Unique and efficient synthesis of [2S-(2R*,3S*,4R*)]-2-amino-1-cyclohexyl-6-methyl-3,4- heptanediol, a popular C-terminal component of many renin inhibitors

Full text of DOI:10.1021/jo960694n

9564
J . Org. Chem. 1996, 61, 9564-9568
Sch em e 1a
Un iqu e a n d Efficien t Syn th esis of
[2S-(2R*,3S*,4R*)]-2-Am in o-1-cycloh exyl-
6-m eth yl-3,4-h ep ta n ed iol, a P op u la r
C-Ter m in a l Com p on en t of Ma n y Ren in
In h ibitor s
Mark A. Schwindt,* Daniel T. Belmont, Mark Carlson,
Lloyd C. Franklin, Valerie S. Hendrickson,
Gregory L. Karrick, Russell W. Poe,
Denis M. Sobieray, and J im Van De Vusse
Chemical Development Department, Parke-Davis
Pharmaceutical Research Division, 188 Howard Avenue,
Holland, Michigan 49424
Received April 16, 1996
In tr od u ction
Many compounds containing 1-amino-2,3-diol func-
tionalities have antihypertensive activity and act as renin
inhibitors.1 Aminodiols have also been reported to have
other biological properties such as a gastro protective
substance2 and inhibitors of the HIV protease and
antiviral activity.3 On a broader scope, a number of
amino sugars and polyhydroxylated alkaloids have been
reported to be useful in the treatment of carbohydrate
dependant metabolic disorders due to their selective
inhibition of glycosidases.4 New synthetic routes to
amino alcohols would be useful in the synthesis of
compounds in all these areas.
a
(a) H2, 2% Rh-10% Pd/C; (b) ClCO2Et, MeHNOMe, HCl; (c)
Our interest in 1-amino-2,3-diols was in the economical
synthesis of [2S-(2R*,3S*,4R*)]-2-amino-1-cyclohexyl-6-
methyl-3,4-heptanediol (9), PD 134373sa segment of CI-
992sa compound Parke-Davis intended for the treatment
of hypertension and congestive heart failure.5 The
original preparation of 9 was a multistep low-yielding
synthesis starting from L-phenylalanine utilizing an OsO4
oxidation of a cis-alkene.6 Many research groups have
sought to overcome the problems associated with the
osmylation oxidation procedure.7
LAH; (d) acetone cyanohydrin, catalytic KCN, TBAI; (e) HCl; (f)
NaOH, ClCO2Et; (g) iBuLi, THF/LiBr; (h) NaB(O2CCH3)3H; (i)
KOH.
Our unique and improved synthesis involved the linear
addition of two inexpensive low molecular weight frag-
ments in a stereochemically controlled manner illustrated
in Scheme 1. Relative chirality was controlled by the
isolation of cyanohydrin 4 and subsequent reduction of
7 to 8. The key synthetic strategy surrounds the lithium
carboxylate of 6 coupling with isobutyllithium. This
route converted BOC-phenylalanine (BOC-Phe) to the
aminodiol 9 in eight steps on a multikilogram scale
(overall yield of 37%). The elegance of the synthetic route
was the combination of steps with the right chemistry
that was reliably prepared on a large scale. The overall
yield and stereoselectivity of this new synthesis were
better than those of previously described routes.
(1) (a) Rosenberg, S. H.; et al. J . Med. Chem. 1993, 36, 449. (b) Ocain,
T. D.; Abou-Gharbia, M. Drugs Future 1991, 16, 37-51. (c) Greenlee,
W. J .; Siegl, P. K. S. Annu. Rep. Med. Chem. 1991, 26, 63-72. (d)
Greenlee, W. J . Med. Res. Rev. 1990, 10, 173-236.
(2) Kawai, A.; Hara, O.; Hamada, Y.; Shioiri, T. Tetrahedron Lett.
1988, 29, 6331.
(3) Thaisrivongs, S.; Turner, S. R.; Strohback, J . W.; Ten Brink, R.
E.; Tarpley, W. G.; McQuade, T. J .; Heinrikson, R. L.; Tomasselli, A.
G.; Hui, J . O.; Howe, W. I. J . Med. Chem. 1993, 36, 941.
(4) Poitout, L.; Merrer, Y. L.; Depezay, J .-C. Tetrahedron Lett. 1994,
35, 3293-3296. Fairbanks, A. J .; Carpenter, N. C.; Fleet, G. W. J .; et
al. Tetrahedron 1992, 48, 3365-3376. Kajimoto, T.; Liu, K. K.-C.;
Wong, C.-H.; et al. J . Am. Chem. Soc. 1991, 113, 6187-6196. Anzeveno,
P. B.; Creemer, L. J .; Liu, P. S.; et al. J . Org. Chem. 1989, 54, 2539-
2542 and references therein.
(5) Patt, W. C.; Hamilton, H. W.; Taylor, M. D.; Ryan, M. J .; Taylor,
D. G., J r.; Connolly, C. J . C.; Doherty, A. M.; Klutchko, S. C.; Sircar,
I.; Steinbaugh, B. A.; Batley, B. L.; Painchaud, C. A.; Rapundalo, S.
T.; Michniewicz, B. M.; Olson, S. C. J . Med. Chem. 1992, 35, 2562-
2572.
(6) Luly, J . R.; BaMaung, N.; Soderquist, J .; Fung, A. K. L.; Stein,
H.; Kleinert, H. D.; Marcotte, P. A.; Egan, D. A.; Bopp, B.; Merits, I.;
Bolis, G.; Greer, J .; Perun, T. J .; Plattner, J . J . J . Med. Chem. 1988,
31, 2264-2276.
(7) (a) Luly, J . R.; Hsiao, C.-N.; BaMaung, N.; Plattner, J . J . J . Org.
Chem. 1988, 53, 6109-6112. (b) Wood, J . L.; J ones, D. R.; Hirschmann,
R.; Smith, A. B., III. Tetrahedron Lett. 1990, 31, 6329-6330. (c) Baker,
W. R.; Condon, S. L. Tetrahedron Lett. 1992, 33, 1581-1584. (d) Chan,
M. F.; Hsiao, C.-N. Tetrahedron Lett. 1992, 33, 3567. (e) Krysan, D.
J .; Haight, A. R.; Menzin, J . A.; Welch, N. Tetrahedron 1994, 50, 6163.
(f) Spero, D. M.; Kapadia, S.; Farina, V. Tetrahedron Lett. 1995, 36,
4543-4546. (g) Krysan, D. J .; Rockway, T. W.; Haight, A. R. Tetrahe-
dron Asymmetry 1994, 5, 625-632.
Resu lts a n d Discu ssion
The key aspects of this new synthesis are (1) direct
isolation of a single diastereomer of cyanohydrin 4, (2)
unique conversion of the acid 5 to the â-amino R-hydroxy
ketone 7, and (3) improved selective reduction of the
ketone 7 to the desired aminodiol diastereomer 8. The
majority of the discussion entails details for these key
steps.
The synthetic process begins with the reduction of
BOC-Phe to the cyclohexyl analogue 1 by standard
methods.8 In the second step the preferred method of
reduction from the amino acid 1 to aldehyde 3 was via
the methoxymethylamide 2. Aldehyde 3 underwent
(8) For a specific example, see: Schuda, P. F.; Greenlee, W. J .;
Chakravarty, P. K.; Eskola, P. J . Org. Chem. 1988, 53, 873-875. We
used a 2% Rh-10% Pd/C catalyst.
S0022-3263(96)00694-9 CCC: $12.00 © 1996 American Chemical Society

Notes
J . Org. Chem., Vol. 61, No. 26, 1996 9565
reaction with cyanide resulting in a mixture of diaster-
eomeric isomers including cyanohydrin 4.
ods to overcome these problems.12 Still, examples of good
yielding ketone formation reactions involving an unpro-
tected heteroatom or “multianionic intermediate” species
are limited and plagued with problems. Insolubility and
aggregation of lithium species have been thought to
contribute to the low yields in these reactions. The
complexities of working with multianionic metal species
in solution have been the subject of a recent paper.13 We
were able to overcome all of these problems. The best
procedure for preparing the ketone 7 involved the fol-
lowing parameters: toluene/heptane solvents, controlled
low temperature to minimize impurities, and the pres-
ence of lithium bromide (LiBr)/tetrahydrofuran (THF) to
modify lithium aggregates. The addition of LiBr/THF
was a key factor that allowed us to prepare the â-amino
R-hydroxy ketone 7 from 6 in good yields.
The mixture of cyanohydrins was purified and isolated
by crystallization to the desired diastereomer in good
yield, an advantage over published alternatives.9 These
other methods do not isolate a single diastereomer but
carry the mixture on through the acid hydrolysis to
obtain crystalline 5 in about 60% yields.9 Both classical
bisulfite and acetone cyanohydrin methods work equally
well on a laboratory scale to prepare crude cyanohydrin
4. The new methodology in the preparation of 4 is the
isolation of the desired pure crystalline diastereomer out
of a hydrocarbon and water solvent mixture. Crystal-
lization from a two-phase mixture of hexane or heptane
and water gave 4 in 65-90% isolated yields. This new
procedure reliably allowed up to a 90% isolated yield of
solid 4 on a laboratory scale.
An experimental design study was initially used to
determine the most important processing parameters of
the coupling reaction to 7. Crude HPLC yields were used
as the measure of response (data analyzed by Design-
Ease software). The experimental design study gave us
a broad overview of the reaction parameters. The design
results quickly gave us an idea of which parameters were
interactive and to what the consequences of these inter-
actions would lead. The design indicated interactions
between (a) addition time and concentration, (b) addition
time and addition temperature, and (c) addition temper-
ature and reaction temperature. The three important
findings from the statistical study are summarized as
follows: (1) There was a high probability that product
purity increased with more equivalents of isobutyl-
lithium, the higher concentration of starting material,
and increased reaction temperature, (2) the addition
temperature seemed to have the lowest probability of
having any effect on the product purity, and (3) the
reaction temperature parameter appeared to favor the
higher temperatures. Because of the interactive param-
eters, prior to the experimental design, it appeared that
a change in conditions resulted in wildly random yields.
We could now focus on the important parameter condi-
tions and avoid those that resulted in drastically lower
yields. The biggest problem was getting the reaction to
go to completion. In all conditions investigated, the
starting material 6 remained as the main impurity. The
major contributor to this problem was thought to be
solvation of various intermediates.
Various solvent systems were explored after it became
clear that the reaction conditions studied would not
produce a complete reaction. An apparent multianionic
lithium salt intermediate precipitated during the iBuLi
addition with the THF/heptane system. Very thick
slurries were produced in the 0.3-0.5 M reaction mix-
tures at <-10 °C using toluene. No precipitation was
observed in reactions using methyl-tert-butyl ether (MTBE)
under similar conditions, an improvement over toluene.
Toluene gave the better yields, but as with all solvent
systems attempted, the starting material was still the
major impurity in the product.
The work with different solvents suggested product/
impurity ratios were influenced by solvent, but a satis-
factory ratio was not achieved. It was noted that the
reaction worked best in an ethereal solvent (diethyl ether,
THF, or MTBE). The literature suggests that lithium
aggregates may be possible and could be modified or
broken up with additives such as lithium chloride (LiCl)
The desired cyanohydrin diastereomer 4 seemed to be
produced with very high selectivity, but significant
epimerization was found to occur over time in solution.
A solid isolation of this compound was desirable due to
the nature of the cyanohydrin to epimerize in solution.
On a laboratory scale the isolation of 4 works very well.
Isolation of 4 by crystallization was aimed to minimize
epimerization on a large scale reaction.10
An epimerization study using a HPLC solution con-
taining pure cyanohydrin 4 illustrated that the single
cyanohydrin diastereomer was stable in solution until a
catalytic amount of potassium cyanide (KCN) was added.
Upon addition of KCN, the cyanohydrin epimerized over
time to a 1:1 mixture of diastereomers. Epimerization
was also verified in an NMR study comparing 13C NMR
spectra, before and after KCN addition. Without the
isolation of the solid 4, a reliable assay of the crude
diastereomeric mixture content was difficult to obtain.
In our case the conversion of cyanohydrin 4 to â-amino
R-hydroxy acid 5 also used acid conditions.9 Use of the
isolated solid 4 allowed conversion in up to 98% yield to
the acid 5. The workup procedures of 5 were much easier
using the solid 4 than with the diastereomeric oily crude
mixtures of 4.
The most complex and intriguing step was the lithium
carboxylate of 6 coupling with isobutyllithium to prepare
ketone 7. The addition of organometallic reagents to
carboxylic acid derivatives is known,11 yet several prob-
lems have limited the use of these reactions. The major
concerns are that organometallic reagents are basic as
well as nucleophilic, resulting in deprotonation rather
than addition and the double addition of the organome-
tallic reagent to the ketone to produce a tertiary alcohol.
Several research groups have recently introduced meth-
(9) (a) Iizuka, K.; Kamijo, T.; Harada, H.; Akahane, K.; Kubota, T.;
Umeyama, H.; Kiso, Y. J . Chem. Soc., Chem. Commun. 1989, 1678.
Harada, H.; Tsubaki, A.; Kamijo, T.; Iizuka, K.; Kiso, Y. Chem. Pharm.
Bull. 1989, 37, 2570. (b) See ref 7a.
(10) The longer reaction and workup times involved in a large scale
reaction were a concern for epimerization. On a 20 kg scale, a 42%
yield was obtained for the combined reaction steps preparing 3 and 4.
We believe the lower large scale yield was also due to larger than
expected amounts of the corresponding alcohol present arising from
overreduced aldehyde 3. Both of these factors contributed to the lower
isolation yield in the crystallization process of 4. Pursuit in optimizing
the scale-up process of this reaction was dropped due to finding a
commercial kilogram quantity source for 5 (Nippon Kayaku Co., Ltd.).
(11) For reviews, see: Shirley, D. A. Org. React. 1954, 8, 28.
J orgenson, M. J . Org. React. 1970, 18, 1.
(12) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.
Hlasta, D. J .; Court, J . J . Tetrahedron Lett. 1989, 30, 1773.; Roemmele,
R. C.; Rapoport, H. J . Org. Chem. 1989, 54, 1866. Whipple, W. L.;
Reich, H. J . J . Org. Chem. 1991, 56, 2911.
(13) Seeback, D. Angew. Chem., Int. Ed. Engl. 1988, 27, 1624.

9566 J . Org. Chem., Vol. 61, No. 26, 1996
Notes
or tetramethylethylenediamine (TMEDA).13 While in-
complete reactions were still observed with the incorpo-
ration of the additives LiCl and TMEDA, a dramatic
improvement in yield and purity was observed with the
addition of LiBr in THF. A strong exotherm was ob-
served while dissolving the LiBr in THF, indicating an
immediate complexing between the salt and the solvent.14
The success of using LiBr/THF as an additive demon-
strated that the incomplete reactions were likely the
result of aggregates.
1/100 using sodium triacetoxyborohydride in both THF
and a mixture of heptane and acetonitrile. Sodium
triacetoxyborohydride was a superior selective reducing
reagent over sodium borohydride or sodium cyanoboro-
hydride in THF resulting in the desired aminodiol 8.
Deprotection of the amine functionality of the amino-
diol 8 to the free amine 9 went as expected. The
ethylcarbamate protecting group was hydrolyzed to the
amine using potassium hydroxide in excellent yield.
Our previous work had shown that reaction byproduct
impurities were reduced in a toluene solvent. A mixed
toluene/THF solvent system was examined after it was
found that LiBr could be solubilized by adding only a
small amount of THF. The isolated yields, when cor-
rected for purity, resulted in conversions of 80-85% of
the starting material 6 to 7. The modification of lithium
aggregates by use of the LiBr/THF combination was
critical for the reaction to go to completion. Impurities
were minimized by using optimum parameters: addition
temperatures (-10 °C), reaction temperatures (10 °C),
and equivalents of iBuLi (4.15). High product conver-
sions were achieved with a 2-3 h reaction time under
these conditions. The toluene/LiBr/THF system was
superior to both the THF/LiBr and the toluene systems.
This provided definite proof that addition of small
amounts of THF/LiBr was critical and favorably affected
the solvation influencing the reaction.
The order of the quench addition influenced the forma-
tion of a tertiary alcohol impurity resulting from double
addition.11,12 The tertiary alcohol impurity was mini-
mized by adding the reaction solution to water while
rapidly stirring the solution at <10 °C. There was also
some indication that the tertiary alcohol impurity was
more easily formed in the toluene systems than in ether
solvents. This may be the result of reduced solubility in
the two-phase quench.
An unknown 1% impurity was observed and traced to
9 made from the method using toluene. It was removed
by a recrystallization of the 9 material, isolated from the
recrystallization filtrate by chromatography, and identi-
fied as a toluene derivative of 9. This impurity was
created by conversion of a small amount of toluene to
phenylmethyllithium which competed with isobutyl-
lithium for the carboxylate in the coupling reaction to
aminodiol 8.15 This could be avoided in the future by use
of MTBE or possibly ethylbenzene.
Reduction of the â-amino R-hydroxy ketone 7 to the
desired aminodiol 8 was known to occur with good
stereoselectivity.7a Reduction studies of 7 with different
solvents and reagents resulted in significant improve-
ments in the selectivity to the aminodiol 8. The resulting
diastereomeric ratio was found to be dependent upon the
reducing reagent and solvent used. A mixture of the two
possible reduction diastereomers of 8 was determined
from a doubling of peaks observed in both 13C NMR and
HPLC data. The ratio of undesired/desired diastereomer
was found to be 60/40 using sodium borohydride in
methanol, 18/82 using sodium borohydride in THF, and
Su m m a r y
A new, simple, large scale route to [2S-(2R*,3S*,4R*)]-
2-amino-1-cyclohexyl-6-methyl-3,4-heptanediol (9) was
established starting from BOC-Phe. Diastereoselectivity
was improved by selective crystallization of pure cyano-
hydrin diastereomer 4 and using a sodium triacetoxy-
borohydride reagent in the reduction of â-amino R-hy-
droxy ketone 7. Aggregate and solvation problems
dealing with multianionic intermediates in the carboxy-
late coupling to â-amino R-hydroxy ketone 7 were over-
come by introducing a toluene/LiBr/THF solvent system.
The synthetic route consists of eight steps starting from
BOC-Phe and results in a 37% overall yield. All but two
steps were run on a multikilogram scale. These two steps
preparing 4 and 5 worked excellently on a large (>150
g) laboratory scale.
We have demonstrated a traditional synthetic ap-
proach to [2S-(2R*,3S*,4R*)]-2-amino-1-cyclohexyl-6-
methyl-3,4-heptanediol (9) by controlling stereochemistry
of each additional chiral center starting from a chiral
amino acid, BOC-Phe. This overall efficient and practical
synthetic strategy should be applicable to a wide variety
of 1-amino-2,3-diols utilizing commercially available
chiral amino acids as starting materials and varying the
lithium alkyl used in the coupling reaction.
Exp er im en ta l Section
(S )-r-[[(1,1-D im e t h y le t h o x y )c a r b o n y l]a m in o ]c y c lo -
h exa n ep r op a n oic Acid (1). N-(tert-Butyloxycarbonyl)-L-phen-
ylalanine (29.9 kg, 112.7 mol) and 1.94 kg of 2% Rh-10% Pd on
carbon catalyst was stirred with 300 L of isopropyl alcohol. The
reaction mixture was pressurized to 50 psig with hydrogen at
about 50 °C until hydrogen consumption ceased. The resulting
slurry was cooled to about 15 °C and filtered. The wet catalyst
was rinsed with 29 L of isopropyl alcohol. The combined filtrates
were concentrated to an oil by vacuum distillation (5-10 mmHg)
at 50-55 °C. All residual IPA was removed by addition of 97 L
of toluene followed by concentration to an oil by vacuum
distillation (5-10 mmHg) at 50-55 °C. Dichloromethane (106.5
kg) was added and stirred, and the resulting solution was used
immediately in the next step. A sample of the solution was
taken; the solvent was removed under vacuum and analyzed to
result in a white plastic resinous oil with a calculated 107% yield,
32.8 kg, of 1 and the following characteristics.
1H NMR (200 MHz, DMSO): δ 12.29 (br s, 1H), 7.00 (d, 1H,
J ) 8.0 Hz), 3.91 (m, 1H), 1.70-0.71 (m, 13H), 1.36 (s, 9H). [R]D
) -9.1° (c ) 1.0, MeOH). HPLC assay: 96.4% by area %,
corrected for toluene; Phenomenex (5 µm, 4.6 × 250 mm) C-18
column at 40 °C; mobile phase, 50:50 acetonitrile:0.5 M aqueous
triethylamine, pH adjusted to 3.0 with H3PO4; flow rate ) 1.0
mL/min; UV detector at 214 nm. Anal. Calcd for C14H25NO4:
C, 61.97; H, 9.29; N, 5.16. Found: C, 60.89; H, 9.81; N, 4.97.
(S)-[1-(1-Cycloh exylm et h yl)-2-(m et h oxym et h yla m in o)-
2-oxoeth yl]ca r ba m ic Acid , 1,1-Dim eth yleth yl Ester (2).
After the dichloromethane solution of 1 (estimated as 100.6 kg/
30.6 kg, 112.7 mol) was cooled to about -40 °C, 13.9 kg (140.1
mol) of N-methylpiperidine was added slowly, maintaining the
batch temperature below -30 °C. Ethyl chloroformate (14.9 kg,
137.3 mol) was added slowly. The reaction mixture was stirred
at about -40 °C for about 2 h. A solution of 13.2 kg (135.3 mol)
(14) This did not occur with LiCl. For related interest, see: Hendrix,
J . C.; Halverson, K. J .; J arrett, J . T.; Lanbury, P. T., J r. J . Org. Chem.
1990, 55, 4517-4518. Claudy, P.; Bonnetot, B.; Mathurin, D.; Turck,
G. Thermochim. Acta 1977, 20, 315-323.
(15) For related chemistry of carboxylate reactions with toluene and
n-butyllithium, see: Lochmann, L.; Trekoval, J . J . Organomet. Chem.
1987, 326, 1-7. Schlosser, M.; Hartmann, J . Angew. Chem., Int. Ed.
Engl. 1973, 12, 508.

Notes
J . Org. Chem., Vol. 61, No. 26, 1996 9567
of N,O-dimethylhydroxylamine hydrochloride and 13.9 kg (140.1
mol) of N-methylpiperidine in 91 kg of dichloromethane was
added slowly to the reaction mixture, maintaining the batch
temperature at about -30 °C. The reaction mixture was slowly
warmed to about 20 °C and stirred for at least 8 h. Toluene (50
kg) and 35 L of water were added to the reaction mixture, and
the layers were separated. The organic layer was washed once
with 35 L of water, three times with 1% hydrochloric acid (16
kg), and once with sodium bicarbonate solution (1.9 kg in 30 L
of water). The organic layer was concentrated by atmospheric
distillation to an oil. Toluene (78 kg) was added, and the solution
was concentrated by vacuum distillation to obtain 32.2 kg (90.8%
yield) of 2 as a solution in about 55 L of toluene. A sample was
taken, and 2 was isolated as a pale brown oil by removing excess
solvents under vacuum for analysis.
°C; mobile phase, 105:90 methanol:0.5 M aqueous triethylamine,
pH adjusted to 2.5 with H3PO4; flow rate ) 1.0 mL/min; RI
detector, UV detector at 210 nm. Anal. Calcd for C10H19NO3‚
HCl: C, 50.52; H, 8.48; N, 5.89. Found: C, 50.61; H, 8.67; N,
5.78.
[R-(R*,S*)]-â-[(Eth oxyca r bon yl)a m in o]-r-h yd r oxycyclo-
h exa n ebu ta n oic Acid (6). The hydrochloride salt 5 (10 kg,
42.1 mol) was dissolved in 33.6 L of water and cooled to 5 °C.
The pH was adjusted to about 9.5 using 1 N sodium hydroxide
(roughly 85 L), maintaining the temperature at about 5 °C. Ethyl
chloroformate (5.0 kg, 46.1 mol) was added. The pH of the
reaction mixture was maintained at about 9.5 by the addition
of more 1 N sodium hydroxide. The reaction mixture was stirred
at 5 °C for 2-3 h. The solution was washed once with 90 L of
toluene to remove impurities, and hydrochloric acid was added
to the water layer, adjusting the pH to about 2.0. The reaction
mixture was extracted by washing the reaction mixture two
times with 90 L of toluene. The toluene layers were combined
and partially concentrated to obtain 10.5 kg of 6 as a colorless
to yellow solution (91.3% yield). The solution was carried on
into the next step. A sample was isolated by removing excess
solvents under vacuum and gave the following physical char-
acteristics.
1H NMR (200 MHz, DMSO): δ 6.87 (d, 1H, J ) 8.4 Hz), 4.44
(m, 1H), 3.70 (s, 3H), 3.06 (s, 3H), 1.77-0.83 (m, 13H), 1.35 (s,
9H). [R]25 ) -10.8° (c ) 1.0, MeOH). GC assay: 91.6%; DB-5
D
(30 m) column; flow rate ) 1.0 mL/min; injection ) 280 °C, FID
detection ) 300 °C; program set at 50 °C for 2 min, 15 °C/min
ramp to 270 °C for 9 min. Anal. Calcd for C16H30N2O4: C, 61.12;
H, 9.62; N, 8.91. Found: C, 61.09; H, 9.89; N, 8.65.
(S)-(2-Cycloh exyl-1-for m yleth yl)ca r ba m ic Acid , 1,1-Di-
m eth yleth yl Ester (3). A solution of 2 (33.7 kg, 107.2 mol, in
30.7 kg of toluene) was concentrated and charged to a solution
of 50.8 kg of sodium bis(2-methoxyethoxy)aluminum hydride
(Vitride) in 95 L of toluene at about -15 °C. After the mixture
stirred at about -15 °C for about 1.5 h, the reaction was
quenched with an aqueous sodium chloride solution (51.3 kg in
286 L of water). The layers were separated, and the organic
layer was washed three times with a 5% hydrochloric acid
solution (136, 68, and 68 kg), two times with dilute sodium
hydroxide solution (1.2 kg of 50% NaOH in 15 L of water), and
once with brine (19 kg of NaCl in 54 L of water) and dried over
31 kg of magnesium sulfate. The solution was filtered and
partially concentrated under vacuum and the resulting unstable
optically active 3 immediately used as a solution. A sample of
the aldehyde 3 was isolated as a clear oil by removing excess
solvents under vacuum and analyzed immediately.16
1H NMR (200 MHz, CDCl3): δ 5.51 (d, 1H, J ) 9.7 Hz), 4.20-
4.00 (m, 2H), 4.15 (q, 2H, J ) 6.7 Hz), 1.16 (t, 3H, J ) 6.7 Hz).
13C NMR (50.3 MHz, CDCl3): δ 175.1, 157.1, 72.1, 61.0, 51.1,
39.8, 34.3, 33.7, 32.9, 26.6, 26.5, 14.4. HPLC assay: 97.72% by
area %; Nucleosil (5 µm, 4.6 × 250 mm) C-18 column, int ) 35
°C, ext ) 40 °C; mobile phase, 50:50 acetonitrile:0.01 M aqueous
tetrabutylammonium hydrogen sulfate, pH adjusted to 3.0 with
H3PO4; flow rate ) 0.9 mL/min; RI detector. Anal. Calcd for
C
13H23NO5: C, 57.13; H, 8.48; N, 5.12. Found: C, 57.10; H, 8.68;
N, 4.94.
(2S,3R)-2-[(E t h oxyca r b on yl)a m in o]-1-cycloh exyl-3-h y-
d r oxy-6-m eth ylh ep ta n -4-on e (7). Lithium bromide (8.5 kg,
99 mol), a 27% toluene solution of 6 (10.5 kg, 38.4 mol), and
13.5 kg of toluene were combined and stirred under a nitrogen
atmosphere. After the batch was cooled to 10 °C, 12.6 kg of THF
was added over about 20 min while controlling the temperature
at <25 °C. The batch was stirred for 20 min and then cooled to
-15 °C. An isobutyllithium solution (22.2% in heptane, 46 kg,
158 mol) was slowly added over 6 h while maintaining the
temperature between -15 and -3 °C. The reaction was
complete after stirring for 2 h at 5-10 °C. A quench solution of
30 kg of ammonium chloride in 90 L of water was prepared and
cooled to 5 °C. The toluene reaction mixture was slowly added
to the water solution using vigorous agitation. The temperature
of the two-phase system was held at 10 °C during the addition,
and the mixture was warmed to 25 °C and stirred for 1 h. After
the layers were separated, the organic layer was concentrated
under vacuum. Acetonitrile (75 kg) was added and the solution
concentrated under vacuum to produce a solution weighing 25.6
kg, containing 10.5 kg of 7 (86.8% yield).
1H NMR (200 MHz, CDCl3): δ 4.87 (d, 1H, J ) 10.1 Hz), 4.31
(m, 1H), 4.03 (s, 1H), 4.01 (q, 2H, J ) 7.1 Hz), 2.53 (m, 2H),
2.17 (m, 1H), 1.90-0.80 (m, 11H), 1.19 (t, 3H, J ) 6.9 Hz), 0.93
(d, 6H, J ) 6.7 Hz). 13C NMR (50.3 MHz, CDCl3): δ 210.0, 156.1,
78.5, 60.9, 50.0, 46.6, 40.8, 34.2, 33.5, 33.0, 26.5, 26.2, 26.1, 24.5,
22.5, 22.4, 14.5. [R]D ) -80.2 (c ) 0.015 g/mL). HPLC assay:
92.7% by wt/wt; Ultrasphere (5 µm, 4.6 × 250 mm) C-18 column,
int ) 40 °C, ext ) 35 °C; mobile phase, 60:40 acetonitrile:0.1 M
aqueous tetrabutylammonium hydrogen sulfate, pH adjusted to
3.0 with H3PO4; flow rate ) 1.5 mL/min; RI detector. Anal.
Calcd for C17H31NO4: C, 65.14; H, 9.97; N, 4.47. Found: C,
63.70; H, 9.67; N, 7.37.
1H NMR (200 MHz, DMSO): δ 9.41 (s, 1H), 7.20 (m, 1H), 3.87
(m, 1H), 1.75-0.81 (m, 13H), 1.37 (s, 9H). [R]25 ) -28.2° (c )
D
1.0, MeOH). GC assay: 79%; DB-5 (30 m) column; flow rate )
2.0 mL/min; injection ) 280 °C, FID detection ) 300 °C; program
set at 50 °C for 2 min, 15 °C/min ramp to 270 °C for 9 min.
[R-(R*,S*)]-1-(Cycloh exylm eth yl)-2-cya n o-2-h yd r oxyca r -
ba m ic Acid , 1,1-Dim eth yleth yl Ester (4). Potassium cyanide
(1.55 g) and tetrabutylammonium iodide (1.83 g) were added to
a solution of water (300 mL), acetone cyanohydrin (94.75 g, 1113
mmol), and 3 (188.7 g, 739 mmol, in about 915 mL of heptane,
assuming 100% yield of 3). The reaction mixture was stirred
overnight at 25 °C. The two-phase system was washed five
times with 300 mL of water. The resulting slurry was filtered,
rinsed with heptane, and dried to obtain 190.9 g of 4 as a white
crystalline solid (91.5% yield, over two steps).
1H NMR (200 MHz, CDCl3): δ 4.84 (d, 1H), 4.60 (m, 2H), 3.82
(m, 1H), 1.80-0.90 (m, 13H), 1.45 (s, 9H). 13C NMR (50.3 MHz,
CDCl3): δ 156.4, 118.5, 81.1, 65.1, 51.9, 34.1, 33.9, 32.3, 28.3,
27.9, 26.1, 25.9. [R]D ) -47.16° (c ) 1.0, methanol). HPLC
assay: 97.4% by area %. Supelco DB (5 µm, 4.6 × 250 mm) C-18
column, int ) 40 °C, ext ) 35 °C; mobile phase, 45:55 acetoni-
trile:water; flow rate ) 1.5 mL/min; RI detector. Anal. Calcd
for C15H26N2O3: C, 63.80; H, 9.28; N, 9.92. Found: C, 64.10;
H, 9.55; N, 9.64.
[R-(R*,S*)]-â-Am in o-r-h ydr oxycycloh exan ebu tan oic Acid,
Mon oh yd r och lor id e (5). The crystalline solid 4 (177.1 g, 627.1
mmol) and 1.25 L of concentrated aqueous hydrochloric acid were
heated at 80 °C for 3 h and cooled to 0 °C. The resulting slurry
was filtered and dried to obtain 132.8 g of 5 as a white crystalline
solid (89.1% yield).
[2S-(2R*,3S*,4R*)]-2-Am in o-1-cycloh exyl-6-m et h yl-3,4-
h ep ta n ed iol (9). To sodium triacetoxyborohydride (6.8 kg, 32.1
mol) in heptane (16 L) was added 7 (10.2 kg, 32.5 mol) dissolved
in acetonitrile (14.7 kg) from the above procedure at 20-28 °C.
The reaction mixture was stirred for 4 h at 20-24 °C and cooled
to 7 °C, the reaction was quenched with 9% aqueous sodium
bicarbonate (60 kg) at 7-15 °C, and the mixture was filtered.
The solid intermediate was washed with heptane (25 L) and
water (20 L) and vacuum dried to obtain 7.1 kg of 8 (69.2% yield).
1H NMR (200 MHz, CDCl3): δ 4.20 (d, 1H, J ) 3.4 Hz), 3.49
(m, 1H), 1.50-0.69 (m, 13H). 13C NMR (50.3 MHz, CDCl3): δ
177.5, 72.6, 54.2, 39.6, 36.2, 35.8, 35.4, 29.1, 28.9, 28.7. IR
(cm-1): 1728, 1072. [R]25 ) -13.4° (c ) 1.08, water). Mp:
D
210.4-213.1 °C. HPLC assay: 100% by area % (RI), 98.1% by
area % (UV); Nucleosil (5 µm, 4.6 × 250 mm) C-18 column at 24
(16) The aldehyde 3 could be stored for several days packed in dry
ice without a change in optical purity measured by optical rotation.
(17) All of the data collected on 9 were identical with established
data reported in the literature.7

9568 J . Org. Chem., Vol. 61, No. 26, 1996
Notes
HPLC assay: 98.1% by area %, Supelco DB (5 µm, 4.6 × 250
mm) C-18 column, int ) 40 °C, ext ) 35 °C; mobile phase, 45:55
acetonitrile:water; flow rate ) 1.5 mL/min; RI detector. Anal.
Calcd for C17H33NO4: C, 64.73; H, 10.54; N, 4.44. Found: C,
64.95; H, 10.75; N, 4.43.
Hz). 13C NMR (50.3 MHz, CDCl3): δ 74.5, 73.3, 48.4, 43.7, 43.3,
34.1, 33.8, 32.9, 26.5, 26.3, 26.1, 24.8, 23.5, 22.0. [R]25D ) -34.5°
(c ) 1.0, methanol). HPLC assay: 99.49% by wt/wt; Supelco DB
(5 µm, 4.6 × 250 mm) C-18 column, int ) 40 °C, ext ) 35 °C;
mobile phase, 35:65 acetonitrile:0.01 M aqueous 1-octanesulfonic
acid, sodium salt, 2 mL of triethylamine/L of water, pH adjusted
to 3.0 with H3PO4; flow rate ) 1.5 mL/min; RI detector. Chiral
HPLC purity assay: 100%, no enantiomer detected; Beckman
ODS (5 µm, 4.6 × 250 mm) C-18 column; mobile phase, 50:50
acetonitrile:water; flow rate ) 1.5 mL/min; UV detector at 214
nm.
To 8 (7.1 kg, 22.5 mol) in 29% aqueous methanol (52 kg) was
added 45% aqueous potassium hydroxide (11.2 kg), and the
reaction solution was heated at reflux for 16 h. To the reaction
solution was added heptane (17 L) at 15 °C, and the mixture
was reheated to reflux. The solution was cooled to 12-20 °C
and the resultant solid collected. The solid product was washed
with 17% aqueous methanol (14 kg) and water (20 L) and
vacuum dried to obtain 5.1 kg of the off-white crystalline solid
9 (93.1% yield).17
Ack n ow led gm en t. We would like to thank J . Deer-
ing, N. Maddox, and W. R. Marsh for analytical support.
1H NMR (200 MHz, CDCl3): δ 3.77 (dt, 1H, J ) 9.5, 3.6 Hz),
3.23 (dd, 1H, J ) 2.2, 3.6 Hz), 3.04 (dt, 1H, J ) 2.1, 6.7 Hz),
1.70-0.80 (m, 16H), 0.96 (d, 3H, J ) 6.6 Hz), 0.92 (3H, J ) 6.5
J O960694N