The isomerisation of (Z)-3-[2H1]-phenylprop-2-enone as a measure of the rate of hydroperoxide addition in Weitz-Scheffer and Julia-Colonna epoxidations

Article abstract of DOI:10.1039/b404389h

(Z)3-[²H₁]-Phenylprop-2-enone is isomerised by hydroperoxide to an equimolar mixture of the (Z)- and (E)-isomers prior to epoxidation. Poly-(L)-leucine (10 mole %) accelerates the addition of hydroperoxide by an order of magnitude an

Full text of DOI:10.1039/b404389h

View Article Online / Journal Homepage / Table of Contents for this issue
The isomerisation of (Z)-3-[2H1]-phenylprop-2-enone as a measure of
the rate of hydroperoxide addition in Weitz–Scheffer and Julia´–
Colonna epoxidations{
David R. Kelly,*a Eva Caroff,b Robert W. Flood,b William Healb and Stanley M. Roberts*b
a Department of Chemistry, Cardiff University, P. O. Box 912, Cardiff, Wales, UK CF10 3TB.
E-mail: KellyDR@Cardiff.ac.uk; Fax: (44) 029-20874030; Tel: 029-20874063
b Department of Chemistry, University of Liverpool, Liverpool, UK L69 3BX.
E-mail: smrsm@liverpool.ac.uk; Tel: 0151-7943501
Received (in Cambridge, UK) 23rd March 2004, Accepted 23rd July 2004
First published as an Advance Article on the web 25th August 2004
(Z)-3-[2H1]-Phenylprop-2-enone is isomerised by hydro-
postulating slow, rate limiting addition of hydroperoxide anion,
followed by fast cyclisation to the epoxide, or fast addition followed
by slow rate-limiting cyclisation.
peroxide to an equimolar mixture of the (Z)- and (E)-isomers
prior to epoxidation. Poly-(L)-leucine (10 mole %) accelerates
the addition of hydroperoxide by an order of magnitude and
sequesters hydroperoxide from THF.
(E)-Chalcone 1a and (Z)-chalcone 1b, under both Weitz–
Scheffer6 and Julia´–Colonna conditions, give the trans-epoxide 3a
(racemic or scalemic respectively), via (E)-chalcone 1a and this
behaviour is typical of other alkenes. Rate studies suggest that
isomerisation of the (Z)-alkene to the (E)-alkene7 occurs by fast
addition of hydroperoxide to give a peroxy-enolate 2, which has a
sufficient lifetime to allow rotation of the a,b-bond before elimina-
tion of hydroperoxide. Definitive conclusions could not be drawn
from the previous studies, because the relative rate of hydroper-
oxide addition to the (Z)- and (E)-alkenes was unknown. Further-
more there was uncertainty as to why the peroxy-enolate derived
from the (Z)-alkene did not cyclise, whereas that from the (E)-
alkene did.
An opportunity to resolve this conundrum was provided by the
observation that phenylprop-2-enone 1c (phenyl vinyl ketone)
was converted into (2R,3)-epoxyphenylpropanone 3c (w 95% ee)8
using PLL, UHP and either DBU or BEMP in THF. Therefore,
the (Z)-deuterio-analogue 1d was synthesised9 which is presumed
to be sterically and electronically unbiased in the addition step and
to undergo free rotation as the peroxy-enolates 2d, 2e.10
Addition of DBU to (Z)-alkene 1d in THF caused isomerization
to the (E)-alkene 1e, due to the reversible addition of the base.11
However, no isomerisation occurred when (Z)-alkene 1d was
dissolved in THF containing BEMP, until UHP was added. Then a
relatively slow isomerization of the (Z)-alkene 1d to the (E)-alkene
1e was observed.12 A plot of the excess of (Z)-alkene 1d (%) against
time (Fig. 1) showed that isomerisation reached the halfway point
after ca. 1 h and was incomplete even after 3 h. Epoxide formation
was near to completion after ten hours.
The Weitz–Scheffer oxidation of a,b-unsaturated ketones 1 by
basic peroxide is a two-stage process involving the initial formation
of a peroxy-enolate 2 followed by ring closure, with loss of
hydroxide ion, to give an epoxide 3 (Scheme 1).1 Julia´ and Colonna
discovered an asymmetric variant of the Weitz–Scheffer reaction,
by adding a peptide, e.g. poly-(L)-leucine (PLL), to aqueous
hydrogen peroxide and substrate (e.g. (E)-chalcone 1a) in a water
immiscible solvent.2 This triphasic system has been recently
complemented by a non-aqueous biphasic protocol in which the
organic solvent contains the substrate, the oxidant and the base.3
The biphasic system exhibits impressively short reaction times
(v 1 h, for more reactive enones) and chalcone epoxide 3a is
produced in remarkably high enantiomeric excess (w 95%) given
that in the absence of catalyst a substantial amount of racemic
chalcone epoxide 3a (ca. 20%)4 is formed in the same time period.
This dichotomy was resolved by a simple experiment.
PLL (100 mg) and UHP (28 mg) were dissolved in THF (1.5 ml)
and stirred for 5 min and then DBU (60 ml) was added. After 5 min
the supernatant THF was removed and the peroxide content of
both the solution and the gel-like PLL was estimated using
potassium iodide and thiosulfate in standard fashion. Only 30% of
the total peroxide was found in the THF solution; the remainder
of the peroxide was associated with the relatively small quantity of
polyleucine gel. Thus one of the catalytic modes of polyleucine is
sequestration of hydroperoxide.
Early studies of the Weitz–Scheffer reaction demonstrated that
the rate of epoxide formation is first order with respect to the
concentration of hydroperoxide or a,b-unsaturated ketone and has
no definite order with respect to hydroxide or hydrogen peroxide
concentration.5 These results can be equally well rationalised, by
Scheme 1 Reagents and conditions: (i) Weitz–Scheffer: H2O2, NaOH,
co-solvent, e.g. CH3OH. (ii) Julia´–Colonna triphasic system: PLL
(20 mole %), 30% H2O2 (30 equiv.), 5 M NaOH (4 equiv.), toluene,
24 h. (iii) Biphasic system: PLL (5–10 mole %), urea–hydrogen peroxide
complex (UHP, 1.3 equiv.), DBU (1.3 equiv.), THF, ½ h. All at rt.
Fig. 1 Isomerisation of (Z)-3-[2H1]-phenylprop-2-enone 1d. Reagents and
conditions: UHP (1.2 equiv.), THF, without and with PLL (10 mole %)
stirred for 10 min and BEMP (1.2 equiv.) added. Reaction monitored by
1H-NMR and HPLC.
{ Electronic supplementary information (ESI) available: experimental
section. See http://www.rsc.org/suppdata/cc/b4/b404389h/
2 0 1 6
C h e m . C o m m u n . , 2 0 0 4 , 2 0 1 6 – 2 0 1 7
T h i s j o u r n a l i s ß T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
When the (Z)-alkene 1d was subjected to biphasic Julia´–Colonna
conditions, the epoxidation reaction was complete after 2 h. HPLC
analysis showed a 78% enantiomeric excess, while 1H-NMR
spectroscopy indicated an equimolar mixture of (2R,3R)-epoxide
3d and (2R,3S)-epoxide 3e isotopomers.13 The excess of (Z)-alkene
1d over (E)-alkene 1e was reduced from 45% to 17% in just five
minutes and complete isomerization was observed after 1 h (Fig. 1).
This isomerisation is at least an order of magnitude faster than that
observed under Weitz–Scheffer conditions.14
Scheme 3 Reagents and conditions: UHP (2 equiv.), DBU (2.1 equiv.), PLL
(20 mole %), THF, 72 h, rt.
There are three main conclusions that can be drawn from these
reactions. Firstly, polyleucine increases the rate of addition of
hydroperoxide to the (Z)-alkene 1d (and presumably 1c, 1e).
Secondly, the peroxy-enolate intermediate 2d undergoes rotation,
even without a steric driving force, and finally epoxide formation is
directed solely by the orientation of the hydroperoxide moiety
relative to the enolate. The increased rate presumably occurs by
stabilisation of the peroxy-enolates 2d, 2e and the latter two
observations are a consequence of the following conformational
constraints. Addition and elimination of hydroperoxide must occur
via the same transition state, with hydroperoxide or the hydro-
peroxy moiety of the peroxy-enolate 2 in alignment with the
orbitals of the p-system. Additionally, when epoxide is formed the
O–O bond must be anti-periplanar to the p-system to enable
overlap with the O–O anti-bonding orbital. Therefore we view the
rapid rate of addition relative to epoxidation, as a failure to achieve
the requisite O–O bond orientation in the majority of the additions.
we conjecture that the orientation of the hydroperoxide moiety of
the peroxy-enolate 2 and facilitation of O–O bond cleavage are due
to hydrogen bonding to the peptide.16 This might be thwarted by
an intramolecular hydrogen bond between the hydroperoxide
proton and the ether or an intermolecular hydrogen bond between
the peptide and the ether. Whatever the explanation, this result
clearly indicates random facial addition of hydroperoxide to this
substrate.
In conclusion, it has been shown that for substrate 1d the first
stage of the Weitz–Scheffer and Julia´–Colonna oxidations is the
reversible conjugate addition of peroxide anion, which is sub-
stantially accelerated by the presence of PLL.
We thank the Swiss National Research Fund, the EPSRC and
DTI (Micrograms to Multikilograms LINK Scheme), the BBSRC,
and Novartis for funding to EC, RWF and WH.
Notes and references
1 E. Weitz and A. Scheffer, Ber., 1921, 54, 2327.
2 S. Banfi, S. Colonna, H. Molinari, S. Julia´ and J. Guixer, Tetrahedron,
1984, 40, 5207.
3 T. Geller and S. M. Roberts, J. Chem. Soc., Perkin Trans. 1, 1999, 1397
and ref. 3 therein.
4 R. Takagi, A. Shiraki, T. Manabe, S. Kojima and K. Ohkata, Chem.
Lett., 2000, 366, ref. 12; R. W. Flood, PhD Thesis, University of
Liverpool, 2001.
5 C. A. Bunton and G. J. Minkoff, J. Chem. Soc., 1949, 665.
6 R. E. Liutz and W. B. Black, J. Am. Chem. Soc., 1953, 75, 5990; S. Patai,
Z. Rappoport, in The Chemistry of the Alkenes, (S. Patai, ed.),
Wiley-Interscience, London, 1964, p. 469.
Scheme 2 Stereochemistry of alkene isomerisation. a and b refer to the
orientation of the hydroperoxy moiety.
7 H. O. House and R. S. Ro, J. Am. Chem. Soc., 1958, 80, 2428.
8 Physical data for (2R,3)-epoxyphenylpropanone 3c lmax (film) 1689; dH
(CDCl3) 2.89 (1 H, dd, J 6.5, 2.5), 3.12 (1 H, dd, J 6.5, 4.5), 4.23 (1 H, dd,
J 4.5, 2.5) 7.49–8.08 (5 H, m). dC(CDCl3) 47.5, 51.2, 128.4, 128.9, 133.9,
135.5, 194.7. Found: C, 72.7; H, 5.4; C9H8O2 requires C, 72.95; H,
5.45%. [a]D22 160.8 (c 0.74, CH2Cl2), mp 64–66 uC. An authentic sample
was obtained in six steps from styryl alcohol using Sharpless asymmetric
epoxidation in the first step.
9 Prepared by hydroalumination of 1-phenylprop-2-yn-1-ol with LiAlH4
in THF, quenching the reaction with D2O (M. J. Kang, J.-S. Jang and
S.-G. Lee, Tetrahedron Lett., 1995, 36, 8829 ) and Dess–Martin
periodinane oxidation (70% overall yield). The ratio (Z)-isomer 1d : (E)-
isomer 1e was 95–90 : 5–10.
There are two distinguishable mechanisms for alkene isomerisa-
tion/epoxidation. The first possibility involves face selective
addition (re or si, Scheme 2), followed by random elimination or
epoxidation while the second invokes random facial addition,
followed by conformationally controlled elimination of hydro-
peroxide or hydroxyl from the a- or b-conformers. The (2R)-
stereochemistry of the epoxides 3d, 3e, in the Julia´–Colonna
reaction, shows that epoxide formation occurs from the b-(R)-2d
and b-(S)-2e conformers, when the O–O-bond is optimally
orientated.
Evidence to distinguish between the two mechanisms was
provided by epoxidation of an a,b-unsaturated ester. These are
usually inert under triphasic Julia´–Colonna epoxidation conditions,
but when tert-butyl (E)-3-benzyloxybut-2-enoate 4 was subjected
to biphasic epoxidation conditions the optically active epoxide 5
(20%, w 85% ee)15 and the hydroperoxide 6 (58% yield) were
isolated (Scheme 3). The hydroperoxide 6 was reduced with
triphenylphosphine to the corresponding alcohol, and converted
into (S)-Mosher’s esters. 1H-NMR spectra showed the presence of
equal amounts of two diastereoisomers, indicating that the alcohol
was racemic. The hydroperoxide 6 did not form racemic or
optically active epoxide 5 when re-subjected to the Julia´–Colonna
conditions.
10 Y. Apeloig, M. Karni and Z. Rappoport, J. Am. Chem. Soc., 1983, 105,
2784.
11 V. K. Aggarwal and A. Mereu, Chem. Commun., 1999, 2311.
12 The isomerization was followed by 1H NMR spectroscopy.
13 In a separate experiment it was confirmed that epimerisation of the
epoxides 3d, 3e by deprotonation/reprotonation did not occur under the
reaction conditions.
14 If every addition occurs with full bond rotation, then the rate of
isomerisation is half the rate of addition. Clearly addition–elimination
may occur without bond rotation, particularly when the peroxy-enolate 2
is bound to poly-leucine. Therefore the rate of isomerisation is 50% of the
lowest possible rate of addition.
15 The absolute configuration is inferred to be (2R,3S) by analogy with
the results from Julia´–Colonna oxidation of a wide variety of a,b-
unsaturated ketones.
16 The role of the peptide in this mechanism is discussed in the following
paper, in this issue (DOI: 10.1039/b404390c).
We interpret this result as follows. Ester enolates are less
stable than ketone enolates, hence the peroxy-enolate was
protonated before b-elimination or epoxide formation. Moreover,
C h e m . C o m m u n . , 2 0 0 4 , 2 0 1 6 – 2 0 1 7
2 0 1 7