Double diastereocontrol in the synthesis of enantiomerically pure polyoxamic acid

Article abstract of DOI:10.1039/a807471b

Polyoxamic acid 4 is prepared by a short and efficient process involving diastereochemically matched cycloaddition of 5-(S)-phenylmorpholin-2-one 1 with (S)-glyceraldehyde acetonide 2, followed by sequential hydrolysis and hydrogenolysis of the adduct.

Full text of DOI:10.1039/a807471b

Double diastereocontrol in the synthesis of enantiomerically pure polyoxamic
acid
Laurence M. Harwood* and Sarah M. Robertson
Department of Chemistry, University of Reading, Whiteknights, Reading, UK RG6 6AD.
E-mail: l.m.harwood@reading.ac.uk
Received (in Liverpool, UK) 24th September 1998, Accepted 26th October 1998
Polyoxamic acid 4 is prepared by a short and efficient
process involving diastereochemically matched cycloaddi-
tion of 5-(S)-phenylmorpholin-2-one 1 with (S)-glycer-
aldehyde acetonide 2, followed by sequential hydrolysis and
hydrogenolysis of the adduct.
from the face opposite the 5-phenyl substituent would be
predicted to furnish cycloadduct 3 (Scheme 2).1
Although a solid, crystals of sufficient quality for X-ray
structural analysis could not be obtained. A combination of
2-dimensional 1H NMR and NOE difference studies supported
the predicted stereochemical outcome of the cycloaddition step,
enhancements of signals corresponding to H2 and H3b on
irradiating H7 proving diagnostic of their mutual syn-relation-
ship, but the stereochemistry at C-9 remained unresolved.‡
Final confirmation of the stereochemical outcome of the ylide
generation and trapping sequence came by conversion of 3 to
polyoxamic acid as its natural (2S,3S,4S) enantiomer. Treatment
of the cycloadduct with aqueous methanolic HCl gave the
deketalized methyl ester which was not isolated but subjected
immediately to hydrogenolysis (H2, Pearlman’s catalyst, aq.
MeOH, TFA) leading to the isolation of a homogeneous
We have demonstrated the use of 5-(S)-phenylmorpholin-2-one
1† as a chiral template in the rapid, diastereo- and enantio-
controlled synthesis of b-hydroxy-a-amino acids.1 In this
process, the chiral azomethine ylide intermediate generated by
condensation with an aldehyde is trapped by excess of the
aldehyde to furnish a cycloadduct which can be subsequently
dismantled to lead to threo-(2S, 3R)-configured products of
high stereochemical integrity (Scheme 1).
In the cycloaddition step, stereochemical discrimination
arises from the chiral azomethine ylide reacting with an achiral
aldehyde dipolarophile. Utilizing a chiral aldehyde in such a
reaction therefore poses the question as to whether there might
be diastereochemical ‘match’ or ‘mismatch’ in either the ylide
generation step or the cycloaddition step between the chiral
reacting partners.
24
material in 95% overall yield with a specific rotation [a]D
+2.4 (c 1.0, H2O), [lit.,2d +2.2 (c 1.0, H2O)] and spectroscopic
data identical with those of authentic polyoxamic acid 4
(Scheme 3).§ In the same manner, the non-natural enantiomer
We now report an efficient synthesis of polyoxamic acid 4,2
the unique polyhydroxyamino acid constituting the side chain
moiety of the antifungal polyoxin antibiotics.3 The key
conversion in our synthesis involves reaction of 5-(S)-phenyl-
morpholin-2-one 1 with excess (S)-glyceraldehyde acetonide 2
(obtained from commercially available 5,6-O-isopropylidene-
H
H
N
Ph
O
O
1
L
-gulono-1,4-lactone.4) in refluxing toluene with removal of
i
water.‡ It would appear that this combination of starting
materials leads to matched diastereocontrol as close examina-
tion of the crude product mixture isolated from reaction
between (S)-2 with 5-(S)-phenylmorpholin-2-one 1 indicated
the presence of only a single product. This could be isolated in
53% yield and showed spectroscopic features consistent with
those expected of a cycloadduct resulting from highly diastereo-
selective reaction of two equivalents of (S)-2 with (S)-1.
By analogy with our earlier rationale for stereocontrol in both
ylide generation and trapping steps, cycloaddition of the
aldehyde dipolarophile to the E-isomer of the azomethine ylide
Bulky acetonide in sterically
least demanding environment
for exo approach of aldehyde
O
(S)
H
O
Phenyl group controls
ring conformation and
geometry of ylid
O
H
(S)
Ph
O
N
H
+
–
O
O
(S)
O
H
H
R
H
H
R
O
H
H
H
H
N
Ph
N
O
Ph
Ph
N
O
R
ii
i
H
O
O
O
O
O
1
iii
O
H
H
O
H
H
H
OH
OH
R
H
O
H
H
Ph
N
Ph
N
iv, v
CO2H
R
H
O
O
H
CO2CH3
H
NH2
H
O
HO
3
Scheme 1 Reagents and conditions: i, RCHO, solvent, reflux; ii, RCHO
(excess); iii, 1 HCl, MeOH, reflux; iv, H2 (5 atm), Pd(OH)2/C, TFA (1
equiv.), aq. MeOH; v, basic ion-exchange resin.
M
Scheme 2 Reagents and conditions: i, (S)-glyceraldehyde acetonide 2 (3
equiv.), toluene, reflux.
Chem. Commun., 1998, 2641–2642
2641

diastereochemically matched reaction of 5-(S)-phenylmorpho-
lin-2-one with (S)-glyceraldehyde acetonide. We are currently
investigating the scope of this process and applying it to the
synthesis of other enantiomerically pure b-hydroxy-a-amino
acids with additional stereogenic centres.
We thank the University of Reading for postgraduate support
under the R.E.T.F. framework (to S. M. R).
O
H
O
O
H
H
O
Ph
N
O
H
H
O
O
H
3
i
Notes and references
OH
H
† We use the trivial morpholin-2-one nomenclature to describe the
2,3,5,6-tetrahydro-4H-oxazin-2-one ring system.
OH
H
H
‡ Cycloaddition procedure: Freshly prepared aldehyde 2 (780 mg, 6 mmol,
3 equiv.) was added to a solution of the morpholin-2-one 1 (177 mg, 1
mmol, 1 equiv.) in dry toluene (60 ml) and the reaction mixture heated to
reflux for 48 h under nitrogen with a Soxhlet extractor containing activated
3Å sieves. Solvent was removed in vacuo to yield a pale yellow oil which
solidified on cooling. Column chromatography, eluting with Et2O–light
petroleum (1:4) and recrystallisation from Et2O furnished 3 as colourless
fine needles (220 mg, 53%), mp 199–202 °C (C22H29NO7 requires C, 63.0;
H, 7.00; N, 3.3. Found C, 62.8; H, 6.85; N, 3.2%); nmax(KBr)/cm21 1737;
dH(250 MHz, CDCl3) 7.47–7.33 (m, 5H), 4.47 (t, J 11.3, 1H), 4.44 (d, J 9.0,
1H), 4.42 (ddd, J 8.9, 6.4, 2.3, 1H), 4.32 (d, J 8.0, 1H), 4.27 (dd, J 11.3, 3.1,
1H), 4.18 (dd, J 9.0, 2.3, 1H), 4.10–4.02 (m, 3H), 3.99 (dd, J 11.3, 3.1, 1H),
(3.91 (t, J 8.9, 1H), 3.84 (dd, J 8.7, 4.2, 1H), 1.40 (s, 6H), 1.26 (s, 3H) and
1.07 (s, 3H); NOE H7?H2 (4.0%)?H3b (3.8%)?H10 (1.2%); dC(100
MHz, CDCl3) 167.4, 135.3, 129.2, 128.9, 128.6, 109.9, 109.3, 96.6, 74.3,
74.1, 73.3, 66.2, 66.0, 60.1, 59.3, 26.3, 26.2, 25.4 and 24.8; m/z (CI) 420
(MH+); [a]D25 +23.2 (c 1.0, CHCl3).
H
H
N
Ph
OH
CO2CH3
HO
ii, iii
OH
H
CO2H
HO
H
OH H NH2
(+)–polyoxamic acid
4
Scheme 3 Reagents and conditions: i, 1
M
HCl, MeOH, reflux; ii, H2 (5
§ Preparation of polyoxamic acid: To a solution of cycloadduct 3 (0.2
atm), Pd(OH)2/C, TFA (1 equiv.), aq. MeOH; iii, basic ion-exchange
resin.
mmol) in MeOH (4 ml) was added 1
M HCl (1 ml) and the reaction mixture
heated to reflux under nitrogen for 1 h. The solvent was removed in vacuo
and the residue transferred to a Fischer–Porter bottle. TFA (15 ml),
Pearlman’s catalyst (70 mg), MeOH (3 ml) and water (0.3 ml) were added,
the solution degassed and subjected to hydrogen at 5 atm. for 48 h. Catalyst
was removed by centrifugation, the solvent removed in vacuo and the crude
mixture purified on a Dowex® basic ion-exchange column to yield 4 as a
colourless powder (31 mg, 95%), mp 152–154 °C (decomp.) (lit., 165–170
°C,2d 162–168 °C2g). Spectroscopic data as in ref. 2(c); [a]D24 +2.4 (c 1.0,
of polyoxamic acid was obtained by reacting (R)-glycer-
aldehyde acetonide (prepared by oxidative cleavage of
1,2:5,6-di-O-isopropylidene-
-mannitol5) with (R)-1, followed
D
by sequential degradation of the cycloadduct. Pure ent-
polyoxamic acid was isolated in excess of 50% yield over the
24
whole sequence, with a specific rotation, [a]D 22.5 (c 1.0,
H2O), [lit.,2d +2.2 (c 1.0, H2O)]; ent-4 [a]D 22.5 (c 1.0, H2O). We thank
24
Professor R. F. W. Jackson for providing copies of 1H and 13C NMR spectra
of polyoxamic acid.
H2O).
Having successfully prepared polyoxamic acid and its
enantiomer it was decided to investigate the synthesis of
diastereoisomers by employing the alternative combination of
reactant enantiomers. However, under the same conditions as
before, reaction of 5-(S)-phenylmorpholin-2-one 1 with (R)-2
resulted in a product mixture consisting of roughly equal
quantities of three products which were found to be diastereoi-
somers of 3 by spectroscopic analysis. Unfortunately none
could be isolated with sufficient purity to permit definitive
structural assignment, nor was it possible to separate the
deprotected acids at the ultimate stage of the synthetic route.
However, the observation of three cycloadducts indicates
diastereochemical mismatch in more than one element of the
ylide generation and trapping sequence, be it reactant ylide
geometry, diastereofacial control or exo/endo approach of the
dipolarophile.
1 L. M. Harwood, J. Macro, D. J. Watkin, C. E. Williams and L. F. Wong,
Tetrahedron: Asymmetry, 1992, 3, 1127; D. A. Alker, G. Hamblett, L. M.
Harwood, S. M. Robertson and C. E. Williams, Tetrahedron, 1998, 54,
6089.
2 For recent synthetic approaches to polyoxamic acid and derivatives, see:
(a) A. Dondoni, S. Franco, F. L. Merchàn, P. Merino and T. Tejero,
Tetrahedron Lett., 1993, 34, 5479; (b) R. F. W. Jackson , N. J. Palmer and
M. J. Wythes, J. Chem. Soc., Chem. Commun., 1994, 95; (c) R. F. W.
Jackson, N. J. Palmer, M. J. Wythes, W. Clegg and M. R. J. Elsegood,
J. Org. Chem., 1995, 60, 6431; (d) B. M. Trost, A. C. Krueger, R. C. Bunt
and J. Zambrano, J. Am. Chem. Soc., 1996, 118, 6520; (e) S. H. Kang and
H-W. Choi, Chem. Commun., 1996, 1521; (f) G. Casiraghi, G. Rassu, P.
Spanu and L. Pinna, Tetrahedron Lett., 1994, 35, 2423; (g) A. K.
Saksena, R. G. Lovey, V. M. Girijavallabhan, A. K. Ganguly and A. T.
McPhail, J. Org. Chem., 1986, 51, 5024.
3 K. Isono, K. Asahi and S. Suzuki, J. Am. Chem. Soc., 1969, 91, 7490.
4 C. Hubschwerlen, Synthesis, 1986, 962; C. Hubschwerlen, J-L. Speclin
and J. Higelin, Org. Synth., 1995, 72, 1.
In conclusion, we have demonstrated that diastereochem-
ically matched and mismatched reactions can occur in the
generation and trapping of azomethine ylides in which both
starting materials are chiral and have established a rapid
5 J. Mann, N. K. Partlett and A. Thomas, J. Chem. Res. (S), 1987, 369.
diastereocontrolled synthesis of polyoxamic acid via
a
Communication 8/07471B
2642
Chem. Commun., 1998, 2641–2642