Efficient control of the diastereoselectivity and regioselectivity in the singlet-oxygen ene reaction of chiral oxazolidine-substituted alkenes by a remote urea NH functionality: Comparison with dimethyldioxirane and m-chloroperbenzoic acid epoxidations

Article abstract of DOI:10.1021/ja010463k

The singlet-oxygen ene reaction and the epoxidation by DMD of chiral oxazolidine-substituted alkenes, equipped with a free urea NH functionality and a conformationally fixed double bond, proceed in high like diastereoselectivity (up to > 95:5); also a high regioselectivity was found for the ¹O₂ ene reaction. Capping of the free NH functionality by methylation erases this like selectivity for both oxidants and significantly reduces the regioselectivity in the ene reaction. These data demonstrate effective hydrogen bonding between the remote urea NH functionality and the oxidant that favors the like attack on the C-C double bond. For ¹O₂, the hydrogen bonding in the exciplex results in preferred hydrogen abstraction from the alkyl group cis to the directing urea functionality.

Full text of DOI:10.1021/ja010463k

7228
J. Am. Chem. Soc. 2001, 123, 7228-7232
Efficient Control of the Diastereoselectivity and Regioselectivity in
the Singlet-Oxygen Ene Reaction of Chiral Oxazolidine-Substituted
Alkenes by a Remote Urea NH Functionality: Comparison with
Dimethyldioxirane and m-Chloroperbenzoic Acid Epoxidations
Waldemar Adam,† Karl Peters,§ Eva-Maria Peters,§ and Simon B. Schambony*,†
Contribution from the Institut fu¨r Organische Chemie, UniVersita¨t Wu¨rzburg,
D-97074 Wu¨rzburg, Germany
ReceiVed February 20, 2001
Abstract: The singlet-oxygen ene reaction and the epoxidation by DMD of chiral oxazolidine-substituted alkenes,
equipped with a free urea NH functionality and a conformationally fixed double bond, proceed in high like
diastereoselectivity (up to >95:5); also a high regioselectivity was found for the 1O2 ene reaction. Capping of
the free NH functionality by methylation erases this like selectivity for both oxidants and significantly reduces
the regioselectivity in the ene reaction. These data demonstrate effective hydrogen bonding between the remote
urea NH functionality and the oxidant that favors the like attack on the C-C double bond. For 1O2, the hydrogen
bonding in the exciplex results in preferred hydrogen abstraction from the alkyl group cis to the directing urea
functionality.
Introduction
amino7 groups has recently received much attention, and high
stereocontrol was achieved through the advent of beneficial
hydrogen bonding between the conformationally aligned allylic
functionality and the singlet-oxygen enophile. However, as
indicated by the low diastereoselectivities obtained in the
photooxygenations of chiral homoallylic alcohols,8 it is essential
that the directing hydrogen-bonding functionality is positioned
in the proximate neighborhood of the double bond and fixed
on one π face of the C-C double bond by conformational
constraint.
Similar directing effects were also found to operate in the
epoxidations of olefins by dimethyldioxirane (DMD)9 and
m-chloroperbenzoic acid (mCPBA).10 Thus, allylic hydrogen-
bonding donors such as the hydroxy and amino groups, which
are favorably aligned by allylic strain, may steer the incoming
The ene reaction of singlet oxygen (1O2) with alkenes
constitutes a convenient route to allylic hydroperoxides and,
after reduction, allylic alcohols; the latter are versatile building
blocks with synthetic utility.1 Much effort has been expended
to achieve control of the regio- as well as the diastereoselectivity
for such oxyfunctionalizations.2
1
The regioselectivity of the O2 ene reaction, that is, the site
of hydrogen abstraction, is governed by several empirical
facts: The so-called cis effect3 predicts that hydrogen abstraction
occurs predominantly from the higher substituted side of the
double bond. In contrast, the gem effect4 and large-group
nonbonding effect5 entail predominant abstraction at the geminal
position, promoted by electron-withdrawing and bulky substit-
uents.
(6) (a) Adam, W.; Nestler, B. J. Am. Chem. Soc. 1992, 114, 6549-
6550. (b) Adam, W.; Nestler, B. J. Am. Chem. Soc. 1993, 115, 5041-
5049. (c) Adam, W.; Wirth, T. Acc. Chem. Res. 1999, 32, 703-710.
(7) (a) Adam, W.; Bru¨nker, H.-G. J. Am. Chem. Soc. 1993, 115, 3008-
3009. (b) Bru¨nker, H.-G.; Adam, W. J. Am. Chem. Soc. 1995, 117, 3976-
3982. (c) Murray, R. W.; Singh, M.; Williams, B. L.; Moncrief, H. J. Org.
Chem. 1996, 61, 1830-1841.
(8) (a) Adam, W.; Bru¨nker, H.-G.; Kumar, A. S.; Peters, E.-M.; Peters,
K.; Schneider, U.; von Schnering, H. G. J. Am. Chem. Soc. 1996, 118,
1899-1905. (b) Adam, W.; Saha-Mo¨ller, C. R.; Schambony, S. B.; Schmid,
K. S.; Wirth, T. Photochem. Photobiol. 1999, 70, 476-483.
In regard to the diastereoselectivity, singlet oxygen as a linear
two-atomic molecule cannot itself transmit steric effects, so that
diastereoselection may only arise through substrate control. In
this context, the directing propensity of allylic hydroxy6 and
† Institut fu¨r Organische Chemie. E-mail: adam@chemie.uni-wuerzburg.de.
Fax: +49(0)931/8884756. Internet: http://www-organik.chemie.uni-
wuerzburg.de.
§ Max-Planck-Institut fu¨r Festko¨rperforschung, Heisenbergstrasse 1,
D-70569 Stuttgart, Germany.
(1) (a) Adam, W.; Richter, M. J. Acc. Chem. Res. 1994, 27, 57-62. (b)
Prein, M.; Adam, W. Angew. Chem., Int. Ed. Engl. 1996, 35, 477-494.
(2) (a) Clennan, E. L. Tetrahedron 2000, 56, 9151-9179. (b) Stratakis,
M.; Orfanopoulos, M. Tetrahedron 2000, 56, 1595-1615.
(3) (a) Orfanopoulos, M.; Grdina, M. B.; Stephenson, L. M. J. Am. Chem.
Soc. 1979, 101, 1, 275-276. (b) Schulte-Elte, K. H.; Muller, B. L.;
Rautenstrauch, B. HelV. Chim. Acta 1978, 61, 2777-2783. (c) Stratakis,
M.; Orfanopoulos, M.; Chen, J. S.; Foote, C. S. Tetrahedron Lett. 1996,
37, 4105-4108.
(4) (a) Orfanopoulos, M.; Foote, C. S. Tetrahedron Lett. 1985, 26, 5991-
5994. (b) Adam, W.; Griesbeck, A. Synthesis 1986, 1050-1052. (c)
Akasaka, T.; Takeuchi, K.; Misawa, Y.; Ando, W. Heterocycles 1989, 28,
445-451.
(5) (a) Orfanopoulos, M.; Stratakis, M.; Elemes, Y. Tetrahedron Lett.
1989, 30, 4875-4878. (b) Orfanopoulos, M.; Stratakis, M.; Elemes, Y. J.
Am. Chem. Soc. 1990, 112, 6417-6419.
(9) (a) Adam, W.; Smerz, A. K. J. Org. Chem. 1996, 61, 3506-3510.
(b) Adam, W.; Smerz, A. K. Bull. Soc. Chim Belg. 1996, 105, 581-599.
(c) Adam, W.; Mitchell, C. M.; Paredes, R.; Smerz, A. K.; Veloza, L. A.
Liebigs Ann. Chem./Recl. 1997, 1365-1369.
(10) (a) Henbest, H. B.; Wilson, R. A. L. J. Chem. Soc. 1957, 1958-
1965. (b) Chamberlain, P.; Roberts, M. L.; Whitham, G. H. J. Chem. Soc.
B 1970, 1374-1381. (c) Berti, G. In Topics in Stereochemistry; Allinger,
N. L., Eliel, E. L., Eds.; Wiley: New York, 1973; Vol. 7, pp 93-251. (d)
Sharpless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12, 63-73.
(e) Rossiter, B. E.; Verhoeven, T. R.; Sharpless, K. B. Tetrahedron Lett.
1979, 4733-4736. (f) Kogen, H.; Nishi, T. J. Chem. Soc., Chem. Commun.
1987, 311-312. (g) Rao, A. S. In ComprehensiVe Organic Synthesis; Ley,
S. V., Ed.; Pergamon Press: Oxford, 1991; Vol. 7, pp 357-387. (h)
Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93, 1307-
1370. (i) Asensio, G.; Mello, R.; Boix-Bernardini, C.; Gonza´lez-Nu´n˜ez,
M. E.; Castellano, G. J. Org. Chem. 1995, 60, 3692-3699.
10.1021/ja010463k CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/07/2001

1
Regio- and DiastereoselectiVity in the O2 Ene Reaction
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7229
Table 1: Diastereo- and Regioselectivity in the Photooxygenationa of Oxazolidine-Substituted Olefins 1 (cf. Scheme 2)
regioselectivityb
diastereoselectivitiesb
entry
1
R1 R2 R3 R4
Ar
Ph
solvent
CDCl3
2
2′ 2′′
3
lk-2:
:
ul-2
lk-2′
:
ul-2′ lk-2′′
:
ul-2′′
1a
Me Me
Me Me
Me Me
Me Me
H
H
H
H
H
H
H
H
93c
96c
-
-
-
-
-
7
94:6c
85:15c
>95:5c
41:59c
-
-
-
-
-
:
2d 1a
Ph
CDCOCD3
4
<5
30
e
3
4
5f
6
7
8
1b
1c
1d
1e
p-NO2 CDCl3
>95c
70c
Me Ph
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
Et
H
H
H
H
H
H
Ph
Ph
Ph
Ph
>95
-
-
<5 >95g
:
-
:
5
5
Me Me
31 69 <5
53
:
h
:
47
5
52
48
E-1f Me Et
Z-1f Et Me
H
H
92
8
-
-
<5 >95
12
-
-
12
76
h
>95i
a Sensitizer was 5,10,15,20-tetrakis(pentafluorophenyl)porphine (TPFPP), the reactions were conducted at -10 °C; unless otherwise stated,
1
conversion and mass balance were >90%. b Determined by H NMR spectroscopy, error (5% of the stated values. c Differentiation between the
f
products 2 and 2′ is not possible when R1 ) R2)Me, because these products are identical. d Mass balance was only 74%. e p-Nitrophenyl. Conversion
was 47%. g The newly formed double bond is E-configured. h Not determined due to small amounts. i The E/Z ratio of the newly formed double
bond was 86:14.
Scheme 1a
oxidant through hydrogen bonding. Nevertheless, only a few
examples of the remote directing group are known for selective
epoxidations. The observed diastereoselectivity for homoallylic
alcohols is significantly lower than for allylic substituents due
to the less efficient alignment of the directing hydrogen-bonding
entity.7c,8a,10h,11
Recently, we reported that alkenyl-substituted oxazolidines,
similar to those that have been used as chiral auxiliaries,12,13
undergo highly diastereoselective singlet-oxygen ene reaction14
as well as epoxidations by mCPBA and DMD.15 The controlling
feature derives from a remotely positioned urea NH functional-
ity, which provides beneficial hydrogen bonding with the
oxidant. In the present contribution, we supply the full synthetic
and mechanistic details on this new remotely directed diaste-
1
reocontrol and demonstrate that the regioselectivity of the O2
ene reaction is also governed by the hydrogen-bonding directing
group of the oxazolidine.
a (i) 1. Molecular sieves (4 Å), CH2Cl2, 20 °C, 3 h. 2. ArNCO (see
Table 1), CH2Cl2, 20 °C, 16 h. (ii) H2, Lindlar catalyst, MeOH, 20 °C,
3 h.
Results
The oxazolidine-substituted alkenes 1a, b, e, f were synthe-
Scheme 2
sized by the condensation of the corresponding aldehyde with
S-phenylglycinol in analogy to the reported procedure,12g fol-
lowed by reaction with the appropriate aryl isocyanate (Scheme
1).16 The cis-disubstituted olefin 1d was obtained by partial
catalytic hydrogenation of the alkyne 1g, which was synthesized
from 2-pentynal, S-phenylglycinol, and phenylisocyanate (Scheme
1). Methylation of the urea functionality in the oxazolidine 1a
yielded the N-methylated derivative 1c. The like relative
configuration of the stereogenic centers of the oxazolidine ring
was assessed for all cases by NOE measurements.
(11) (a) Kobayashi, Y.; Uchiyama, H.; Kanbara, H.; Sato, F. J. Am. Chem.
Soc. 1985, 107, 5541-5543. (b) Kocˇovsky´, P.; Stary´, I. J. Org. Chem. 1990,
55, 3236-3243.
(12) (a) Colombo, L.; Gennari, C.; Poli, G.; Scolastico, C. Tetrahedron
Lett. 1985, 26, 5459-5462. (b) Cardani, S.; Poli, G.; Scolastico, C.; Villa,
R. Tetrahedron 1988, 44, 5929-5938. (c) Hussain, A.; Wyatt, P. B.
Tetrahedron 1993, 49, 2123-2130. (d) Colombo, L.; Di Giacomo, M.;
Brusotti, G.; Milano, E. Tetrahedron Lett. 1995, 36, 2863-2866. (e) Agami,
C.; Couty, F.; Lam, H.; Mathieu, H. Tetrahedron 1998, 54, 8783-8796.
(f) Garc´ıa-Valverde, M.; Nieto, J.; Pedrosa, R.; Vicente, M. Tetrahedron
1999, 55, 2755-2762. (g) Agami, C.; Couty, F.; Hamon, L.; Venier, O. J.
Org. Chem. 1997, 62, 2106-2112.
(13) (a) Kanemasa, S.; Suenaga, H.; Onimura, K. J. Org. Chem. 1994,
59, 6949-6954. (b) Porter, N. A.; Rosenstein, I. J.; Breyer, R. A.; Bruhnke,
J. D.; Wu, W.-X.; McPhail, A. T. J. Am. Chem. Soc. 1992, 114, 7664-
7676.
The oxazolidines 1 were photooxygenated at low temperature,
and the resulting hydroperoxides 6 were reduced in situ, usually
after complete conversion of the starting material. The reaction
sequence afforded the corresponding regioisomeric allylic
alcohols 2, 2′, and 2′′ and, in some cases, the spiro-dioxolanes
3 (Scheme 2). The product distribution and the diastereomeric
ratios of the allylic alcohols were determined from the crude
1
reaction mixture by H NMR spectroscopy and are given in
Table 1. In the case of the substrates 1a, b, d and E-and Z-1f,
the allylic alcohols 2 were formed in high like diastereoselec-
tivity and good yields in the nonpolar chloroform (entries 1, 3,
5, 7, 8). The diastereoselectivity depends strongly on the reaction
medium, as demonstrated by the significantly lower selectivity
(14) Adam, W.; Peters, K.; Peters, E.-M.; Schambony, S. B. J. Am. Chem.
Soc. 2000, 122, 7610-7611.
(15) Adam, W.; Schambony, S. B. Org. Lett. 2001, 3, 79-82.
(16) (a) Goldberg, E. P.; Nace, H. R. J. Am. Chem. Soc. 1953, 75, 6260-
6262. (b) Hatam, M.; Goerlich, J. R.; Schmutzler, R.; Gro¨ger, H.; Martens,
J. Synth. Commun. 1996, 26, 3685-3698.

7230 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Adam et al.
Scheme 3
Scheme 5a
a (i) MeI, KOH, DMSO, 20 °C, 16 h. (ii) Activated Al2O3, n-hexane,
20 °C, 16 h.
Scheme 4
Scheme 6a
Table 2: Diastereoselectivities in the DMD Epoxidation of
Oxazolidine-Substituted Olefins 1 (cf. Scheme 4)a
a (i) 1. K2CO3, CDCl3, 20 °C, 2 h. 2. PhNCO, CDCl3, 20 °C, 3 h.
diastereoselectivityb
Scheme 7a
entry
R1 R2 R3 R4
Ar
Ph
lk-4
:
ul-4
1
2
3
4
5
6
7
1a
1b
1c
1d
1e
Me Me
Me Me
Me Me
H
H
H
H
H
H
>95
>95
26
:
:
:
:
:
:
:
5
5
74
5
26
5
5
c
p-NO2
Me Ph
Et
H
H
H
H
H
H
Ph
Ph
Ph
Ph
>95
Me Me
74
>95
>95
E-1f Me Et
Z-1f Et Me
H
H
1
a Conversion and mass balance were >90%. b Determined by H
NMR spectroscopy, error (5% of the stated values. c p-Nitrophenyl.
a (i) Activated Al2O3, n-hexane, 20 °C, 16 h.
of the N-methylated derivative 1c afforded mainly the unlike
epoxide ul-4c.
of the photooxygenation of the olefin 1a in the polar acetone.
Furthermore, the diastereoselectivity is dramatically reduced for
the olefin 1e that bears no abstractable allylic hydrogen atom
positioned cis to the oxazolidine moiety (entry 6). A slight unlike
preference was found for the N-methylated olefin 1c.
Also the epoxidation of the olefins 1a-c, e by m-chloro-
perbenzoic acid (mCPBA) shows the same trend in diastereo-
selectivity as DMD. Thus, like selectivity for 1a, b, e and unlike
for 1c are observed, but these diastereoselectivities are lower
than the ones found for the corresponding DMD epoxidations.
These results15 are given in the Supporting Information section.
Besides the high diastereoselectivity in the photooxygenation
of olefins 1, the site of hydrogen abstraction (regioselectivity)
is also usually well defined. In the ene reaction of olefins 1a,
b, d, e and E- and Z-1f, hydrogen abstraction occurs predomi-
nantly from one of the methyl groups and only a small amount
from the aminal position. The hydroperoxides derived from the
latter H abstraction mode cyclize in situ to the dioxolanes 3
(Scheme 3). Here, the N-methylated derivative 1c exhibits again
a lower selectivity; as much as 30% of the dioxolanes 3c was
found in the reaction mixture (entry 4). For the isomeric olefins
E- and Z-1f, the ene reactivity of the two geminal alkyl groups
of the double bond (twix versus twin reactivity17) differs; the
abstraction occurs in both cases predominantly at the twix alkyl
group, that is, the one located cis to the lone oxazolidine
substituent (entries 7 and 8).
Configurational Assignments
To assess the relative configuration of the allylic alcohol
lk-2a, single crystals were grown and submitted to X-ray
analysis (cf. Figure 1 in the Supporting Information).18 The like
configuration of the corresponding epoxide lk-4a was established
by base-catalyzed rearrangement to lk-2a (Scheme 5).19 Fur-
thermore, epoxide lk-4a was methylated to yield the epoxide
lk-4c (Scheme 5), and thus, also its relative like configuration
was established. The epoxides 4d, e, and E-4f were synthesized
independently as epimeric mixtures from the corresponding
racemic epoxy aldehydes 5, and their absolute configuration
was assigned by independent synthesis of the unlike isomers
from the optically active epoxy aldehydes (2R)-5d-f (Scheme
6). The like configuration of the allylic alcohol lk-2d was
determined by base-catalyzed rearrangement of epoxide lk-4d
(Scheme 7).
The olefins 1 were epoxidized with dimethyldioxirane
(DMD) to yield the corresponding epoxides 4 (Scheme 4). The
diastereomeric ratios of the epoxides were determined by
1H NMR spectroscopy from the crude product mixture and are
given in Table 2. The epoxidations of the olefins 1a, b, d, and
E- and Z-1f proceeded highly like diastereoselectively; not
even traces of the unlike diastereomer could be detected in the
spectra (entries 1, 2, 4, 6, 7). The reaction of the cis-dimethyl-
substituted olefin 1e is less selective, but still a significant like
preference was observed (entry 5). In contrast, the epoxidation
Discussion
The present results provide compelling evidence for an
attractive hydrogen-bonded interaction between the attacking
(18) Peters, K.; Peters, E.-M.; Adam, W.; Schambony, S. B. Z.
Kristallogr. 2000, 215, 213-214.
(19) (a) Joshi, V. S.; Damodaran, N. P.; Dev, S. Tetrahedron 1968, 24,
5817-5830. (b) Va´zquez, J. T.; Chang, M.; Nakanishi, K. J. Org. Chem.
1988, 53, 4797-4800.
(17) Adam, W.; Bottke, N.; Krebs, O. J. Am. Chem. Soc. 2000, 122,
6791-6792.

1
Regio- and DiastereoselectiVity in the O2 Ene Reaction
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7231
Scheme 8
Scheme 10
olefins 1a, b, d-f by such lone abstraction (entries 1-3 and
5-7 in Table 1). The unfavorable lone abstraction is due to
fixation of the oxazolidine moiety by the hydrogen bonding in
the energy-favored like exciplex, in which the C-H bond of
the aminal (lone) hydrogen atom is in the nodal plane of the
C-C double bond (cf. Scheme 8). Since a perpendicular
arrangement of this C-H bond is necessary for hydrogen
abstraction,20 the aminal hydrogen atom at the lone site is not
available for abstraction.
For the methylated derivative 1c (entry 4 in Table 1),
hydrogen bonding is prevented, and the oxazolidine moiety
rotates quite freely; consequently, the dioxolane 3c is formed
to a significant extent (30%). It is important to emphasize that
the observed regioselectivity does not arise from the so-called
cis effect, which predicts that hydrogen abstraction takes place
from the higher substituted side of the double bond.3 If this
were the case, a similar regioselectivity would be expected for
the olefin 1a with a nonmethylated urea functionality and its
methylated derivative 1c, which is not the case (entries 1 and 4
in Table 1).
Since the singlet-oxygen ene reaction involves a perepoxide-
like geometry (Schemes 8 and 9), it was of mechanistic interest
to compare the photooxygenation results of the oxazolidines 1
with those of the DMD epoxidations. Indeed, both oxyfunc-
tionalizations display strikingly similar trends, in that the high
diastereoselectivity observed in the singlet-oxygen reaction of
substrates 1a, b, d, f is found also for the DMD epoxidations
(entries 1, 2, 4, 6, and 7 in Table 2). Here again, hydrogen
bonding between the urea NH functionality and the DMD
oxidant operates in the like transition state (Scheme 10). This
favorable bonding reduces the energy of the transition state,
which leads to the high (>95:5) diastereoselectivities given in
Table 2. Methylation of the NH group as in olefin 1c (entry 3)
prevents such hydrogen bonding, and instead, the urea group
imposes a steric constraint such that a moderate (26:74) unlike
selectivity is observed.
The reason for the relatively low selectivity observed in the
case of the cis-dimethyl-substituted olefin 1e (entry 5) may be
explained in terms of less effective fixation of the double bond.
Calculations (B3LYP/6-31G*) reveal two possible ground-state
conformers for the model compounds 7a and 7e (Scheme 11),
of which conformer A forms a hydrogen bond to the dioxirane
but conformer B does not. In the case of the olefin 7a, conformer
A is nearly exclusively populated (structure A is favored by
3.2 kcal/mol), and expectedly, only the like epoxide should
result. In contrast, for the substrate 7e both conformers are nearly
equally populated (structure B is favored by 0.36 kcal/mol),
and both may react with the oxidant equally likely. Applied to
the experimental results (entries 1 and 5 in Table 2), the lower
selectivity in the epoxidation of oxazolidine 1e compared to 1a
arises from the superposition of the like-selective attack on
conformer A and a completely unselective attack on conformer
B (Scheme 11). Despite this less effective fixation of the double
Scheme 9
singlet-oxygen enophile and the NH group of the urea func-
tionality in the photooxygenation of the oxazolidines 1a, b, d,
f (entries 1, 3, 5, 7, and 8 in Table 1). Hydrogen bonding is
only feasible in the transition state for the like exciplex, whereas
such interaction is prevented for the unlike attack (Scheme 8).
The hydrogen bonding between the urea-NH moiety and the
negatively polarized, terminal oxygen atom of singlet oxygen
lowers the energy of the corresponding like transition state, and
therefore, the attack on this π face of the double bond is favored.
That hydrogen bonding plays an important role in the
transition state of the like exciplex is confirmed by the
dependence of the diastereoselectivity on the reaction medium,
which is exemplified for oxazolidine 1a. In acetone, a hydrogen-
bonding acceptor itself, the hydrogen bonding between the
oxazolidine substrate and the singlet-oxygen enophile is less
effective, and consequently, the diastereoselectivity of the ene
reaction in this solvent is significantly lower than in the nonpolar
chloroform (entry 1 and 2 in Table 1). Furthermore, methylation
of the NH functionality as in oxazolidine 1c prevents hydrogen
bonding, and hence, a poor diastereoselectivity is found; in fact,
the opposite (unlike) diastereomer is slightly favored (entry 4).
The urea hydrogen bonding not only controls the diastereo-
selectivity of the singlet-oxygen attack but also promotes a high
regioselectivity as demonstrated in the hydrogen abstraction of
the diastereomeric olefins E- and Z-1f. As exemplified for the
E-1f diastereomer (Scheme 9), due to the hydrogen bonding,
along this like trajectory the terminal oxygen atom of 1O2 points
toward the urea functionality. Clearly, hydrogen abstraction from
the twin ethyl group is prohibited, and instead, removal from
the twix and lone positions is expected. Thus, the geometrical
fixation in the like exciplex due to hydrogen bonding promotes
abstraction predominantly from the allylic substituent located
cis to the oxazolidine (twix position); thus, the twix-methyl group
in substrate E-1f and the twix-ethyl group for the olefin Z-1f
are ene-active (Scheme 9).
Furthermore, the hydrogen bonding in the exciplex hinders
effectively the possible abstraction of the aminal (lone) hydrogen
atom, as witnessed by the minor amounts of the dioxolanes 3
(cf. Scheme 3), which are formed in the ene reaction of the
(20) (a) Herz, W.; Juo, R. J. Org. Chem. 1985, 50, 618-627. (b) Poon,
T. H. W.; Pringle, K.; Foote, C. S. J. Am. Chem. Soc. 1995, 117, 7611-
7618.

7232 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Adam et al.
Scheme 11
Unexpected is the fact that the diastereoselectivities for the
mCPBA epoxidations15 (cf. the Supporting Information) show
the same trend as DMD (Table 2), but are generally lower.
Usually mCPBA is more diastereoselective than DMD, for
example, in the epoxidations of chiral allylic alcohols.8a This
has been explained in terms of less efficient hydrogen bonding
in the DMD case, especially because the reaction medium is
more polar for this reagent (acetone versus chloroform). In the
peracid epoxidation of the oxazolidines 1, hydrogen bonding
also favors the like transition state (Scheme 10), but for the
peracid the distance between the hydrogen-bond acceptor (the
carbonyl oxygen atom)21 and the transferred oxygen atom (2.4
Å)22 is significantly larger than for the dioxirane (1.5 Å)23 and
for the singlet oxygen (1.2 Å).24 Since the distance between
the C-C double bond and the hydrogen atom of the NH group
is only 2.6 Å (calculated for the model compound 7a),
presumably the peracid does not fit as well into the cavity
formed by the alkenyl group and the urea functionality.
Therefore, the hydrogen bonding is less effective for mCPBA
Scheme 12
1
than for DMD or O2, and consequently, a lower diastereose-
lectivity is found for the peracid oxidant.
In summary, the present results clearly demonstrate that the
remotely located urea-NH functionality of chiral carbamoyl-
substituted oxazolidines is highly effective in controlling the
diastereoselectivity, as well as the regioselectivity, of the singlet-
oxygen ene reaction. The controlling feature is the efficient
hydrogen bonding between the urea functionality and the
oxidant, which besides singlet oxygen also operates well in the
epoxidation by dimethyldioxirane and m-chloroperbenzoic acid.
The control of diastereoselectivity through hydrogen bonding
is well established in the oxyfunctionalization by these oxi-
dants;6,7,9,10 however, such high directivity has hitherto not been
observed by a hydrogen-bonding functionality that far away
from the C-C double bond of the substrate. Furthermore, it
should be emphasized that the present oxazolidine chiral
auxiliary exerts better diastereoselectivity control for DMD
than mCPBA epoxidation reactions, an unprecedented fact. Since
the oxazolidine ring system is usually readily cleaved by
hydrolysis,12f,g,25 the present methodology offers promising
prospects in the preparation of optically active building blocks
for asymmetric synthesis.
bond in oxazolidine 1e, the moderate (74:26) like diastereose-
lectivity in its epoxidation clearly shows that hydrogen bonding
also operates for this substrate.
This is mechanistically significant in view of the completely
unselective ene reaction of the substrate 1e with singlet oxygen
(entry 6 in Table 1). Thus, the two possible regioisomeric allylic
alcohols 2e′ and 2e′′ were formed each as ca. 1:1 diastereomeric
mixture (entry 6 in Table 1), although the substrate bears a free
NH functionality available for hydrogen bonding. This remark-
able lack of selectivity may be accounted for in terms of two
Acknowledgment. The generous financial support by the
Deutsche Forschungsgemeinschaft (Schwerpunktprogramm “Per-
oxidchemie: Mechanistische und pra¨parative Aspekte des
Sauerstofftransfers”) and the Fonds der Chemischen Industrie
(Doctoral Fellowship 1998-2000 for S.B.S.) is gratefully
appreciated. We dedicate this work to Prof. Dr. A. M. Braun
(Karlsruhe), on the occasion of his 60th birthday.
1
different singlet-oxygen attacks. When the O2 approaches the
double bond with the terminal oxygen located on the side of
the urea functionality (path a in Scheme 12), good hydrogen
bonding is expected; however, hydrogen abstraction from the
allylic methyl groups is not possible, because they point away
Supporting Information Available: Complete experimental
procedures (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
1
from the abstracting terminal oxygen atom of O2 in this like
exciplex. The only allylic hydrogen on the same side of the
terminal oxygen atom is the one at the aminal position.
Nevertheless, no abstraction occurs from this site due to fixation
of the oxazolidine ring by hydrogen bonding, which places the
aminal hydrogen atom in the nodal plane of the π system.
Instead, the exciplex dissociates, and no ene reaction occurs.
When the 1O2 attacks along the approach in path b (Scheme
12), that is, the terminal oxygen atom points toward the two
allylic methyl groups, the NH functionality is too far away for
hydrogen bonding with the singlet oxygen, and there is no
π-facial selection; therefore, low if any diastereoselectivity is
observed. The slight (69:31) preference in regioselectivity for
the formation of the allylic alcohol 2e′′ over its regioisomer
2e′ (entry 6 in Table 1) is caused by the large-group-nonbonding
effect, which favors geminal abstraction to a bulky substituent.2,5
JA010463K
(21) (a) Bach, R. D.; Este´vez, C. M.; Winter, J. E.; Glukhovtsev, M. N.
J. Am. Chem. Soc. 1998, 120, 680-685. (b) Adam, W.; Bach, R.;
Dmitrenko, O.; Saha-Mo¨ller, C. R. J. Org. Chem. 2000, 65, 6715-6728.
(22) Plesnicar, B.; Tasevski, M.; Azman, A. J. Am. Chem. Soc. 1978,
100, 743-746.
(23) Suenram, R. D.; Lovas, F. J. J. Am. Chem. Soc. 1978, 100, 5117-
5122.
(24) Greenwood, N. N.; Earnshaw, A. Chemie der Elemente; VCH:
Weinheim, 1990; p 784.
(25) (a) Garner, P.; Park, J. M.; Malecki, E. J. Org. Chem. 1988, 53,
4395-4398. (b) Dondoni, A.; Fantin, G.; Fogagnolo, M.; Pedrini, P. J. Org.
Chem. 1990, 55, 1439-1446. (c) Wu, M.-J.; Pridgen, L. N. J. Org. Chem.
1991, 56, 1340-1344. (d) Fleischhacker, W.; Riedl, T.; Vo¨llenkle, H.; Noe,
C. R. Arch. Pharm. 1996, 329, 41-49.