Greatly simplified procedures for the synthesis of α-amino acids by the direct alkylation of pseudoephedrine glycinamide hydrate

Article abstract of DOI:10.1021/jo990341z

A modified procedure for the synthesis of highly enantiomerically enriched α-amino acids is described that involves the direct alkylation of pseudoephedrine glycinamide hydrate (1 · H₂O) followed by hydrolysis. The modified procedure was developed to overcome several inconvenient aspects of our earlier reported procedure. Advantages of the new method include (1) a greatly simplified one-step synthesis of the alkylation substrate (1 · H₂O) by the direct combination of glycine methyl ester hydrochloride with pseudoephedrine in the presence of lithium tert-butoxide, (2) the use of the weaker base lithium hexamethyldisilazide (LHMDS) in lieu of lithium diisopropylamide (LDA) for the enolization reaction, (3) a protocol for the direct alkylation of 1 · H₂O without the need for prior drying of the alkylation substrate, and (4) a one-step alkylation procedure that generates LHMDS and anhydrous lithium chloride simultaneously from the reaction of lithium metal with n-hexyl chloride in the presence of hexamethyldisilazane.

Full text of DOI:10.1021/jo990341z

3322
J . Org. Chem. 1999, 64, 3322-3327
Gr ea tly Sim p lified P r oced u r es for th e Syn th esis of r-Am in o Acid s
by th e Dir ect Alk yla tion of P seu d oep h ed r in e Glycin a m id e Hyd r a te
Andrew G. Myers,*,† Patrick Schnider, Soojin Kwon, and Daniel W. Kung
Division of Chemistry and Chemical Engineering, California Institute of Technology,
Pasadena, California 91125
Received February 24, 1999
A modified procedure for the synthesis of highly enantiomerically enriched R-amino acids is described
that involves the direct alkylation of pseudoephedrine glycinamide hydrate (1‚H2O) followed by
hydrolysis. The modified procedure was developed to overcome several inconvenient aspects of our
earlier reported procedure. Advantages of the new method include (1) a greatly simplified one-step
synthesis of the alkylation substrate (1‚H2O) by the direct combination of glycine methyl ester
hydrochloride with pseudoephedrine in the presence of lithium tert-butoxide, (2) the use of the
weaker base lithium hexamethyldisilazide (LHMDS) in lieu of lithium diisopropylamide (LDA) for
the enolization reaction, (3) a protocol for the direct alkylation of 1‚H2O without the need for prior
drying of the alkylation substrate, and (4) a one-step alkylation procedure that generates LHMDS
and anhydrous lithium chloride simultaneously from the reaction of lithium metal with n-hexyl
chloride in the presence of hexamethyldisilazane.
In prior work, a method for the asymmetric synthesis
alkylation products 2 that is suitable for both large- and
small-scale applications.
of R-amino acids was described that involved the alkyl-
ation of an enolate dianion derived from anhydrous
pseudoephedrine glycinamide (1) followed by hydrolysis
of the resulting alkylation products (2, Scheme 1).1 The
alkylation substrate (1) was prepared by the base-
catalyzed condensation of glycine methyl ester with
pseudoephedrine followed by crystallization of the prod-
uct, pseudoephedrine glycinamide hydrate (1‚H2O), and
drying.2 Although the method offered several advantages
over existing procedures for the alkylative assembly of
R-amino acids,1 there were aspects of the procedure that
were felt to provide opportunities for improvement. These
are indicated by the bold italics within Scheme 1, which
also summarizes the improvements detailed in this work.
Among the problematic features of the prior method were
the steps to prepare the alkylation substrate 1, which
included the preparation and distillation of the free-base
form of glycine methyl ester, the need to use anhydrous
lithium chloride to promote the clean condensation of the
latter product with pseudoephedrine, and a separate step
to dry the crystalline hydrate (1‚H2O) to afford anhydrous
1 as a hygroscopic solid.2 The protocol for alkylation of
anhydrous 1 required the thorough drying of both 1 and
lithium chloride, an additive that promotes a rapid, clean,
and highly diastereoselective alkylation reaction. The
procedure also demanded that the number of equivalents
of the base used for enolization, lithium diisopropylamide
(LDA), be measured carefully, for quantities in excess of
1.95 equiv led to partial cleavage of pseudoephedrine
from the substrate 1. In this work, we describe a modified
sequence that addresses each of these concerns, providing
a greatly simplified procedure for the preparation of the
A Mod ified P r oced u r e for th e P r ep a r a tion of
P seu d oep h ed r in e Glycin a m id e Hyd r a te (1‚H2O):
Step 1. Pseudoephedrine glycinamide hydrate (1‚H2O)
is a highly crystalline, free-flowing solid that is stable to
air and moisture and is derived from two inexpensive
commodity chemicals, glycine methyl ester hydrochloride
and pseudoephedrine, both of which are also free-flowing,
nonhygroscopic crystalline solids. The earlier procedure
that was developed for the synthesis of 1‚H2O involved
the base-catalyzed condensation of pseudoephedrine with
glycine methyl ester in the presence of anhydrous lithium
chloride. This procedure required the initial neutraliza-
tion of glycine methyl ester hydrochloride followed by
distillation of the resultant liquid free base. The propen-
sity of glycine methyl ester to polymerize upon standing
led us to focus on developing a simpler synthesis of 1‚
H2O that avoided using this starting material. Attention
was turned to procedures for the neutralization of glycine
methyl ester hydrochloride in situ, in the presence of
pseudoephedrine. This, in turn, provided an opportunity
to solve another problemsthe need to use anhydrous
lithium chloride in the condensation reactionsby using
a lithium salt for the neutralization step. In this way,
lithium chloride would be formed as a byproduct of the
neutralization reaction, potentially obviating the need to
add this hygroscopic solid to the reaction mixture (in
earlier work, lithium chloride had been shown to be a
necessary additive in order to prevent the self-condensa-
tion of glycine methyl ester, as well as over-glycylation
of the product 1).2
An initial experiment toward this end was quite
promising. Addition of 1.7 equiv of solid lithium meth-
oxide in one portion to a stirring, ice-cooled suspension
of pseudoephedrine (1 equiv) and glycine methyl ester
hydrochloride (1.2 equiv) in tetrahydrofuran (THF) led
to complete consumption of glycine methyl ester within
7 h at 0 °C. Approximately 20% of pseudoephedrine
remained. Addition of water and recrystallization, as
† Address correspondence to this author at Department of Chemistry
and Chemical Biology, Harvard University, 12 Oxford St., Cambridge,
MA 02138.
(1) (a) Myers, A. G.; Gleason, J . L.; Yoon, T. J . Am. Chem. Soc. 1995,
117, 8488. (b) Myers, A. G.; Gleason, J . L.; Yoon, T.; Kung, D. W. J .
Am. Chem. Soc. 1997, 119, 656.
(2) Myers, A. G.; Yoon, T.; Gleason, J . L. Tetrahedron Lett. 1995,
36, 4555.
10.1021/jo990341z CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/16/1999

Direct Alkylation of Pseudoephedrine Glycinamide Hydrate
J . Org. Chem., Vol. 64, No. 9, 1999 3323
Sch em e 1
and recrystallization, as before, provides analytically
pure, crystalline 1‚H2O in 73-76% yield. For comparison,
our earlier published procedure required a reaction time
of ∼8 h, involved the use of 2 equiv of anhydrous lithium
chloride (as well as the free-base form of glycine methyl
ester), and provided 1‚H2O in 68-76% yield.
A New En oliza tion P r otocol a n d a Meth od for th e
Dir ect Alk yla tion of P seu d oep h ed r in e Glycin a m id e
Hyd r a te (1‚H2O): Step 2. As stated in the introduction
and summarized graphically in Scheme 1, there were
several aspects of our original alkylation procedure that
were targeted for modification. These included the need
to ensure the dryness of both lithium chloride and
pseudoephedrine glycinamide, and the need to carefully
measure the number of equivalents of LDA used in the
enolization reaction. The latter requirement arose be-
cause of a side reaction that occurred whenever >1.95
equiv of base was used, leading to the partial cleavage
of pseudoephedrine from 1. Through further experimen-
tation, we have learned that it is possible to substitute
the weaker base lithium hexamethyldisilazide (LHMDS)
for LDA in the enolization reaction and that this base
does not induce the cleavage reaction observed with the
latter reagent, even when used in excess. The yield and
diastereoselectivity of alkylation reactions conducted with
LHMDS in lieu of LDA were not affected by the substitu-
tion. Furthermore, the presence of an excess (>2 equiv)
of LHMDS in the reaction does not greatly diminish the
yield of alkylation product with many alkyl halides;
therefore, it is not necessary to titrate nor to accurately
measure the LHMDS used in the reaction, provided that
at least 2 equiv is present. LHMDS is conveniently
available from commercial suppliers in several forms,
both neat and in solution, and exhibits excellent shelf
stability.
With an improved protocol for the enolization of 1,
efforts turned toward streamlining the entire alkylation
procedure. In particular, a means of simplifying the
dehydration of 1‚H2O was sought. Although 1‚H2O is an
air-stable, free-flowing solid, anhydrous 1 is hygroscopic
and must be handled with the exclusion of moisture.
Furthermore, the dehydration of 1‚H2O constitutes an
additional step in the synthetic sequence. The possibility
that the dehydration step might be omitted entirely was
considered. Specifically, it was proposed to use the
hydrate directly in the alkylation reaction with the
inclusion of an extra equivalent of base to deprotonate
the water of hydration. Upon investigation, we have
found that the proposed protocol is not only feasible but,
in many cases, represents an improved procedure as well
as a one of greater convenience. Because both lithium
alkoxides and lithium halides are known to influence the
aggregation state and reactivity of enolates, it was
deemed prudent to investigate the effect of lithium
chloride under the new conditions, in which 1 equiv of
lithium hydroxide is formed. The diastereoselectivity of
the alkylation of 1‚H2O using 3.2 equiv of LHMDS as
base and benzyl bromide (1.2 equiv) as the electrophile
was measured as a function of the concentration of added
lithium chloride (Table 1). The maximum diastereose-
lectivity (93% de) was observed when g3 equiv of lithium
chloride was used and was the same value as that
previously obtained for this reaction using anhydrous 1,
lithium chloride, and LDA as base.
before, afforded crystalline 1‚H2O in 50-55% yield. Use
of the more soluble base lithium tert-butoxide led to
further improvement in the reaction, providing what is
presently the optimum procedure.3 Although lithium tert-
butoxide is a hygroscopic solid, it is easily handled and
is more convenient to manipulate, e.g., in benchtop
weighing, than anhydrous lithium chloride (used in the
prior method). The procedure entails the addition of the
base to the reaction mixture in a single portion. The
addition is conducted at ambient temperature and,
although it produces a mild exotherm, external cooling
has been found to be unnecessary, as tested thus far on
scales as large as 60 g. Typically, after addition of the
base, the reaction mixture clears within 15 min; g95%
of pseudoephedrine is consumed within 1-2 h. Workup
(3) This procedure is also effective for the preparation of pseu-
doephedrine sarcosinamide (nonhydrated), in 71% yield (from sarcosine
hydrochloride) after recrystallization from toluene.
In the optimized procedure for the alkylation of 1‚H2O,
anhydrous lithium chloride (3.2 equiv) is first flame-dried

3324 J . Org. Chem., Vol. 64, No. 9, 1999
Myers et al.
Ta ble 1. Alk yla tion Dia ster eoselectivity a s a F u n ction of
Ad d ed Lith iu m Ch lor id e
during the reaction, but also because it avoids a second
precipitation that can occur when approximately 2 equiv
of LHMDS is added to the mixture of lithium chloride
and 1‚H2O and the resulting solution is allowed to stand.
This can occur in larger-scale experiments where the
addition of the base is slowed by the sheer volume that
must be transferred; precipitation then occurs before the
final 1.2 equiv of base can be added. This precipitation
does not occur when the suggested two-stage addition
protocol (1.5 equiv followed by rapid addition of 1.7 equiv,
see above) is followed. It is also not observed in small
scale experiments (<30 mmol), presumably because the
addition of base is relatively rapid. Both protocols afford
products with identical, high diastereoselectivities and
yields.
entry
enolate concn (M)
equiv LiCl
de (%)
1
2
3
4
5
6
7
8
0.16
0.16
0.16
0.16
0.16
0.16
0.16
0.24
0
2
3
4
5
6
10
3.2
74
88
90
93a
93
93
93
93b
The new procedure for the direct alkylation of 1‚H2O
using LHMDS as base is operationally much simpler
than our earlier published procedure and reliably affords
good yields of products with a wide range of alkyl halides
as substrates (Table 2). An exception to this generaliza-
tion may be alkyl halide substrates that are sluggish to
react by nucleophilic displacement, but which undergo
rapid dehydrohalogenation. Isobutyl iodide is suspected
to be a case in point (entries 4 and 5, Table 2). The yield
of this alkylation product using the new method is only
40-62%, whereas the earlier procedure afforded the
same product in 70% yield (albeit requiring 2 d at 23 °C).
The culprit here is suspected to be lithium hydroxide,
but this has not been established. For most standard
alkyl halide substrates, such as the remaining examples
of Table 2, the modified procedure is more convenient,
more reliable, and tends to proceed more cleanly and to
a greater degree of completion than the earlier published
procedure.
a
Recommended for larger-scale (g30 mmol 1‚H2O) experiments.
Recommended for smaller-scale (<30 mmol 1‚H2O) experiments.
b
in vacuo in a round-bottom flask and then is allowed to
cool under an inert atmosphere. Sufficient THF is added
to the solid lithium chloride so as to prepare a solution
that would be ∼3 M in lithium chloride, were it dissolved.
The resulting slurry of lithium chloride in THF is stirred
at ambient temperature for at least 15 min during which
time visible, but largely incomplete, solubilization occurs.
At this point, 1‚H2O (1 equiv) is added, leading to a rapid
clearing of the slurry. This solubilizing effect is notewor-
thy, for neither LiCl nor 1‚H2O is soluble in THF
independently.4 The clear solution of lithium chloride and
1‚H2O has been observed to spontaneously form a thick
slurry under certain circumstances. LHMDS can be
added and a solution of enolate obtained in either case,
but working with the thick slurry is inconvenient.
Precipitation appears to occur whenever the local con-
centration of lithium chloride in the solution is depleted
relative to the hydrate. For this reason the hydrate is
never added prior to the lithium chloride, which always
leads to precipitation. Rather, the hydrate is added to
the prestirred slurry of lithium chloride (3 equiv) in THF.
For larger-scale experiments (g30 mmol), we recommend
two additional modifications designed to minimize the
precipitation of the mixture of lithium chloride and 1‚
H2O. First, the volume of THF is increased 3-fold
(producing a final solution of hydrate that is ∼0.33 M)
and, second, the number of equivalents of lithium chlo-
ride is increased to 4, which is approximately the point
of saturation of this reagent at these concentrations. In
all procedures, approximately 10 min after solubilization
of the mixture of lithium chloride and 1‚H2O occurs, the
reaction solution is cooled to 0 °C and, 20 min later, a
commercial, 1.0 M solution of LHMDS (3.2 equiv) in THF
is added rapidly to the cold solution. This addition leads
to a mild exotherm. Careful monitoring has shown that
the exotherm occurs primarily during the addition of the
first equivalent of base, presumably involving the depro-
tonation of water. In the optimized procedure, 1.5 equiv
of LHMDS is added carefully to the cold solution of
hydrate and lithium chloride. After sufficient time has
elapsed for the solution to cool once again to 0 °C, the
remaining base is added very rapidly. This procedure is
advisable not only for optimum temperature control
A P r oced u r e for th e Gen er a tion of LHMDS‚LiCl
In Situ fr om Lith iu m Meta l, Hexa m eth yld isila za n e,
a n d Hexyl Ch lor id e a n d Its Use for th e En oliza -
tion -Alk yla tion of 1‚H2O. The only remaining un-
solved problem identified within Scheme 1 is the need
for anhydrous lithium chloride in the alkylation reaction
mixture. As detailed above, the alkylation protocol is
typically initiated by flame-drying reagent grade anhy-
drous lithium chloride in vacuo. Although this is easily
done, we noted that the optimum number of equivalents
of lithium chloride and the number of equivalents of
LHMDS employed in the enolization-alkylation reaction
were approximately the same (∼3 equiv each). Recogniz-
ing that the commercial synthesis of alkyllithium re-
agents is accompanied by the formation of 1 equiv of
lithium chloride (typically removed during processing),
we considered a scenario whereby a solution of LHMDS
and lithium chloride could be formed in situ by the
reaction of lithium metal with an n-alkyl chloride in the
presence of hexamethyldisilazane. A related procedure
has been described by Einhorn and Luche for the
(4) Seebach and co-workers have previously shown that lithium
halides can increase the solubility of peptides in ethereal solvents by
as much as 100-fold; see: Seebach, D.; Thaler, A.; Beck, A. K. Helv.
Chim. Acta 1989, 72, 857.

Direct Alkylation of Pseudoephedrine Glycinamide Hydrate
J . Org. Chem., Vol. 64, No. 9, 1999 3325
Ta ble 2. Dia ster eoselective Alk yla tion of P seu d oep h ed r in e Glycin a m id e Hyd r a te (1‚H2O)
preparation of LDA.5 The procedure envisioned was
readily implemented, with great success, and was opti-
mized using n-hexyl chloride so as to form the liquid
byproduct hexane. In the optimized protocol, lithium wire
(6.8 equiv) is washed with hexanes, cut into ∼1-cm pieces,
and the pieces are suspended in sufficient THF to produce
a solution that is ultimately ∼0.86 M in LHMDS.6
Reagent-grade n-hexyl chloride (3.6 equiv) and hexa-
methyldisilazane (3.7 equiv) are then added at 23 °C.
After a short induction period, the lithium metal begins
to react, dissolving in the process. This is accompanied
by a mild exotherm, which is controlled with a water or
ice-water bath so as to maintain an internal reaction
temperature that is below 35 °C. After 2-10 h, depending
upon the quality of the lithium metal used and the
reaction scale, dissolution is essentially complete and the
reaction mixture is cooled in an ice bath. At this point,
solid 1‚H2O is added to the basic solution in a single
portion. The ensuing reaction with the base is only mildly
exothermic, and dissolution of the substrate occurs
quickly forming a homogeneous enolate solution, after
which time the electrophile is added. The yields and
diastereoselectivities for this procedure are quite com-
parable to those of the procedure reported above. The
method has been tested on a 50-g scale using commercial
sources of anhydrous THF and HMDS, as received, and
been found to be both efficient and convenient to imple-
ment.
Exp er im en ta l Section
Gen er a l Exp er im en ta l P r oced u r es. All reactions were
conducted under a positive pressure of argon. Lithium tert-
butoxide can be handled and weighed in the atmosphere, but
should be stored under an inert atmosphere to avoid the
absorption of carbon dioxide and water from the air. Com-
mercial alkyl halides were purified by passage through
activated basic alumina (Brockman Grade 1) or were distilled
from calcium hydride. Commercial solutions of lithium hex-
amethyldisilazide (1.0 M in THF, Aldrich) were used without
titration. Reagent grade anhydrous lithium chloride was dried
by heating under vacuum (150 °C, 0.5 mmHg, 12 h) and was
flame-dried immediately prior to use, as described.
Gen er a l P r oced u r es for Deter m in a tion of th e Dia -
ster eom er ic Excess of Alk yla tion P r od u cts. The diaster-
eomeric excess (de) of an alkylation product was determined
by acetylation followed by capillary GC analysis (Chirasil-Val
column, length 25 m, inj 250 °C, det 275 °C). The procedure
for acetylation of the products is as follows. Acetic anhydride
(1 mL) and 4-(dimethylamino)pyridine (ca. 2 mg) were added
to a solution of the alkylation product (10-20 mg) in pyridine
(1 mL). The resulting solution was stirred at 23 °C for 1 h
and then was diluted with water (20 mL). The aqueous solution
was extracted with ethyl acetate (2 × 20 mL), and the
combined organic layers were washed with saturated aqueous
sodium bicarbonate solution (25 mL). The organic layer was
dried over anhydrous sodium sulfate, filtered, and concen-
(5) Einhorn, J .; Luche, J . L. J . Org. Chem. 1987, 52, 4124.
(6) Einhorn and Luche recommend that lithium be added in portions
in larger-scale experiments in order to control the reaction exother-
micity.5

3326 J . Org. Chem., Vol. 64, No. 9, 1999
Myers et al.
trated in vacuo. The residue was dissolved in ethyl acetate
(10 mL), and the resulting solution was analyzed by capillary
GC analysis.
reaction flask had been immersed in an ice bath. The rate of
addition of the first 300-mL portion of base (300 mmol, 1.5
equiv, to the 340-mL mark) was modulated such that the
internal reaction temperature did not exceed 3 °C (addition
time ∼30 min). When the reaction mixture had again cooled
to 0 °C, the stopcock of the addition funnel was opened fully
to allow for rapid addition of the second portion of base (340
mL, 340 mmol, 1.70 equiv). The latter addition produced a
modest exotherm (8 °C maximum internal temperature). After
20 min, allyl bromide (18.2 mL, 210 mmol, 1.05 equiv) was
injected slowly into the orange enolate solution over a period
of 15 min. The internal temperature did not exceed 5 °C during
the addition. After the addition of allyl bromide was complete,
the reaction mixture was stirred at 0 °C for 1 h. Water (500
mL) was added at this point, and the resulting biphasic
mixture was carefully acidified to pH 0 by the addition of
aqueous hydrochloric acid solution (6 M, 300 mL). The acidified
aqueous solution was transferred to a 4-L separatory funnel
and was extracted with ethyl acetate (750 mL). The organic
layer was separated and extracted sequentially with single
300-mL portions of 3 M and 1 M aqueous hydrochloric acid
solution, respectively. The aqueous layers were combined and
cooled to an internal temperature of 5 °C by stirring in an ice-
water bath. The cold solution was cautiously basified to pH
14 by the addition of 50% aqueous sodium hydroxide solution
(200 mL). The temperature of the solution was maintained at
or below 25 °C during basification. The basified solution was
extracted sequentially with one 800-mL portion and three 250-
mL portions of dichloromethane. The combined organic ex-
tracts were dried over anhydrous solid potassium carbonate
and then were filtered and concentrated in vacuo. The de of
the crude reaction product was shown to be 93% by capillary
GC analysis of an acetylated sample (190 °C, major tR ) 22.7
min, minor tR ) 19.6 min) as described in the general
procedures. The oily, pale yellow residue solidified upon
standing in vacuo (0.5 mmHg). The solid was dissolved in hot
toluene (100 mL); crystallization occurred upon cooling the
resulting solution to 23 °C. After standing overnight at 23 °C,
the crystallization flask was cooled to -20 °C for 6 h. The
crystals were collected and washed sequentially with 25 mL
of cold (0 °C) toluene and four 100-mL portions of ether. The
crystals were dried in vacuo (0.5 mmHg) at 23 °C for 14 h to
provide 32.7 g (62.3%) of diastereomerically pure 2g. The
mother liquors were concentrated, and the liquid residue was
dissolved in hot toluene (25 mL). Cooling to 23 °C and then at
-20 °C (3.5 d) afforded a second crop of pure 2g (4.4 g, 8.4%,
de >99%, mp 72-73 °C). Anal. Calcd for C15H22N2O2: C, 68.67;
H, 8.45; N, 10.68. Found: C, 68.44; H, 8.70; N, 10.68.
Spectroscopic data were identical to those previously reported.1
La r ge-Sca le Alk yla tion of (R,R)-(-)-P seu d oep h ed r in e
Glycin a m id e H yd r a t e (1‚H 2O) Usin g Lit h iu m H exa -
m eth yld isila zid e a n d Lith iu m Ch lor id e Gen er a ted in
Sit u fr om Lit h iu m Met a l, 1-Ch lor oh exa n e, a n d H exa -
m eth yld isila za n e. Syn th esis of [1R(S),2R]-N-(2-Hyd r oxy-
1-m e t h yl-2-p h e n e t h yl)-2-a m in o-N -m e t h yl-4-p e n t e n a -
m id e (2g). In this experiment, tetrahydrofuran, hexameth-
yldisilazane, and 1-chlorohexane were used as received from
commercial suppliers, and no provisions were made to dry the
reaction apparatus. A 3-L, three-necked, round-bottom flask
fitted with an inlet adapter connected to a source of vacuum
and argon, a magnetic stirring bar, and a thermometer was
charged with tetrahydrofuran (800 mL). The flask was alter-
nately evacuated and flushed with argon three times. Lithium
wire (3.2 mm diameter, 98+%, sodium content 0.5-1%, 9.44
g, 1.36 mol, 6.80 equiv) was freed of oil by briefly rinsing in
pentane and then was cut into 2-5 cm pieces, and the pieces
were added to the reaction vessel under a stream of argon.
Hexamethyldisilazane (156 mL, 740 mmol, 3.70 equiv) and
1-chlorohexane (99.0 mL, 720 mmol, 3.60 equiv) were added
sequentially to the reaction mixture, and the reaction vessel
was immersed in a water bath at 23 °C. After a few minutes,
the reaction temperature increased to 35 °C and the solution
became cloudy, presumably due to the precipitation of lithium
chloride.6 Ice was added to the water bath to maintain an
internal temperature of ∼30 °C. Approximately 20 min after
Im p r oved , On e-St ep Syn t h esis of (R,R)-(-)-P seu -
d oep h ed r in e Glycin a m id e Hyd r a te (1‚H2O). A 2-L, three-
necked, round-bottom flask fitted with an inlet adapter
connected to a source of vacuum and argon, a thermometer,
and a glass stopper was charged with (R,R)-(-)-pseudoephe-
drine (60.0 g, 363 mmol, 1 equiv) and glycine methyl ester
hydrochloride (59.2 g, 472 mmol, 1.30 equiv). Tetrahydrofuran
(510 mL) was added to the solid mixture, and the whole was
stirred at 23 °C with a magnetic stirring bar for 15 min,
producing a fine suspension. The glass stopper was briefly
removed and lithium tert-butoxide powder (40.7 g, 508 mmol,
1.40 equiv) was added to the suspension in a single portion
from a weighing boat. The latter addition produced a modest
exotherm (40 °C maximum internal temperature, no provision
for external cooling was necessary). Within 15 min, the
reaction mixture became a homogeneous, pale yellow solution.
After stirring at ambient temperature for a total period of 2
h, water (450 mL) was added to the reaction solution, and the
mixture was concentrated in vacuo to remove the bulk of the
tetrahydrofuran. The largely aqueous concentrate was trans-
ferred to a 2-L separatory funnel and was extracted with a
1-L portion of dichloromethane. Solid sodium chloride was
added to the aqueous layer to the point of saturation, and the
resulting solution was extracted with four 300-mL portions of
dichloromethane. The combined organic extracts were dried
over anhydrous solid potassium carbonate and then were
filtered and concentrated by rotary evaporation. The liquid
residue was further concentrated under high vacuum (0.5
mmHg, 10 h). The resulting syrup was dissolved in hot
tetrahydrofuran (300 mL), and water (12 mL) was added to
the warm solution. Upon cooling (23 °C), 1‚H2O crystallized
within minutes. The crystallization flask was allowed to stand
overnight at 23 °C and then was cooled to -20 °C for 3 h. The
crystals were collected by vacuum filtration on a 600-mL glass
frit, washing with two 200-mL portions of ether. After drying
in vacuo (0.5 mmHg) for 10 h, 1‚H2O was obtained as a
nonhygroscopic, free flowing, crystalline solid (63.9 g, 73.3%,
mp 85-87 °C).7 Anal. Calcd for C12H20N2O3: C, 59.98; H, 8.39;
N, 11.66. Found: C, 59.96; H, 8.67; N, 11.75. Spectroscopic
data were identical to those previously reported.1
La r ge-Sca le Alk yla tion of (R,R)-(-)-P seu d oep h ed r in e
Glycin am ide Hydr ate (1‚H2O) Usin g Com m er cial Lith iu m
Hexa m eth yld isila zid e a s Ba se. Syn th esis of [1R(S),2R]-
N-(2-Hyd r oxy-1-m eth yl-2-p h en eth yl)-2-a m in o-N-m eth yl-
4-p en ten a m id e (2g). An oven-dried, 3-L, three-necked, round-
bottom flask fitted with two glass stoppers and an inlet adapter
connected to a source of vacuum and argon was charged with
anhydrous lithium chloride (33.9 g, 800 mmol, 4.00 equiv). The
flask was evacuated, and a gentle flame was applied to further
dry the solid lithium chloride. After cooling to 23 °C in vacuo,
the flask was flushed with argon and then was fitted with a
mechanical stirrer and an oven-dried 1-L pressure-equalizing
addition funnel, marked at 340 and 640 mL. Tetrahydrofuran
(600 mL) was added to the flask, and the resulting suspension
was stirred for 20 min at 23 °C. Solid (R,R)-(-)-pseudoephe-
drine glycinamide hydrate (48.1 g, 200 mmol, 1 equiv) was
added to the suspension in eight portions over a period of 10
min through one of the necks by way of a powder funnel. The
resulting slightly cloudy solution was cooled by immersing the
reaction flask in an ice bath. The pressure-equalizing addition
funnel was charged with a commercial 1 M solution of lithium
hexamethyldisilazide in tetrahydrofuran (640 mL, 640 mmol,
3.20 equiv). The argon inlet adapter was transferred from the
neck of the flask to the neck of the addition funnel, and the
open neck was fitted with a septum pierced with a thermo-
couple to monitor the internal temperature of the reaction
mixture. Dropwise addition of the base was initiated when the
substrate solution had cooled to 0 °C, ∼35 min after the
(7) Prolonged (>1 d) drying in vacuo (0.5 mmHg) may cause partial
dehydration of the product. Rehydration occurs upon exposure to the
atmosphere.

Direct Alkylation of Pseudoephedrine Glycinamide Hydrate
J . Org. Chem., Vol. 64, No. 9, 1999 3327
the reagents had been added, cooling was discontinued, and
the reaction mixture was stirred at 23 °C for 10 h, until the
lithium metal appeared to have been completely consumed.8
The resulting suspension was cooled to 0 °C in an ice-water
bath, and solid (R,R)-(-)-pseudoephedrine glycinamide hydrate
(48.1 g, 200 mmol, 1 equiv) was added in one portion. The
reaction mixture was stirred vigorously at 0 °C for 2.5 h to
allow the crystalline hydrate to react, forming a fine, orange
enolate suspension. Allyl bromide (18.2 mL, 210 mmol, 1.05
equiv) was added dropwise via syringe over a period of 10 min,
so as to maintain the temperature of the reaction suspension
below 8 °C. After completed addition, the resulting solution
was stirred at 0 °C for 1 h. Water (500 mL) was added slowly
to the reaction mixture, and the resulting biphasic mixture
was carefully acidified to pH 0 by the addition of aqueous
hydrochloric acid solution (6 M, 300 mL), maintaining the
temperature below 10 °C. The mixture was transferred to a
4-L separatory funnel, and ethyl acetate (750 mL) was added.
The organic layer was separated and subsequently extracted
sequentially with 300-mL portions of 3 M and 1 M aqueous
hydrochloric acid solution, respectively. The aqueous layers
were combined and cooled to 5 °C by stirring in an ice-water
bath. To the cold aqueous solution was added 50% aqueous
sodium hydroxide solution (160 mL) at such a rate as to
maintain the temperature of the solution at or below 25 °C.
The basified solution (pH 14) was extracted sequentially with
one 500-mL portion and three 250-mL portions of dichlo-
romethane. The combined organic extracts were dried with
anhydrous solid potassium carbonate and then were filtered
and concentrated in vacuo. The de of the crude reaction product
was shown to be 92% by capillary GC analysis of an acetylated
sample (190 °C, major tR ) 22.2 min, minor tR ) 19.5 min), as
described in the general procedures. Toluene (200 mL) was
added to the cold reaction product, a pale yellow liquid, and
the resulting solution was concentrated in vacuo (rotary
evaporator, then at 0.5 mmHg for 2 h), affording a solid
residue. The solid was dissolved in hot toluene (100 mL), and
the resulting solution was allowed to cool to 23 °C. Crystal-
lization occurred after 1 h. The crystallization flask was
allowed to stand for an additional 3 h at 23 °C and then was
cooled to -20 °C for 1 h. The crystals were collected and
washed with three 100-mL portions of ether and then were
dried in vacuo (0.5 mmHg) at 23 °C for 24 h to provide 29.4 g
(56.0%) of diastereomerically pure product (2g). The mother
liquors were concentrated and subjected to recrystallization
twice, as above, to afford an additional 2.2 g of diastereomeri-
cally pure product (4.2%, de > 99%, mp 71-73 °C). Anal. Calcd
for C15H22N2O2: C, 68.67; H, 8.45; N, 10.68. Found: C, 68.48;
H, 8.68; N, 10.76. Spectroscopic data were identical to those
previously reported.1
Ack n ow led gm en t. Financial support from the Na-
tional Science Foundation is gratefully acknowledged.
P.S. acknowledges postdoctoral fellowship support from
the Swiss National Science Foundation and the Ciba-
Geigy-J ubila¨ums-Stiftung. D.W.K. acknowledges an
NSF predoctoral fellowship. We thank Dr. J ohn Greaves
for HRMS data.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and analytical data for the products of Table 2 and copies
of 1H and 13C NMR spectra of compounds 2d and 2j. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(8) Small amounts of undissolved lithium metal may remain and
do not influence the subsequent alkylation reaction; however, care
should be exercised during the quenching of the reaction. Sonication
has been reported to accelerate the rate of reaction of lithium metal
with diisopropylamine.5
J O990341Z