An efficient large-scale process for the human leukocyte elastase inhibitor, DMP 777

Article abstract of DOI:10.1021/op015507o

This report describes a new convergent, selective, and economical synthesis of DMP 777, ending with the coupling of the chiral β-lactam half of the molecule (1) to the chiral amine as the isocyanate (2). Other steps involve the coupling of the β-lactam 3 to the phenolic moiety under phase-transfer conditions, followed by resolution of the resulting piperazine derivative using a chiral acid, and recycling of the undesired enantiomer also under phase-transfer conditions. The chiral amine 4 was produced efficiently starting from (R)-α-methylbenzylamine and the corresponding butyrophenone.

Full text of DOI:10.1021/op015507o

Organic Process Research & Development 2002, 6, 54−63
An Efficient Large-Scale Process for the Human Leukocyte Elastase Inhibitor,
DMP 7771
Louis Storace,* Luigi Anzalone, Pat N. Confalone, Wayne P. Davis, Joseph M. Fortunak, Mark Giangiordano,
James J. Haley, Jr., Kenneth Kamholz, Hui-Yin Li, Philip Ma, William A. Nugent, Rodney L. Parsons, Jr.,
Patrick J. Sheeran, Charlotte E. Silverman, Robert E. Waltermire, and Christopher C. Wood
Chemical Process Research and DeVelopment Department, Pharmaceutical Research Institute, Bristol-Myers Squibb
Pharma Company, Deepwater, New Jersey 08023-0999, U.S.A.
Abstract:
focused on identifying efficient conditions for elaboration
of 3 into racemic 1, and its subsequent resolution. The
precursor to isocyanate 2, amine 4, is not an article of
commerce; therefore, a process from simple materials had
to be devised. We will discuss in turn the synthesis of
fragments 1 and 2 and conclude by describing a procedure
for their efficient coupling.
This report describes a new convergent, selective, and economi-
cal synthesis of DMP 777,1 ending with the coupling of the chiral
â-lactam half of the molecule (1) to the chiral amine as the
isocyanate (2). Other steps involve the coupling of the â-lactam
3 to the phenolic moiety under phase-transfer conditions,
followed by resolution of the resulting piperazine derivative
using a chiral acid, and recycling of the undesired enantiomer
also under phase-transfer conditions. The chiral amine 4 was
produced efficiently starting from (R)-r-methylbenzylamine and
the corresponding butyrophenone.
Contemporary pharmaceuticals often contain multiple
chiral centers which provide a significant challenge to the
process chemist. Fortunately, an array of tools is now
available to address absolute stereochemistry, including
classical resolution, chiral auxiliaries, asymmetric catalysis,
and enzyme resolution. It is increasingly evident that no
single method is consistently superior. A thorough approach
to process development requires the consideration of all
available chiral technologies to provide the most efficient
route to a given drug candidate.
DMP 777, an inhibitor of leukocyte elastase, could
prevent degradation of the structural proteins elastin and
collagen, a process implicated in cystic fibrosis and rheu-
matoid arthritis.2
Retrosynthetic analysis indicates that DMP 777 should
be accessible by coupling two enantiopure fragments, â-
lactam 1 and isocyanate 2 (Figure 1). Potentially, â-
lactam 1 may derive from the commercially available
racemic lactam 3.3 Consequently route-development efforts
Figure 1. Retrosynthetic analysis of DMP 777
A short convergent process required a practical syn-
thesis and resolution of the â-lactam portion (Scheme 1).
The reaction conditions had to be balanced to provide the
requisite basic elimination of propionic acid from 3 to give
the reactive 3,3-diethylazetin-2-one, 5, without hydrolysis
of lactam. Phase-transfer catalytic conditions using a tetra-
butylammonium salt gave the most efficient reaction, with
an isolated yield of 88%.4 About 25 chiral acids were
screened for utility in the resolution. Diacetoneketo-
gulonic acid (DAG) gave by far the best resolution, with
the R-enantiomer DAG salt crystallizing out in high ee and
thus overcoming a 10-bond separation between amine
group and the chiral center. The S-isomer was then easily
precipitated, DAG was recovered, and the R-isomer was
racemized under phase-transfer conditions similar to those
for its original synthesis (conversion of 3 to 1-racemate).
DAG is an intermediate in an industrial-scale ascorbic
* Corresponding author. Telephone: 302-376-1984. E-mail: storacel@
yahoo.com. The research described was carried out at the former DuPont Merck
Pharmaceutical Company followed by the former DuPont Pharmaceuticals
Company.
(1) Sheeran, P. J.; Anzalone, L.; Li, H.-Y.; Fortunak, J. M.; Storace, L. U.S.
Patent 6,194,569, 2001. This work was presented at the 220th American
Chemical Society National Meeting, August 21, 2000.
(2) Macdonald, S. J. F.; Clarke, G. D. E.; Dowle, M. D.; Harrison, L. A.;
Hodgson, S. T.; Inglis, G. G. A.; Johnson, M. R.; Shah, P.; Upton,
R. J.; Walls, S. B. J. Org. Chem. 1999, 64, 5166-5175 and references
therein.
(3) Reference to previous process: Cvetovich, R. J.; Chartrain, M.; Hartner,
F. W., Jr.; Roberge, C.; Amato, J. S.; and Grabowski, E. J. J. J. Org. Chem.
1996, 61, 6575-6580.
(4) A reference for use of PTC in synthesis of â-lactam derivatives:
Murakami, M.; Aoki, T.; Nagata, W. Heterocycles 1990, 30, 567-
581. DAG reference: Mohasci, E.; Leimgruber, W. Organic Syn-
theses; John Wiley
pp 826-829.
& Sons: New York, 1988; Collected Vol. 6,
54
•
Vol. 6, No. 1, 2002 / Organic Process Research & Development
10.1021/op015507o CCC: $22.00 © 2002 American Chemical Society and The Royal Society of Chemistry
Published on Web 12/15/2001

acid process and so was available in bulk. For an alterna-
tive synthesis of 1, see ref 5.
impractical from a production perspective when compared
to a simple Grignard reagent.
Scheme 3. Chiral amine, D-phenylglycinol approach
Scheme 1. Chiral lactam
The starting propionyloxylactam 3 and the benzoylpiper-
azide 6 were obtained from commercial sources. Initially, 6
was synthesized by the two methods indicated (Scheme 2).
The convenient one-step DCC coupling was unusual because
it had to be heated to reflux to obtain a clean high-yield
reaction. DCC couplings are typically carried out at 0-25
°C. In this case, the bifunctional nature of the acid caused
some reaction of DCC at the phenol end. Howerver, on
heating, equilibria favored formation of the desired product.
Some pilot-plant facilities avoid the use of DCC because it
is a skin sensitizer. The three-step pivaloyl chloride route to
the benzoylpiperazide avoids DCC and is also efficient since
none of the intermediates need to be isolated.
One underutilized method of enantioselective synthesis of
amines involves the use of phosphinylimines.8 Phosphinyl-
imines are readily made from arylaldoximes and chlorophos-
phines, or from aldehydes and phosphinamides. Typically,
reaction of phosphinylimines with a dialkylzinc reagent in
the presence of a chiral moderator provides phosphinamides
in high ee and moderate yields. One of the main advantages
of this route is that chromatography is not necessary. Phos-
phinamides are easily isolated, purified, and crystallized. The
best ligand is the morpholine derivative of norephedrine, as
reported by Soai. The reaction (Scheme 4) is slow, requiring
16 h at 0 °C followed by completion at room temperature for
several hours, which allows time for a variety of byproducts
to form. The main byproducts were derived from reduction:
simple reduction of phosphinylimine to an N-benzylphos-
phinamide, and reductive formation of diphenylphosphine
oxide which adds to phosphinylimine in conjugate fashion.
Other ligands including quinines, prolinol derivatives, and
other norephedrine derivatives (all vicinal amino alcohols)
mediated the reaction but gave inferior results. In general
the phosphinylimine route giVes acceptable selectiVity only
with aromatic phosphinylaldimines, and not with ketimines.
Scheme 2. Benzoylpiperazide
A number of routes to the chiral amine 4 were considered,
including the published routes and an aminotransferase
enzymatic resolution of the racemic amine. Five different
enantioselective routes were successfully demonstrated as
summarized in Schemes 3-7. The fifth route, clearly the
most efficient from a raw material and operational cost basis,
was ultimately selected for scale up. The enantioselective
processes include routes based on chiral auxiliaries as well
as asymmetric catalysis and provided reaction enantioselec-
tivities ranging from 86 to 96% ee. Formation and isolation
of the hydrochoride salt of 4 upgraded the ee to about 99%.
The hydrochoride salt of 4 was needed for conversion to
isocyanate in the next step.
Scheme 4. Chiral amine, phosphinylimine-alkylzinc route
The D-phenylglycinol approach6,7 (Scheme 3) worked well
using allylmagnesium chloride as nucleophile, giving addi-
tion in 89% de. The allyl group could be reduced and the
chiral auxiliary hydrogenolyzed in one step. Use of propyl-
magnesium chloride provided a slow reaction rate, byprod-
ucts, and poor yield. Propylcerium dichloride7 gave addition
in 94% de, and 70% isolated yield, but was relatively
(5) Lipase route: Roberge, C.; Cvetovich, R. J.; Amato, J. S.; Pecore, V.;
Hartner, F. W., Jr.; Greasham, R.; Chartrain, M. J. Ferment. Bioeng. 1997,
83(1), 48-53.
(6) Li, H.-Y.; Anzalone, L.; Ma, P. U.S. Patent 5,932,749, 1999.
(7) Wu, M. J.; Pridgen, L. J. Org. Chem. 1991, 56, 1340-1344.
Another method for the synthesis of chiral amines uses a
chiral sulfur atom as an auxiliary.9 Sulfinylimine 7 was
synthesized using a commercially available menthyl sulfinate
Vol. 6, No. 1, 2002 / Organic Process Research & Development
•
55

(Scheme 5). It was purified by silica gel chromatography.
Diastereoselective reduction of the imine gave the desired
sulfinamide 8, which could be readily isolated and purified.
Deprotection under mild acidic conditions gave 4. This was
the only route in which the chiral auxiliary was recoverable.
The need for chromatography of 7 inhibited further develop-
ment of this route.
The (R)-phenylethylamine route (Scheme 7) was deter-
mined to be the most efficient, due to yield, low cost of
starting materials and reagents, and operational efficiency.6
Starting from butyrophenone derivative 11, the imine was
produced as a 2:1 E/Z mixture (The isomer ratio was
conveniently monitored by HPLC and NMR). Success of
this route was dependent on a number of factors. The E-imine
12 reduces selectively, using Raney nickel, to the required
diastereomer of 13. Under optimum reaction conditions, the
Z-imine is reduced very slowly to the undesired diastereomer
but isomerizes more rapidly to the E-isomer. Finally, under
optimum palladium-catalyzed reduction conditions, the
benzyl group containing the less substituted aromatic ring
is highly selectively cleaved. Similar steric-controlled selec-
tive reductions have been reported.11
Scheme 5. Chiral amine, sulfinylimine route
Scheme 7. Chiral amine, method of choicea
A catalytic asymmetric hydrogenation route using a
DuPhos catalyst10 (Scheme 6) was also demonstrated. The
required enamide 9 was prepared by addition of propylmag-
nesium bromide to piperonylonitrile followed by acylation
with acetic anhydride. The crude enamide was purified by
flash chromatography. Asymmetric hydrogenation of the
enamide was carried out using (MeDuPHOS)Rh(COD)+ as
catalyst in methanol solvent. The hydrogenation proceeded
at an 1800:1 substrate-to-catalyst ratio (0.05 mol % catalyst)
under mild conditions (30 psi hydrogen, room temperature).
The desired amide 10 was produced quantitatively in 96.5%
enantiomeric excess. A single crystallization afforded pure 10
in 86% yield and >99% enantiomeric excess. The apparent
need for chromatography after the first step could possibly
be overcome with further research, but scale-up of this route
was inhibited.
a 1. TiCl4 (0.55 equiv), (R)-PEA (1.2 equiv), Et3N (2 equiv), toluene. 2.
Ra-Ni, H2, toluene-EtOH, 150 psi, 25 °C, 4 h, then 70 °C, 2 h. 3. 10%Pd/C
(10 wt %), H2, AcOH, toluene-EtOH. 4. HCl in i-PrOH-toluene; i-PrOH-
n-heptane.
Even though the efficient synthesis of isocyanates from
amine HCl salts and phosgene in refluxing toluene (Scheme
8) has been known for many decades, a surprise awaited us.
Small amounts of iron dissolved in liquid phosgene (from a
steel cylinder) induced racemization of the isocyanate!
Transfer of phosgene as a gas avoided this problem, leaving
iron behind in the cylinder. In control experiments, it was
also found that pyridine inhibits the effect of iron (probably
through coordination with iron) and that ferric chloride
racemizes the isocyanate at reflux in toluene. The mechanism
of racemization has not been investigated but may in-
volve formation of a π complex between the electron-rich
aromatic ring of 2 and an iron cation, or by a mechanism
similar to that of the known racemization of R-methyl-
benzylamine by Pd.12 Phosgene equivalents such as diphos-
gene, triphosgene, and so forth were shown to be useful
Scheme 6. Chiral amine, DuPHOS routea
a (MeDuPHOS)Rh+:1,2-bis((2R,5R)-2,5-dimethylphospholano)benzene-
(cyclooctadiene)rhodium (I) tetrafluoroborate.
(10) Burk, M. J.; Casy, G.; Johnson, N. B. J. Org. Chem. 1998, 63, 6084; Zhu,
G.; Zhang, X. J. Org. Chem. 1998, 63, 9590; Zhang, F.-Y.; Pai, C.-C.;
Chan A. S. C. J. Am. Chem. Soc. 1998, 120, 5808. Review of DuPHOS
ligands and their uses see: Burk, M. J. Chemtracts: Org. Chem. 1998,
11, 787.
(11) Bringmann, G.; Geisler, J.-P.; Geuder, T.; Kunkel, G.; Kinzinger, L. Liebigs
Ann. Chem. 1990, 795-805.
(8) Soai, K. J.; Hatanaka, T.; Miyazawa, T. J. Chem. Soc., Chem. Commun.
1992, 1097-1098. Stec, W. J. Synthesis 1978, 521-3, and 1982, 270-
272. Hutchins, R. O.; Abdel-Magid, A.; Stercho, Y. P.; Wambsgan, A. J.
Org. Chem. 1987, 52, 704-706. Brown, C.; Hudson, R. F.; Maron, A.;
Record, K. A. F. J. Chem. Soc., Chem. Commun. 1976, 663-664.
(9) Hua, D. H.; Miao, S. W.; Chen, J. S.; Iguchi, S. J. Org. Chem. 1991, 56,
4-6 and references therein. Davis, F. A.; Reddy R. E.; Portonovo, P. S.
Tetrahedron Lett. 1994, 35, 9351-9354.
(12) Murahashi, S.-I.; Noriaki, Y.; Tsumiyama, T.; Kojima, T. J. Am. Chem.
Soc. 1983, 105, 5002-5011.
56
•
Vol. 6, No. 1, 2002 / Organic Process Research & Development

but were considered to be less practical or efficient than the
phosgene method.
Experimental Section
For characterization of known compounds 1-4, and DMP
777 see ref 3.
Scheme 8. Chiral isocyanate
3,3-Diethyl-4-{4-[(4-methylpiperazinyl)carbonyl]phen-
oxy}azetidin-2-one (Racemic 1). Water (219 L) was charged
to a 300-gal vessel, followed by potassium carbonate (137
kg, 993 mol), tetrabutylammonium hydrogen sulfate (415
g, 1.2 mol; about 4.8 mol % relative to 3), 6 (56 kg, 244
mol; 0.97 mol equiv relative to 3), and isopropyl acetate (330
L). The mixture was warmed to 45 °C, and a solution of
3,3-diethyl-4-oxoazetidin-2-yl propanoate 3 (59 kg, 85% w/w
content, 251 mol; health caution: compound 3 is a mutagen)
in isopropyl acetate (56 L) added over 1.2 h. The 6 dissolves
as reaction proceeds. After being held at 45 °C for 1 h, the
reaction was observed to be complete by HPLC (<0.2% of
total area remaining for 6 versus the completion criterion of
less than 5% of 6). The lower (aqueous) layer was separated,
and the organic phase was concentrated under reduced
pressure (100 mm/40 °C) to about 250 L. Methyl tert-butyl
ether (250 L) was added, and the mixture was stirred for
10 h at 15 °C while the product slowly crystallized. After
cooling to -10 °C over 2 h, the mixture was filtered, and
then the wet cake was returned to the vessel and slurried in
290 L of methyl tert-butyl ether at 20 °C. The slurry was
cooled to -5 °C and filtered (2% yield loss in the filtrate).
After drying at 50 °C and 50 mm in a vacuum oven to
constant weight, 75.6 kg of product was obtained as a white
solid, 98.0% w/w purity, 88% yield (corrected for purity of
both starting material and product). A sample was purified
for use as analytical reference material by dissolving it in
warm toluene, washing the solution thoroughly with water,
drying, and then crystallizing the product by addition of tert-
butyl methyl ether. mp 105-107 °C.
Resolution. (4S)-3,3-Diethyl-4-{4-[(4-methylpiperazinyl)-
carbonyl]phenoxy}azetidin-2-one 1. 2,3:4,6-Di-O-diiso-
propylidene-2-keto-L-gulonic acid (DAG, 28.3 kg, 96.5 mol)
was slurried in 266 L of isopropyl acetate and heated with
stirring to 70 °C in a 200-gal, glass-lined vessel. This gave
a pale-yellow, slightly hazy solution. A 300-gal, glass-lined
vessel was charged with 532 L of isopropyl acetate followed
by a small DAG charge (5.7 kg, 19.3 mol, to ensure that no
chemical degradation of the racemic amine occurs with
heating.). The racemic amine (66.8 kg, 193 mol) was then
charged and the resulting slurry heated to 70 °C to give a
clear solution. The DAG solution was then added, at a
constant rate to the racemic amine solution over 2.5 h. The
suspension was cooled to 20 °C over 3 h and held with
stirring for an additional 11 h. A stir time of >5 h is required
to reach the maximum ee possible for 1 in the liquors. The
salt was filtered by centrifugation and the wet cake washed
with 25 kg of isopropyl acetate. This solid was later dried
to give 62.9 kg of 1-enantiomer DAG salt (92% ee). The
ee of the combined liquors and wash was 93% for 1. The
liquors were transferred back to the 300-gal vessel (which
had been cleaned with water and then a methanol boil-out)
and washed with 20% aqueous potassium bicarbonate
solution. The layers were separated, and the organic layer
In preparation for pilot-plant-scale reactions of the final,
convergent step (Scheme 9) a number of solvents and
catalysts (including DBU) were evaluated. These reactions
were sometimes complicated by formation of variable
amounts of symmetrical urea 14, which required extra
purification procedures and lowered isolated yields to 70-
80%. Furthermore, there was mass spectroscopic evidence
for a 1:2 adduct between DBU and the isocyanate,13 which
was proposed to form “symmetrical urea” during aqueous
workup. The working hypothesis used in development of an
efficient coupling process was that the catalyst needs to be
of sufficient basicity to deprotonate the lactam while having
poor nucleophilicity, and that the resulting anion needs to
react with isocyanate much faster than it would decompose
by elimination of phenolate. Base, cation, solvent, and
temperature were the relevant variables examined. It was
found that the reaction required a base with conjugate acid
pKa of >20. The more strongly N-coordinating lithium
cation provided the necessary lactam anion stability. Only
trace levels of symmetrical urea are formed when lithium
tert-butoxide (5 mol %) is used as catalyst and the stoi-
chiometry carefully controlled. This catalyst offered the
most robust process and allowed optimization of volume,
solvent, temperature, and so forth. A clean reaction made
it possible to crystallize the coupled product directly out of
the quenched and washed reaction solution in high yield and
purity.
Scheme 9. Final, convergent, step; coupling to produce
DMP 777
This entire synthesis has been carried out safely, reliably,
and volume efficiently on pilot-plant scale. A total of 190
kg of DMP 777 were prepared by this process.
(13) DBU-isocyanate 1:2 adducts: Oediger, H.; Moeller, F. Ger. Offen.
2640964, Bayer A-G, 1978; Chem. Abstr. 1978, 88, 190910.
Vol. 6, No. 1, 2002 / Organic Process Research & Development
•
57

was washed with 188 kg of 15% aqueous sodium chloride
solution. The isopropyl acetate solution was concentrated
by distillation at about 100 mmHg/40 °C. A total of
460 L of isopropyl acetate was collected out of a total of
approximately 1000 L. Vacuum was released, and 235 kg
of heptanes was charged to the vessel. Vacuum, heating, and
distillation were resumed until a final volume of approxi-
mately 110 L was reached. GC analysis showed the
composition of the solvent to be 60:40 isopropyl acetate-
heptanes with the target being 25-40% isopropyl acetate.
After cooling to 20 °C and stirring for 9 h, an additional
60 kg of heptanes was added and the mixture stirred at
20 °C for 2 h. Analysis of a sample showed the liquors to
be about 79% ee 1. Quantitative HPLC showed the liquors
contained only about 350 g of 1 (0.19 wt % 1 in 207 kg of
liquors). The precipitated 1 was filtered and then washed
with 100 kg of heptanes. After drying to a constant weight,
31.0 kg of material was obtained. The dried product had an
assay of 100%. KF determination showed 0.29% water by
weight, and the ee was 94%. The yield, corrected for assay
of both starting material and product, was 91.6%. This
material was useable for the next step. A sample was
recrystallized from toluene and then again from chlorobutane;
it was then slurried in 20 °C tert-butyl methyl ether to >99%
ee, mp 121-123 °C. Chiral HPLC method: column, Chiracel
OD 4.6 mm × 25 cm, 10 µm; detector: 240 nm; 40 °C;
flow rate 1.0 mL/min; mobile phase: A, hexane; B, ethanol;
C, 2-propanol; linear gradient ratio of A:B:C: 90:10:0 at
t ) 0; to 80:20:0 at 10 min; to 75:20:5 at 15 min; to
65:20:15 at 20 min; to 60:20:20 at 25 min; stop time
35 min; post-time 5 min. Retention times: toluene 2.9 min;
1, 13.0 min; ent-1, 18.4 min.
Racemization of ent-1-DAG to rac-1. Water (290 L) was
charged to a 300-gallon vessel, followed by potassium
carbonate (64 kg, 464 mol) and 34 kg of sodium chloride.
ent-1-DAG (72.5 kg, 117 mol) was added and then isopropyl
acetate (442 kg). The mixture was agitated for 15 min, and
then the aqueous layer was withdrawn and charged to a
200-gal reactor for later DAG recovery. This solution was
extracted with 316 kg of isopropyl acetate, and the extract
was combined with the organic solution in the 300-gal
reactor, whereupon 2.6 kg of 6 (10 mol %; added to inhibit
decomposition), tetrabutylammonium bromide (188 g), and
aqueous potassium carbonate (64 kg in 103 L water) were
also added to the 300-gal vessel. The mixture was then heated
to 60 °C until the ee of the mixture was <5% or the impurity
levels began to rise (typically 2-3 h). If the total area %
response for rac-1 reaches 90% or less, then racemization
should be stopped to avoid material loss due to poor crys-
tallization efficiency. The mixture was then cooled to 20-25
°C, and the phases were separated. The isopropyl acetate
layer was washed with 60 L of 25% aqueous potassium
carbonate solution and then with 75 kg of 20% aqueous
sodium chloride solution. The isopropyl acetate layer was
concentrated at 40-50 °C/40-50 mm to about one-third to
one-fourth of its original volume. To this solution at 30 °C
was added 87 kg of tert-butyl methyl ether. The solution
was cooled to 20 °C over 1 h and then held for 2 h. The
suspension was then cooled to -10 °C over 1 h and held at
-10 °C for 2 h. The crystallization was complete when the
solution contained e3% w/w content of rac-1. The rac-1
was isolated by centrifugation with recycling of the mother
liquors, and the cake was then washed with 15 kg of tert-
butyl methyl ether. The rac-1 was dried (40 °C/50 mm) to
constant weight, 34 kg (99 area %, 99 wt %, 3.7% ee, 86%
yield).
Synthesis of 1-(4-Hydroxybenzoyl)-4-methylpiperazine,
6, using DCC. N-methylpiperazine (226 mL, 2.0 mol) was
added slowly (15 min) to a slurry of 4-hydroxybenzoic acid
(229 g, 2.0 mol) in 2 L of ethyl acetate at 50-70 °C. A
solution of dicyclohexylcarbodiimide (438 g, 2.1mol) in
0.4 L of ethyl acetate was added over 1.5 h to the salt slurry
at 70-78 °C. The mixture was refluxed 1 h and followed
by gradient HPLC: C18, 20-95% acetonitrile-H2O-0.05%
TFA, 250 nm; product at 2.6 min, starting acid at 4.5 min,
DCC adduct at 9.7 min. The mixture was cooled, 1.1 L of
2 M aqueous HCl was added, and the mixture was stirred
1.5 h. It was then filtered and the DCU cake rinsed
thoroughly with water (3 × 250 mL). The layers were
separated. The pH of the water layer (pH 1) was then adjusted
to 13 with NaOH and heated to hydrolyze any ester by-
products (2 area %) and then readjusted to pH 9. n-Butyl
alcohol (2 L) was added. The layers were separated at
75 °C (4% yield loss to aqueous layer). The organic layer
was azeotropically dried by distillation of butyl alcohol
(1 L). It was then filtered hot to remove salt (2 wt %). The
filtrate was concentrated by distillation to half its starting
volume. The product crystallized as the concentrate was
cooled to 0 °C. The product was collected by filtration, rinsed
with butyl alcohol, and dried in a vacuum oven at 95 °C.
Yield 380 g, 84% (97 wt % pure; 100 area %). About 4%
1
yield remained in the filtrate. mp 184-186 °C. H NMR
(400 MHz, DMSO-d6) δ 2.17 (s, 3H), 2.28 (br, 4H), 3.46
(br, 4H), 6.79 (d, J ) 8.6 Hz, 2H), 7.23 (d, J ) 8.6 Hz,
2H), 9.83 (br s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 45.6
(CH3), 44.4 (very broad, CH2 × 2), 54.5 (CH2 × 2), 114.8
(CH × 2), 126.0 (C), 129.1 (CH × 2), 158.7 (C-O), 169.2
(CdO). CIMS (M + H)+, 221.
Preparation of 1-(4-Hydroxybenzoyl)-4-methylpipera-
zine, 6, Using Pivaloyl Chloride. Isopropyl acetate (120 kg),
was charged to a 100-gal reactor, followed by triethylamine
(29 kg, 287 mol). The solution was cooled to 10 °C and
p-hydroxybenzoic acid (18.0 kg, 130 mol) added in portions.
The resulting salt slurry was warmed to 30 °C, held for
0.75 h, then cooled to 0-5 °C. Pivaloyl chloride (34.6 kg,
287 mol), was added over 1.0-2.0 h while controlling the
temperature at 0-5 °C. The reaction was held at 0-10 °C
until complete (about 3 h). Reaction progress was followed
by HPLC until the ratio of ester/mixed anhydride to
4-hydroxybenzoic acid was >30/1. The reaction mass at
8-12 °C was washed twice with 65 L of water. The solution
of ester/mixed anhydride (4-[(2,2-dimethylpropanoyl)oxy-
carbonyl]phenyl 2,2-dimethylpropanoate), at 0-5 °C was
treated with N-methylpiperazine (10.8 kg, 106 mol) while
maintaining at 0-10 °C. The reaction was warmed to
20-25 °C and held until complete (ester/amide to ester/
58
•
Vol. 6, No. 1, 2002 / Organic Process Research & Development

anhydride ratio >90/1 by HPLC). Water (65 L) was then
added and the pH adjusted to 8.4-8.6 with 30% sodium
hydroxide. The contents of the reactor were warmed to
40 °C, and the lower aqueous phase was separated. The
organic phase was washed with 68 kg of 7% aqueous sodium
bicarbonate at 40 °C, followed by 65 L of water and then
65 L of 30% aqueous sodium chloride (all at 40 °C). The
organic solution of the ester/amide (4-[(4-methylpiperazinyl)-
carbonyl]phenyl 2,2-dimethylpropanoate) was cooled to
20-25 °C, and sodium methoxide (25 kg of 25% solution
of in methanol, 116 mol) was added, while the temperature
was maintained at 25-35 °C. The reaction was held at
35-40 °C until complete (<1 area % of ester/amide) by
HPLC. It was then cooled to 20-25 °C, 24 L of water added,
and the pH adjusted to 8.4-8.6 with 37% HCl. After aging
at 20-25 °C for 0.5 h and then at 5-10 °C for 1.0 h, the
reaction was filtered and the wet cake washed with 39 kg of
i-PrOAc. The solvent-wet solid was slurried in 58 L of water.
After filtration it was dried in a vacuum tray drier at 45 °C/
50 mmHg to constant weight of 19.5 kg 6 (98 wt %), 67%
yield. HPLC method: column, Zorbax SB-Phenyl, 4.6 mm
× 15 cm, 5 µm; flow rate, 1.0 mL/min; 20 °C; 240 nm;
mobile phase 0.1% TFA/water and acetonitrile; linear
gradient; initial 5% CH3CN, hold 5 min, then to 25% CH3CN
at 10 min, hold 5 min, then to 50% CH3CN at 20 min, then
to 100% CH3CN at 25 min; stop time 35 min; post time
5 min. Retention times: 6, 5.2 min; p-hydroxybenzoic acid,
11.6 min; ester/amide, 17.2 min; ester/anhydride, 26.3 min.
2-[(1-(2H-Benzo[3,4-d]1,3-dioxolan-5-yl)but-3-enyl)-
amino]-2-phenyl(2R)ethan-1-ol. A 2 M solution of allyl-
magnesium chloride in THF (9.4 L) was added to a cold
solution (10-15 °C) of the imine (2.02 kg) in THF (9.0 L)
over a period of about 2 h. The rate of addition was
controlled to maintain the temperature below 30 °C. The
resulting mixture was aged for 1 h (<5 area % of imine, by
HPLC), cooled to 5-10 °C, and quenched by adding it
slowly to a 30% aqueous acetic acid solution (13 L) while
keeping the temperature below 30 °C. The organic phase
was separated and treated with 20% aqueous NaOH solution
to pH 8-10. The layers were separated, and the organic
solution was washed with 10% aqueous NaCl and then
concentrated to an oil under reduced pressure. Acetonitrile
(15 L) was added followed by tartaric acid (1 equiv, 1.1 kg).
The mixture was warmed to 50 °C, aged for 1 h, and then
cooled to 20 °C over 2-4 h. After 1-2 h at this temperature,
the product was filtered, washed with acetonitrile (10 L),
and dried to constant weight under vacuum at 45-50 °C to
give the benzyl glycinol tartrate salt, 2.57 kg, 96 wt %, 81%
yield as an off-white solid. 1H NMR (300 MHz, DMSO-d6)
δ 2.41 (m, 1H), 2.58 (m, 1H), 3.57 (m, 2H), 3.77 (m, 2H),
4.23 (s, OCH × 2, tartrate), 4.96 (m, 2H), 5.55 (m, 1H),
5.96 (s, 2H), 6.72 (dd, J ) 8.0 and 1.4 Hz, 1H), 6.79 (d,
J ) 8.0 Hz, 1H), 6.89 (d, J ) 1.4 Hz, 1H), 7.19-7.29 (m,
5H). 13C NMR (75 MHz, DMSO-d6) δ 40.0 (CH2), 60.0
(NCH), 62.0 (NCH), 64.8 (OCH2), 72.5 (OCH × 2, tartrate),
101.2 (CH2), 108.0 (CH), 108.2 (CH), 117.9 (CH2), 121.4
(CH), 127.6 (CH), 128.1 (CH × 2), 128.5 (CH × 2), 135.2
(CH), 136.3 (C), 141.1 (C), 146.6 (C-O), 147.6 (C-O),
174.0 (CO2 × 2, tartrate).
1
Ester/mixed anhydride: H NMR (400 MHz, CDCl3) δ 1.25
(s, 9H), 1.37 (s, 9H), 7.17 (d, J ) 8.8 Hz, 2H), 8.14 (d, J )
8.8 Hz, 2H). 13C NMR (100 MHz, CDCl3) δ 26.9 (CH3),
27.0 (CH3), 38.6 (C), 39.2 (C), 121.7 (CH), 126.6 (C), 131.8
(CH), 155.5 (C-O), 171.7 (CdO), 176.5 (CdO), 185.8
(CdO). Ester/amide: 1H NMR (300 MHz, CDCl3) δ 1.3
(s, 9H), 2.3 (s, 3H), 2.4 (br, CH2 × 2), 3.6 (br, CH2 × 2),
7.1 (d, J ) 8.6 Hz, 2H), 7.4 (d, J ) 8.6 Hz, 2H).
1-(2H-Benzo[3,4-d]1,3-dioxolan-5-yl)(1R)butylamine 4.
A nitrogen-sparged and vacuum-degassed solution of the
benzyl glycinol tartrate (2.5 kg) in methanol (9 L) and acetic
acid (4 L) was pressure-transferred to a slurry of “wet” 10%
palladium on carbon (∼50% water content, 0.8 kg) in
methanol (8 L) and acetic acid (4.0 L). The resulting slurry
was hydrogenated at ambient pressure and 20-25 °C for
24-48 h. Progress of the reaction was followed by HPLC.
The propenyl side chain reduces rapidly, followed by
“debenzylation”. Once the reaction was complete, the catalyst
was removed by filtration and washed with methanol. The
combined filtrates were then concentrated under reduced
pressure to a residue, which was partitioned between toluene
(3 L) and 1 N aqueous HCl (5 L). The aqueous phase was
separated and basified to pH 12-13 with 30% aqueous
NaOH solution in the presence of toluene (6 L). The layers
were separated, and the aqueous layer was extracted with
toluene (4 L). The combined organic solutions were washed
with 20% aqueous NaCl, clarified through a Celite pad, and
analyzed for product content. The toluene solution was
cooled to 10-15 °C and treated with 1 equiv 6 N HCl
in isopropyl alcohol while maintaining a temperature
below 20 °C. The slurry was aged 1 h at 20-25 °C and
then filtered. The solid was washed with toluene and dried
to give 4 HCl salt, 2.05 kg, 100 wt %, 99.9% ee, 82% yield.
Phosphinylimine Route. Piperonaldoxime (66 g, 0.40
mol) was dissolved in methylene chloride (660 mL) and then
Chiral Amine, D-Phenylglycinol Approach. (3E)-4-(2H-
Benzo[3,4-d]1,3-dioxolan-5-yl)-3-aza-2-phenyl(2R)but-3-
en-1-ol. A solution of piperonal (2.3 kg), D-phenylglycinol
(2.1 kg), and p-toluenesulfonic acid (2.3 g) in toluene (12 L)
was heated to reflux. Once the theoretical amount of water
had distilled (3-4 h; a Dean-Stark trap was used), the
reaction was analyzed (<5% piperonal as determined by 1H
NMR). The solution was cooled to 80-85 °C. Heptane (7 L)
was added slowly, and the resulting solution was cooled
further to 5-10 °C, and aged for 1 h. Product precipitation
begins at about 60 °C. The product was isolated by filtration
and dried under vacuum at 50-55 °C to give 3.85 kg of
1
imine (95.4% yield of 100% purity). H NMR (300 MHz,
CDCl3) δ 2.46 (br, 1H, OH), 3.86 (dd, J ) 11 and 4.4 Hz,
1H), 3.95 (dd, J ) 11 and 8.4 Hz, 1H), 4.45 (dd, J ) 8.4
and 4.4 Hz, 1H), 6.00 (s, 2H), 6.81 (d, J ) 8.0 Hz, 1H),
7.10 (d, J ) 8.0 Hz, 1H), 7.20-7.45 (m, 6H) 8.24 (s, 1H,
NdCH). 13C NMR (75 MHz, CDCl3) δ 67.8 (OCH2), 76.0
(NCH), 101.5 (CH2), 106.8 (CH), 108.1 (CH), 124.9 (CH),
127.4 (CH × 2), 127.5 (CH), 128.6 (CH × 2), 130.8 (C),
140.8 (C), 148.3 (C-O), 150.1 (C-O), 161.9 (NdCH).
Vol. 6, No. 1, 2002 / Organic Process Research & Development
•
59

triethylamine (0.41 mol) and heptane (200 mL) were added;
the solution was cooled to -45 °C. Chlorodiphenylphosphine
(79 mL, 0.40 mol) in 100 mL methylene chloride was added
dropwise over 25 min while maintaining the above tem-
perature. Hexanes (500 mL) was added while warming the
mixture to room temperature. The TEA-HCl was filtered
off. The filtrate was concentrated on a rotary evaporator, the
residue dissolved in 1.5 L of hot toluene, and 3 g of Darco
added. After filtration, the filtrate was concentrated by rotary
evaporation (distilled off 1 L of toluene). Heptane (800 mL)
was added, and N-(diphenylphosphinyl)piperonaldimine (128
g, 90% yield) was collected as a yellow powder, mp 138
(CH3), 19.4 (CH2), 41.9 (d, J ) 4 Hz CH2), 55.4 (d, J ) 1
Hz NCH), 100.9 (OCH2), 106.6 (CH), 108.0 (CH), 119.9
(CH), 128.2 (d, J ) 13 Hz, CH × 2), 128.4 (d, J ) 13 Hz,
CH × 2), 131.6 (d, J ) 3 Hz, CH), 131.7 (d, J ) 3 Hz,
CH), 131.8 (d, J ) 85 Hz, C-P), 131.8 (d, J ) 10 Hz, CH
× 2), 132.5 (d, J ) 10 Hz, CH × 2), 133.5 (d, J ) 82 Hz,
C-P), 137.9 (d, J ) 5 Hz, C), 146.4 (C-O), 147.6 (C-O).
31P NMR (162 MHz, CDCl3, H3PO4 external) δ 23.5. ESMS
(M + H)+, 394.
4. The phosphinamide (1.0 g), was dissolved in methanol
(4 mL) and 2.0 mL of TFA added. The solution was heated
to reflux until completion (3 h). The solution was concen-
trated on a rotovap and the residue partitioned with water
and toluene. The toluene layer was discarded. The pH of
the water layer was adjusted to 12 with NaOH and the
product extracted with i-PrOAc to give the (R)-amine, 4
(0.34 g, 80% yield, 90% ee).
1
°C. H NMR (300 MHz, CDCl3) δ 6.07 (s, 2H), 6.90 (d, J
) 7 Hz, 1H), 7.45 (m, 7H), 7.60 (s, 1H), 7.92 (m, 4H), 9.15
(d, J ) 30 Hz, 1H, PNdCH). 13C NMR (75 MHz, CDCl3)
δ 101.9 (CH2), 107.3 (CH), 108.3 (CH), 128.4 (d, J ) 13
Hz, CH × 4), 128.8 (CH), 131.5 (d, J ) 10 Hz, CH × 4),
131.6 (d, J ) 3 Hz, CH × 2), 148.6 (C-O), 152.5 (C-O),
172.3 (d, J ) 7 Hz, CdNP), several peaks were not resolved.
CIMS (M + H)+, 350.
Byproduct of the Propylzinc Reaction, the Phosphine
Oxide Addition Product.
The phosphinylimine (11.0 g, 90%, 30 mmol) and the
morpholine derivative of (1R, 2S)-norephedrine (10 g, 45
mmol) were added to a mixture of dry toluene (100 mL)
and heptane (100 mL); the mixture was cooled to -15 °C,
and dipropylzinc (13 mL, 94 mmol) was added via syringe
while maintaining -15 °C. The mixture was stirred 15 h at
0 °C and then 24 h at 19 °C. Quantitative HPLC showed
67% reaction yield of the desired phosphinamide [HPLC
method: C18, 25 cm, 60% acetonitrile-water (0.05% TFA),
225 nm] with 69 area % product (6.5 min), 3% starting
material (5.5 min), 3% reduction product (4.0 min), and 8%
of the phosphine oxide addition product (4.3 min, see below).
The mixture was poured into cold 5% citric acid (500 mL)
and the solid phosphine oxide addition product collected by
filtration. The layers were separated (the chiral moderator
can be recovered from the aqueous layer), and the organic
phase was shaken with 5% citric acid and then 5% NaHCO3,
dried (Na2SO4), stirred with 12 g of silica gel, filtered, and
rotovapped to give 10 g of residue. A sample of the product
was purified by preparatory TLC for chiral HPLC analysis:
Chiralcel OD, 15% ethanol/hexanes, flow rate 0.7 mL/min,
(R)-enantiomer at 7.6 min, (S)- at 9.2 min, which showed
89% ee (R)-isomer. The product was dissolved in hot toluene
(50 mL) and heptane added (50 mL). The product crystallized
slowly with cooling to 0 °C and was collected by filtration
to give the phosphinamide (6.0 g, 90 area %, 89% ee, 46%
yield). The reaction times and isolation were not optimized.
A product sample was purified for characterization by flash
chromatography (5% i-PrOH/30% ether/hexanes) followed
by recrystallizations from 1-chlorobutane and then toluene/
hexanes, mp 140 °C. IR 3436, 3190. NMR shows the phenyl
rings are diastereotopic. 1H NMR (300 MHz, CDCl3) δ 0.82
(t, J ) 7 Hz, 3H), 1.20 (m, 2H), 1.70 (m, 1H), 1.90 (m,
1H), 3.22 (dd, J ) 6 and 10 Hz, 1H, NH), 4.07 (m, 1H,
NCH), 5.94 (s, 2H), 6.55 (d, J ) 8 Hz, 1H), 6.67 (s, 1H),
6.68 (d, J ) 8 Hz, 1H), 7.34 (td, J ) 7 and 3 Hz, 2H),
7.4-7.5 (m, 4H), 7.76 (dd, J ) 7 and 12 Hz, 2H), 7.86 (dd,
J ) 7 and 12 Hz, 2H). 13C NMR (75 MHz, CDCl3) δ 13.7
IR 3425. NMR shows the phenyl rings are diastereotopic
(four different rings). H NMR (300 MHz, CDCl3) δ 4.5
1
(br, 1H, NH), 5.13 (q, J ) 11 Hz, 1H, NCH), 5.82 (s, 2H,
OCH2), 6.40 (dd, J ) 8 and 2 Hz, 1H), 6.54 (d, J ) 8 Hz,
1H), 6.76 (s, 1H), 7.17-7.43 (m, 13H), 7.55 (m, 5H, ortho
to P), 8.06 (dd, J ) 10 and 7 Hz, 2H, ortho to P). 13C NMR
(not all peaks were resolved; 75 MHz, CDCl3) δ 53.4 (d,
J ) 39 Hz, PCH), 100.9 (OCH2), 107.7 (CH), 109.3 (CH),
122.7 (CH), 127.6-129.0 (m), 131.0-132.5 (m).31P NMR
(162 MHz, CDCl3) δ 26.2 (d, JPP ) 27 Hz), 34.0 (d, J ) 27
Hz). CIMS (M + H)+, 552.
Chiral Amine: Sulfinylimine Route. (1E)-2-(2H-Ben-
zo[d]1,3-dioxolen-5-yl)-1-aza-1-[(4-methylphenyl)(R)sul-
finyl]pent-1-ene, 7. Piperonylonitrile (25 g, 0.17 mol) was
dissolved in toluene (100 mL) and reacted with n-propyl-
magnesium chloride (0.18 mol) at 70 °C/30 min. After
cooling and addition of THF (75 mL), (+)-menthyl p-
toluenesulfinate solid (25 g, 0.085 mol) was added at -50
°C. The mixture was slowly warmed to 15 °C and then
quenched with cold 5% citric acid. After thorough washing,
the toluene solution was concentrated on a rotovap to give
54 g of oil containing the desired sulfinylimine 7, menthol,
piperonal, and so forth. The reactions were followed by
HPLC [C18, 80% acetonitrile-H2O (0.05% TFA), flow rate
1 mL/min, 225 nm; sulfinylimine at 4.0 min]. The product
was purified by flash chromatography (silica gel, 20-50%
t-BME-hexanes and then again with chloroform) and
yielded 20 g (70% yield). The product was a 1:1 mixture of
two geometric isomers. 1H NMR (300 MHz, CDCl3) δ 1.02
(t, J ) 7 Hz, 3H), 1.63 (m, 2H), 2.40 (s, 3H), 3.13 (br, 2H),
6.00 (s, 2H), 6.80 (d, J ) 7 Hz, 1H), 7.32 (d, J ) 7 Hz,
2H), 7.38 (br d, 1H), 7.42 (br, 1H), 7.75 (d, J ) 7 Hz, 2H).
13C NMR (due to configurational mobility, not all peaks
60
•
Vol. 6, No. 1, 2002 / Organic Process Research & Development

resolved; 75 MHz, CDCl3) δ 14.1 (CH3), 21.4 (ArCH3), 22.1
(CH2), 35.1 (br, CH2), 101.7 (CH2), 107.8 (CH), 107.9 (CH),
123.1 (CH), 125.3 (CH × 2), 129.7 (CH × 2), 131.7, 141.6,
144.0, 148.1. CIMS (M + H)+, 330.
CDCl3) δ 13.8 (CH3), 17.9 (CH2), 40.2 (CH2), 101.7 (OCH2),
107.7 (CH), 107.8 (CH), 124.1 (CH), 131.8 (C), 148.0
(C-O), 151.4 (C-O), 198.4 (CdO).
5-((1E)(3R)-2-Aza-3-phenyl-1-propylbut-1-enyl)-2H-
benzo[d]1,3-dioxolene 12. A 100-gal, glass-lined reactor was
charged sequentially with toluene (144 kg), (R)-(+)-R-
methylbenzylamine (12.9 kg, 106 mol, 1.25 equiv), triethyl-
amine (22.2 kg, 219 mol, 2.5 equiv), and 11 (16.6 kg, 86.4
mol, 1 equiv) and then cooled to 0 °C. Titanium (IV) chloride
(10.0 kg, 52.7 mol, 1.2 equiv) was charged subsurface over
1-2 h with vigorous stirring, while maintaining the tem-
perature at less than 20 °C. After addition was complete,
the reaction mass was heated to a gentle reflux (111 °C)
with continued stirring for 6 h. After the reaction was
complete, (2.8 area % of 11 relative to both isomers of 12)
it was cooled to 20-25 °C and filtered to remove precipitated
triethylamine hydrochloride and titanium dioxide. The filter
cake was washed with 58 kg of toluene. The combined
toluene solutions were washed at 5-10 °C with 10% aqueous
sodium hydroxide solution (55 kg) and twice with 50-L
portions of water. The solution was passed through a
cartridge filter and then concentrated at atmospheric pressure
(110 °C) to approximately one-fourth volume to afford a
solution of 12 in toluene (52.6 kg, 48.3 wt %; solution yield
of 83.6%, corrected for the purity of starting 11 of 98 wt %).
(2H-Benzo[d]1,3-dioxolen-5-yl)(1R)butyl)[(4-methyl-
phenyl)(R)sulfinyl]amine 8. Diastereoselective reduction of
the chiral su1finylimine 7 (1.8 mmol, 0.6 g) in THF at -40
°C using DIBAlH (3.2 mmol) gave, after workup with citric
acid and toluene extraction, 0.52 g of sulfinamide 8 (90 area
% by HPLC, retention time 3.8 min; see above for method).
1H NMR (300 MHz, CDCl3) δ 0.91 (t, J ) 7 Hz, 3H), 1.23
(m, 2H), 1.74 (m, 1H), 1.99 (m, 1H), 2.38 (s, 3H), 4.15 (d,
1H, J ) 6 Hz, NH), 4.31 (m, 1H, NCH), 5.92 (s, 2H), 6.6
(m, 3H), 7.19 (d, J ) 7 Hz, 2H), 7.51 (d, J ) 7 Hz, 2H).
CIMS (M + H)+, 332. The chiral auxiliary was removed in
warm 1% TFA/MeOH to give the desired amine 4 in 86%
ee, (R) (chiral HPLC).
DuPhos Route. N-(2H-benzo[3,4-d]1,3-dioxolen-5-yl-
(R)butyl)acetamide 10 by asymmetric hydrogenation of
enamide 9, that is, N-(1-(2H-benzo[3,4-d]1,3-dioxolen-5-yl)-
but-1-enyl)acetamide. A Fisher-Porter tube was charged
with 9 (13.0 g, 61.5 mmol), 1,2-bis-((2R,5R)-2,5-dimethyl-
phospholano)benzene (cyclooctadiene)-rhodium (I) tetra-
fluoroborate (20.0 mg, 0.033 mmol), and methanol (75 mL)
under nitrogen. The system was flushed with hydrogen,
pressured to 30 psig H2, and stirred for 68 h at room
temperature. The solution was filtered though a pad of silica,
and the solvent was removed at reduced pressure. The
enantiomeric excess of the crude product was determined to
be 96.5% by chiral GC analysis (L-Val, 160 °C, isothermal).
The product was crystallized from ether/heptane at 3 °C to
afford pure acetamide 10 (11.26 g, 86% yield) as a white
1
All lots were >99 area % by HPLC. H NMR (300 MHz,
CDCl3) δ 0.90 (t, J ) 7 Hz, 3H), 1.45 (m, 2H), 1.50 (d, J )
7 Hz, 3H), 2.67 (t, J ) 7 Hz, 2H), 4.85 (q, J ) 7 Hz, 1H),
5.99 (s, 2H), 6.80 (d, J ) 8.7 Hz, 1H), 7.30 (m, 5H) 7.45
(s, 1H), 7.47 (d, J ) 8.7 Hz, 1H).
(2H-benzo[3,4-d]1,3-dioxolen-5-yl(1R)butyl)((1R)-1-
phenylethyl)amine 13. A slurry of 12 (34.3 kg of a 50.3
w/w % solution in toluene; 17.25 kg, 58.4 mol content) and
Raney nickel (2 kg of type 2800, in 88.5 kg of ethanol) was
hydrogenated at 150 psi hydrogen pressure at 25 °C for 4 h
to completely reduce the 12. The temperature was then
increased to 70 °C to increase the rate of isomerization of
the ketimine geometric isomer of 12, and the hydrogenation
was continued at 75 psi for another 2.5 h. After the reaction
was complete by HPLC analysis, the catalyst was filtered
and washed with 10 kg of ethanol. The ethanol wash was
combined with the reaction solution to give 13 in toluene-
ethanol (132.7 kg, 12.1 wt %, 89% de, for a calculated yield
of 16.1 kg, or 92%). A sample of 13 was isolated as the
mandelic acid salt for use as analytical reference material:
1H NMR (300 MHz, DMSO-d6) δ 0.72 (t, J ) 7.4 Hz, 3H),
1.01 (m, 1H), 1.14 (m, 1H), 1.23 (d, J ) 6.6 Hz, 3H), 1.45
(m, 1H), 1.63 (m, 1H), 3.19 (t, J ) 7.2 Hz, 1H, NCH), 3.49
(q, J ) 6.6 Hz, 1H, NCH), 4.93 (s, 1H, OCH), 6.00 (s, 2H,
O2CH2), 6.50 (dd, J ) 7.7 and 1.5 Hz, 1H), 6.80 (d, J ) 7.7
Hz, 1H), 6.91 (s, 1H), 7.2-7.5 (m, 10H). HPLC method:
column, Waters Symmetry-C18, 150 × 3.9 mm, 5 µm; 40 °C;
detection, 285 nm; flow rate: 0.8 mL/min; mobile phase:
acetonitrile and 0.05 M aqueous KH2PO4, pH adjusted to
6.0 with sodium hydroxide; linear gradient: initial composi-
tion 65:35 buffer-acetonitrile, to 30:70 at 10 min; retention
times: 11, 3.0 min; toluene, 4.0 min; Z-isomer of 12, 6.5
min; 12, 9.2 min; 13, 7.2 min; 13-diastereomer, 5.5 min.
(R)-r-Propylpiperonylamine Hydrochloride (4‚HCl).
A 50-gal, stainless steel reactor was charged sequentially with
25
1
crystalline solid, [R]D ) +134.6 (c ) 1, chloroform). H
NMR (300 MHz, CDCl3) δ 0.87 (t, J ) 7 Hz, 3H), 1.28 (m,
2H), 1.91 (m, 2H), 1.93 (s, 3H), 4.80 (m, 1H), 5.65 (br d,
1H), 5.90 (s, 2H), 6.74 (s, 3H). Upon chiral GC analysis as
above, we were unable to observe the minor enantiomer
under conditions where 0.5% was readily observable,
indicating that enantiomeric excess was >99%.
Chiral Amine: (R)-(+)-Phenylethylamine Route. 1-(2H-
Benzo[3,4-d]1,3-dioxolen-5-yl)butan-1-one 11. A 12-L
reaction flask was charged with 1,2-methylenedioxybenzene
(1.50 kg, 12.3 mol), dichloroethane (5 L), and n-butyric
anhydride (2.43 kg, 15.3 mol); the resulting mixture was
cooled to -15 °C. Boron trifluoride (1.15 kg, 17 mol) was
then charged slowly during 1-2 h, while maintaining the
temperature at or below -5 °C. After addition was complete,
the mixture was stirred at -5 to -10 °C for 3 h. The reaction
was then quenched into a sodium acetate solution (9.1 kg
sodium acetate trihydrate in 11 L of water). The bottom,
organic layer was washed with 5% NaOH (5 L) and then
water (2 × 2.5 L). The organic phase was concentrated
to an oil (∼3 L) which was crystallized from cold heptane
(3.5 L). The solid was filtered and washed with cold heptane
(2 L) to give 11 (2.03 kg, 100 wt %, 86% yield). mp
1
52-54 °C. H NMR (400 MHz, CDCl3) δ 0.99 (t, J ) 7.4
Hz, 3H), 1.75 (h, J ) 7.4 Hz, 2H), 2.86 (t, J ) 7.4 Hz, 2H),
6.03 (s, 2H), 6.84 (d, J ) 8.1 Hz, 1H), 7.43 (d, J ) 1.7 Hz,
1H), 7.56 (dd, J ) 1.7 and 8.1 Hz, 1H). 13C NMR (100 MHz,
Vol. 6, No. 1, 2002 / Organic Process Research & Development
•
61

13 solution (132.7 kg with a content of 16.1 kg), acetic acid
(10 kg) and 50% water-wet 10% palladium-on-carbon (5.3
kg). The stirred suspension was hydrogenated at 150 psi at
a temperature of 20-25 °C for 8 h. After reaction was
complete, the catalyst was filtered and washed with 10 kg
of ethanol to give a solution of 4 (152 kg). The filtrate was
combined with another batch to give 301 kg of solution with
an assayed w/w content of 6.97% (21.0 kg) 4. This solution
was concentrated in a 100-gal, glass-lined vessel to a volume
of about 40-50 L and diluted with ethanol (17 kg) and
toluene (170 kg). This solution was washed twice with 90
kg of 10% aqueous sodium hydroxide solution, followed by
2 × 55 L of water. After charging 24 kg of isopropyl alcohol,
HCl gas (5.5 kg, 150 mol) was sparged below the surface
of the stirred solution at 20 °C, and a slurry formed. After
stirring at 20 °C for 1 h, the precipitated solid was filtered
and the filter cake washed with isopropyl alcohol-heptane
(66 kg:86 kg) to give crude 4‚HCl (96% ee). The wet cake
was then slurried in 157 kg of isopropyl alcohol-heptane
(99 kg/58 kg). Filtration at 15 °C and drying to a constant
weight gave 16.4 kg of 4‚HCl (99.9% purity by HPLC area,
99.0% ee). mp 239-240 °C. HPLC method: as above but
initial composition of 90:10 buffer-acetonitrile, with gradient
to 30:70 at 10 min; flow rate: 1.0 mL/min; retention times:
13, 14 min; 4, 5.7 min. Chiral HPLC method: column,
CrownPak CR+, 4.0 mm × 15 cm, 5 µm; detection: 285
nm; flow rate: 1.3 mL/min; 50 °C; mobile phase: 84:15:1
water:methanol:70% perchloric acid (premixed); retention
times: 4, 16.4 min; (S)-enantiomer, 13.0 min.
1-(2H-Benzo[d]1,3-dioxolen-5-yl)(R)butanisocyanate 2.
In 77 L of toluene was slurried 33.6 mol (7.7 kg) of the above
HCl salt, and 136 mL (1.7 mol, 5 mol %) of pyridine added.
The solution was heated to 100-110 °C. Phosgene (4.0 kg,
40 mol) was introduced as a gas (from the vapor space of
the phosgene cylinder) below the surface of the toluene. The
slurry temperature was kept at 100-110 °C during the
transfer. The reaction mass was held at reflux (typically 106-
109 °C) for an additional 30 min and then cooled to 90 °C.
The solution was sparged at this temperature with subsurface
nitrogen to purge remaining unreacted phosgene (a packed
column, counter-current scrubber containing 5% NaOH was
used). In all batches the reaction was complete, and no
additional charge of phosgene was required. The solution
was then cooled to 0-10 °C and transferred to a stirred
solution of cold aqueous 5% sodium bicarbonate in a 100-
gal glass-lined vessel while maintaining 0-10 °C. The
toluene layer was washed with water at 5-10 °C and then
distilled at 45-50 °C reduced pressure to a water content
of <50 ppm (determined by Karl-Fischer titration). The
solution yield of 2 was 95.4%. For characterization, the
product was isolated in pure form by distillation, bp 90-92
°C/0.1 mmHg. HPLC method: column, Zorbax SB-Phenyl,
4.6 mm × 15 cm, 5 µm; flow rate, 1.0 mL/min; 40 °C; 235
nm; mobile phase: buffer solution: 0.05 M aqueous am-
monium acetate, methanol, and acetonitrile; a linear gradient
was applied over 25 min by varying the amounts of aqueous
buffer and acetonitrile, while holding the amount of methanol
constant at 20%; initial 50% buffer, 30% CH3CN to 60%
CH3CN at 25 min; stop time 25 min; post time 5 min.
Retention times: 6, 2.6 min; 4, 3.6 min; toluene, 4.8 min;
1, 5.9 min; olefinic byproduct, 8.9 min; 2, 10.1 min;
symmetrical urea, 11.0 min. HPLC chiral method: column,
Chiralpak AD, 4.6 mm × 25 cm, 10 µm; 10 °C; 245 nm;
mobile phase: hexane-methanol 98:2; flow rate, 1.0 mL/
min; stop time, 10 min; retention times: toluene, 3.5 min; 2
((S)-enantiomer), 6.2 min; 2 ((R)-enantiomer), 6.9 min.
N-(1-(2H-Benzo[d]1,3-dioxolen-5-yl)(1R)butyl)((4S)-
3,3-diethyl-4-{4-[(4-methylpiperazinyl)carbonyl]phenoxy}-
2-oxoazetidinyl)carboxamide DMP 777. Toluene (210 kg)
was charged to a reaction vessel followed by 40 kg of 1
(113.0 mol), and the stirred mixture was chilled to about 5
°C. A Karl-Fischer determination of the supernatant solution
showed a water content of 140 ppm. Azeotropic distillation
can be done if necessary at 45 °C/55-60 mmHg to reduce
the water level below the designated acceptable limit of
350 ppm. A toluene solution of 2 (153 kg of a 17.1% w/w
solution, 119 mol) was added, and the reaction slurry was
cooled to -10 °C. The catalyst (4.0 kg, 1 M lithium tert-
butoxide in hexanes, 5.7 mol, 5.0 mol % relative to 1) was
added. A modest exotherm was observed which caused the
temperature of the reaction mass to go up to -5 °C over 6
min. The jacket temperature was held constant at -15 °C.
All of the 1 dissolved within 30 min to give a light-yellow
solution. HPLC analysis showed the reaction was complete
(completion criteria: <0.1 area % 2 and 0.1-1.2% remaining
unreacted 1). The reaction was quenched with 0.35 kg of
glacial acetic acid (5.7 mol) and washed with 43 kg of 2%
aqueous sodium chloride solution while warming to 20 °C.
After separation of the aqueous layer, the organic layer was
concentrated to slightly less than one-half the original volume
by distilling at 40 °C/60 mmHg to approximately 200 L.
Analysis of the toluene solution by HPLC showed (by area)
<0.1% 2, 0.3% 1, 0.06% symmetrical urea 14, 0.3% 4, 0.1%
6, 0.4% of the methyl carbamate derived from reaction of
traces of residual methanol in the apparatus with 2, 0.2%
tert-butyl carbamate derived from reaction of tert-butyl
alcohol with 2, 0.15% of olefinic elimination byproduct of
2, and 3.3% of diastereomer derived from reaction of ent-1
(carried along as an impurity in 1) with 2. The warm (40-
45 °C) toluene solution was filtered through a 0.5 µm, in-
line cartridge filter into a 200-gal, glass-lined vessel. After
warming the product solution to 50 °C, tert-butyl methyl
ether (58 kg) was added, followed by heptanes (119 kg,
added over 1 h) while keeping the temperature at about 50
°C. The product precipitated during the heptanes addition.
The mixture was held at 45 °C for 1 h. After cooling at a
steady rate to -15 °C over 7 h, the mixture was centrifuged,
and the cake was washed with a solution of 33 kg of toluene
in 61 kg of heptanes which had been precooled to -20 °C.
The wet cake was dried to constant weight in a vacuum oven
at 40 °C/50 mmHg to yield 59.6 kg DMP 777, 93% yield.
The filtrate and washes were combined (365 kg) and assayed
by HPLC to determine the overall mass balance of the
reaction. It contained 4.5% yield of DMP 777 (2.6 kg), 2.0
kg of its diastereomer (R,R and S,S), 20 g of 6, 210 g of 4,
210 g of 14, 40 g of 1, 470 g of methyl carbamate, 250 g of
tert-butyl carbamate, and 50 g of the olefin byproduct carried
through from the 2 solution. HPLC method: column, Zorbax
SB-C18, 25 cm × 4.6 mm, 5 µm; 50 °C; detector, 235 nm;
flow: 1.1 mL/min; mobile phase: A ) 30% methanol/water
62
•
Vol. 6, No. 1, 2002 / Organic Process Research & Development

(0.02 M trifluoroacetic acid and triethylamine, pH 7.4), B
) acetonitrile; linear gradient, 30% B to 70% B at 15 min;
then, flow rate 1.5 mL/min and linear gradient to 95% B at
19 min; stop time, 20 min; retention times: 6, 2.6 min; 4,
4.4 min; 1, 5.2 min; methyl carbamate, 6.7 min; toluene,
8.1 min; symmetrical urea, 10.6 min; tert-butyl carbamate,
11.0 min; olefin, 11.6 min; 2, 11.7 min; diastereomer of
DMP 777, 14.5 min; DMP 777, 14.9 min. For published
chiral HPLC methods see ref 14.
Acknowledgment
We thank Henry Rapoport for helpful discussions.
Received for review October 9, 2001.
OP015507O
(14) Zagrobelny, J.; Matuszewski, B. K.; Kline, W. F.; Vincent, S. H. J. Pharm.
Biomed. Anal. 1998, 17, 1057-1064. Williams, R. C.; Riley, C. H.;
Sigvardson, K. W.; Fortunak, J. M.; Ma, P.; Nicolas, E. C.; Unger, S. E.;
Krahn, D. F.; Bremner, S. L. J. Pharm. Biomed. Anal. 1998, 17, 917-
924.
Vol. 6, No. 1, 2002 / Organic Process Research & Development
•
63