous
NaHCO
dried over MgSO
solid. This solid was dissolved in 100 mL of CH
10 mL, 100 mmol) was added at 22 °C. After 2 h, the mixture was
concentrated in vacuo and purified by silica gel chromatography
gradient from ethyl acetate to 4:1 ethyl acetate/MeOH) to afford H N-
Thr(t-Bu)-Bu as a light orange oil. This oil was dissolved in 100 mL
of CH Cl , HOBT (6.2 g, 40 mmol) and EDC (7.7 g, 40 mmol) were
added, and the reaction was subsequently cooled to 0 °C. Fmoc-t-Leu
14 g, 40 mmol) and Et N (7.0 mL, 50 mmol) were introduced, and
3
. The organic layer was washed with 1 × 100 mL of brine,
4
, filtered, and concentrated in vacuo to afford a white
Cl , and piperidine
2
2
(
(
2
2
2
1
priori, the exact identity of the most effective catalyst.
(
3
Nonetheless, the hypotheses presented herein are important,
because they bear notable implications vis- a´ -vis mechanisms
of other asymmetric processes promoted by this class of ligands.
For instance, our initial studies indicate that, in a similar fashion
as detailed previously, in the Zr-catalyzed alkylations of imines
the resulting solution was stirred for 12 h at 22 °C. The reaction was
quenched upon addition of 50 mL of a 10% citric acid solution. The
organic layer was removed and subsequently washed with 1 × 100
mL of an aqueous 10% citric acid solution and 2 × 100 mL solutions
3
of saturated aqueous NaHCO . The organic layer was washed with 1
5
× 50 mL of brine, dried over MgSO
vacuo to afford a white solid. The resulting solid was dissolved in 50
mL of CH Cl , and piperidine (8.0 mL, 81 mmol) was added at 22 °C.
4
, filtered, and concentrated in
by dialkylzincs and Cu-catalyzed conjugate additions of
6
dialkylzincs to enones, the presence and the stereochemical
2
2
After 2 h, the mixture was concentrated in vacuo and purified by silica
gel chromatography (gradient from ethyl acetate to 4:1 ethyl acetate/
(
22) For a study on the effect of electron density on alkoxide ligand
exchange in Ti complexes, see: Campbell, C.; Bott, S.; Larsen, R.; Van
DerSluys, W. Inorg. Chem. 1994, 33, 4950-4958.
MeOH) to afford H
resulting solid was dissolved in 20 mL of CH
5
1
2
N-t-Leu-Thr(t-Bu)-Bu as a white solid. The
Cl and treated with
-methoxysalicylaldehyde (3.5 mL, 28 mmol) and MgSO
(100 mg ×
0 mmol) for 16 h at 22 °C. The solution was filtered and concentrated
(23) It is less likely that dissociation of the AA2 carbonyl from 12 would
2
2
yield an active Ti complex. This would lead to the formation of a
coordinatively unsaturated, but neutral, Ti center, which may only weakly
bind the imine substrate and not provide efficient activation towards cyanide
addition. In a similar fashion, association of an Oi-Pr instead of the AA1
carbonyl is possible but would yield a less Lewis acidic transition metal
center.
4
2
in vacuo to afford a yellow solid. Recrystallization from MeOH/H O
(10:1) provided a yellow solid (8.3 g, 17 mmol, 35% yield). IR (neat,
NaCl): 3322 (br), 2964 (s), 2927 (s), 2865 (m), 1664 (s), 1497 (s),

11598 J. Am. Chem. Soc., Vol. 123, No. 47, 2001
Josephsohn et al.
1
356 (m), 1275 (s), 1084 (m), 1041 (m), 776 (m), 777 (m) cm-1. 1H
flask was charged with Ti(Oi-Pr)
4
(60 µL of a 0.10 M solution in
NMR (CDCl , 400 MHz): 12.06 (1H, s, OH), 8.20 (1H, s, HCdN),
.19 (1H, d, J ) 3.8, aromatic H), 6.92-6.89 (3H, m, aromatic H and
Thr NH), 6.74 (1H, s, aromatic H), 4.24-4.16 (2H, m, Thr-RH and
Thr-âH), 3.74 (3H, s, OCH ), 3.50 (1H, s, Tle-RH), 3.34-3.24 (1H,
m, NHCH Pr), 3.18-3.08 (1H, m, NHCH Pr), 1.44-1.36 (2H, m,
NHCH CH Et), 1.33-1.19 (12H, m, Thr C(CH and NH(CH CH
CH ), 1.03 (9H, s, Tle C(CH ), 0.85 (6H, apparent t, Thr âCH and
NH(CH ). C NMR (CDCl , 100 MHz): 170.9, 169.7, 167.2,
3
toluene, 0.010 mmol), and the resulting yellow solution was stirred
for 10 min at 22 °C. Subsequently, the imine (0.10 mmol) was added
as a solid. The reaction vessel was capped with a septum, sealed with
Teflon tape, and transferred from the glovebox to a room maintained
at 4 °C. The flask was equipped with a balloon of argon, and TMSCN
(27 µL, 0.20 mmol) was added by syringe to the stirred solution. i-PrOH
(15 µL, 0.20 mmol) in toluene (1.0 mL) was added dropwise by a
syringe pump over 20 h (50 µL/h), with additional stirring for another
7
3
2
2
2
2
)
3 3
)
2 2
2
-
3
)
3 3
3
1
3
2
)
3
CH
3
3
1
4
C
56.0, 153.1, 121.2, 119.1, 118.9, 116.2, 85.2, 76.4, 67.1, 58.3, 37.0,
4 h. Reaction was quenched through addition of wet Et O (5 mL) and
2
0.2, 36.2, 32.5, 29.4, 28.4, 21.2, 18.6, 14.9. HRMS (ES) calcd for
passed through a plug of silica gel with 3 mL of CH Cl . Purification
2
2
+
26
H N O
44 3 5
(M + H ), 478.3281; found, 478.3284.
by silica gel chromatography and conversion and enantioselectivity
levels were determined by chiral HPLC analysis (chiralpak AD, 90:10
EtOAc/hexanes, 1 mL/min, 254 nm).
Diphenylmethylaminophenylacetonitrile (6). The general proce-
dure was followed with 5-methoxy-salicyl-t-Leu-Thr(t-Bu)-NHBu (7)
5
-Methoxy-salicyl-t-Leu-(D)-Thr(t-Bu)-NHBu (10). IR (neat,
NaCl): 3326 (br), 2961 (s), 1652 (s), 1589 (m), 1459 (s), 1362 (m),
280 (s), 1086 (m), 1035 (m), 991 (m), 783 (m), 733 (m) cm-1. H
NMR (CDCl , 400 MHz): δ 12.07 (1H, s), 8.21 (1H, s), 7.05 (1H, d,
J ) 5.7), 6.93-6.88 (2H, m), 6.75 (1H, d, J ) 2.4), 4.30 (1H, dd, J )
1
1
3
(49 mg, 0.10 mmol), Ti(Oi-Pr) (0.10 mmol, 0.20 mL of a 0.50 M
4
3
3
.9, 5.7), 4.07 (1H, dq, J ) 6.4, 3.8), 3.73 (3H, s), 3.52 (1H, s), 3.55-
.28 (1H, m), 3.21-3.13 (1H, m), 1.48-1.41 (2H, m), 1.37-1.28 (2H,
solution in toluene), benzaldehyde diphenylmethyl imine (5) (270 mg,
1.0 mmol), and TMSCN (270 µL, 2.0 mmol) in toluene (2.5 mL).
i-PrOH (150 µL, 2.0 mmol) in toluene (2.0 mL) was added over 20 h
with additional stirring for 10 h. The unpurified reaction mixture was
passed through a plug of silica with CH Cl (5.0 mL), and volatiles
1
3
3
m), 1.20 (9H, s), 1.02 (9H, s), 0.89-0.84 (6H, m). C NMR (CDCl ,
1
1
1
00 MHz): δ 170.3, 169.0, 166.1, 155.3, 152.3, 120.3, 118.2, 118.1,
15.3, 84.3, 75.5, 66.1, 57.2, 56.1, 39.3, 35.2, 31.6, 28.3, 27.3, 20.2,
2
2
+
7.3, 13.8. HRMS (ES) calcd for C26
H
44
N
3
O
5
(M + H ), 478.3281;
were removed in vacuo. The resulting pale yellow solid was recrystal-
found, 478.3272.
-Methoxy-salicyl-t-Leu-Gly-NHBu (9). IR (neat, NaCl): 3307 (br),
955 (s), 2867 (m), 1652 (s), 1488 (s), 1274 (m), 1161 (m), 1035 (m)
lized from 5:1 hexanes/CH Cl to afford 230 mg of 6 as a white solid
2
2
5
(82% yield, >99% ee). (The unpurified reaction mixture gave 99%
conversion and 97% ee by HPLC.) IR (CCl , NaCl): 3327 (m), 3087
(s), 3069 (s), 3031 (s), 2848 (m), 2231 (w), 1948 (m), 1879 (m), 1810
2
4
-
1
1
cm . H NMR (CDCl
3
, 400 MHz): δ 12.02 (1H, s), 8.21 (1H, s),
-
1 1
6
.93 (1H, d, J ) 2.9), 6.90 (1H, s), 6.77-6.61 (1H, m), 6.75 (1H, d,
(m), 1608 (m), 1501 (s), 1451 (s), 1187 (m), 929 (m) cm . H NMR
(CDCl , 400 MHz): δ 7.58-7.20 (15H, m), 5.24 (1H, s), 4.59 (1H, d,
J ) 12.0), 2.14 (1H, d, J ) 12.0). C NMR (CDCl
J ) 2.7), 6.34-6.27 (1H, m), 3.94 (1H, dd, J ) 5.6, 16.1), 3.80 (1H,
dd, J ) 5.6, 16.1 Gly RH), 3.77 (3H, s), 3.55 (1H, s), 3.22-3.17 (2H,
3
13
3
, 100 MHz): δ
m), 1.45-1.22 (4H, m), 1.02 (9H, s), 0.85 (3H, apparent t, 7.1). 13
C
143.4, 141.7, 135.6, 129.7, 129.6, 129.4, 128.6, 128.4, 128.1, 127.9,
+
NMR (CDCl
3
, 100 MHz): δ 171.2, 168.7, 166.8, 155.1, 152.5, 120.6,
127.7, 119.4, 66.2, 53.0. HRMS (ES) calcd for C21
299.1548; found, 299.1549. Anal. Calcd for C21
6.08; N, 9.39. Found: C, 84.22; H, 6.19; N, 9.28. [R]
(c 5.0, CHCl ).
H
18
N
2
(M + H ),
1
18.2, 115.5, 83.8, 56.1, 43.4, 39.5, 35.4, 31.7, 27.3, 20.2. HRMS (ES)
18 2
H N : C, 84.53; H,
+
24
calcd for C20
H N O
32 3 4
(M + H ), 378.2393; found, 378.2380.
D
-64.2 ( 0.1°
5
-Methoxy-salicyl-t-Leu-Thr(t-Bu)-OMe (11). IR (neat, NaCl):
3
3
1
417 (br), 2971 (s), 1749 (s), 1674 (s), 1495 (s), 1273 (s), 1204 (m),
Study of Nonlinear Effects. In a glovebox, 98 µL (0.010 mmol) of
a Ti(Oi-Pr) solution (60 µL Ti(Oi-Pr) in 2.0 mL of toluene) was added
-
1 1
078 (m), 777 (w) cm . H NMR (CDCl , 400 MHz): δ 12.17 (1H,
3
4
4
s), 8.27 (1H, s), 6.92 (1H, d, J ) 2.7), 6.89 (1H, s), 6.79 (1H, d, J )
.8), 6.68 (1H, d, J ) 8.8), 4.38 (1H, dd, J ) 9.0, 1.6) 4.18 (1H, dq,
J ) 6.4, 1.8), 3.75 (3H, s), 3.59, (3H, s), 3.54 (1H, s), 1.19 (3H, d, J
to 200 µL (0.010 mmol) of a 0.05 M solution of (L)-Ligand 7 (72 mg
in 2.6 g of toluene) and allowed to stir at 22 °C for 5 min. Imine 5 (27
mg, 0.10 mmol) was added, followed by 300 µL of toluene. The flask
was sealed with a septum and moved from the glovebox to a cold room
maintained at 4 °C. The flask was equipped with a syringe containing
1 mL of a solution of i-PrOH (15 µL in 1.0 mL toluene) affixed to a
digital (electronically controlled) syringe pump (50 µL/h addition rate).
The reaction vessel was allowed to equilibrate to +4 °C for 5 min
with stirring, at which time 22 µL (0.20 mmol) of TMSCN was added
through syringe and the addition of i-PrOH was initiated by the syringe
pump. After 18 h, the reaction was quenched with wet ether and
concentrated in vacuo. The yellow solid was dissolved in ethyl acetate
2
1
3
)
3
6.2), 1.05 (9H, s), 1.02 (9H, s). C NMR (CDCl , 100 MHz): δ
1
7
71.2, 170.9, 166.7, 155.3, 152.4, 120.4, 118.3, 118.0, 115.4, 83.8,
4.2, 67.2, 58.0, 56.1, 52.35, 35.1, 28.4, 27.3, 21.8. HRMS (ES) calcd
+
for C19
H N
29 2
O
6
(M + H - C(CH
3
)
3
), 381.2026; found, 381.2038.
5
-Methoxy-salicyl-t-Leu-NHBu (8). IR (neat, NaCl): 3310 (br),
2
7
959 (s), 2865 (m), 1649 (s), 1492 (s), 1273 (s), 1160 (m), 1028 (m),
-
1 1
77.8 (w) cm . H NMR (CDCl , 400 MHz): δ 12.18 (1H, s), 8.21
3
(
5
3
1H, s), 6.94 (1H, d, J ) 2.9), 6.90 (1H, s), 6.77 (1H, d, J ) 2.9),
.91-5.84 (1H, m), 3.75 (3H, s), 3.52 (1H, s), 3.37-3.28 (1H, m),
.15-3.07 (1H, m), 1.48-1.40 (1H, m), 1.33-1.25 (2H, m), 1.01 (9H,
2 2
and passed through a plug of silica gel with 3 mL of CH Cl . The
13
1
s), 0.86 (3H, t, J ) 7.3). C NMR (CDCl
3
, 100 MHz): δ 170.0, 166.5,
resulting solution was concentrated in vacuo and analyzed by H NMR
spectroscopy and chiral HPLC (see general procedure above), indicating
the formation of the desired amino nitrile in 93% ee (normalized (set
to 1) for plotting purposes). Additional data, as illustrated in Figure 2,
were collected in a similar fashion.
1
2
3
54.9, 152.6, 120.5, 118.2, 117.9, 115.4, 84.0, 56.1, 39.2, 35.1, 31.8,
+
7.3, 20.3, 13.9. HRMS (ES) calcd for C18
21.2178; found, 321.2172.
29 2 3
H N O (M + H ),
Procedure for synthesis of benzaldimine 5: Benzaldehyde (50 mmol,
.0 mL), diphenylmethylamine (50 mmol, 8.6 mL), and MgSO (1.0
5
4
Measurement of Initial Rate Difference between Reactions of
5-Methoxy-salicyl-t-Leu-Thr(t-Bu)-Bu (7) and 5-Methoxy-salicyl-
t-Leu-Thr(t-Bu)-OMe (11). In a glovebox, an oven-dried ReactIR flask
was charged with 200 µL (0.010 mmol) of a 0.05 M solution of ligand
7 (0.12 g in 5.0 mL toluene), 400 µL (0.10 mmol) of a 0.25 M solution
of imine 5 (1.0 g in 15 mL of toluene), and 200 µL (0.010 mmol) of
g) in benzene (100 mL) were stirred for 8 h at 22 °C. The solution was
filtered and concentrated in vacuo. The product was purified by
2
recrystallization (Et O, hexanes) to provide a white solid (12.6 g, 46
mmol, 93% yield).
Benzaldehyde Diphenylmethyl Imine (5). IR (CCl
4
, NaCl): 3081
(
7
w), 3062 (w), 3037 (w), 2842 (w), 1646 (m), 1495 (m), 1457 (m),
a 0.05 M solution Ti(Oi-Pr) (74 µL of Ti(Oi-Pr)
4
4
in 5.0 mL of toluene).
-
1 1
46 (m), 690 (s) cm . H NMR (CDCl
3
, 400 MHz): δ 8.40 (1H, d, J
Upon addition of Ti(Oi-Pr) , the solution turned from yellow to bright
4
)
2.4), 7.84-7.81 (2H, m), 7.41-7.20 (13H, m), 5.58 (1H, d, J )
orange. The flask was removed from the glovebox, and the IR probe
was submerged into the reaction mixture. A 1 mL syringe containing
a solution of i-PrOH (18 µL, 0.24 mmol in 1 mL of toluene), connected
to an electronically controlled syringe pump (50 µL/h addition rate),
was attached to the reaction flask. ReactIR data collection was initiated
immediately after TMSCN (22 µL, 0.20 mmol) was added to the
reaction mixture and addition of the i-PrOH solution was started through
a syringe pump. IR spectra were collected at 1 min intervals for 2 h
13
2.4). C NMR (CDCl
129.1, 129.1, 128.3, 127.6, 78.6. HRMS (ES) calcd for C20
271.1361; found, 271.1354. Anal. Calcd for C20 17N: C, 88.52; H,
6.31; N, 5.16. Found: C, 88.59; H, 6.34; N, 5.04.
Procedure for Ti-Catalyzed Cyanide Addition to Imines. In a
3
, 100 MHz): δ 161.4, 144.5, 137.0, 131.4, 129.2,
H17N,
H
glovebox, chiral ligand (0.010 mmol) was placed into a flame-dried
round-bottomed flask and was dissolved in toluene (0.50 mL). The

EnantioselectiVe Ti-Catalyzed Strecker Reaction
J. Am. Chem. Soc., Vol. 123, No. 47, 2001 11599
and were imported into the QuantIR software where the spectra were
analyzed applying the calibration curve (see the Supporting Informa-
tion).
Eyring Plot Measurements. In a glovebox, an oven-dried ReactIR
flask was charged with 100 µL (0.012 mmol) of a 0.12 M solution of
ligand 7 (400 mg in 8.1 g toluene), 100 µL (0.12 mmol) of a 1.2 M
solution of imine 5 (5.0 g in 18 g of toluene), 24 µL (0.012 mmol) of
to the reaction mixture and addition of the i-PrOH solution was started
through a syringe pump. IR spectra were collected at 1 min intervals
for 2 h and imported into the QuantIR software where the spectra were
analyzed by applying the calibration curve (see the Supporting
Information).
Measurement of Kinetic Isotope Effects. Chiral ligand 7 (25 mg,
0.05 mmol) was dissolved in i-PrOD and allowed to stir under an
atmosphere of Ar for 2 days. At this time, the solution was concentrated
in vacuo. Analysis of 1H NMR of the recovered ligand indicated
complete (>98%) disappearance of amide and phenol proton signals.
A ReactIR vessel was charged with 4.8 mg (0.010 mmol) of deuterated
7, 27 mg (0.10 mmol) of imine 5, 500 µL of toluene, and 98 µL (0.010
a 0.50 M solution of Ti(Oi-Pr)
4
(30 µL of Ti(Oi-Pr)
4
in 2.0 mL of
toluene), and 280 µL of toluene. Upon addition of Ti(Oi-Pr)
4
, the
solution turned from yellow to bright orange. The flask was removed
from the glovebox, and the IR probe was immersed into the mixture
which is then submerged into an ethylene glycol bath maintained at
-
15 °C by a circulation bath. A 1 mL volume of a solution of i-PrOH
18 µL, 0.24 mmol in 1.0 mL of toluene) was placed in a syringe
mmol) of a 0.10 M solution of Ti(Oi-Pr)
4
in toluene. Upon addition of
(
Ti(Oi-Pr) , the solution turned from yellow to bright orange in color.
4
connected to an electronically controlled syringe pump (50 µL/h
addition rate) attached to the flask. TMSCN (22 µL, 0.20 mmol) was
added, and i-PrOH (electronically controlled addition) addition was
initiated. IR spectra were collected at 1 min intervals for 2-4 h and
imported into the QuantIR software where the spectra were analyzed
applying the calibration curve (see the Supporting Information).
Representative Measurement of Reaction Rate. (See the Sup-
porting Information for full and further details.) In a glovebox, an oven-
dried ReactIR flask was charged with 250 µL (0.030 mmol) of a 0.12
M solution of ligand 7 (400 mg in 8.1 g of toluene), 250 µL (0.30
mmol) of a 1.2 M solution of imine 5 (5.0 g in 18 g of toluene), and
The flask was removed from the glovebox, and the IR probe was
submerged into the mixture. A 1 mL syringe containing a solution of
i-PrOD (15 µL, 0.20 mmol in 1 mL of toluene), connected to an
electronically controlled syringe pump (50 µL/h addition rate), was
attached to the flask. Data collection was carried out in the manner
described previously.
Acknowledgment. This research was supported by the
National Institutes of Health (GM-57212). We thank Professors
D. L. McFadden (Boston College), L. T. Scott (Boston College),
W. von E. Doering (Harvard), and C. L. Perrin (University of
California, San Diego) and Mr. James R. Porter (Boston
College) for helpful discussions.
3
0 µL (0.030 mmol) of a 1.0 M solution of Ti(Oi-Pr)
4
(60 µL of Ti-
, the solution
(Oi-Pr) in 2 mL of toluene). Upon addition of Ti(Oi-Pr)
4
4
turned from pale yellow to bright orange. The reaction flask was
removed from the glovebox, and the IR probe was submerged into the
reaction mixture. A 1 mL volume of a solution of i-PrOH (46 µL, 0.60
mmol in 1 mL of toluene) was placed in a syringe connected to an
electronically controlled syringe pump (50 µL/h addition rate) which
was in turn attached to the reaction vessel. ReactIR data collection
was initiated immediately after TMSCN (75 µL, 0.60 mmol) was added
Supporting Information Available: Detailed data on vari-
ous measurements (19 pages, PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
JA011875W