Direct conversion of (S)-3-hydroxy-γ-butyrolactone to chiral three- carbon building blocks

Full text of DOI:10.1021/jo9818226

1
036
J . Org. Chem. 1999, 64, 1036-1038
antiviralagents14 such as (6) and the thromboxane syn-
Dir ect Con ver sion of
thase inhibitor15 (7). Despite the existence of these
(
S)-3-Hyd r oxy-γ-bu tyr ola cton e to Ch ir a l
Th r ee-Ca r bon Bu ild in g Block s
Guijun Wang and Rawle I. Hollingsworth*,†
Department of Chemistry, Michigan State University,
East Lansing, Michigan 48824
Received September 8, 1998
Many pharmaceutical compounds contain three-carbon
chiral substructures, the development of routes to which
is often the most difficult aspect of the large scale
synthesis of such compounds. Because of this, there is a
constant ongoing effort to identify synthetic routes to
these three-carbon chiral building blocks. Another strat-
egy, the so-called “chiral pool” approach, is to identify
natural sources containing these substructures and carve
them out to provide the substructures in a form that
could be integrated easily into a drug synthesis scheme.
Such building blocks include chiral glycidols, halo diols,
and aminodiols. Three key three-carbon building blocks
are (R)-glycidol (1), (R)-1-bromo-2,3-dihydroxypropane
methods, there is a need for others that might have
certain advantages. For instance, they might have the
potential to yield some compounds more directly, give
better results at larger scale, not require the use of high
pressures or hydrogen, give superior optical purity, offer
some potential price advantages, or not produce certain
waste materials especially those containing heavy metals.
Because of the high degree of oxygenation in the
desired three-carbon fragments, carbohydrates are a
logical choice when the strategy is to carve out the desired
chiral substructure from some readily available, natu-
rally occurring, chiral raw materials. They have the
highest density of chiral functional groups of all naturally
occurring materials. Many of them, such as starch and
lactose, are readily available and are very cheap, renew-
able feedstock. It is, however, extremely difficult to
convert carbohydrate raw materials into chiral small
molecules in good yield by a simple process. This is
because several steps of protection and functionalization
typically have to be followed. In addition to this, carbo-
hydrates are appreciably soluble only in water. This has
proven to be a serious limitation to the chemical trans-
formations they can be subjected to. They are also very
sensitive both to acids and bases, and pH conditions are
also a major constraint. Despite these difficulties, there
are some applications where carbohydrates have been
used as a source of small chiral building blocks. One of
these methods with some practical value is outlined in
Scheme 1. In this process, fructose or mannose is reduced
to mannitol by catalytic hydrogenation.16 This is then
converted to a di-O-isopropylidene acetal which can be
oxidized to yield two molecules of D-isopropylidene-
glyceraldehyde, a valuable three-carbon synthon, by
treatment with sodium periodate17 or lead tetraacetate.
The drawback of this approach is that the carbohydrate
(2), and (S)-3-amino-1,2-dihydroxypropane (3). Current
routes to (R)-glycidol (1) and its (S)-isomer utilize strate-
gies such as catalytic oxidations with peroxides and chiral
transition metal complexes,1 enzymatic resolutions of
racemic esters using lipases to selectively deacylate one
enantiomer,3 or using glycerol kinase to selectively
,2
-5
6
phosphorylate one isomer. Another common method is
to treat a chiral 1,2-propanediol with a leaving group
such as a halide or tosylate ester in the 3-position with
7
base. The chemistries of compound 1 and 2 are intercon-
8
nected. The availability of an easy route to 2 constitutes
a straightforward route to 1 since this transformation is
easily effected by treatment of 2 with a mild base such
as silver oxide or potassium carbonate.9 The amino diol
,10
3
is a substructure that appears in a large number of
classes of important drugs. These include the â-block-
ers1
1-13
such as Propanalol (4) and Atenolol (5) and
†
Phone: 517-353-0613. Fax: 517-353-9334. E-mail: rih@
argus.cem.msu.edu.
(
(
1) Byun, H.-S.; Bittman, R. J . Org. Chem. 1994, 59, 668-671.
2) Kolb, H. C.; Bennani Y. L.; Sharpless, K. B. Tetrahedron:
Asymmetry 1993, 4, 133-141.
3) Bednarski, M. D.; Chenault, H. K.; Simon, E. S.; Whitesides, G.
M. J . Am. Chem. Soc. 1987, 109, 1283-1285.
4) Fuganti, C.; Grasselli, P.; Servi, S.; Lazzarini, A.; Casati, P.
Tetrahedron 1988, 44, 2575-2582.
5) Chenault, H. K.; Chafin, L. F.; Liehr, S. J . Org. Chem. 1998, 63,
(
(
18
(
4
039-4045.
(
(
6) Chen, J .; Shum, W. Tetrahedron Lett. 1993, 34, 7663-7666.
7) Baldwin, J . J .; McClure, D. E.; Gross, D. M.; Williams, M. J .
(12) Fuganti, C.; Grasselli, P.; Seneci, P. F.; Servi, S. Tetrahedron
Lett. 1986, 27, 2061-2062.
Med. Chem. 1982, 25, 931-936.
(
(
8) Hanson, R. M. Chem. Rev. 1991, 91, 437-475.
(13) Bevinakatti, H. S.; Banerji, A. A. J . Org. Chem. 1992, 57, 6003-
9) Kolb, H. C.; Sharpless, K. B. Tetrahedron 1992, 48, 10515-
6005.
1
1
5
0530.
(14) Hol y´ , A. Collect. Czech. Chem. Commun. 1984, 49, 2148-2166.
(15) Wang, L. H.; Matijevic-Aleksic, N.; Ruan, K. H.; Wu, K. K.;
Kulmacz, R. J . J . Biol. Chem. 1996, 271, 19970-19975.
(16) Makkee, M.; Kieboom, A. P. G.; van Bekkum, H. Carbohydr.
Res. 1985, 138, 237-245.
(
10) Furrow M. E.; Schaus, S. E.; J acobsen, E. N. J . Org. Chem.
998, 63, 6776-6777.
(
173-5176.
11) Terao, Y.; Murata, M.; Achiwa, K. Tetrahedron Lett. 1988, 29,
1
0.1021/jo9818226 CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/15/1999

Notes
J . Org. Chem., Vol. 64, No. 3, 1999 1037
Sch em e 1
Sch em e 2
has to be modified and protected before the actual
reaction to generate the building block is applied. In
addition to this, lead is toxic, periodate is relatively
expensive, and large quantities are required because of
its high molecular weight. An alternative set of reagents
for carrying out this oxidation is sodium hypochlorite and
ruthenium(III) chloride. In this case, the protected glyc-
eric acid is the product.19
There is one significant advance in the use of an
unprotected carbohydrate as a raw material for producing
a chiral building block that could be used to gain access
to compounds such as 1-3. This is the development of a
route to (S)-3-hydroxy-γ-butyrolactone (8), a very valu-
able four-carbon chiral synthon for which a simple, one-
step, high-yield route starting from starch, maltose,
lactose, maltodextrins, and other 4-linked D-pyranoses
to a dioxolane ring could also restrict the group suf-
ficiently to limit access to the conformation required for
elimination. The reaction sequence is shown in Scheme
2
.
The lactone20 was first converted to the protected
amide 10. The first product of the Hoffmann reaction on
the protected amide was the protected aminodiol (11).
This was formed with virtually quantitative conversion
as judged by NMR spectroscopic analysis of the crude
reaction mixture although some product was lost on
extraction and concentration. The isopropylidene acetal
has only recently been introduced into the literature.2
0-22
(S)-3-Hydroxy-γ-butyrolactone readily undergoes a vari-
ety of chemical transformations to yield chiral substituted
tetrahydrofurans, amides, lactams, nitriles, and epoxides.
This high degree of flexibility as a chiral intermediate
makes it a very valuable synthon for the pharmaceutical
industry. A simple one-carbon degradation method should
provide the desired transformation of this intermediate
to a three-carbon structure, thus providing routes for
obtaining the building blocks of interest. In principle, this
chain degradation is possible from (S)-3-hydroxy-γ-bu-
tyrolactone by any one of the classical routes such as the
Hoffman, Curtius, Lossen, or Schmidt reactions which
involve rearrangement of acyl species with migration to
1
1 is a very useful form of 3 since it allows modification
of the amino group without interference from the hy-
droxyl functions. It also has a much lower boiling point
than 3, and this makes purification by distillation pos-
sible even with very modest vacuums. The acetal group
was readily removed by treatment of 11 with slightly
more than an equivalent of acid to give 3 as a salt. The
protected aminodiol (11) could be easily converted to the
bromodiol 2 by treating it with nitrous acid in the
presence of bromide ion. The corresponding chloro com-
pound (12) was prepared in good yield by using hydro-
chloric acid and sodium chloride instead.
In summary, we have developed a general route for
converting the (S)-3-hydroxy-γ-butyrolactone to other
useful chiral three-carbon building blocks including (S)-
3-amino-1,2-dihydroxypropane as its isopropylidene ac-
etal, (R)-1-bromo-2,3-dihydroxypropane, and the corre-
sponding chloro compound. The latter two compounds are
glycidol equivalents. The chemical yields are high, and
the enantiomeric purities are extremely high. The pro-
pensity for the γ-hydroxyl group to participate has been
removed by tying it up in a cyclic acetal. No â-elimination
was observed. The same reaction schemes can be used
to obtain the other enantiomers of the above three-carbon
building blocks by using (R)-3-hydroxy-γ-butyrolactone
as starting material.
an electron-deficient nitrogen and loss of the acyl carbon.
Of these, the Hoffman reaction23 seemed the most at-
tractive because the starting compound is a primary
amide. This primary amide can easily be obtained by
treating the lactone with ammonia. There are potential
complications, however. Some of these are peculiar to this
starting material and include â-elimination and the
participation of the γ-hydroxy group instead of migration
to re-form the lactone. There are also a myriad of side
reactions, such as dialkyl urea formation, that normally
attend the Hoffmann reaction. Protection of the hydroxy
groups was necessary, and the use of an isopropylidene
function was explored because of the ease of installation
and removal. In addition, confining the â-alkoxy fragment
(
17) (a) J ackson, D. Y. Synth. Commun. 1988, 18, 337-341. (b)
Kawakami, Y.; Asai, T.; Umeyama, K.; Yamashita, Y. J . Org. Chem.
1
982, 42, 3581-3585.
18) J ung, M. E.; Shaw, T. J . J . Am. Chem. Soc. 1980, 102, 6304-
311.
19) Emons, C. H. H.; Kuster, B. F. M.; Vekemans, J . A. J . M.;
(
6
Exp er im en ta l Section
(
Sheldon, R. A. Tetrahedron: Asymmetry 1991, 2, 359-362.
(
S)-4-(2,2-Dim eth yl)-1,3-d ioxola n e Aceta m id e (10). (S)-
(
20) Hollingsworth, R. I. Biotech. Annu. Rev. 1996, 2, 281-287.
21) Huang, G.; Hollingsworth, R. I. Tetrahedron 1998, 54, 1355-
3
-Hydroxy-γ-butyrolactone (204 g, 2 mol) was converted to the
(
amide by treatment at room temperature for 14 h with 440 mL
of 30% ammonium hydroxide (3.4 mol). The solution was then
concentrated to a syrup at ∼50° C under reduced pressure until
no more water could be removed. Acetone (2 L) and 2,2-
1
360.
(
(
22) Hollingsworth, R. I. U.S. Patent 5,292,939, 1994.
23) (a) Wallis, E. S.; Lane, J . F. Org. React. 1946, 3, 267-306. (b)
Sy A. O.; Raksis, J . W. Tetrahedron Lett. 1980, 21, 2223-2226.

1
038 J . Org. Chem., Vol. 64, No. 3, 1999
Notes
dimethoxypropane (420 g, 2 mol) were added. Sulfuric acid (6
mL) was then added, and the mixture was protected from
moisture with a calcium chloride drying tube, heated at 60 °C
for 30 min, and stirred at room temperature for 12 h. Sodium
carbonate (50 g) was added and the mixture stirred for 1 h.
Methanol (400 mL) was then added and the mixture filtered and
concentrated to dryness. The amide (10) crystallized on concen-
trating. The crystals were washed with hexane and then acetone
and were used without further purification. Conversion was
essentially quantitative. A small amount when recrystallized
(m, 1 H), 3.47 (dd, 1 H, J ) 12, 4.8 Hz), 3.41 (dd, 1H, J ) 12, 5.7
Hz), 3.00 (dd, 1H, J ) 13.2, 3.0 Hz), 2.78 (dd, 1H, J ) 13.2, 9.3
Hz; 13C NMR (D
O, 75 MHz) δ ppm 63.3, 58.6, 37.2.
R)-3-Br om o-1,2-p r op a n ed iol (2). The amine 11 (10 g) was
2
(
dissolved in 400 mL of water. HBr solution (50 mL, 47% aqueous
solution) and 52 g of sodium bromide were added to the solution
which was then cooled to 10 °C. Sodium nitrite (70 g) was added
to the mixture, and it was stirred at room temperature for 20 h
after which time NMR spectroscopy indicated complete conver-
sion of the aminodiol to the bromodiol. The mixture was
neutralized by sodium bicarbonate, then most of the water was
removed by rotary evaporation, and the residue was taken up
in chloroform. The chloroform solution was dried over sodium
sulfate, and removal of the solvent gave the bromodiol 2 as a
2
3
from acetone gave white crystals, mp 106-108 °C. 10: [R] 589
1
)
-15.4 (CHCl
3
, c ) 1); H NMR (CDCl
3
, 300 MHz) δ ppm 6.10
(
s, 1H), 5.65 (s, 1H), 4.43 (m, 1H), 4.14 (dd, 1H, J ) 8.1, 6.3 Hz),
3
2
.63 (dd, 1H, J ) 8.1, 6.8 Hz), 2.55 (dd, 1H, J ) 15.3, 7.5 Hz),
.46 (dd, 1H, J ) 15.3, 4.8 Hz), 1.42 (s, 3H), 1.35 (s, 3H); 13
, 75 MHz) δ ppm 172.9, 109.5, 72.2, 69.0, 40.1, 26.9,
C
23
yellow liquid. The yield was 10.3 g (87%). 2: [R] 589 ) - 4.00
NMR (CDCl
3
17b
25
1
(
(
3
CHCl
CDCl
3
3
, c ) 1) (lit. [R] 589 ) -3.94, c ) 5.07, CHCl ); H NMR
-
1
2
1
5.5; IR (CHCl
3
, NaCl window) cm 3361, 3190, 2994, 2898,
661, 1633, 1431, 1419, 1372; MS (EI, MH ) calcd 160.0974,
3
, 300 MHz) δ ppm 3.91 (m, 1H), 3.75 (dd, 1H, J ) 11.4,
+
2
1
.6 Hz), 3.64 (dd, 1H, J ) 11.4, 6.0 Hz), 3.44 (m, 2H) (lit. , H
found 160.0968.
S)-4-(2,2-Dim eth yl)-1,3-d ioxola n -4-ylm eth yla m in e (11)
NMR, CDCl , 400 MHz, 3.98-3.89, 3.78, 3.69, 3.51, 3.46 ppm);
3
(
13
2
C NMR (CDCl
00 MHz, 71.4, 64.3, 34.9 ppm).
R)-3-Ch lor o-1,2-p r op a n ed iol (12). The procedure was
3 3
,75 MHz) δ ppm 71.4, 64.3, 34.6 (lit. , CDCl ,
The amide (10) (79.5 g, 0.5mol) was treated with a 10-12%
sodium hypochlorite solution (500 mL) and the mixture stirred
until all of the solid had dissolved (∼5 min). Sodium hydroxide
1
(
similar to that used in the preparation of the bromodiol. The
amine 11 (2.62 g, 0.02 mol) was dissolved in 10 mL of water.
Sodium chloride (8.78 g, 0.15 mol) and concentrated hydrochloric
acid (20 mL, 0.2 mol) diluted with 10 mL of water were added
to the mixture. Sodium nitrite (10.4 g, 0.15mol) was then added
over a period of 10 min. The mixture was then stirred for 24 h,
after which time an analysis of the reaction by NMR spectros-
copy indicated complete conversion to the chlorodiol. The mixture
was then concentrated to dryness, and the product was extracted
with chloroform 3 or 4 times. The extracts were combined and
(80 g dissolved in 500 mL of water) was added to the mixture,
and the solution was warmed to 50-60 °C and kept at this
temperature for 24 h by which time conversion to amine 11 was
1
completed. H NMR spectroscopy indicated 100% conversion of
1
0 to 11. The amine 11 was isolated as a light yellow liquid by
extraction of the mixture with ether which upon standing gave
2
4
colorless crystals. This was reported to be a liquid, probably
because it was not isolated in as pure a state as we have here.
The yield was 56.5 g (86%). 11: mp 54-56 °C, bp 61 ( 2 °C, 15
2
4
23
Torr; 98-100 °C, 100 Torr (lit. bp 62-65 °C, 15 Torr); [R] 589
dried with sodium sulfate. Removal of solvent gave chlorodiol
2
4
20
) +15.0, neat); 1H NMR
)
+0.9 (CHCl
3
, c ) 1) (lit. [R]
, 300 MHz) δ ppm 4.13 (m, 1 H), 4.00 (dd, 1 H, J ) 8.1,
D
23
1
2
2 as a light yellow liquid (1.81 g, 82%): [R] 589 ) -7.2 (H O,
(CDCl
3
25
20
O; lit. [R]25589 ) -7.4, c )
26
c ) 5) (lit. [R]
1
3
D
) -6.8, c ) 5, H
2
6
4
3
6
2
.6 Hz), 3.67 (dd, 1H, J ) 8.1, 6.3 Hz), 2.85 (dd, 1H, J ) 13.2,
1
, H
.48 (m, 4H) (lit.
2
O); H NMR (D O, 300 MHz) δ ppm 3.86 (m, 1H), 3.68-
2
.2 Hz), 2.78 (dd, 1H, J ) 13.2, 6.0 Hz), 1.40 (s, 3H), 1.34 (s,
25
1
H NMR, D
2
O, δ (DSS), 3.85-4.0, 3.6-3.5
1
3
H), 1.31 (s, 2H); C NMR (CDCl
3
, 75 MHz) δ ppm 109.1, 77.4,
, 109.1, 77.5, 67.0, 44.8, 26.8,
13
ppm); C NMR (CDCl
3
, 75 MHz) δ ppm 71.7, 63.6, 45.8. All
6.9, 44.7, 26.8, 25.3 (lit.24 CDCl
5.4).
3
products were >99.5% optically pure by chiral GC.
(
S)-3-Am in o-1,2-p r op a n ed iol (3). Compound 11 (1 g) was
Ack n ow led gm en t. This work was supported by
Synthon Corporation and the Michigan State University
Research Excellence Fund.
treated with 2 mL of concentrated hydrochloric acid and 2 mL
of water. The solution was heated on a steam bath for 30 min.
The solvent was removed by rotatory evaporation to yield a light
yellow syrup, which upon cooling to room temperature gave
white crystals of the hydrochloride salt. The conversion was
quantitative. 3: [R] 589 ) -23.3 (H
28.8, c ) 2.0, 5 N HCl); H NMR (D
J O9818226
2
3
24
20
2
O, c ) 1) (lit. [R]
D
)
1
-
2
O, 300 MHz) δ ppm 3.76
(25) Crans, D. C.; Whitesides, G. M. J . Am. Chem. Soc. 1985, 107,
7
019-7027.
(
24) Danklmaier, J .; Hoenig, H. Liebigs Ann. Chem. 1988, 1149-
(26) Ellis, M. K.; Golding, B. T.; Watson, W. P. J . Chem. Soc., Chem.
1
154.
Commun. 1984, 23, 1600-1602.