A cleavable linker strategy for optimising enolate alkylation reactions of a polymer-supported Evans' oxazolidin-2-one

Article abstract of DOI:10.1039/b713966g

A cleavable linker strategy has been used to optimise the enolate alkylation reactions of a recyclable l-tyrosine derived polymer-supported oxazolidin-2-one for the asymmetric synthesis of a series of chiral α-alkyl acids. The Royal Society of Chemistry.

Full text of DOI:10.1039/b713966g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A cleavable linker strategy for optimising enolate alkylation reactions
of a polymer-supported Evans’ oxazolidin-2-one{
Rachel Green,a Andrew T. Merrittb and Steven D. Bull*a
Received (in Cambridge, UK) 11th September 2007, Accepted 13th November 2007
First published as an Advance Article on the web 29th November 2007
DOI: 10.1039/b713966g
These results were clearly inferior to those reported previously
for alkylation of lithium enolates of related N-propionyl-
oxazolidin-2-ones in ‘solution phase’.4 This prompted Kotake
et al. to propose that the poor diastereocontrol observed for (S)-1
might be a result of the oxazolidin-2-one being attached to
polymer via its stereodirecting C4-benzyl group. Therefore, they
prepared new types of Wang-supported N-acyloxazolidin-2-one
(S)-3, whose oxazolidin-2-one fragments were attached to polymer
via their remote C5-positions, leaving an unhindered C4-benzyl
group free for stereocontrol. This strategy proved successful, with
lithium5 and sodium6 enolates of a range of N-acyloxazolidin-2-
ones (S)-3 being used in asymmetric enolate alkylation reactions to
afford eleven chiral a-alkyl acids in 84–97% ee and 50–70% yield.6
Their polymers could be recycled with no loss of stereocontrol,
although a decrease in yield of chiral acids produced between each
reaction cycle was observed.5
A cleavable linker strategy has been used to optimise the
enolate alkylation reactions of a recyclable L-tyrosine derived
polymer-supported oxazolidin-2-one for the asymmetric synth-
esis of a series of chiral a-alkyl acids.
A number of polymer-supported Evans’ oxazolidin-2-ones have
been developed for asymmetric synthesis,1 however their use for
asymmetric enolate alkylation reactions has proven to be
problematic.2 For example, Burgess and Lim initially reported
that alkylation of the lithium enolate of an L-tyrosine derived
Wang-supported N-propionyloxazolidin-2-one (S)-1 with five
equivalents of benzyl bromide at 0 uC, followed by treatment
with LiBH4, gave 2-methyl-3-phenylpropan-1-ol (R)-2a in 66%
yield and 90% ee (Scheme 1).3 However, they described that
extended reaction times led to alcohol (R)-2a in lower ee and
reduced yield, with losses in diastereocontrol dependent on the
amount of base used, and lower yields being due to ‘‘loss of the
propionyl group from the oxazolidin-2-one prior to cleavage’’.
Furthermore, alkylation of the lithium enolate of (S)-1 with allyl
bromide, followed by LiBH4 cleavage, afforded chiral alcohol (R)-
2b in lower 81% ee and a much poorer 25% yield.3
We were unconvinced that the poor performance of (S)-1 in
enolate alkylation reactions was due to the oxazolidin-2-one being
attached to polymer via its stereodirecting C4-benzyl substituent,
since this hypothesis failed to explain why diastereoselectivity was
dependent on both the length of reaction and the amount of base
present. Furthermore, L-tyrosine derived polymer-supported
oxazolidin-2-ones had been used previously in other types of
asymmetric transformation to afford different classes of chiral
product in high de.1
Consequently, we decided to investigate the enolate alkylation
reactions of an L-tyrosine derived oxazolidin-2-one attached to
polymer-support via an orthogonally cleavable linker that would
enable polymer-supported intermediates to be identified as
required. Therefore, 2-chlorotrityl chloride resin was treated with
three equivalents of oxazolidin-2-one (S)-47 and ten equivalents of
diisopropylethylamine (DIPEA) in CH2Cl2–THF–DMF, to give
O-linked-NH-oxazolidin-2-one polymer (S)-5 with a reproducible
loading of 0.96–1.16 mmol g21. The exclusive formation of (S)-5
was confirmed via its treatment with ten equivalents of LDA and
twenty equivalents of BnBr in THF at 278 uC, which gave
polymer (S)-8 that was cleaved with 1% trifluoroacetic acid
(TFA)–5% triisopropylsilane (TIS) in CH2Cl2 to afford
N-benzyloxazolidin-2-one (S)-7 as a single product, with no
O-benzyloxazolidin-2-one being present. The N-propionylation
reaction of polymer (S)-5 was also optimised using this TFA
cleavage approach, which revealed that five equivalents of
propionic anhydride, Et3N and LiCl in THF were required to
ensure complete conversion of (S)-5 to N-propionyl polymer
(S)-6a. These conditions were then employed to prepare a series of
four N-acyloxazolidin-2-one polymers (S)-6a–d with loadings of
between 0.91–1.15 mmol g21 (Scheme 2).
Scheme 1 Enolate alkylation reactions of L-tyrosine derived polymer
(S)-1 and phenylnorstatine derived polymer (S)-3.
aDepartment of Chemistry, University of Bath, Claverton Down, Bath,
UK BA2 7AY. E-mail: s.d.bull@bath.ac.uk
bGlaxoSmithKline Research and Development, Medicines Research
Centre, Gunnels Wood Road, Stevenage, Hertfordshire, UK SG1 2NY
{ Electronic supplementary information (ESI) available: Experimental
details. See DOI: 10.1039/b713966g
508 | Chem. Commun., 2008, 508–510
This journal is ß The Royal Society of Chemistry 2008

View Article Online
In order to explain these results we proposed that base catalysed
epimerisation of (S,aR)-10a was responsible for the poor de in
these reactions, rather than poor diastereoselectivity in the initial
enolate alkylation reaction. In order to confirm this hypothesis,
diastereoisomerically pure (S,aR)-10a (.95% de)10 was treated
with ten equivalents of LHMDS for 12 h at 215 uC, followed by
cleavage with TFA, which gave epimerised (S,aR)-9a (75% de) and
NH-oxazolidin-2-one (S)-4 in a ratio of 9 : 1 (Scheme 3).
Therefore, it was clear that excess LHMDS could deprotonate
(S,aR)-10a resulting in an enolate that was either protonated to
afford (S,aR)-10a in inferior de, or decomposed to regenerate
(S)-5, thus explaining the erosion in de and yield of recovered
(S,aR)-9a over time. Consequently, it was decided to modify the
reaction conditions so that the excess LHMDS responsible for
generating the enolate of (S,aR)-10a would be removed before
addition of benzyl bromide, thus preventing any excess LHMDS
from being available to epimerise (S,aR)-10a. Therefore, (S)-6a
was encased in an IRORI mini-KanTM and treated with ten
equivalents of LHMDS in THF at 215 uC (or 0 uC) for 30 min,
before the solvent/excess LHMDS was removed via cannula. The
resultant polymer was then re-suspended in THF at 215 uC (or
0 uC), and treated with twenty equivalents of benzyl bromide in
pre-chilled THF at 215 uC (or 0 uC). TFA cleavage of the
resultant resin revealed that (S,aR)-9a had been formed in 70–75%
yield and in a much improved .95% de, along with 20–25% NH-
oxazolidin-2-one (S)-4 and 0–5% N-benzyloxazolidin-2-one (S)-7.
All attempts to optimise these conditions further in an attempt to
reduce the amount of (S)-5 formed proved unsuccessful.
Therefore, these conditions were used for reaction of the lithium
enolates of 150 mg batches of N-acyloxazolidin-2-ones (S)-6a–d
with a range of electrophiles (benzyl bromide, methyl iodide and
allyl iodide) to afford nine a-alkylated resins 10a–i. Twenty
milligrams of each resin 10a–i were then cleaved with 1% TFA and
the cleavage products analysed via HPLC, which revealed that
complete consumption of N-acyloxazolidin-2-ones (S)-6a–d had
occurred to afford a-alkylated products 9a–i in good 87–99% de
(Table 2). The remaining resin 10a–i (approx. 130 mg) was then
hydrolysed using five equivalents of LiOOH11 in THF–H2O to
afford their corresponding chiral a-alkyl acids 11a–i in acceptable
42–69% yields (calculated over three steps for N-acylation, enolate
alkylation and side-chain cleavage) (Table 2). The enantiomeric
excess of chiral acid (R)-11a was determined as 97% ee via chiral
HPLC analysis, which was identical in value to the diastereoiso-
meric excess of (S,aR)-9a of 97% de, thus confirming that LiOOH
hydrolysis of (S,aR)-10a had occurred without racemisation.
Consequently, it was concluded that the ee values of a-alkyl-acids
Scheme 2 Formation and N-acylation of 2-chlorotrityl polymer-
supported oxazolidin-2-one (S)-5.
The enolate alkylation reaction of (S)-6a was then investigated,
via treatment with varying amounts of base,8 at different
temperatures, using different amounts of benzyl bromide, followed
by subsequent TFA cleavage to determine product ratios (Table 1).
These results revealed that a competing enolate decomposition
pathway was occurring on polymer support to afford significant
amounts of NH-oxazolidin-2-one (S)-5, and small amounts of
N-benzyloxazolidin-2-one (S)-8.9 Whilst the amount of (S)-5
formed could be reduced to 15% by carrying out the reaction at
215 uC, this lower temperature required excess base and/or
extended reaction times to ensure complete consumption of (S)-6a.
However, changing to these conditions resulted in a significant
reduction in both the de and yield of (S,aR)-10a.
Table 1 Investigating enolate alkylation reactions of (S)-6a using
TFA cleavage to examine product ratios
Ratio of products from
TFA cleavage reaction
Reaction conditions
6a
26
20
0
9a
4
41
15
42
38
7
0
3
0
5
3 eq. LDA, 0 uC;
33 (85% de)
62 (85% de)
58 (76% de)
57 (68% de)
5 eq. BnBr, 30 min
3 eq. LHMDS, 215 uC;
5 eq. BnBr, 30 min
3 eq. LHMDS, 215 uC;
5 eq. BnBr, 12 h
10 eq. LHMDS, 215 uC;
20 eq. BnBr, 12 h
0
Scheme 3 Epimerisation of (S)-10a using excess LHMDS.
Chem. Commun., 2008, 508–510 | 509
This journal is ß The Royal Society of Chemistry 2008

View Article Online
The enolate alkylation results described herein clearly demon-
strate that the poor de values previously reported for L-tyrosine
derived polymer-supported oxazolidin-2-ones are not a result of
the chiral auxiliary being attached to polymer via its C4-
stereodirecting group. Furthermore, our new stepwise reaction
protocol enables the cheap and readily available L-tyrosine derived
polymer (S)-512 to be used for asymmetric enolate reactions with
comparable levels of stereocontrol and performance to Kiso’s
phenylnorstatine derived polymer (S)-3.13
Table 2 De values of chiral N-a-alkyloxazolidin-2-ones 9a–i and
yields of chiral acids 11a–i
In conclusion, we have used a cleavable linker strategy to
optimise the performance of a polymer-supported L-tyrosine
derived oxazolidin-2-one for enolate alkylation reactions, thus
enabling the asymmetric synthesis of a series of nine chiral a-alkyl
acids. We anticipate that this type of cleavable linker approach will
prove extremely useful for optimising the performance of other
types of chiral auxiliary on polymer support.
Reactants
(polymer/R1X)
De of TFA cleavage
products 9a–i
Yield of chiral
acids 11a–i
We would like to thank the EPSRC (RG), GlaxoSmithKline
(RG), the Royal Society (SDB) and the University of Sydney
(SDB) for funding, and the mass-spectroscopic facility at Swansea
for its assistance.
6a/BnBr
6a/AllylBr
6b/MeI
(S,aR)-9a, 97% de
(S,aR)-9b, 92% de
(S,S)-9c, 87% de
(S,S)-9d, 95% de
(S,aR)-9e, 90% de
(S,aR)-9f, 99% de
(S,S)-9g, 90% de
(S,S)-9h, 99% de
(S,S)-9i, 99% de
(R)-11a, 69%
(R)-11b, 61%
(S)-11c, 45%
(S)-11d, 53%
(R)-11e, 48%
(R)-11f, 58%
(S)-11g, 42%
(S)-11h, 52%
(S)-11i, 60%
6c/MeI
6c/BnBr
6c/AllylBr
6d/MeI
6d/BnBr
6d/AllylBr
Notes and references
1 For an excellent review see: C. W. Y. Chung and P. H. Toy,
Tetrahedron: Asymmetry, 2004, 15, 387.
2 See: (a) S. M. Allin and S. J. Shuttleworth, Tetrahedron Lett., 1996, 37,
8023; (b) S. P. Bew, S. D. Bull and S. G. Davies, Tetrahedron Lett., 2000,
41, 7577; (c) S. P. Bew, S. D. Bull, S. G. Davies, E. D. Savory and
D. J. Watkin, Tetrahedron, 2002, 58, 9387; (d) S. M. Allin, C. A. Johnson
and A. Timm, Tetrahedron Lett., 2005, 46, 2495.
11b–i were identical to the de values of their corresponding
a-alkyloxazolidin-2-ones 9b–i.
3 K. Burgess and D. Lim, Chem. Commun., 1997, 785.
We then demonstrated that the polymer could be efficiently
recycled by carrying out four sequential polymer-supported
asymmetric enolate alkylation reactions (comprising N-acylation,
alkylation and side chain cleavage) on the same batch of resin
(S)-5. Characterisation of the a-alkyloxazolidin-2-ones 10 pro-
duced in these consecutive enolate alkylation reactions was
achieved by cleaving 20 mg portions of resin with 1% TFA and
4 (a) D. A. Evans, M. D. Ennis and D. J. Mathre, J. Am. Chem. Soc.,
1982, 104, 1737; (b) D. L. J. Clive, K. S. Keshava Murthy, A. G. H.
Wee, J. S. Prasad, G. V. J. da Silva, M. Majewski, P. C. Anderson,
C. F. Evans, R. D. Haugen, L. D. Heerze and J. R. Barrie, J. Am.
Chem. Soc., 1990, 112, 3018.
5 T. Kotake, S. Rajesh, Y. Hayashi, Y. Mukai, M. Ueda, T. Kimura and
Y. Kiso, Tetrahedron Lett., 2004, 45, 3651.
6 T. Kotake, Y. Hayashi, S. Rajesh, Y. Mukai, Y. Takiguchi, T. Kimura
and Y. Kiso, Tetrahedron, 2005, 61, 3819.
1
analysing the cleavage products by H NMR spectroscopy. This
7 R. Green, P. J. M. Taylor, S. D. Bull, T. D. James, M. F. Mahon and
A. Merritt, Tetrahedron: Asymmetry, 2003, 12, 2619.
revealed that the same batch of resin (S)-5 could be recycled four
times to afford chiral a-alkyl-acids (R)-11a (69%, 95% de for
(S,aR)-9a), (S)-11g (45%, 87% de for (S,S)-9g), (S)-11c (43%, 89%
de for (S,S)-9c) and (R)-11a (38%, 96% de for (S,aR)-9a)
respectively, with no losses in diastereocontrol when compared
to results obtained using ‘fresh’ resin (S)-5 (see Table 2). Indeed,
the first and fourth reaction cycles produced a-benzyl propionic
acid (R)-11a with essentially identical de values of 95% and 97% de
respectively. However, the yield of chiral a-alkyl acids produced
after each progressive reaction cycle did decrease, with the fourth
reaction cycle producing (R)-11a in only 38% yield, which was
much lower than the 69% yield produced in the first reaction cycle
using virgin polymer. 1H-NMR spectroscopic analysis of the NH-
oxazolidin-2-one (S)-5 obtained from cleavage of 20 mg portions
of resin at the end of each reaction cycle (after LiOOH hydrolysis)
revealed increasingly complex spectra, that we propose are a result
of progressive accumulation of small amounts of different
N-alkyloxazolidin-2-ones formed from the enolate decomposition
pathway of each reaction cycle.
8 Although sodium enolates of N-acyloxazolidin-2-ones are more stable
than lithium enolates (ref. 4a), using NaHMDS as a base gave no
reduction in products arising from the enolate decomposition pathway.
9 S. D. Bull, S. G. Davies, S. Jones and H. J. Sanganee, J. Chem. Soc.,
Perkin Trans. 1, 1999, 387.
10 Diatereomerically pure polymer (S,aR)-10a (.95% de) was prepared via
asymmetric synthesis of (4S)-4-(4-hydroxybenzyl)-3-((2R)-2-methyl-3-
phenylpropionyl)oxazolidin-2-one in ‘solution phase’, followed by its
attachment to 2-chlorotrityl resin.
11 TFA cleavage studies revealed that lower concentrations of LiOOH
resulted in incomplete hydrolysis of polymers 10a–i, whilst treatment
with LiOH resulted in competing formation of endocyclic cleavage
products, see: D. A. Evans, T. C. Britton and J. A. Ellman, Tetrahedron
Lett., 1987, 28, 6141.
12 Polymer-supported oxazolidin-2-one (S)-3 is prepared from the non-
natural b-amino acid phenylnorstatine which is very expensive; (R,R)-
N-Boc-3 from CNH Technologies (1 g for £195.00), (S,S)-N-Boc-3 from
PepTech Corporation (1 g for £395.00).
13 It is surprising that polymers (S)-3 do not suffer from problems
associated with base-catalysed epimerisation of the a-stereocentres of
their a-alkylated products, with reactions employing excess NaHMDS
(or LDA) for enolate generation having been reported to afford chiral
a-alkyl acids of up to 97% ee. See ref. 5 and 6.
510 | Chem. Commun., 2008, 508–510
This journal is ß The Royal Society of Chemistry 2008