First synthesis of L-ascorbic acid (vitamin C) from a non-carbohydrate source

Article abstract of DOI:10.1039/a806062b

The enantiopure cis-1,2-dihydrocatechol 2, which is obtained by microbial oxidation of chlorobenzene, has been converted, via 3,5-O-benzylidene-L-gulonolactone (8), into L-ascorbic acid (1).

Full text of DOI:10.1039/a806062b

View Article Online / Journal Homepage / Table of Contents for this issue
First synthesis of L-ascorbic acid (vitamin C) from a
non-carbohydrate source
Martin Banwell,* Simon Blakey, Gwion Harfoot and Robert Longmore
Research School of Chemistry, Institute of Advanced Studies, The Australian National
University, Canberra, ACT 0200, Australia
Received (in Cambridge) 3rd August 1998, Accepted 20th August 1998
The enantiopure cis-1,2-dihydrocatechol 2, which is
Cl
Cl
obtained by microbial oxidation of chlorobenzene, has
O
O
OR
been converted, via 3,5-O-benzylidene-L-gulonolactone (8),
into L-ascorbic acid (1).
ii
OR
O
-Ascorbic acid (vitamin C, 1), a biologically significant
2 R = H
3 R,R = C(Me)2
4
i
O
OH
Cl
H
O
OH
O
O
OH
OH
H
H
O
O
Cl
OH
H
HO
O
O
OR
1
2
HO
BnO
reducing agent, is implicated in a variety of key physiological
processes1 including, for example, the production of collagen.2
It is also considered to be important in the prevention of vari-
ous chronic diseases such as cancer, cerebral apoplexy, diabetes,
atopic dermatitis, myocardial infarction and AIDS.3 The com-
pound has been the subject of many synthetic studies4,5 and is
produced commercially from -glucitol via a six step reaction
sequence including an initial microbiological oxidation pro-
cess.6 In connection with other work, we required access to
certain 17O-, 13C- and/or 2H-labelled samples of -ascorbic acid
and none of the existing syntheses, all of which involve the use
of carbohydrate-based starting materials, seemed suitable for
our purposes. Consequently, we now report the first total syn-
thesis of the title compound from non-carbohydrate sources.7,8
The reaction sequence described herein should be especially
amenable to the preparation of labelled ascorbates which could
be used for probing their in vivo functions.
OH
5
6 R = Bn
7 R = H
v
R
O
O
OH
O
OMe
OMe
H
H
ix
x
O
1
O
Compound 11
Ph
O
O
H
H
PO
O
8
9
P = R = H
P = H, R = OH
10 P = TBDMS, R = H
11 P = TBDMS, R = OH
12
vii
viii
The present synthesis (Scheme 1) employs enantiopure
(>99% ee) cis-1,2-dihydrocatechol 2† (obtained by microbial
oxidation of chlorobenzene) as starting material and was
inspired by the seminal contributions of Hudlicky and co-
workers who have pioneered the use of such microbial oxid-
ation products in the chemoenzymatic synthesis of, inter alia,
various monosaccharides.7,9 Thus, the known10 acetonide
derivative, 3, of diol 2 was converted into the previously
reported10,11 mono-epoxide 4 by standard methods. Reaction of
this last compound with benzyl alcohol and a catalytic amount
of trifluoromethanesulfonic acid (TfOH) afforded a stereo-
chemically pure nucleophilic ring-cleavage product which, by
analogy with related reactions of this epoxide,11 is assigned
structure 5‡ {50%, mp 93–94 ЊC, [α]D ϩ23.2 (c 0.2, CHCl3)}.12
Confirmation of this assignment followed from the conversion
of compound 5 into the known13 2,3-O-isopropylidene--
gulono-1,4-lactone (7). Thus, subjection of chloroalkene 5 to
reaction with ozone in methanol and subsequent in situ reduc-
tion of the intermediate hydroperoxide with sodium cyano-
borohydride at pH 3 gave the -gulonolactone derivative 6
{74%, mp 120–122 ЊC, [α]D ϩ53.4 (c 0.8, CHCl3)}. Compound
6 could not be de-benzylated under standard conditions but
by using transfer hydrogenolysis techniques14 (with sonication)
conversion into compound 7 {53%; mp 143–146 ЊC; lit.,13 mp
Scheme 1 Reagents and conditions: (i) (CH3O)2C(CH3)2 (1.5 mol
equiv.), CH2Cl2, p-TsOH (1.5 mol%), 20 ЊC, 0.33 h; (ii) MCPBA,
NaHCO3, CH2Cl2, 18 ЊC, 5 h; (iii) C6H5CH2OH (1.05 mole equiv.),
TfOH (5 mol%), CH2Cl2, 20 ЊC, 0.25 h; (iv) O3 (excess), CH3OH,
Ϫ78 ЊC → 20 ЊC then HCl (2 M solution in CH3OH), NaBH3CN
(8 mole equiv.), 20 ЊC, 3 h; (v) cyclohexa-1,4-diene (10 mole equiv.), 10%
Pd on C, CH3CH2OH, sonication, 18 ЊC, 8 h; (vi) C6H5CHO (4 mole
equiv.), TfOH (5 mol%), 18 ЊC, 4 h; (vii) TBDMSCl (1.1 mole equiv.),
imidazole (3.0 mole equiv.), DMAP (3 mol%), THF, 0 ЊC → 18 ЊC,
19 h; (viii) Dess–Martin periodinane (1.1 mole equiv.), CH2Cl2, 18 ЊC,
1.5 h; (ix) CH3CO2H–H2O (7:3), 70–75 ЊC, 4 h; (x) CH3COCl (5 mol%),
(CH3)2CO, 18 ЊC, 2 h then K2CO3 (5 mole equiv.), (CH3O)2SO2 (5 mole
equiv.), 56 ЊC, 3 h.
143–146 ЊC; [α]D ϩ74 (c 0.5, EtOH); lit.,13 [α]D ϩ30 (c 2,
EtOH)§} was achieved. Reaction of acetonide 7 with neat benz-
aldehyde and TfOH as catalyst resulted in a trans-acetalisation
reaction and formation of the benzylidene acetal 8 {47%;
mp 188–190 ЊC; lit.,5b mp 188–189 ЊC; [α]D ϩ57 (c 1.0, DMF);
lit.,5b [α]D ϩ61.1 (DMF)} which was identical, in all respects,
with an authentic sample5,15 prepared (in <10% yield) from
commercially available -gulonic-γ-lactone. Theoretically, the
acquisition of acetal 8 constitutes a formal total synthesis of
J. Chem. Soc., Perkin Trans. 1, 1998, 3141–3142
3141

View Article Online
-ascorbic acid since Crawford and Breitenbach5 have con-
verted the former compound into the latter via sequential oxid-
ation and acetal hydrolysis steps. However, despite numerous
attempts we were not able to effect the previously reported5
manganese dioxide16-promoted oxidation of compound 8 to
compound 9. In order to examine alternate oxidants, the pri-
mary hydroxy group of diol 8 was protected as the correspond-
ing TBDMS-ether so as to give compound 10 {74%; mp 176 ЊC;
[α]D ϩ55.5 (c 0.4, CHCl3)}. After extensive experimentation
with a variety of potential oxidants it was established that the
Dess–Martin periodinane17 effects smooth conversion of com-
pound 10 into a variable mixture of ketone hydrate 11 and a
regioisomer which is presumed to derive from silyl-group
migration. Acetic acid promoted hydrolysis of this unstable
mixture then provided -ascorbic acid (1) in 66% yield (from
compound 10). For the purposes of comprehensive spectro-
scopic characterisation, compound (1) was converted, via a
simple one-pot procedure,18 into the stable derivative (12) {80%,
mp 99–100 ЊC; lit.,18 mp 100–101 ЊC; [α]D ϩ9.1 (c 0.6, CHCl3);
[α]D (authentic sample) ϩ9.3 (c 0.6, CHCl3)}. This latter
material was identical, in all respects, with an authentic sample
prepared from commercial -ascorbic acid.
ships to S. B. and G. H. Dr Gregg Whited of Genencor Inter-
national Inc. (Palo Alto, CA) is warmly thanked for providing
generous samples of the cis-1,2-dihydrocatechol 2. Helpful
discussions with Professor R. J. Ferrier (Victoria University of
Wellington) are gratefully acknowledged.
Notes and references
† Available commercially from Genencor International Inc., 925
Page Mill Road, Palo Alto, CA 94304-1013, USA.
1
‡ All new compounds had spectroscopic data (IR, H and 13C NMR,
m/z) consistent with the assigned structure. Satisfactory combustion
analysis data were obtained for new compounds.
§ The specific rotation recorded for an authentic sample of compound
(7) {[α]D ϩ74 (c 0.5, EtOH)} matched that obtained on our material
thus suggesting that the value recorded in the literature13 may be in
error.
1 P. Ge and K. L. Kirk, J. Org. Chem., 1997, 62, 3340 and references
cited therein.
2 F. Mahmoodian, A. Gosiewska and B. Peterkofsky, Arch. Biochem.
Biophys., 1996, 336, 86.
3 (a) R. C. Rose and A. M. Bode, FASEB J., 1993, 1135; (b) B. N.
Ames, M. K. Shigenaga and T. M. Hagen, Proc. Natl. Acad. Sci.
USA, 1993, 90, 7915.
4 M. Csiba, J. Cleophax, S. Petit and S. D. Gero, J. Org. Chem., 1993,
58, 7281 and references cited therein.
Experimental
5 (a) T. C. Crawford and R. Breitenbach, J. Chem. Soc., Chem. Com-
mun., 1979, 388; (b) T. C. Crawford, US Patent 4,111,958 (1978)
(Chem. Abstr., 1979, 90, 138150p).
6 P. Collins and R. Ferrier, Monosaccharides – Their Chemistry and
Their Roles in Natural Products, J. Wiley and Sons, Chichester, 1995,
pp. 311–312.
7 For an excellent review on “modern methods of monosaccharide
synthesis from non-carbohydrate sources” see: T. Hudlicky, D. A.
Entwistle, K. K. Pitzer and A. J. Thorpe, Chem. Rev., 1996, 96, 1195.
8 Syntheses of the ( )-modifications of 4-thioascorbic acid and 4-
azaascorbic acid from non-carbohydrate starting materials have
been reported. See (a) H.-D. Stachel, J. Schachtner and H. Lotter,
Tetrahedron, 1993, 49, 4871; (b) H.-D. Stachel, K. Zeitler and
H. Lotter, Liebigs Ann. Chem., 1994, 1129.
9 T. Hudlicky and A. J. Thorpe, Chem. Commun., 1996, 1993 and
references cited therein.
10 T. Hudlicky, J. Rouden, H. Luna and S. Allen, J. Am. Chem. Soc.,
1994, 116, 5099.
11 M. G. Banwell, N. Haddad, T. Hudlicky, T. C. Nugent, M. F.
Mackay and S. L. Richards, J. Chem. Soc., Perkin Trans. 1, 1997,
1779.
12 The bromo-analogue of compound 5 has recently been reported:
T. Hudlicky, K. A. Abboud, J. Bolonick, R. Maurya, M. L. Stanton
and A. J. Thorpe, Chem. Commun., 1996, 1717.
13 T. Ravindranathan, S. V. Hiremath, D. Rajagopala Reddy and
A. V. Rama Rao, Carbohydr. Res., 1984, 134, 332.
14 A. M. Felix, E. P. Heimer, T. J. Lambros, C. Tzougraki and
J. Meienhofer, J. Org. Chem., 1978, 43, 4194.
15 M. Matsui, M. Okada and M. Ishidate, Yakugaku Zasshi, 1966, 86,
110 (Chem. Abstr., 1966, 64, 15967b).
16 J. Attenburrow, A. F B. Cameron, J. H. Chapman, R. M. Evans,
B. A. Hems, A. B. A. Jansen and T. Walker, J. Chem. Soc., 1952,
1094.
17 (a) R. E. Ireland and L. Liu, J. Org. Chem., 1993, 58, 2899; (b)
V. Samano and M. J. Robins, J. Org. Chem., 1990, 55, 5186.
18 M. G. Kulkarni and S. R. Thopate, Tetrahedron, 1996, 52, 1293.
19 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923.
Compound 6
A solution of chloroalkene 5 (100 mg, 0.32 mmol) in methanol
(15 cm3) was cooled to Ϫ78 ЊC and a stream of ozone was
bubbled through the solution until a blue colouration persisted.
The solution was then purged with dioxygen for 0.25 h before
being allowed to warm to room temperature. The resulting
solution was concentrated to about one-third of the original
volume then treated with one drop of methyl orange indicator
followed by sufficient HCl (2 M solution in methanol) to estab-
lish pH 3. Sodium cyanoborohydride (160 mg, 2.5 mmol) was
added in four roughly equal portions over a period of 3 h while
ensuring pH 3 was maintained by appropriate additions of
HCl. The reaction was quenched with acetone (10 cm3) then
filtered through CeliteTM and the filtrate concentrated under
reduced pressure to give a light yellow oil. Subjection of this
material to flash chromatography19 (1:1 hexane–ethyl acetate
elution) and concentration of the appropriate fractions (Rf 0.2
in 3:2 hexane–ethyl acetate) afforded a white solid. Recrystal-
lisation (CH2Cl2–hexane) of this material then gave the title
compound 6 (73 mg, 74%) as white needles, mp 118–120 ЊC
(Found: Mϩ 308.1262; C, 62.0; H, 6.6. C16H20O6 requires Mϩ
᝽
᝽
308.1260; C, 62.3; H, 6.5%); νmax (KBr)/cmϪ1 3473 and 1788;
δH (300 MHz, CDCl3) 7.42–7.28 (5H, complex m), 4.85 (1H, d,
J 11.4 Hz), 4.80 (2H, broadened s), 4.68 (1H, m), 4.64 (1H, d,
J 11.4 Hz), 3.91 (1H, dd, J 12.1 and 2.7 Hz), 3.84 (1H, dt, J 8.2
and 3.1 Hz), 3.77 (1H, dd, J 12.1 and 3.1 Hz), 2.04 (1H, br s),
1.45 (3H, s, CH3), 1.37 (3H, s, CH3); δC (75 MHz, CDCl3) 173.8
(C), 137.6 (C), 128.5 (CH), 128.0 (CH), 127.9(7) (CH), 114.1
(C), 80.3 (CH), 78.7 (CH), 75.9 (CH), 75.8 (CH), 73.4 (CH2),
61.3 (CH2), 26.7 (CH3), 25.8 (CH3); m/z (EI, 70 eV) 308 (8%),
250 [5, (M Ϫ CH3COCH3)ϩ ] and 91 (100, C7H7ϩ).
᝽
Acknowledgements
The Research School of Chemistry is thanked for generous
financial support including the provision of Summer Student-
Communication 8/06062B
3142
J. Chem. Soc., Perkin Trans. 1, 1998, 3141–3142