Polymer supported oxazolidin-2-ones derived from L-serine - A cautionary tale

Article abstract of DOI:10.1016/S0040-4039(00)01301-0

The capacity of N-Boc-4-hydroxymethyl-oxazolidin-2-ones to undergo rapid O-O and N-O acyl transfer makes these serine derived chiral auxiliaries unsuitable for attachment to polymers. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01301-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 7577–7581
Polymer supported oxazolidin-2-ones derived from
-serine—a cautionary tale
L
Sean P. Bew, Steven D. Bull and Stephen G. Davies*
The Dyson Perrins Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QY, UK
Received 14 June 2000; accepted 28 July 2000
Abstract
The capacity of N-Boc-4-hydroxymethyl-oxazolidin-2-ones to undergo rapid O–O and N–O acyl
transfer makes these serine derived chiral auxiliaries unsuitable for attachment to polymers. © 2000
Elsevier Science Ltd. All rights reserved.
There is currently much interest within the synthetic community directed towards the
development of polymer supported chiral auxiliaries for the asymmetric synthesis of libraries of
homochiral compounds.1 Within this area, a number of homochiral polymer bound oxazolidin-
2-ones derived from tyrosine and serine and attached through the side-chain hydroxyl function-
ality have been described,2 although the key issue of polymer recyclability has not been
addressed.3
As recognised by Allin and Shuttleworth,2a serine derived oxazolidin-2-ones have the potential
for polymer support via attachment of the side-chain hydroxyl group of 1 to Merrifield resin.
They have reported that polymer 3 may be employed for the asymmetric synthesis of carboxylic
acid 4 in 42% yield and 96% e.e. (Scheme 1).2a
Scheme 1. Allin and Shuttleworth’s synthesis of polymer bound oxazolin-2-one derived from L-serine. Reagents and
conditions: (i) KH (1.5 equiv.), DMF, 0°C; Merrifield resin, 18-crown-6 (cat.), 80°C, 5 days; (ii) dil. HCl, CH2Cl2, D,
6 h; (iii) (CH3CH2CO)2O, DMAP (10%), Et3N, THF, D, 4 days; (iv) LDA (2 equiv.), THF, 0°C; BnBr (2 equiv.);
NH4Cl (aq.); (v) LiOH, THF, H2O
* Corresponding author.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01301-0

7578
In the SuperQuat series,4 generation of the fully protected intermediate 5 from
L
-serine
proceeded as planned (Scheme 2). However, treatment of N-Boc-O-TBDMS-SuperQuat 5 with
TBAF did not afford the desired N-Boc-oxazolidin-2-one 9 but instead gave homochiral
O-Boc-oxazolidin-2-one 8 as a single product in a high yield ([h]2D3 −53.2, c 0.66, CHCl3). Clearly
TBAF mediated O-deprotection of the silyl protecting group of 5 afforded alkoxide intermediate
6 which readily underwent intramolecular attack at the N-Boc carbonyl group to generate 7,
resulting in protecting group migration to afford O-Boc-oxazolidin-2-one 8. Attempts to
O-benzylate (as a model reaction for Merrifield resin) the intermediate alkoxide anion 6 before
N-O-Boc protecting group migration could occur were also unsuccessful since TBAF deprotec-
tion of 5 followed by the immediate addition of a large excess of benzyl bromide afforded
N-benzyl-O-Boc-oxazolidin-2-one 10 (Scheme 2).
Scheme 2. Reagents and conditions: (i) MeMgBr (4 equiv.), THF, −78°C; (ii) KOtBu, THF; (iii) Boc2O, DMAP (cat.),
NEt3, CH2Cl2; (iv) TBAF, THF; (v) NH4Cl (aq.); (vi) BnBr
There is clear literature precedent for this type of N–O acyl exocyclic rearrangement occurring
for serine derived N-acyl oxazolidin-2-ones. McCombie et al. have shown that N-benzoyl-oxazo-
lidin-2-one 13, prepared in situ via treatment of (rac)-cis-epoxide 11 with benzoylisocyanate 12,
is unstable and undergoes an N–O benzoyl migration to form (rac)-O-benzoyl-oxazolidin-2-one
14 (Scheme 3).5
Scheme 3.

7579
It is also known that alkoxides of serine derived oxazolidin-2-ones have the capacity to
undergo a competing endocyclic rearrangement via intramolecular attack of the alkoxide anion
at the oxazolidin-2-one carbonyl. For example, Katsumura et al. have shown that alkaline
hydrolysis of homochiral (S)-N-benzyl-O-benzoyl-oxazolidin-2-one 15 (LiOH in THF/H2O;
Cs2CO3 in MeOH; or K2CO3 in MeOH) results in racemic N-benzyl-oxazolidin-2-one 16, via
intramolecular attack of the primary alkoxide at the C2 carbonyl of the oxazolidin-2-one
(Scheme 4).6 This implied that our structural assignment of 8 and 10 might be incorrect since
alkoxide 6 had the capacity to undergo this type of endocyclic rearrangement to afford
oxazolidin-2-one 17.
Scheme 4.
In order to resolve this structural problem an authentic sample of oxazolidin-2-one 10 was
prepared according to the protocol described in Scheme 5. Thus, treatment of (rac)-epoxide 18
with benzylisocyanate in the presence of a catalytic amount of bis-(dibutylchlorotin)oxide
afforded epoxy-carbamate 19 which was cyclised to oxazolidin-2-one 20 via treatment with
n-BuLi at −20°C. Subsequent treatment of 20 with NaH at room temperature led to equilibra-
tion of 20 to the thermodynamic rearranged product 21 which was O-Boc protected to generate
(rac)-10 (Scheme 5).
Scheme 5. Reagents and conditions: (i) Benzylisocyanate, cat. [(n-Bu)2SnCl]2O, CH2Cl2; (ii) n-BuLi (1 equiv.), THF,
−20°C; (iii) NaH, THF, rt; (iv) Boc2O, DMAP, CH2Cl2
Clearly the above endo- or exo-cyclic migrations potentially render serine derived SuperQuat
oxazolidin-2-ones unsuitable for attachment to polymer support via the hydroxymethyl side-
chain. Furthermore, the report by Allin and Shuttleworth (Scheme 1) of the generation of an
alkoxide from 1 (KH, DMF, 0°C) and clean attachment of Merrifield resin {DMF, 18-crown-6
(cat.), 80°C, 5 days} without apparent racemisation (endocyclic rearrangement), or Boc migra-
tion (exocyclic rearrangement), is remarkable and merited investigation. Firstly authentic
samples of (rac)-N-benzyl-O-Boc-23 and (S)-N-Boc-O-benzyl-4-hydroxymethyloxazolidin-2-
ones 25 were prepared as shown in Scheme 6. Thus, (rac)-glycidol was treated with benzyliso-
cyanate, and the resulting carbamate deprotonated with n-BuLi to afford N-benzyl

7580
oxazolidinone 22, which was protected on oxygen to afford (rac)-N-benzyl-O-Boc-oxazolidin-2-
one 23 (Scheme 6). Alternatively, reduction of O-benzyl- -serine with borane, followed by
L
treatment with CDI, afforded O-benzyl-oxazolidin-2-one 24, which was protected on nitrogen to
afford (S)-N-Boc-O-benzyl-oxazolidin-2-one 25 (Scheme 6).
Scheme 6. Reagents and conditions: (i) Benzylisocyanate, cat. [(n-Bu)2SnCl]2O, CH2Cl2; (ii) n-BuLi (1 equiv.), THF,
−20°C; (iii) Boc2O, DMAP, CH2Cl2; (iv) BH3, THF; (v) CDI, CH2Cl2
The infra-red carbonyl absorptions cited by Allin and Shuttleworth for the polymer bound
auxiliary 2 (1749 and 1685 cm−1) were clearly incompatible with the N-benzyl-O-Boc derivative
23 (1744 and 1722 cm−1) or the N-Boc-O-benzyl-derivative 25 (1802 and 1713 cm−1). The
chemistry in Scheme 1 was then reinvestigated. (R)-N-Boc-4-hydroxymethyloxazolidin-2-one 1
was prepared by reduction of the known ester (S)-26 with NaBH4 and shown to be enantiomer-
1
ically pure by H NMR chiral shift experiments ([h]2D3 −45.6, c 0.83, CHCl3) with (+)-Eu(tfc)3 in
comparison with (rac)-1.6 Treatment of homochiral (R)-1 with KH in DMF at 0°C in DMF,
followed by excess BnCl (as a model reaction for Merrifield resin) and 18-crown-6 (cat.) in DMF
at room temperature, led to the clean formation of homochiral (S)-N-benzyl-O-Boc derivative
23 (Scheme 7). This result is consistent with the mechanism described in Scheme 3 whereby the
alkoxide of (R)-1 undergoes a very rapid N- to O-Boc migration to generate a nitrogen anion,
which is subsequently benzylated. Since N-benzyl-O-Boc derivative (S)-23 was isolated in
homochiral form ([h]2D3 +10.5, c 1.54, CHCl3) the observed exocyclic migration described in
Scheme 3 must occur at a much faster rate than the corresponding endocyclic attack of the
alkoxide at the oxazolidin-2-one carbonyl described in Scheme 3. No trace of the N-Boc-O-ben-
zyl derivative (R)-25 was observed in this reaction.
Scheme 7. Reagents and conditions: (i) NaBH4, EtOH, 0°C; phosphate buffer pH 4.0; (ii) KH, DMF, 0°C;
18-crown-6 (cat.), BnCl

7581
In light of these observations we must conclude that the material described by Allin and
Shuttleworth as the O-bound homochiral polymeric auxiliary 2 was in fact N-bound polymeric
auxiliary 27. In order to investigate the possibility that this class of ester might prove useful for
the asymmetric synthesis of homochiral acid 4 we prepared N-benzyl-O-acyl-oxazolidin-2-one
28 ([h]2D3 +10.0, c 1.35, CHCl3) from homochiral (S)-23 ([h]D23 +29.0, c 1.66, CHCl3) according to
the protocol described in Scheme 8 and investigated its performance under enolate alkylation
conditions. Thus, deprotonation of O-acyl-N-benzyl-28 with 2 equivalents of LDA at −78°C in
THF followed by addition of 2 equivalents of benzyl bromide resulted in a mixture of the
epimeric ketones 29/30 in good yield, with no evidence of the expected product of enolate
benzylation 31/32, an authentic sample (0% d.e.) of which was prepared by acylation of the
hydroxyl functionality of 22 with (rac)-4. Employing conditions in which the enolate of 28 was
generated at 0°C, afforded a more complex mixture of reaction products containing the epimeric
ketones 29/30, with no evidence of any 31/32 having been formed.
Scheme 8. Reagents and conditions: (i) 2 M HCl; (ii) (CH3CH2CO)2O, DMAP, Et3N, CH2Cl2; (iii) LDA (2 equiv.),
−78°C, THF; BnBr
In light of these results further work is necessary to explain the success of Allin and
Shuttleworth in employing polymer supported oxazolidin-2-ones for the generation of carboxylic
acid 4 with a 96% e.e.2a
All new compounds were fully characterised including elemental analysis or HRMS.
The successful attachment of the SuperQuat chiral auxiliary derived from tyrosine via its
phenolic hydroxyl will be reported elsewhere.
References
1. Obrecht, D.; Villalgordo, J. M. Solid-Supported Combinatorial and Parallel Synthesis of Small Molecular-Weight
Compound Libraries; Tetrahedron Organic Chemistry Series; Elsevier: Oxford, 1998; Vol. 17.
2. (a) Allin, S. M.; Shuttleworth, S. J. Tetrahedron Lett. 1996, 37, 8023. (b) Purandare, A. V.; Natarajan, S.
Tetrahedron Lett. 1997, 38, 8777; (c) Burgess, K.; Lim, D. J. Chem. Soc., Chem. Commun. 1997, 785. (d) Phoon,
C. W.; Abell, C. Tetrahedron Lett. 1998, 39, 2655.
3. In a recent report on the use of a polymer supported oxazolidin-2-one derived from
L
-tyrosine for asymmetric
.
1,3-dipolar cycloadditions the authors reported on ‘a diminished enantioselectivity of up to 14% e.e.’ upon
recycling; Faita, G.; Paio, A.; Quadrelli, P.; Rancati, F.; Seneci, P. Tetrahedron Lett. 2000, 41, 1265.
4. (a) Davies, S. G.; Sanganee, H. J. Tetrahedron: Asymmetry 1995, 6, 671. (b) Bull, S. D.; Davies, S. G.; Jones, S.;
Polywka, M. E. C.; Prasad, R. S.; Sanganee, H. J. Synlett 1998, 519. (c) Bull, S. D.; Davies, S. G.; Jones, S.;
Sanganee, H. J. J. Chem. Soc., Perkin Trans 1 1999, 387.
5. McCombie, S. W.; Nagabhushan, T. L. Tetrahedron Lett. 1987, 28, 5395.
6. Katsumura, S.; Kondo, A.; Han, Q. Chem. Lett. 1991, 1245, and references cited therein.