Enantioselective enolate protonation: Matching chiral aniline and substrate acidity

Article abstract of DOI:10.1021/jo990897m

A comparison of chiral anilines 1a-f in the asymmetric protonation of enolate 15 shows that the optimum ΔpK(a) value (chiral acid vs protonated enolate) for the highest enantioselectivity is ca. 3 (Table 2). An extension of this concept to amino acid enolates was possible, and 1e was found to give the best enantioselectivity (85% ee) with the alanine-derived N-lithioenolate 5a (Table 3). Changes in aniline pK(a) due to variation of substituents at the aniline nitrogen were evaluated briefly, but these changes did not show consistent trends in the enantioselectivity vs pK(a).

Full text of DOI:10.1021/jo990897m

J . Org. Chem. 1999, 64, 7863-7870
7863
En a n tioselective En ola te P r oton a tion : Ma tch in g Ch ir a l An ilin e
a n d Su bstr a te Acid ity
E. Vedejs,* A. W. Kruger, and E. Suna
Chemistry Department, University of Wisconsin, Madison Wisconsin 53706
Received J une 2, 1999
A comparison of chiral anilines 1a -f in the asymmetric protonation of enolate 15 shows that the
optimum ∆pKa value (chiral acid vs protonated enolate) for the highest enantioselectivity is ca. 3
(Table 2). An extension of this concept to amino acid enolates was possible, and 1e was found to
give the best enantioselectivity (85% ee) with the alanine-derived N-lithioenolate 5a (Table 3).
Changes in aniline pKa due to variation of substituents at the aniline nitrogen were evaluated
briefly, but these changes did not show consistent trends in the enantioselectivity vs pKa.
Enolate desymmetrization by protonation with chiral
acids has reached practical levels of enantioselectivity in
a number of studies.1 In the best examples, ee values in
the range of 95-99% have been demonstrated, and
catalytic as well as stoichiometric procedures are known
at this level of enantioselectivity. Progress with the
mechanistic aspects of the proton-transfer step has been
slower. Enolates have several structural options in solu-
tions that contain other species such as lithium amides,
lithium halides, enolate dimers, mixed aggregates, and
so on.2 Experimental variables are also potentially com-
plex. In our own work, the optimized procedures have
typically been developed by empirical means.1b,3 The
effort required to isolate and to understand individual
variables has inevitably taken much more time than the
empirical optimization. On the other hand, the need to
improve our understanding of the key variables is clear.
Despite extraordinary levels of enantioselectivity for the
best applications, the available chiral acids tend to be
effective only for a narrow range of enolate substrates.
As discussed in a recent review by Fehr,1a highly
enantioselective proton transfer between a chiral acid
and a prochiral enolate is likely only if the rate of proton
transfer is slow. A strongly exothermic R-carbon proto-
nation event would probably be unselective, especially if
the chiral acid is so potent that proton-transfer ap-
proaches diffusion control. On the other hand, the ideal
chiral acid A*-H for a given enolate substrate E(-) must
be strong enough to completely protonate the enolate and
to resist the reverse reaction where the chiral carbonyl
product E-H is deprotonated reversibly by the anion
A*(-). This is essential if formation of racemic E-H is
to be avoided.
Complete protonation of the enolate requires proton
transfer on a convenient laboratory time scale and
therefore involves both the kinetic and the thermody-
namic acidity of the chiral acid. The latter term can be
evaluated by comparing the pKa’s of the chiral acid and
the protonated enolate, while the former (kinetic acidity)
can be approximated by assuming that the rate of proton
transfer will increase as the ∆pKa (the value of [pKaE-H]-
[pKaA*-H]) increases.4 However, there has been limited
information regarding the acceptable range of ∆pKa
values (chiral acid vs protonated enolate) that allow
enantioselective proton transfer. In a prior investigation
from this laboratory, evidence was presented that ∆ pKa
should be ca. 3, based on the behavior of two enolates
that differ in the degree of anion stabilization. However,
some of the earliest promising examples of enolate
desymmetrization (Duhamel et al.) used chiral carboxylic
acids as the proton donors, and the ∆pKa relevant to these
experiments appears to be considerably larger.5
We have considered the possibility that an optimum
∆pKa range might be identified for different classes of
enolate substrates in asymmetric protonation using
structurally similar chiral acids based on the chiral
diamine skeleton 1. Previous studies have established
that the commercially available p-chloro derivative 1b
functions as an excellent chiral “acid” in the enantiose-
lective protonation of strongly basic â,γ-unsaturated
amide enolates.1b If ∆pKa is an important variable, then
the issue can be probed by studying derivatives of 1
(Scheme 1) where the substituent X is varied systemati-
cally to change acidity. Alternatively, chiral acid pKa can
be adjusted by varying the aniline nitrogen substituent.
In principle, this could be an easier approach starting
from the aniline 2, but enantiomerically pure 2 is not
easy to prepare.6 On the other hand, a dimethoxy
analogue 3a is readily available via an asymmetric
hydrogenation approach and the N-methyl derivative 3b
is comparable to 1a as a chiral acid in the protonation of
(1) (a) Review: Fehr, C. Angew. Chem., Int. Ed. Engl. 1996, 35, 2567.
(b) Vedejs, E.; Lee, N.; Sakata, S. T. J . Am. Chem. Soc. 1994, 116,
2175. (c) Asensio, G.; Aleman, P.; Cuenca, A.; Gil, J .; Medio-Simon,
M. Tetrahedron: Asymmetry 1998, 9, 4073. Takeuchi, S.; Nakamura,
Y.; Ohgo, Y.; Curran, D. P. Tetrahedron Lett. 1998, 39, 8691. Asensio,
G.; Aleman, P.; Gil, J .; Domingo, L. R.; Medio-Simon, M. J . Org. Chem.
1998, 63, 9342. Yanagisawa, A.; Kikuchi, T.; Kuribayashi, T.; Yama-
moto, H. Tetrahedron 1998, 54, 10253. Yanagisawa, A.; Inanami, H.;
Yamamoto, H. J . Chem. Soc., Chem. Commun. 1998, 1573. Aboulhoda,
S. J .; Reiners, I.; Wilken, J .; Henin, F.; Martens, J .; Muzart, J .
Tetrahedron: Asymmetry 1998, 9, 1847. Mikami, K.; Yamaoka, M.;
Yoshida, A. Synlett 1998, 607. Yanagisawa, A.; Ishihara, K.; Yama-
moto, H. Synlett 1997, 411 and references therein. (d) Martin, J .; Lasne,
M-C.; Plaquevent, J -C.; Duhamel, L. Tetrahedron Lett. 1997, 38,
7181.
(4) Eigen, M. Angew. Chem., Int. Ed. Engl. 1964, 3, 1.
(5) Duhamel, L.; Plaquevent, J . C. J . Am. Chem. Soc. 1978, 100,
7415. Duhamel, L.; Duhamel, P.; Launay, J -C.; Plaquevent, J -C. Bull.
Soc. Chim. Fr. 1984, II, 421.
(6) Vedejs, E.; Trapencieris, P.; Suna, E. J . Org. Chem. 1999, 64,
6724.
(2) Seebach, D. Angew. Chem., Int. Ed. Engl. 1988, 27, 1624.
Williard, P. G.; Hintze, M. J . J . Am. Chem. Soc. 1987, 109, 5539.
Henderson, K. W.; Dorigo, A. E.; Liu, Q-Y.; Williard, P. G.; Schleyer,
P. von R.; Bernstein, P. R. J . Am. Chem. Soc. 1996, 118, 1339.
(3) Vedejs, E.; Lee, N. J . Am. Chem. Soc. 1995, 117, 891.
10.1021/jo990897m CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/23/1999

7864 J . Org. Chem., Vol. 64, No. 21, 1999
Vedejs et al.
Ta ble 1. p Ka (DMSO) of p-Su bstitu ted An ilin es 8 a n d
Sch em e 1
Ch ir a l Acid s 1
aniline 8
chiral acid 1
pKa(DMSO)
chiral acid
p-substituent X
pKa(DMSO)a
1
H
Cl
CF3
CO2Et
Ts
30.7
29.4
27.0
26.5d
24.9e
20.9
29.0b
27.7c
25.3b
24.8b
23.3b
19.3b
1a
1b
1c
1d
1e
1f
NO2
a
b
pKa(DMSO) from ref 8b unless noted otherwise. Estimated
as described in text. c Reference 1b. Calculated from σ- ) 0.74
d
p
and F ) 5.67 using the Hammett equation, pKa p-CO R - pKa p-H
) ∆pKa ) Fσp according to ref 9. e The pKa value for ph2enylsulfonyl
is listed as an estimate for toluenesulfonyl.
on variations in the aniline ring substituent in 1. The
pKa values of the series of chiral diamines 1 can be
estimated from the known pKa’s of the parent anilines
(Table 1) by comparison with the value determined by
Bordwell and Satish for the commercially available 1b.
The measured pKa for 1b is 27.7 (DMSO conditions),8a
while the parent p-chloroaniline 8b has a DMSO pKa of
29.4.8b The difference of 1.7 pKa units reflects the net
contribution by the N-methyl group and the tetrahy-
droisoquinoline subunit. We will assume that the same
correction of 1.7 pKa units can be applied to the other
anilines 8 as an approximate way to estimate the pKa
values of the structurally related diamines 1. Thus,
diamines having a range of DMSO pKa values from ca.
19-29 would be available for study as the aromatic
substituent in 1 is modified from strong acceptors such
as X ) NO2 (1f) to the unsubstituted aniline 1a (X ) H).
Because the variable substituent X is relatively far from
the aniline nitrogen, the asymmetric protonation of
enolates would encounter little if any change in steric
effects as the pKa is changed. Electronic factors would
also be less than in the examples 3c or 3d (where the
nitrogen substituents were altered), although a trend
toward sp2 hybridized nitrogen might be expected in 1f
and perhaps also in 1d due to delocalization involving
the nitrogen electron pair and the para acceptor group.
The synthesis of the chiral diamines (Scheme 2) began
with the commercially available 1b. Conversion to the
aminal 9 was easily carried out using isobutyraldehyde
in the presence of acetic acid. With the N-H bonds
temporarily blocked, 9 could be transformed into the
Grignard reagent 10. Forcing conditions were necessary,
but mechanically activated magnesium in refluxing THF
proved sufficient for essentially complete chlorine-
magnesium exchange.10 Upon quenching the Grignard
solution with aqueous ammonium chloride, 11 was
recovered in high yield. Alternatively, electrophilic trap-
ping with diethyl carbonate11a or p-toluenesulfonyl
fluoride11b gave the ester 12 or the sulfone 13 in 71% and
85% yield, respectively. Hydrolytic cleavage of the ami-
nals 11, 12, or 13 with dilute HCl gave the desired
diamines 1a , 1d , and 1e. The unsubstituted diamine 1a
amide enolates. Conversion of 3a to 3c has already been
reported, and 3d can be made in a similar way (Cbz
protection of the secondary amine followed by N-benzo-
ylation and deprotection).6 The DMSO pKa values for 3c
and 3d can be estimated as ca. 12 and 19, respectively,
by comparison with the parent aniline derivatives (N-
phenylsulfonylaniline, pKa 11.95; N-benzoylaniline, pKa
) 18.77) studied by Bordwell et al.7 However, preliminary
attempts to find possible applications with carboxylic acid
derived enolates were not promising. Thus, the sulfona-
mide 3c gave <10% ee in test reactions with ester
enolates 4 or 5a , and only marginally improved results
were obtained with enolates 6 or 7 (36% ee and 28% ee,
respectively). The analogous amide 3d was ineffective in
all cases (<5% ee with 4, 6-7). Only the N-sulfamoyl
derivative 3e gave a promising result among the anilines
containing an electron-withdrawing substituent at ni-
trogen (58% ee with lactone enolate 7), but similar
experiments with 4a (15% ee) and 6 (12% ee) were not
encouraging. Furthermore, 3e was difficult to make and
to purify (N-Cbz protection of 3a followed by treatment
with Me2NSO2Cl/pyridine resulted in a mixture of dis-
proportionation products; see experimental).
The asymmetric protonations with 3c suggest a modest
trend for improved enantioselectivity with the less basic
enolates, as might be expected if the optimum ∆pKa value
should be relatively small. However, no trend is evident
with 3e, and the lack of any significant enantioselectivity
with 3d is problematic for an investigation of related
chiral acids having varied pKa. Other electron-withdraw-
ing nitrogen substituents can be considered, but the
preliminary results indicate that comparisons will be
difficult because steric and electronic changes near the
aniline nitrogen may be large enough to obscure the
possible role of ∆pKa.
(8) (a) The pKa value was previously reported in ref 3. (b) Bordwell,
F. G.; Algrim, D. J . J . Am. Chem. Soc. 1988, 110, 2964.
(9) Hoefnagel, A. J .; Monshouwer, J . C.; Snorn, E. C. G.; Wepster,
B. M. J . Am. Chem. Soc. 1973, 95, 5350. For a review of Hammett
parameters, see: Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991,
91, 165.
An alternative approach to matching the pKa values
of chiral acid and enolate substrate was pursued, based
(10) Baker, K. V.; Brown, J . M.; Hughes, N.; Skarnulis, A. J .; Sexton,
A. J . Org. Chem. 1991, 56, 698.
(7) (a) Bordwell, F. G.; Hughes, D. L. J . Am. Chem. Soc. 1985, 107,
4737. (b) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456. (c) Bordwell,
F. G.; Liu, W.-Z. J . Phys. Org. Chem. 1998, 11, 397.
(11) (a) Frye, L. L.; Sullivan, E. L.; Cusack, K. P.; Funaro, J . M. J .
Org. Chem. 1992, 57, 697. (b) Whitmore, F. C.; Loder, D. J . Organic
Syntheses; Wiley & Sons: New York, 1943; Collect. Vol. II, pp 282.

Enantioselective Enolate Protonation
J . Org. Chem., Vol. 64, No. 21, 1999 7865
Ta ble 2. Asym m etr ic P r oton a tion of Am id e En ola te 15
w ith 1
Sch em e 2
chiral acid
X
pKa(DMSO)
ee (%)
1a
1b
1c
1d
1e
1f
H
Cl
CF3
CO2Et
Ts
29.0a
27.7b
25.3a
24.8a
23.3a
19.3a
90
97b
93
40
37
0
NO2
a
b
Estimated; see Table 1 discussion. Reference 1b.
of the chiral acid was increased (1a ), and complete fading
of the enolate color was slower in this experiment
compared to the others.
The DMSO pKa of the amide 16 is too high for accurate
measurement, but a value of ca. 31 has been estimated.13
Thus, the optimum chiral acid 1b among the available
p-substituted anilines 1 has a ∆pKa value ) ca. 3
compared to the protonated carbonyl product 16. Reason-
ably high enantioselectivities are also obtained with 1a
and with 1c, suggesting that ∆pKa can be in the range
of 2-5. Of course, the measured pKa values are strictly
relevant only to DMSO conditions, while the protonation
experiments were conducted in THF (ion pair conditions).
The ion pair pKa’s for the chiral acids and for the carbonyl
product would probably be five or more pKa units smaller
in THF than in DMSO.14 On the other hand, the ∆pKa
values should vary relatively little, assuming similar
solvent effects within a family of related structures. If
this is correct, then it may be possible to anticipate the
best chiral acid among the derivatives of 1 for a given
enolate by evaluating estimated DMSO pKa values.
Among the enolates considered earlier, the lactone-
derived 7 should have the lowest pKa value, previously
estimated as pKa ) ca. 20 in DMSO.3 According to the
pKa’s in Table 2, only the most acidic p-nitro aniline 1f
would have any chance for effecting the direct proton
transfer to 7. However, treatment of 7 with 1f gave
racemic lactone product. Since the estimated ∆pKa ) ca.
1, this is not a surprising result. Proton transfer should
be reversible, and racemization could occur if the initial
protonation event is enantioselective. More strongly basic
enolates are needed to match the chiral aniline series 1
with ∆pKa in the range of 3-5 where optimum enanti-
oselectivity is expected.
We could find no pKa data for enolates such as 4, 5, or
6, but it is possible that 5 would be the most strongly
basic among these substrates due to electron repulsion
in the dianion (more accurately, the dilithiated amide).
This enolate was therefore studied in asymmetric pro-
tonation experiments using several of the more readily
available anilines 1. As shown in Table 3 (entries 1-4),
the optimum chiral acid proved to be the p-toluenesulfo-
nyl derivative 1e (85% ee). The less acidic analogues 1d
and 1b gave similar, but lower, enantioselectivities, while
the most acidic 1f afforded racemic methyl N-benzoyl-
alaninate. Structural modifications in the nitrogen sub-
stituent or the ester O-alkyl group in alanine-derived
was also prepared by a nickel-catalyzed dechlorination
with LiAlH4.
For access to the more acidic p-nitro analogue 1f, the
direct nitration of 1a was briefly explored. However, this
reaction proved difficult to control for the desired regio-
isomer. Better results were obtained by nitrating the
aminal 11 with nitronium tetrafluoroborate at -40 °C.
This gave the mono-nitro derivative 14, and hydrolysis
to 1f proceeded without complications (55% overall yield).
The site of nitration was confirmed by X-ray crystal-
lography.
The trifluoromethyl diamine 1c was prepared as previ-
ously described via a Bischler-Napieralski cyclization
sequence, followed by reduction and resolution.12 The
resolution requires multiple crystallizations of the tar-
trate salt,12b so the quantity of 1c (>99% ee) obtained by
this method was sufficient only for a few test experiments
involving enolate protonation.
Because the estimated pKa values for several of the
diamines 1 should be relatively high (pKa ) ca. 25 or
above for 1a -1d ), the asymmetric protonation studies
began with the strongly basic amide enolate 15. Conver-
sion to amide 16 was carried out by treatment with 1 at
-78 °C, followed by warming to 0 °C and quenching with
dilute NH4Cl. The amide was recovered, and the enan-
tioselectivity was established by HPLC assay using a
chiral stationary phase (HPLC/CSP). As summarized in
Table 2, the original lead compound 1b was the most
highly enantioselective proton donor. Lower enantiose-
lectivity resulted for diamines 1c-f, and the most acidic
p-nitro derivative 1f gave racemic product. This was also
the fastest reaction in terms of discharge of the orange-
red enolate color (seconds at - 78 °C). A smaller decrease
in enantioselectivity was observed when the pKa value
(13) The pKa ) ca. 31 for 16 is reported in ref 3 based on
measurements by F. G. Bordwell and A. V. Satish.
(14) Buncel, E.; Menon, B. J . Org. Chem. 1979, 44, 317. Olmstead,
W. N.; Bordwell, F. G. J . Org. Chem. 1980, 45, 3299. Streitwieser, A.,
J r. Acc. Chem. Res. 1984, 17, 353. Kaufman, M. J .; Gronert, S.; Bors,
D. A.; Streitwieser, A., J r. J . Am. Chem. Soc. 1987, 109, 602. Kaufman,
M. J .; Streitwieser, A., J r. J . Am. Chem. Soc. 1987, 109, 6092. Krom,
J . A.; Petty, J . T.; Streitwieser, A. J . Am. Chem. Soc. 1993, 115, 8024.
Fachetti, A.; Streitwieser, A. J . Org. Chem. 1999, 64, 2281.
(12) (a) Ott, H.; Hardtmann, G.; Denzer, M., Frey, A. J .; Gogerty,
J . H.; Leslie, G. H.; Trapold, J . H. J . Med. Chem. 1968, 11, 7. (b) Suna,
E.; Trapencieris, P. Chem. Heterocycl. Comp., in press.

7866 J . Org. Chem., Vol. 64, No. 21, 1999
Vedejs et al.
Ta ble 3. Asym m etr ic P r oton a tion of Dilith ia ted Am in o
Acid Ester s 5
of any enantioselectivity in the case of 1f in Table 2 as
well as in Table 3 suggests that pKa is not the only reason
1f is ineffective. A hybridization change at aniline
nitrogen due to delocalization with the p-nitro group is
one possible explanation for the large difference between
1f and the other chiral diamines.16 A similar effect may
be the reason the p-ethoxycarbonyl example 1d differs
so much from 1c. The hybridization argument could also
explain why 3d is an ineffective chiral acid, but there
are other variables to consider in this case. In any event,
the pKa trends in Table 2 support a qualitative correla-
tion between enantioselectivity and the DMSO-based
∆pKa of the proton donor and the carbonyl product, but
enantioselectivity with a given enolate also depends on
other factors.
The results of Table 3 describe the best asymmetric
protonations for an alanine-derived substrate reported
to date. However, more effort will be needed to obtain
high enantioselectivity and broader substrate tolerance.5
The latter problem remains a difficult challenge for the
asymmetric protonation of enolates. Mechanistic issues
also remain to be clarified. Additional experimentation
will be required to probe details of the proton transfer
process, and it would be premature to propose transition
state models or to invoke direct proton transfer from
aniline nitrogen to enolate carbon until further evidence
is available.
entry
R1
C6H5
C6H5
C6H5
C6H5
OBn
R2
R3
chiral acid
ee (%) (S)
1 (5a )
2 (5a )
3 (5a )
4 (5a )
5 (5b)
6 (5c)
7 (5d ) R-naphthyl
8 (5e)
9 (5f)
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
Et
CH3
CH3
CH3
CH3
CH3
CH3
CH3
Et
1b
1d
1e
1f
1e
1e
1e
1e
1e
1e
1e
73
71
85
0
50
78
83
81
47
80
65
mesityl
C6H5
C6H5
C6H5
t-Bu
CH3
CH3
10 (5g)
11 (5h ) C6H5
Bn
enolates resulted in lower enantioselectivity (entries
5-9). Other amino acid environments were not explored
in detail, but promising enantioselectivity was observed
in two cases using the optimum proton donor 1e (entries
10, 11).
Based on the DMSO pKa estimate for 1e (23.3) and an
optimum ∆pKa ) ca. 2-4, the pKa for singly protonated
5 might correspond to a range of 25-27. If the lithiated
amido subunit C(OLi)dN in 5 is treated as an unsatur-
ated substituent, then the closest literature analogy
having a known pKa value could be ethyl phenylacetate
(PhCH2CO2Et; DMSO pKa ) 22.6), with phenyl as the
unsaturated group.7b Evidently, C(OLi)dN is not as
effective as phenyl in stabilizing the enolate, but there
is some net stabilization, given that ethyl acetate has a
pKa of ca. 30.15
Exp er im en ta l
One final series of experiments was performed using
enolate 4a as the substrate and the most important chiral
acids 1b and 1e. Racemic products resulted in each case.
Modest enantioselectivities in the range of 37-50% ee
were obtained with the analogous enolate 4b as the
substrate and with 1e as the chiral proton source, but
1b gave 16% ee under similar conditions. These results
were not deemed sufficiently promising to warrant a
more detailed investigation.
Gen er a l. HPLC analysis was performed on a Gilson system
using chiral stationary phases with detection by UV. All air-
and/or moisture-sensitive reactions were run under an atmo-
sphere of nitrogen in oven- or flame-dried glassware. Materials
were purified as follows. Tetrahydrofuran and diethyl ether
were freshly distilled from sodium benzophenone ketyl under
N2. Organolithium reagents were titrated using the menthol/
phenanthroline procedure.
Ch ir a l Acid s. Diamine derivatives 1c12b and 3a -3d 6 are
available in >99% ee as previously described. Diamine 1b has
also been reported,12a but the following procedure was used to
ensure high enantiomer purity. The commercially available
(-)-1-[5′-chloro-2′-(methylamino)-phenyl]-1,2,3,4-tetrahydroiso-
quinoline (-) tartrate (5.0 g, 14.4 mmol, Aldrich) was refluxed
in 800 mL methanol for 1 h. After filtration of the hot solution,
the filtrate was allowed to cool to room temperature followed
by cooling to -20 °C overnight. Filtration yielded a cake of
white solid. The free diamine was isolated by partitioning the
tartrate salt between 40 mL ether and a solution of potassium
hydroxide (2.4 g, 43.0 mmol) in 40 mL water with vigorous
stirring. After dissolution was complete (2-3 h), the aqueous
layer was washed with ether (40 mL). The combined ether
extracts were dried (MgSO4) and evaporated to a solid 2.7 g
(70%). Pure 1b was obtained by recrystallization from ether
(mp 99-99.5 °C; lit.12a mp 98-99 °C).
(S )-1-(2-N ,N -Dim e t h ylsu lfa m oyla m in o)p h e n yl-6,7-
d im eth oxy-1,2,3,4-tetr a h yd r oisoqu in olin e (3e). Conver-
sion of 3a to the N-benzyloxycarbonyl derivative at the
secondary nitrogen has been reported.6 To a solution of the
N-Cbz protected diamine (2.5 g) in pyridine (100 mL, distilled
from CaH2) was added N,N-dimethylsulfamoyl chloride (Acros,
6.4 mL, 59.7 mmol) and the mixture was stirred at room
temperature for 48 h. Pyridine was evaporated (aspirator,
below 40 °C), and the oily residue was dissolved in CHCl3 (150
mL), washed with 1 N HCl (3 × 30 mL), water, and brine,
and dried (Na2SO4). After filtration and solvent removal
(aspirator), the resulting oil was purified by flash column
Su m m a r y
The data summarized in Table 2 provide some support
for the notion that asymmetric protonation of the amide
enolate 15 is optimal when ∆pKa (chiral acid vs 16) is
ca. 3. In this range of ∆pKa, the proton transfer process
is essentially complete and irreversible on the time scale
of the protonation experiment, as required to minimize
the formation of racemic product by enolate equilibration.
The knowledge of DMSO pKa’s was helpful in the selec-
tion of chiral acids for other purposes, as in the amino
acid enolate protonations summarized in Table 3, but it
was by no means sufficient. Entries 1 and 2 differ less
than might have been expected if ∆pKa is the dominant
factor, and other variables must play a role in the
observed enantioselectivies. A similar conclusion can be
made by comparing some of the enantioselectivities in
Table 2. Thus, the behavior of 1d (p-ethoxycarbonyl) and
1c (p-trifluoromethyl) is rather different even though the
difference in effective pKa’s is probably small compared
to the approximation inherent in extending the numbers
from DMSO to THF solution. Furthermore, the absence
(15) There is disagreement on the pKa(DMSO) of ethyl acetate (a)
pKa(DMSO) ) ca. 30: Bordwell, F. G.; Branca, J . C.; Bares, J . E.; Filler,
R. J . Org. Chem. 1988, 53, 780. (b) pKa(DMSO) ) 27.45: Arnett, E.
M.; Harrelson, J . A., J r. J . Am. Chem. Soc. 1987, 109, 809.
(16) Delocalization in the p-nitroaniline 1f is consistent with pla-
narization of the N-methyl bond and the adjacent aromatic carbons
(1.3° dihedral angle; see Supporting Information).

Enantioselective Enolate Protonation
J . Org. Chem., Vol. 64, No. 21, 1999 7867
chromatography (column size, 150 × 50 mm) with petroleum
ether-EtOAc 5:2 (fractions 1-45, 50 mL ea), and then
petroleum ether-EtOAc 1:1 (fractions 46-68, 50 mL ea).
Fractions 9-16 contained solid Me2NSO2NMe2 (1H NMR:
single peak at 2.82 ppm in CDCl3), 1.03 g. Fractions 19-36
gave a mixture of starting material and product in a ratio of
3:7 as a colorless oil that was used in the deprotection step
below. Fractions 41-59 contained 1.11 g of material having
no N-methyl signals in the 1H NMR spectrum, consistent with
replacement of a NMe2 fragment by a second diamine mol-
ecule.
The crude oil from fractions 19-36 (1.03 g) was dissolved
in 30 mL of glacial acetic acid and 10% Pd-C (400 mg) was
added. The reaction was stirred under H2 at room temperature.
After 8 h, additional 10% Pd-C was added (200 mg) and
hydrogenolysis was continued for 6 h. The reaction mixture
was filtered through Celite with methanol rinsing, and the
solvent was removed (aspirator, < 30 °C). The residue was
dissolved in water (10 mL), and the solution was basified by
slow addition of NH4OH. The white crystalline precipitate
formed was filtered, dried in vacuo (ca. 1 mmHg) over P2O5,
and recrystallized twice from EtOAc-hexane to give 660 mg
of crystalline sulfonamide 3e containing ca. 10% of diamine
3a (1H NMR assay). Two recrystallizations yielded pure
sulfonamide 3e (560 mg, 24% overall); analytical TLC on silica
gel, 1:1 hexane/EtOAc, Rf ) 0.17; analytical HPLC, CHIRAL-
CEL OD (60:40 hexane/2-propanol, 0.6 mL/min) tR ) 14.5 min
(S) isomer); by comparison with a racemic sample, the (R)
isomer elutes at 18.8 min. 3e: mp 156 °C, dec, colorless
needles. [R]D) +8.1 (c ) 1, CHCl3). Anal. Calcd: C, 58.28; H,
6.45; N, 10.73, S, 8.19. Found: C, 58.23; H, 6.40; N, 10.62; S,
8.11. IR (CDCl3 film, cm-1) 3325, N-H; 1330, SO2N; NMR
(CDCl3, ppm) δ 7.6-6.95 (4 H, m), 6.62 (1 H, s), 6.22 (1 H, s),
5.11 (1 H, s), 3.84 (3 H, s), 3.60 (1 H, s), 3.39-3.28 (1 H, m),
3.19-3.10 (2 H, m), 2.80-2.68 (1 H, m), 2.39 (6 H, s).
(-)-1-[2′-(Meth yla m in o)p h en yl]-1,2,3,4-tetr a h yd r oiso-
qu in olin e (1a ). To a suspension of anhydrous NiCl2 (1.59 g,
12.3 mmol) and 1b (1.16 g, 4.10 mmol) in 50 mL dry THF at
-40 °C under N2 were slowly added LiAlH4 in THF (Aldrich,
1.0 M; 12.3 mL, 12.3 mmol), and the mixture was stirred for
10 min at -40 °C. The reaction mixture was warmed to RT
gradually and stirred for 48 h at RT. The resulting black
mixture was quenched with saturated Na2SO4 (50 mL), the
inorganic salts were removed by filtration through a Celite
pad, the filter was washed well with THF, and the THF was
evaporated (aspirator) to give a yellow oil. The oil was
redissolved in ether, dried (Na2SO4) and concentrated, and the
crude yellow oil was purified by flash column chromatography
(silica gel, 2 × 15 cm) (gradient elution; 30% EtOAc in hexane
with 1% NEt3 to 50% EtOAc in hexane with 1% NEt3) to give
a yellow oil containing some solid (0.70 g, 71%). The yellow
oil was crystallized from hexane to give a pale-yellow powder,
analytical TLC on silica gel, 30% EtOAc in hexane with 1%
NEt3, Rf ) 0.21. Pure material was obtained by recrystalliza-
tion from hexane, mp 77-79 °C. Molecular ion calcd for
60, 1:9 EtOAc/hexane 1% triethylamine, Rf ) 0.60. Molecular
ion calcd for C20H23ClN2: 326.15506; found m/e ) 326.1541,
error ) 3 ppm; base peak ) 283 amu; IR (KBr, cm-1) 3058,
1
dC-H; 2840-2975, C-H; 270 MHz H NMR (CDCl3, ppm) δ
7.30-7.13 (4 H, m), 7.04 (1 H, dd, J ) 8.8, 2.5 Hz), 6.86 (1 H,
dd, J ) 2.5, 1.2 Hz), 6.44 (1 H, d, J ) 8.6 Hz), 5.11 (1 H, s),
3.37 (1 H, d, J ) 9.6 Hz), 3.20-3.06 (1 H, m), 3.06 (3 H, s),
2.80-2.60 (3 H, m), 2.15-1.96 (1 H, m), 1.02 (3 H, d, J ) 6.9
Hz), 0.97 (3 H, d, J ) 6.6 Hz). 13C NMR (68 MHz, {H},
DEPT135, CDCl3, ppm) δ 141.3 s, 134.8 s, 134.6 s, 129.9 d,
129.5 d, 127.8 d, 127.4 d, 127.0 d, 125.0 d, 121.9 s, 120.0 s,
110.9 d, 87.8 d, 53.9 d, 46.1 t, 40.9 d, 33.6 q, 28.9 t, 19.9 q,
18.8 q.
1-[2′-(Meth yla m in o)p h en yl]-1,2,3,4-tetr a h yd r oisoqu in -
olin e Isobu tylid en e Am in a l (11). A dry 50 mL flask with
condenser, containing a 1" stir bar and magnesium turnings
(600 mg, 25 mmol, Baker and Adamson) was flame dried under
nitrogen flush. Mechanical activation (dry stirring) of the
magnesium turnings was done for at least 5 h following the
literature precedent.10
To another flask was added the aminal 9 (201 mg, 0.62
mmol). The solid was dissolved in 2 mL THF and transferred
via cannula into the activated magnesium, followed by rinsing
with THF (2 × 1 mL). Then 1,2-dibromoethane (0.1 mL, 1.1
mmol) was added, and after bubble evolution was evident, the
solution was heated to reflux. After 15 h, the Grignard solution
containing 10 was cooled to room temperature and 10 mL
saturated NH4Cl was added. After filtration through Celite,
the layers were separated and the aqueous phase was ex-
tracted with ether (2 × 30 mL). The organic extracts were
combined, dried (MgSO4), filtered, and evaporated (aspirator)
to afford 11, 190 mg (100%), as an oil. Analytical TLC on EM
silica gel 60, 7:3 hexane/EtOAc, Rf ) 0.64. Molecular ion calcd
for C20H24N2: 292.1940; found m/e ) 292.1932, error ) 3 ppm;
base peak ) 249 amu; IR (neat, cm-1) 2908, dC-H; 300 MHz
NMR (CDCl3, ppm) δ 7.32-7.08 (5 H, m), 6.92 (1 H, dt, J )
7.4, 1.2 Hz), 6.55-6.49 (2 H, m), 5.16 (1 H, s), 3.39 (1 H, d, J
) 9.2 Hz), 3.18-3.06 (1 H, m), 3.09 (3 H, s), 2.86-2.72 (1 H,
m), 2.68-2.62 (2 H, m), 2.10 (1 H, d sept, J ) 9.2, 6.6 Hz),
1.03 (3 H, d, J ) 6.6 Hz), 0.98 (3 H, d, J ) 6.6 Hz).
1-[5′-Ca r boeth oxy-2′-(m eth yla m in o)p h en yl]-1,2,3,4-tet-
r a h yd r oisoqu in olin e Isobu tylid en e Am in a l (12). The
Grignard reagent 10 was prepared as described above from
magnesium turnings (600 mg, 25 mmol) and aminal 9 (255
mg, 0.78 mmol). After 15 h, the stark black Grignard solution
was transferred via cannula into diethyl carbonate11a (0.280
mL, 2.4 mmol), stirred for 15 h, and then quenched with
saturated NH4Cl. The organic layer was separated and washed
with brine. The combined aqueous extracts were basified to
pH 9 with 1 M NaOH and extracted with ether. All organic
extracts were combined, dried (MgSO4), filtered, and evapo-
rated (aspirator). The residue was purified by flash chroma-
tography on EM silica gel 60 (14 × 1 cm), 3:17 EtOAc/hexane
1% triethylamine eluent, 5 mL fractions; fractions 6-12, 251
mg of 12 (88%); analytical TLC on EM silica gel 60, 1:9 EtOAc/
hexane 1% triethylamine, Rf ) 0.28. Pure material was
obtained by crystallization from ethanol, mp 134.8-135.0 °C.
Molecular ion calcd for C23H28N2O2: 364.21509; found m/e )
C
16H18N2: 238.14705; found m/e ) 238.1470, error ) 0 ppm;
base peak ) 132 amu; IR (neat, cm-1) 3310, N-H; 1586, Cd
C; 200 MHz NMR (acetone-d6, ppm) δ 7.25-6.80 (5 H, m), 6.71
(1 H, d, J ) 7.7 Hz), 6.58 (1 H, dd, J ) 8.0, 1.2 Hz), 6.54 (1 H,
d, J ) 7.7 Hz), 6.25 (1 H, br s), 5.03 (1 H, s), 3.3-3.13 (1 H,
m), 3.06-2.82 (2 H, M), 2.8-2.6 (1 H, m), 2.63 (3 H, s). 13C
NMR (68 MHz, {H}, CDCl3, ppm) δ 148.5, 137.4, 134.8, 130.6,
128.7, 128.5, 126.7, 126.3, 126.1, 125.6, 115.4, 110.3, 61.64,
42.5, 30.1, 29.4.
364.2165, error ) 4 ppm; base peak ) 321 amu; IR (KBr, cm-1
)
1
1697, CdO; 300 MHz H NMR (CDCl3, ppm) δ 7.80 (1 H, dd,
J ) 8.4, 2.1 Hz), 7.66 (1 H, br s), 7.37 (1 H, br d, J ) 7.4 Hz),
7.33-7.19 (2 H, m), 7.13 (1 H, br d, J ) 7.4 Hz), 6.51 (1 H, d,
J ) 8.6 Hz), 5.17 (1 H, s), 4.27-4.16 (2 H, m), 3.46 (1 H, d, J
) 9.7 Hz), 3.20-3.05 (1 H, m), 3.12 (3 H, s), 2.79-2.61 (3 H,
m), 2.17-2.04 (1 H, m), 1.28 (3 H, t, J ) 7.2 Hz), 1.04 (3 H, d,
J ) 6.6 Hz), 1.00 (3 H, d, J ) 6.6 Hz).
1-[5′-Ch lor o-2′-(m eth yla m in o)p h en yl]-1,2,3,4-tetr a h y-
d r oisoqu in olin e Isobu tylid en e Am in a l, 9. To a solution of
1b (2.86 g, 10.50 mmol) in 180 mL of MeOH was added
isobutyraldehyde (1.9 mL, 21 mmol, 2 equiv) and catalytic
acetic acid (0.1 mL), and the mixture was stirred at room
temperature for 2 h. After removal of solvent (aspirator), the
residue was purified by plug filtration chromatography on EM
silica gel 60, 1:4 EtOAc/hexane 1% triethylamine eluent to
yield 3.6 g of crude solid. Pure material was obtained by
crystallization from ether/hexane (crop 1, 1.98 g; crop 2, 0.94
g, 85%), mp 114.5-115.5 °C. Analytical TLC on EM silica gel
1-[5′-Tolu en esu lfon yl-2′-(m eth yla m in o)p h en yl]-1,2,3,4-
tetr a h yd r oisoqu in olin e Isobu tylid en e a m in a l (13). The
Grignard solution containing 10 was prepared as described
above from magnesium turnings (715 mg, 30 mmol) and
aminal 9 (1.24 g, 3.8 mmol). After 9 h, the mixture was cooled
to 0 °C and p-toluenesulfonyl fluoride11b (741 mg, 4.2 mmol,
1.1 equiv, Aldrich, purity 98%) was added as a solution in 10
mL THF. The reaction was allowed to warm to room temper-
ature and was stirred for 13 h. After the addition of 10 mL

7868 J . Org. Chem., Vol. 64, No. 21, 1999
Vedejs et al.
water, the biphasic solution was extracted with ether (4 × 40
mL). The combined ether extracts were dried (MgSO4) and
evaporated (aspirator) to a foam (1.73 g). The residue was
purified by flash chromatography on EM silica gel 60 (7 × 4
cm) 7:3 (100 mL) to 1:1 hexane/EtOAc eluent (10 mL fractions;
fractions 2-3, 107 mg 11, 10%; fractions 9-17, 1.43 g 13, 85%);
analytical TLC on EM silica gel 60, 7:3 hexane/EtOAc, Rf )
0.27. Molecular ion calcd for C27H30N2O2S: 446.20285; found
m/e ) 446.2005, error ) 5 ppm; base peak ) 405 amu; IR
(CHCl3, cm-1) 1592, CdC; 300 MHz NMR (CDCl3, ppm) δ 7.66
(2 H, d, J ) 8.4 Hz), 7.62 (1 H, dd, J ) 9.1, 2.3 Hz), 7.44 (1 H,
dd, J ) 2.3, 1.2 Hz), 7.34-7.16 (5 H, m), 7.12 (1 H, br d, J )
7.4 Hz), 6.51 (1 H, d, J ) 8.8 Hz), 5.10 (1 H, s), 3.45 (1 H, d,
J ) 9.3 Hz), 3.15-3.05 (1 H, m), 3.08 (3 H, s), 2.68-2.58 (3 H,
m), 2.34 (3 H, s), 2.14-1.95 (1 H, m), 1.01 (3 H, d, J ) 6.8 Hz),
0.97 (3 H, d, J ) 6.6 Hz).
evaporated (aspirator) to give an orange oil. 1H NMR analysis
of this material indicated a small amount of dinitration product
was present, so the crude residue was purified by flash
chromatography on EM silica gel 60 (5 × 3 cm, fraction 8-22)
with CH2Cl2 eluent to give 526 mg (55%) of 1f. Analytical TLC
on EM silica gel 60, CH2Cl2, Rf ) 0.30. Pure material was
obtained by crystallization from benzene/pentane (mp 172-
173.5 °C, orange cubes, 454 mg). Molecular ion calcd for
C
16H17N3O2: 283.13210; found m/e ) 283.1315, error ) 2 ppm;
IR (neat, cm-1) 3226, N-H; 1296, NO2; 300 MHz NMR (CDCl3,
ppm) δ 8.14 (1 H, dd, J ) 9.2, 2.6 Hz), 7.95 (1 H, d, J ) 2.6
Hz), 7.29-7.10 (3 H, m), 7.05 (1 H, td, J ) 7.1, 2.4 Hz), 6.77
(1 H, d, J ) 7.7 Hz), 6.50 (1 H, d, J ) 9.2 Hz), 5.14 (1 H, s),
3.33-3.24 (1 H, m), 3.15-3.05 (2 H, m), 2.85-2.78 (1 H, m),
2.79 (3 H, d, J ) 5.1 Hz), 2.06 (1 H, br s).
N-2,4,6-Tr im eth ylben zoyla la n in e Meth yl Ester (5c). To
2,4,6-trimethylbenzoic acid (733 mg, 4.46 mmol) in 10 mL CH2-
Cl2 was added oxalyl chloride (0.40 mL, 4.46 mmol) and 1 drop
DMF. After 3 h, the solvents were evaporated and 15 mL CH2-
Cl2 and racemic alanine methyl ester hydrochloride (625 mg,
4.46 mmol) were added. With a room-temperature bath for
cooling, triethylamine (1.55 mL, 11.2 mmol) was added drop-
wise over 5 min and the solution was stirred at room temper-
ature for 1 h. After the addition of 15 mL of 1.5 N HCl, the
layers were separated and the aqueous layer extracted with
CH2Cl2 (2 × 30 mL). The combined organic extracts were
washed with water/saturated Na2CO3 (1:1, 10 mL), dried
(MgSO4), and evaporated (aspirator) to a golden oil. The
residue was purified by flash chromatography on EM silica
gel 60 (14 × 1.7 cm, 10 mL, fraction 4-8, 1.06 g, 71%), 7:3
hexane/EtOAc eluent; analytical TLC on EM silica gel 60, 7:3
hexane/EtOAc, Rf ) 0.26. Pure 5c was obtained by crystal-
lization from ethyl acetate, mp 88.5-89.5 °C. Molecular ion
calcd for C14H19NO3: 249.13650; found m/e ) 249.1370, error
) 2 ppm; base peak ) 147 amu; IR (neat, cm-1) 3263, N-H;
1751, CdO; 1639, CdO; 300 MHz NMR (CDCl3, ppm) δ 6.84
(2 H, s), 6.18 (1 H, d, J ) 7.0 Hz), 4.83 (1 H, dq, J ) 7.4, 7.4
Hz), 3.78 (3 H, s), 2.29 (6 H, s), 2.27 (3 H, s), 1.51 (3 H, d, J )
7.4 Hz). 13C NMR (CDCl3, ppm) δ 173.3 s, 169.9 s, 138.6 s,
134.3 d, 128.2 s, 128.2 s, 52.5 d, 47.9 q, 21.1 q, 19.0 q, 18.5 q.
N-1-Na p h th oyl a la n in e m eth yl ester (5d ). To 1-naph-
thoic acid (735 mg, 4.27 mmol) in 10 mL CH2Cl2 and 1 drop
DMF was added oxalyl chloride (0.417 mL, 4.70 mmol). After
6 h, the solvent was evaporated and the residue was dissolved
in 15 mL CH2Cl2. After the addition of racemic alanine methyl
ester hydrochloride (596 mg, 4.27 mmol, Aldrich), triethyl-
amine (0.892 mL, 6.4 mmol) was added dropwise while cooling
with a water bath. After 1 h, 15 mL 1.5 N HCl was added and
extracted with CH2Cl2 (2 × 30 mL). The CH2Cl2 layers were
combined and extracted with 1:1 saturated Na2CO3/water (10
mL) and the aqueous layer was back extracted with 10 mL
CH2Cl2. The combined CH2Cl2 layers were dried (MgSO4) and
1-[5′-Ca r boeth oxy-2′-(m eth yla m in o)p h en yl]-1,2,3,4-tet-
r a h yd r oisoqu in olin e (1d ). A 250 mL flask was charged with
12 (1.56 g, 4.29 mmol) and 150 mL of 10% HCl and the solution
was heated to 60 °C under a nitrogen stream. After 3 h, the
mixture was cooled to room temperature, diluted with 60 mL
of ether, neutralized with saturated sodium bicarbonate, and
basified to pH 9 with 1 M NaOH. The layers were separated
and the aqueous layer extracted (3 × 60 mL) with ether. The
combined ethereal extracts were dried (MgSO4), filtered, and
evaporated (aspirator) to give 1d (1.46 g) as a yellow oil. The
residue was purified by flash chromatography on EM silica
gel 60 (15 × 4 cm), 1:4 acetone/hexane 5% triethylamine eluent
(15 mL fractions); fractions 10-26, 1.09 g 1d , 82%; analytical
TLC on EM silica gel 60, 1:4 EtOAc/hexane 5% triethylamine,
Rf ) 0.12. Molecular ion calcd for C19H22N2O2: 310.16815;
found m/e ) 310.1683, error ) 0 ppm; IR (CHCl3, cm-1) 1692,
1
CdO; 3325, N-H; 300 MHz H NMR (CDCl3, ppm) δ 7.91 (1
H, dd, J ) 8.6, 1.9 Hz), 7.75 (1 H, d, J ) 1.9 Hz), 7.20-7.05 (2
H, m), 7.05-6.92 (1 H, m), 6.76 (1 H, br d, J ) 8.0 Hz), 6.55 (1
H, br s), 6.52 (1 H, d, J ) 8.7 Hz), 5.10 (1 H, s), 4.37-4.25 (2
H, m), 3.35-3.20 (1 H, m), 3.15-3.00 (2 H, m), 2.85-2.60 (1
H, m), 2.69 (3 H, br d, J ) 3.1 Hz), 2.02 (1 H, br s), 1.36 (3 H,
t, J ) 7.1 Hz).
1-[5′-Tolu en esu lfon yl-2′-(m eth yla m in o)p h en yl]-1,2,3,4-
tetr a h yd r oisoqu in olin e (1e). To 13 (437 mg, 0.98 mmol) was
added 50 mL of 10% HCl and the mixture was heated to 60
°C under a nitrogen stream for 3 h. After cooling to room
temperature, the solution was neutralized with saturated Na2-
CO3 and taken to pH 10 with 1M NaOH. After extraction with
CH2Cl2 (3 × 15 mL), the combined organic extracts were dried
(MgSO4) and evaporated (aspirator) to give a yellow oil. The
residue was purified by flash chromatography on EM silica
gel 60 (12 × 1.5 cm) 1:1 hexane/EtOAc eluent (10 mL
fractions); fractions 8-13, 345 mg 1e, 90%; analytical TLC on
EM silica gel 60, 1:1 hexane/EtOAc, Rf ) 0.32. Molecular ion
calcd for C23H24N2O2S: 392.1559; found m/e ) 392.1542, error
) 4 ppm; IR (neat, cm-1) 3323, N-H; 1146, SO2; 300 MHz
NMR (CDCl3, ppm) δ 7.78-7.70 (3 H, m), 7.54 (1 H, d, J ) 2.2
Hz), 7.25 (2 H, d, J ) 8.1 Hz), 7.20-7.10 (2 H, m), 6.98 (1 H,
td, J ) 7.0, 2.6 Hz), 6.72 (1 H, br s), 6.64 (1 H, d, J ) 7.7 Hz),
6.53 (1 H, d, J ) 8.8 Hz), 5.07 (1 H, s), 3.28-3.18 (1 H, m),
3.12-2.96 (2 H, m), 2.85-2.71 (1 H, m), 2.68 (3 H, s), 2.38 (3
H, s), 2.04 (1 H, br s).
1-[5′-Nit r o-2′-(m et h yla m in o)p h en yl]-1,2,3,4-t et r a h y-
d r oisoqu in olin e (1f). To 1a (987 mg, 3.38 mmol) in 30 mL
of acetonitrile at -40 °C was added a solution of nitronium
tetrafluoroborate (497 mg, 3.21 mmol, 0.95 equiv, Aldrich) in
35 mL of acetonitrile over 1.5 h, and the mixture was stirred
at this temperature for an additional 30 min. The solution was
warmed to 0 °C and quenched with saturated NaHCO3
solution (50 mL). The reaction was extracted with ether (3 ×
50 mL). The combined ether extracts were dried (MgSO4) and
evaporated (aspirator) to give an orange-black solid. To this
crude residue was added 50 mL 1.5 N HCl and 50 mL THF,
and the mixture was stirred for 10 h and then was heated to
55 °C under a constant nitrogen stream for 2 h. After cooling
to room temperature, the reaction was neutralized with
saturated Na2CO3 solution and extracted with ether (4 × 50
mL). The combined ether extracts were dried (MgSO4) and
1
evaporated to a white solid (904 mg, 82%). H NMR analysis
showed the presence of 8% of an isomeric product, probably
derived from 2-naphthoic acid. The residue was purified by
flash chromatography on EM silica gel 60 (20 × 3 cm, 15 mL,
fractions 11-15, 545 mg, 49%), 7:3 (100 mL) to 1:1 hexane/
EtOAc eluent; analytical TLC on EM silica gel 60, 7:3 hexane/
EtOAc, Rf ) 0.21. Pure material was obtained by crystalliza-
tion from ethyl acetate/hexane, mp 130.5-131.5 °C. Molecular
ion calcd for C15H15NO3: 257.10520; found m/e ) 257.1055,
error ) 1 ppm; base peak ) 155 amu; IR (neat, cm-1) 3280,
N-H; 1743, CdO; 1643, CdO; 300 MHz NMR (CDCl3, ppm) δ
8.36 (1 H, d, J ) 8.6 Hz), 7.95-7.86 (2 H, m), 7.67 (1 H, dd, J
) 7.0, 1.1 Hz), 7.60-7.44 (3 H, m), 6.56 (1 H, d, J ) 6.3 Hz),
4.93 (1 H, dq, J ) 7.4, 7.4 Hz), 3.82 (3 H, s), 1.59 (3 H, d, J )
7.4 Hz).
Der a cem iza tion of Na p r oxen Diisop r op yl Am id e (16);
Rep r esen ta tive p r oced u r e for Ta ble 2. To racemic 163
(50.0 mg, 0.16 mmol) in 2 mL of THF was added sec-BuLi (202
µL, 0.28 mmol, 1.38 M in cyclohexane, 1.75 equiv; Aldrich,
titrated using 1,10-phenanthroline as indicator with menthol
as the proton donor at -78 °C in THF), and the solution was
stirred at -78 °C for 1 h. After dropwise addition of the chiral
diamine 1d (99 mg, 0.32 mmol, 2 equiv) in 1.0 mL THF over

Enantioselective Enolate Protonation
J . Org. Chem., Vol. 64, No. 21, 1999 7869
Ta ble 4. HP LC Assa y Meth od s
conditions, retention time
material
chiral stationary phase
(flow rate ) 1 mL/min)
BnOCH(CH3)CO2BHT (4a )
Chiralcel OD (Daicel)
(R,R)-Whelk-O 1 (Regis)
Chiralcel OJ (Daicel)
95:5 hexane/EtOH, 12 min (minor), 17 min (major)
99:1 hexane/IPA, 5.0 min, 5.7 min
BnOCH(CH3)CO2Et (4b)
BzNHCH(Ph)CO2Me (6)
2-(2-naphthyl)-6,6-dimethylvalerolactone
naproxen N,N-diisopropyl
amide (16)
93:7 hexane/IPA, RT 9.3 min (S), 10.6 min (R)
9:1 hexane/IPA, 11.2 min (major), 15.1 min (minor)
EtOH, 0.5 mL/min, 16 min (major), 48 min (minor)
9:1 hexane/IPA, 6.4 min (S), 9.5 min (R)3
9:1 hexane/IPA, 15 min (S), 25 min (R)
Chiralcel OD (Daicel)
(R,R)-Whelk-O 1 (Regis)
Pirkle type (S,S) â gem 1 (Regis);
(R,R)-Whelk-O 1 (Regis)
Ta ble 5. HP LC Meth od s for th e r-Am in o Acid Der iva tives
conditions, retention time
Table 3 entry/substrate
chiral stationary phase
(flow rate ) 1 mL/min)
entries 1-4 (5a )
entry 5 (5b)
entry 6 (5c)
entry 7 (5d )
entry 8 (5e)
entry 9 (5f)
(R,R)-Whelk-O 1 (Regis)
R-Burke
17/3 hexane/IPA, 20.9 min (R), 25.7 min (S)a
19/1 hexane/IPA, 14.0 min (R), 14.7 min (S)b
17/3 hexane/IPA, 27.6 min (R), 34.2 min (S)c
7/3 hexane/IPA, 28.3 min (R), 35.6 min (S)c
9/1 hexane/IPA, 11.5 min (R), 13.0 min (S)c
11.5/1 hexane/IPA, 7.9 min (R), 8.6 min (S)c
9/1 hexane/IPA, 11.6 min (R), 12.7 min (S)d
39/1 hexane/IPA, 31.9 min (R), 34.4 min (S)e
(R,R)-Whelk-O 1
(R,R)-Whelk-O 1
R-Burke
R-Burke
entry 10 (5g)
entry 11 (5h )
R-Burke
R-Burke
a
b
Stereochemical assignment from comparison with authentic methyl N-benzoyl-(R)-alaninate. Stereochemical assignment from
comparison with authentic methyl N-Cbz-(S)-alaninate. c Stereochemical assignment by analogy to entries 1-4 based on similar structure
d
and chromatographic behavior. Stereochemical assignment by analogy to entries 1-4 and entry 11 based on similar structure and
chromatographic behavior. e Stereochemical assignment from authentic methyl N-benzoyl-(S)-phenylalaninate.
5 min, the solution was stirred for 30 min at -78 °C, followed
by warming to 0 °C over 1 h and quenching with 1 mL
saturated NH4Cl. The reaction was partitioned between 10 mL
ether and 10 mL 1.5 N HCl, and the aqueous layer was washed
with 10 mL ether. The combined ether extracts were dried
(MgSO4) and evaporated (aspirator) to afford 16 (50 mg, 100%)
entry 11 (5g; methyl N-benzoylphenylalaninate)17g were de-
racemized by the same method; see Table 5.
To bromomesitylene (55 µL, 0.359 mmol, 2.5 equiv, Aldrich)
in 0.5 mL THF at -78 °C was added t-BuLi (391 µL, 0.704
mmol, 4.9 mmol, 1.80 M in pentane, Aldrich), and the solution
was stirred for 30 min with concomitant formation of a white
precipitate. In a separate flask, 5a (29 mg, 0.14 mmol) was
dissolved in 0.5 mL THF and added via cannula to the
mesityl-Li. The flask was rinsed with an additional 0.5 mL
THF (added by syringe and then added via cannula to the
mesityl-Li solution. After the solution was stirred for 1 h, 1e
(1.25 mL of a 0.35 M solution in THF, 0.44 mmol, 173 mg)
was added over 5 min and the solution was stirred for 30 min
at -78 °C, followed by warming to -20 °C over 20 min. The
reaction was quenched with 1 mL saturated NH4Cl and
partitioned between 10 mL 1.5 N HCl and 10 mL ether, which
produced an insoluble orange oil that crystallized on standing.
After separation of the ether layer, the orange crystals (or oil)
and aqueous layer were washed with 10 mL ether. The
combined ether extracts were dried (MgSO4) and evaporated
(aspirator) to afford 29 mg of material that was greater than
90% 5a by 1H NMR assay and contained less than 10%
residual mesitylene. HPLC analysis was conducted as shown
in Table 2 to afford a 7.5:92.5 ratio of peaks for the (R) and
(S) enantiomers (85% ee). To recover 1e, the aqueous layer
and the orange crystals (or oil) (dissolved in THF) were basified
to pH 10 with NaOH solution and extracted with ether (2 ×
20 mL). The combined ether extracts were dried (MgSO4) and
evaporated (aspirator) to give 176 mg of a golden foam (1e,
>95%).
1
in >95% purity by H NMR assay. After silica plug filtration
(EtOAc eluent), HPLC analysis as described in Table 4
indicated 40% ee favoring the second eluting enantiomer (R)-
16. In a similar experiment, the reaction mixture was quenched
with a THF/saturated NH4Cl solution at -78 °C, but the same
result was obtained. To recover the 1d , the aqueous layer
above was basified to pH 10 using a 1 M NaOH solution and
extracted with ether (3 × 15 mL). The combined ether layers
were dried (MgSO4) and evaporated to afford 75 mg (76%) 1d .
Der a cem iza tion of 2-(2-Na p h th yl)-6,6-d im eth ylva le-
r ola cton e. To bromomesitylene (0.105 mM, 1.75 equiv, Ald-
rich) in 0.5 mL THF at -78 °C was added t-BuLi (0.204 mM,
3.4 equiv, 1.7 M in pentane, Aldrich), and the mixture was
stirred for 30 min with formation of a white precipitate. In a
separate flask, racemic 2-(2-naphthyl-6,6-dimethylvalerolac-
tone was dissolved in 0.5 mL THF and added via cannula to
the mesityllithium, followed by rinsing with an additional 0.5
mL THF, to afford the bright-yellow enolate 7. After stirring
for 1 h, a solution of sulfonamide 3e6 in THF (0.5 mL) was
added over 5 min via syringe followed by a 0.5 mL THF rinse.
The enolate color faded as 3e was added, and a colorless
mixture was obtained when addition was complete. The
reaction was stirred for 30 min at -78 °C, allowed to warm to
-20 °C over 45 min, and quenched with saturated NH4Cl (1
mL). Partitioning between 5 mL 1 N HCl and 10 mL ether
produced an oil, insoluble both in ether and water (HCl salt
of 3e). After separation of the ether layer, the oil and aqueous
layer were washed with 10 mL ether. The combined ether
extracts were dried (MgSO4) and evaporated (aspirator) to
afford >95% of a material contaminated with mesitylene
residue. The enantiomeric excess (58% ee) was determined by
HPLC (see Table 4).
Der a cem iza tion of La cta tes; Rep r esen ta tive P r oce-
d u r e. The procedure for the deracemization of ethyl O-benzyl
lactate 4b18 is given as a representative example. The BHT
lactate ester 4a 19 was deracemized by a similar method, but
(17) (a) Berkowitz, D. B. Synth. Commun. 1990, 20, 1819. (b) Kolar,
A. J .; Olsen, R. K. Synthesis 1977, 457. (c) Fuji, T.; Itaya, T.; Yoshida,
S.; Matsubara, S. Chem. Pharm. Bull. 1989, 37, 3119. (d) Elliot, I. W.
J . Org. Chem. 1962, 27, 3302. (e) Vollmar, A.; Dunn, M. S. J . Org.
Chem. 1960, 25, 387. (f) Brunner, H.; Knott, A.; Kunz, M.; Thalham-
mer, E. J . Organomet. Chem. 1986, 308, 55. (g) Wightman, R. H.;
Staunton, J .; Battersby, A. R.; Hanson, K. R. J . Chem. Soc., Perkin
Trans. 1 1972, 2355.
(18) Heathcock, C. H.; Pirrung, M. C.; Young S. D; Hagen, J . P.;
J arvi, E. T.; Badertscher U.; Ma¨rki, H.-P.; Montgomery, S. H. J . Am.
Chem. Soc. 1984, 106, 8161.
(19) Mislow, K.; O’Brien R. E.; Schaefer, H. J . Am. Chem. Soc. 1962,
84, 1940.
Der a cem iza tion of Am in o Acid Der iva tives; Rep r e-
sen ta tive P r oced u r e for Ta ble 3. The procedure for the
deracemization of 5a 17a is shown as an example. The other
amino acid substrates for Table 3 including entry 5 (5b; N-Cbz
methyl alaninate),17b entry 6 (5c; methyl N-mesitoylalaninate),17c
entry 7 (5d ; methyl N-R-naphthoylalaninate), entry 8 (5e; ethyl
N-benzoylalaninate),17d entry 9 (5f; tert-butyl N-benzoylalan-
inate),17e entry 10 (5g; methyl R-benzamidobutyrate),17f and

7870 J . Org. Chem., Vol. 64, No. 21, 1999
Vedejs et al.
with sec-butyllithium as the base. To hexamethyldisilazane
(63.2 µL, 0.30 mmol, 2.0 equiv, Aldrich) in 0.5 mL THF at -20
°C was added dropwise n-BuLi (187 µL, 0.30 mmol, 2.0 equiv,
Aldrich), and the solution was stirred for 20 min. After cooling
this solution to -78 °C, 4b was added (0.81 mL of a 0.19 M
solution in THF, 35 mg, 0.15 mmol) and the resulting solution
was warmed to -40 °C for 2 h. After cooling to -78 °C, 1e
(0.85 mL of a 0.48M solution in THF, 161 mg, 0.41 mmol) was
added and the solution was stirred for 30 min at this temper-
ature. After warming to -20 °C over 30 min, 1 mL saturated
NH4Cl was added. The resulting biphasic solution was parti-
tioned between 10 mL ether and 10 mL 1.5 N HCl, and the
aqueous layer was washed with 10 mL ether. The combined
ether layers were dried (MgSO4) and evaporated to give 30
mg (85%) ethyl O-benzyl lactate. HPLC analysis was conducted
as in Table 4 to afford 45% ee favoring the second eluting
enantiomer (R).
HP LC Meth od s, Gen er a l. Standard procedure for the
HPLC analyses included equilibrating column to constant
baseline. The enantiomer separation was validated by testing
the racemic mixture at the beginning of each session. Detection
was by UV at 240/254 nm or 220/240 nm depending on analyte,
and enantiomer ratios were obtained using both wavelengths.
All flow rates are 1.0 mL/min unless otherwise stated.
Ack n ow led gm en t. This work was supported by
NIH (GM 44724).
Su p p or tin g In for m a tion Ava ila ble: NMR spectra of new
compounds and X-ray data tables for 1f. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O990897M