Stereochemistry in the ene reactions of singlet oxygen and triazolinediones with allylic alcohols. A mechanistic comparison

Article abstract of DOI:10.1021/jo990258b

The ene reaction of singlet oxygen with (Z)-4-methylpent-3-en-2-ol- 2,5,5,5-d₄ (1-OH-d4) in nonpolar solvents exhibits a 90% threo diastereoselectivity in the adduct derived from the major syn perepoxide intermediate, but also a moderate threo diastereoselectivity in the adduct derived from the minor anti perepoxide. Photooxygenation of 2,4-dimethylpent- 3-en-2-ol (2) exhibits a significant solvent dependence in the syn/anti methyl stereoselectivity, with nonpolar solvents promoting syn methyl reactivity, while polar solvents promote anti methyl reactivity. These results are in agreement with a steering effect between hydroxyl and singlet oxygen in the rate-determining step of the reaction. N-Phenyltriazolinedione addition to the chiral allylic alcohol 4-methylpent-3-en-2-ol (1-OH) is highly threo diastereoselective in nonpolar solvents, with a solvent dependent variation in the threo/erythro ene products. On the other hand, the nonfunctionalized chiral alkene 2,4-dimethyl-2-hexene (1-Et) exhibits poor diastereoselectivity. Reaction of PTAD with 1-OH-d4 in nonpolar solvents, exhibits a significant threo diastereoselectivity from the syn aziridinium imide intermediate, and a moderate threo diastereoselectivity from the anti intermediate. These results are consonant with a steering effect between the hydroxyl and the electrophile, as proposed in the case of singlet oxygen addition to allylic alcohols 1-OH and 2. In contrast to the analogous ¹O₂ ene reaction, a solvent independent ratio syn/anti ~50/50 was found in the addition of MTAD to 2. The intermolecular kinetic isotope effect in the reaction of 2 with MTAD (k(H)/k(D) = 1.15 ± 0.02), is consistent with formation of the intermediate in fast step, indicative that the steering effect during the formation of aziridium imide is not important in the reaction kinetics. This energetic profile is in contrast to triazolinedione addition to the secondary allylic alcohol 1-OH, where the high threo selectivity and the slight inverse kinetic isotope effect of k(H)/k(D) = 0.98 ± 0.02 are consonant with the formation of the intermediate in the rate- determining step. An explanation for the increased reactivity of the syn methyl in the addition of MTAD to 2 (~50%) is offered.

Full text of DOI:10.1021/jo990258b

4130
J . Org. Chem. 1999, 64, 4130-4139
Ster eoch em istr y in th e En e Rea ction s of Sin glet Oxygen a n d
Tr ia zolin ed ion es w ith Allylic Alcoh ols. A Mech a n istic Com p a r ison
Georgios Vassilikogiannakis,† Manolis Stratakis,*,‡,§ Michael Orfanopoulos,*,† and
Christopher S. Foote*,§
Department of Natural Sciences, University of Cyprus, 1678 Nicosia, Cyprus, Department of Chemistry,
University of Crete, Iraklion 71409, Greece, and Department of Chemistry and Biochemistry,
University of California, Los Angeles California 90095-1569
Received February 11, 1999
The ene reaction of singlet oxygen with (Z)-4-methylpent-3-en-2-ol-2,5,5,5-d4 (1-OH-d 4) in nonpolar
solvents exhibits a 90% threo diastereoselectivity in the adduct derived from the major syn
perepoxide intermediate, but also a moderate threo diastereoselectivity in the adduct derived from
the minor anti perepoxide. Photooxygenation of 2,4-dimethylpent-3-en-2-ol (2) exhibits a significant
solvent dependence in the syn/anti methyl stereoselectivity, with nonpolar solvents promoting syn
methyl reactivity, while polar solvents promote anti methyl reactivity. These results are in
agreement with a steering effect between hydroxyl and singlet oxygen in the rate-determining step
of the reaction. N-Phenyltriazolinedione addition to the chiral allylic alcohol 4-methylpent-3-en-
2-ol (1-OH) is highly threo diastereoselective in nonpolar solvents, with a solvent dependent
variation in the threo/erythro ene products. On the other hand, the nonfunctionalized chiral alkene
2,4-dimethyl-2-hexene (1-Et) exhibits poor diastereoselectivity. Reaction of PTAD with 1-OH-d 4
in nonpolar solvents, exhibits a significant threo diastereoselectivity from the syn aziridinium imide
intermediate, and a moderate threo diastereoselectivity from the anti intermediate. These results
are consonant with a steering effect between the hydroxyl and the electrophile, as proposed in the
1
case of singlet oxygen addition to allylic alcohols 1-OH and 2. In contrast to the analogous O2 ene
reaction, a solvent independent ratio syn/anti ∼ 50/50 was found in the addition of MTAD to 2.
The intermolecular kinetic isotope effect in the reaction of 2 with MTAD (kH/kD ) 1.15 ( 0.02), is
consistent with formation of the intermediate in fast step, indicative that the steering effect during
the formation of aziridium imide is not important in the reaction kinetics. This energetic profile is
in contrast to triazolinedione addition to the secondary allylic alcohol 1-OH, where the high threo
selectivity and the slight inverse kinetic isotope effect of kH/kD ) 0.98 ( 0.02 are consonant with
the formation of the intermediate in the rate-determining step. An explanation for the increased
reactivity of the syn methyl in the addition of MTAD to 2 (∼50%) is offered.
In tr od u ction
selectivity have been reported.5 A syn preference due to
a phenyl-directing effect has been recently recognized in
the photooxygenation of â,â-dimethylstyrene.6 Triazo-
linediones exhibit Markovnikov type selectivity.7 With
trisubstituted alkenes8 and â,â-dimethylstyrene6 a high
preference for syn addition to the more-substituted side
has also been found.
In their reactions with cis alkenes the major ene
products arise from allylic hydrogen abstraction next to
the bulkiest group.9 Furthermore, a geminal selectivity10
was observed with respect to a bulky alkyl substituent
or a functional group at allylic or vinylic position. The
degree of selectivity depends solely on the size of the
The ene reaction of singlet oxygen (1O2)1 and triazo-
linediones (RTAD, R ) methyl or phenyl)2 with alkenes
exhibits a fancinating variety of regio- and stereoselec-
tivities. Although both electrophiles react with alkenes
through structurally similar intermediates (perepoxide
and aziridinium imide or AI, respectively), the degree and
the type of stereoselectivities are often different, prima-
rily due to the different activation parameters of the
reactions and to the bulkiness of the electrophiles. In the
reaction of 1O2 with trisubstituted alkenes3 and enol
ethers,4 the more reactive side of the olefin is the more
substituted (cis effect). Only a few cases of anti “cis effect”
† University of Crete.
(3) (a) Schulte-Elte, K. H.; Muller, B. L.; Rautenstrauch, V. Helv.
Chim. Acta 1978, 61, 2777-2783. (b) Orfanopoulos, M.; Grdina, M.
B.; Stephenson, L. M. J . Am. Chem. Soc. 1979, 101, 275-276.
(4) (a) Rousseau, G.; Leperchec, P.; Conia, J . M. Tetrahedron Lett.
1977, 18, 2517-2520. (b) Lerdal, D.; Foote, C. S. Tetrahedron Lett.
1978, 19, 3227-3230.
‡ University of Cyprus.
§ UCLA.
(1) (a) Wasserman, H. H.; Murray, R. W. Singlet Oxygen; Academic
Press: New York, 1979. (b) Frimer, A. A.; Stephenson, L. M. Singlet
Oxygen, Reactions, modes and products; Frimer, A. A., Ed.; CRC
Press: Boca Raton, FL, 1985. (c) Foote, C. S.; Clennan, E. L. In Active
Oxygen in Chemistry; Foote, C. S., Valentine, J . S., Greenberg, A.,
Liebman, J . F., Eds.; Chapman and Hall: London, 1995; pp 105-140.
(d) Prein, M.; Adam, W. Angew. Chem., Int. Ed. Engl. 1996, 35, 477-
494.
(2) (a) Pirkle, W. H.; Stickler, J . C. J . Chem. Soc., Chem. Commun.
1967, 760-761. (b) Seymour, C. A.; Greene, F. D. J . Am. Chem. Soc.
1980, 102, 6384-6385. (c) Orfanopoulos, M.; Smonou, I.; Foote, C. S.
J . Am. Chem. Soc. 1990, 112, 3607-3614. (d) Radl, S. Adv. Heterocycl.
Chem. 1997, 67, 119-205.
(5) Stratakis, M.; Orfanopoulos, M. Tetrahedron Lett. 1995, 36,
4291-4294.
(6) Stratakis, M.; Orfanopoulos, M.; Foote, C. S. J . Org. Chem. 1998,
63, 1315-1318.
(7) Cheng, C.-C.; Seymour, C. A.; Petti, M. A.; Greene, F. D.; Blount,
J . F. J . Org. Chem. 1984, 49, 2910-2916.
(8) Elemes, Y. Unpublished results.
(9) (a) Orfanopoulos, M.; Stratakis, M.; Elemes, Y. Tetrahedron Lett.
1989, 30, 4875-4878. (b) Elemes, Y.; Stratakis, M.; Orfanopoulos, M.
Tetrahedron Lett. 1989, 30, 6903-6906.
10.1021/jo990258b CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/12/1999

Stereochemistry in Ene Reactions
J . Org. Chem., Vol. 64, No. 11, 1999 4131
of electron-poor allylic alcohols,19 however, the threo
selectivity was attributed to stereoelectronic effects,
rather to intramolecular hydrogen bonding.
Sch em e 1. Dia ster eoselectivity in th e Rea ction of
1O2 w ith Ch ir a l F u n ction a lized Alk en es
On the other hand, when the allylic substituent does
not coordinate to oxygen (e.g., carboxylic acid derivatives,
sulfones, halides, etc.)20 or the reaction favors the forma-
tion of the erythro adduct, electronic repulsions between
oxygen and functionality were proposed as the factor
responsible for that dramatic change in diastereoselec-
tivity. Similarly, oxygen-functionality electronic repul-
sions were postulated to control the regioselectivity in
the photooxygenation of alkenes bearing an electron-
withdrawing group at the â-position.21 A hydroxyl-singlet
oxygen steering effect has been also reported to control
the π-facial selectivity in the [4 + 2] cycloaddition of 1O2
with chiral naphthyl alcohols22 and the reaction pathways
in the photooxygenation of 1-(alkoxymethyl)-2-aryl-1-tert-
butyl-2-methoxyethylene.23
Although there are many examples of diastereoselec-
tive addition of singlet oxygen to alkenes and dienes,
there is lack of information concerning the analogous
triazolinedione reactions. A few examples have been
reported so far in the [4 + 2] cycloaddition of MTAD to
chiral dienols24 and the 2,2-dimethyloxazolidine deriva-
tives of sorbic acid.25
In the photooxygenation of substrate 1-OH, two inter-
mediates, syn and anti, can be formed. There is lack of
information about the contribution of each intermediate
to the observed diastereoselectivity. We examined the
diastereoselectivity resulting from abstraction of allylic
hydrogen from the syn and anti methyls of 1-OH by
specific labeling of the syn methyl. Furthermore, to test
accurately the validity of the proposed hydroxyl steering
effect, we studied the syn/anti methyl reactivity in the
photooxygenation of a tertiary allylic alcohol. Since
triazolinediones have similar chemical reactivity to sin-
glet oxygen, we also examined for comparison the ene
stereoselectivity and diastereoselectivity in the addition
of triazolinediones to secondary and tertiary allylic
alcohols, and the energy profile of these reactions as well.
electrophile, with triazolinediones affording more highly
regioselective reactions. These results were rationalized
in terms of nonbonded interactions.10b,11
With alkenes bearing an electronwithdrawing group
at the â-position a high degree of geminal selectivity has
1
been observed for O2 additions,12 while various selectivi-
ties are observed in the few examples studied so far of
PTAD addition.13 It was found from primary and second-
ary isotope effects that the energy profiles in the ene
reactions of the two electrophiles with R,â-unsaturated
esters are different.14 Also, in their reactions with allyl-
silanes15 a mechanistic dichotomy occurs, with PTAD
1
giving cis ene products but O2 leading to trans products.
Furthermore, in the photooxygenations of chiral alk-
enes 1-X (Scheme 1), coordination of 1O2 to hydroxyl16 or
amino17 functionalities present at the allylic position
leads to highly threo diastereoselective ene reactions.
Although for 1-X a diastereoselective reaction would be
expected due to the chirality in the adjacent carbon to
the double bond, the small difference in steric hindrance18
(methyl versus hydroxyl or amino group) could not by
itself be responsible for the increased diastereomeric
excess (>90%). It was suggested that in the intermediate
exciplex which has the structural requirements of a
perepoxide, oxygen coordinates to hydroxyl (steering
effect). The favorable interaction occurs in a transition
state which has minimized the 1,3-allylic strain and leads
to threo allylic hydroperoxides. In the photooxygenation
Resu lts
Ster eoselectivity a n d Dia ster eoselectivity in th e
En e Rea ction s of Sin glet Oxygen a n d P TAD w ith
Ch ir a l Secon d a r y Allylic Alcoh ols. Although the
steering effect between hydroxyl and oxygen has been
proposed by Adam’s group in the photooxygenation of
chiral allylic alcohols and amines,16,17 they postulated
that reaction occurs on the more-substituted side of the
olefin. To test this hypothesis we examined the diaste-
reoselectivity of the regioselectivity of allylic alcohol
(10) (a) Clennan, E. L.; Chen, X. J . Org. Chem. 1988, 53, 3124-
3126. (b) Orfanopoulos, M.; Stratakis, M.; Elemes, Y. J . Am. Chem.
Soc. 1990, 112, 6417-6419. (c) Orfanopoulos, M.; Elemes, Y.; Stratakis,
M. Tetrahedron Lett. 1990, 31, 5775-5778. (d) Clennan, E. L.; Chen,
X.; Koola, J . J . J . Am. Chem. Soc. 1990, 112, 5193-5199. (e) Clennan,
E. L.; Koola, J . J .; Oolman, K. A. Tetrahedron Lett. 1990, 31, 6759-
6762.
(11) Orfanopoulos, M.; Stratakis, M.; Elemes, Y.; J ensen, F. J . Am.
Chem. Soc. 1991, 113, 3180-3181.
(19) Adam, W.; Renze, J .; Wirth, T. J . Org. Chem. 1998, 63, 226-
227.
(12) Adam W.; Richter, M. J . Tetrahedron Lett. 1993, 34, 8423-
8426 and references cited therein.
(20) (a) Linker, T.; Fro¨hlich, L. J . Am. Chem. Soc. 1995, 117, 2694-
2697. (b) Adam, W.; Brunker, H. G.; Kumar, A. S.; Peters, E.-M.; Peters,
K.; Schneider, U.; von Schering, H. G. J . Am. Chem. Soc. 1996, 118,
1899-1905.
(21) Stratakis, M.; Orfanopoulos M. Tetrahedron Lett. 1997, 38,
1067-1070.
(22) Adam, W.; Peters, E.-M.; Peters, K.; Prein, M.; von Schering
H. G. J . Am. Chem. Soc. 1995, 117, 6686-6690.
(23) Matsumoto, M.; Kobayashi, H.; Matsubara, J .; Watanabe, N.;
Yamashita, S.; Oguma, D.; Kitano, Y.; Ikawa, H. Tetrahedron Lett.
1996, 37, 397-400.
(13) (a) Hoye, T. R.; Bottorff, K. J .; Caruso, A. J .; Dellaria, J . F.; J .
Org. Chem. 1980, 45, 5, 4287-4292. (b) Akasaka, T.; Misawa, Y.; Goto,
M.; Ando, W. Tetrahedron 1989, 45, 6657-6666.
(14) Elemes, Y.; Foote, C. S. J . Am. Chem. Soc. 1992, 114, 6044-
6050.
(15) (a) Dubac, J .; Laporterie, A. Chem. Rev. 1987, 87, 319-334.
(b) Adam, W.; Schwarm, M. J . Org. Chem. 1988, 53, 3129-3130.
(16) (a) Adam, W.; Nestler, B. J . Am. Chem. Soc. 1992, 114, 6549-
6550. (b) Ibid. 1993, 115, 5041-5049.
(17) (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.
(24) Adam, W.; Gla¨ser, J .; Peters, K.; Prein, M.; J . Am. Chem. Soc.
1995, 117, 9190-9193.
(25) Adam, W.; Gu¨thlein, M.; Peters, E.-M.; Peters, K.; Wirth, T. J .
Am. Chem. Soc. 1998, 120, 4091-4093.
(18) Kropf, H.; Reichwaldt, R. J . Chem. Res. (S) 1987, 412-414.

4132 J . Org. Chem., Vol. 64, No. 11, 1999
Vassilikogiannakis et al.
Ta ble 1. Ch a n ge in th e Th r eo/Er yth r o Ra tio of P TAD
Ad d ition to 1-X in Sever a l Solven ts
Sch em e 2. P r ep a r a tion of Allylic Alcoh ol 1-OH-d 4
entry
substrate
solvent
threo/erythro
1-OH, by performing the photooxygenation of the selec-
tively labeled allylic alcohol 1-OH-d 4, where the allylic
methyl syn to the hydroxyl was selectively labeled as
CD3. Synthesis of 1-OH-d 4, was accomplished by reduc-
tion of (Z)-4-methylpent-3-en-2-one-5,5,5-d3, prepared in
94% geometrical purity according to a literature proce-
dure (Scheme 2).26 To simplify the NMR integrations, the
methine hydrogen next to the hydroxyl was labeled as
D, by reducing the precursor ketone with LiAlD4.
In chloroform the syn/anti methyl reactivity was 84/
16 which indicates a “cis effect”, as found for nonfunc-
tionalized trisubstituted alkenes.3 Considering that the
reaction proceeds through the formation of a perepoxide
as intermediate, we define as syn-perepoxide the inter-
mediate where the oxygen is placed syn to the hydroxyl,
and as anti-perepoxide where the oxygen is placed on the
less-substituted side of the double bond (anti to the
hydroxyl). Integration of the allylic methyl resonances
affords the threo/erythro ratio from the syn-perepoxide
(D-abstraction), whereas integration of the olefinic pro-
tons affords the threo/erythro ratio from the anti-pere-
poxide (H-abstraction). The diastereoselectivity in the
photooxygenation of 1-OH-d 4 in choroform is shown in
Scheme 3.
The syn-threo/syn-erythro ratio was found to be 77/7
in chloroform, while the anti-threo/anti-erythro ratio was
13/3. When the reaction was run in acetonitrile, the syn/
anti methyl reactivity was 82/18, identical within experi-
mental error to that found in chloroform. However, the
ratio of diastereomers from each intermediate has
changed. The syn-threo/syn-erythro ratio was 63/19, and
the anti-threo/anti-erythro 13/5. These results show an
increase of the erythro adduct from both the syn and the
anti intermediates on going from the nonpolar chloroform
to the polar acetonitrile.
1
2
3
4
5
6
1-OH
1-OH
1-OH
1-OH
1-Et
benzene
CHCl3
acetone
MeOH
CHCl3
91/9
86/14
66/34
62/38
56/44a
58/42a
1-Et
acetone
a
The threo/erythro stereochemistry of the two diastereomers
was not assigned.
product, tentatively identified from decoupling experi-
ments, was the threo isomer. By irradiating the methyl
group doublet (1.24 ppm) at the chiral center which is
common for the two diastereomers, the methine hydrogen
next to hydroxyl appears as two doublets (at 4.26 ppm,
J 1 ) 5.5 Hz and at 4.22 ppm with J 2 ) 6.6 Hz),
corresponding to the two diastereomers in a ratio similar
to that measured from integration of the diastereotopic
allylic methyls at 1.75 and 1.83 ppm, respectively. The
major adduct with the higher coupling constant was
assigned the threo structure as one would expect it to
have a higher coupling constant than the erythro (see
the Newman projections below). The threo diastereomer
is represented by the most stable threo rotamer which
is expected to have a high coupling constant because of
the trans arrangement of the hydrogens. The erythro
isomer, whose two more stable rotamers have the hy-
drogens at gauche to each other, is expected to have a
lower coupling constant compared to that of the threo
isomer. We consider that in the more stable rotamers,
hydroxyl is gauche to the urazole ring due to an intramo-
lecular hydrogen bond.
To study the stereochemistry of triazolinediones addi-
tion to allylic alcohols, we first examined the dependence
of the threo/erythro diastereoselectivity on solvent polar-
ity, in the addition of PTAD to the chiral allylic alcohol
1-OH.27 The results (Table 1) were compared with the
diastereoselectivity observed in the same reaction of the
nonfunctionalized chiral alkene 1-Et, where hydroxyl has
been replaced by an ethyl group. The reaction proceeded
smoothly at 25 °C without formation of significant
byproducts.
Similar results in the diastereoselectivity of MTAD
addition to several allylic alcohols has been reported
independently by Adam and co-workers.28 The relative
configuration of the major adduct was determined to be
threo by X-ray analysis of the doubly benzoylated adduct
of MTAD with 1-OH.29
To find the syn/anti allylic methyl reactivity in the
reaction of 1-OH and assess further the threo/erythro
ratio from the syn- and anti-AI intermediates, the reac-
1
The threo/erythro ratios were measured by H NMR
integration of the appropriate peaks. For 1-Et, the only
well-defined signals were the allylic hydrogens next to
the urazole substituent. For 1-OH, the allylic methyls
and the olefinic hydrogens signals are well defined for
both isomers in CDCl3 (as NMR solvent). The major
(26) Sum, F. W.; Weiler, L. Can. J . Chem. 1979, 57, 1431-1440.
(27) Some preliminary results have been reported: Stratakis, M.;
Vassilikogiannakis, G.; Orfanopoulos, M. Tetrahedron Lett. 1998, 39,
2393-2396.
(28) Adam, W.; Nestler, B.; Pastor, A.; Wirth, T. Tetrahedron Lett.
1998, 39, 2625-2628.
(29) Peters, K.; Peters, E.-M.; Adam, W.; Nestler, B. Z. Kristallog-
raphie 1998, 213, 305-306.

Stereochemistry in Ene Reactions
J . Org. Chem., Vol. 64, No. 11, 1999 4133
Sch em e 3. P r od u cts F or m ed in th e P h otooxygen a tion of 1-OH-d 4 in Ch lor ofor m
Sch em e 4. P r od u cts F or m ed by Ad d ition of P TAD to 1-OH-d 4 in Ch lor ofor m
tion of 1-OH-d 4 with PTAD was carried out in chloroform
oxygen with two different allylic hydrogens on the same
side of the double bond, as occurs in 1-OH. Such interac-
tion is well-known to stabilize the transition state for
perepoxide formation.30
and acetone. The syn/anti methyl reactivity was 60/40
in chloroform and 55/45 in acetone. Considering that the
reaction proceeds through the formation of an aziri-
dinium imide as intermediate, we define as syn-AI the
intermediate where the PTAD is placed syn to the
hydroxyl, and as anti-AI where the PTAD is placed anti.
Integration of the allylic methyl resonances affords the
threo/erythro ratio from the syn-AI (D-abstraction), and
integration of the olefinic protons affords the threo/
erythro ratio from the anti-AI (H-abstraction). The dias-
tereoselectivity in chloroform is shown in Scheme 4.
When the reaction was run in the more polar acetone,
the syn-threo/syn-erythro ratio was 37/18, and the anti-
threo/anti-erythro 27/18. As noted in the photooxygen-
ation of 1-OH-d 4, the presence of a D geminal to hydroxyl
Synthesis of isomeric (E)- and (Z)-2,4-dimethylpent-3-
en-2-ol-5,5,5-d3 (2E-d 3 and 2Z-d 3), was accomplished
through the stereoselective preparation of (E)- and (Z)-
ethyl 3-methylbuten-2-oate-4,4,4-d3, following known
literature procedures (Scheme 5).26,31 Preparation of 2E-
d 3 was accomplished in 90% geometric purity by the
stereoselective addition of (CD3)2CuMgI to ethyl tetrolate
followed by MeMgI addition to the resulting R,â-unsatur-
ated ester. For the 2Z-d 3 isomer, the phosphate ester,
derived from deprotonation of ethyl acetoacetate with
sodium hydride and subsequent O-alkylation with diethyl
chlorophosphate, reacted with (CD3)2CuLi to produce in
a highly stereoselective manner the ester-d3, which was
then treated with MeMgI to form the desired alcohol in
97% Z geometrical purity.
1
simplifies the analysis of the adducts in the H NMR.
Syn /An ti Meth yl Ster eoselectivity of a Ter tia r y
Allylic Alcoh ol w ith 1O2 a n d MTAD. To test the
validity of the hydroxyl steering effect proposed in the
photooxygenation of the secondary alcohol 1-OH, and to
study further the stereochemistry of triazolinedione
addition to allylic alcohols, we examined the syn/anti
Photooxygenations were carried out at room temper-
ature in the presence of 10-3 M 2,6-di-tert-butylphenol
as a radical scavenger. The xenon Variac Eimac Cermax
300 W lamp was equipped with a filter to cut off the
wavelengths below 588 nm. The reaction of allylic alco-
1
stereoselectivity in the reaction of O2 and MTAD with
1
hols 2E-d 332 and 2Z-d 3 with O2 proceeded smoothly in
the tertiary allylic alcohol 2. For this purpose, we
synthesized 2Z-d 3 and 2E-d 3, respectively, by labeling
the syn and the anti methyl group with respect to the
functionality. Both electrophiles afford exclusively the
ene adducts in quantitative yield. On the more-substi-
tuted side of the double bond, oxygen is capable of
interacting only with one allylic hydrogen. This experi-
ment eliminates the possibility of having interaction of
a variety of solvents to give only the ene products. The
(30) (a) Stephenson, L. M. Grdina, M. B.; Orfanopoulos, M. Acc.
Chem. Res. 1980, 13, 419-425. (b) Hurst, J . R.; Wilson, S. L.; Schuster,
G. B. Tetrahedron 1985, 41, 2191-2197.
(31) Anderson, R. J .; Corbin, V. L.; Cotterrell, G.; Cox, G. R.;
Henrick, C. A.; Schaub, F.; Siddall, J . B. J . Am. Chem. Soc. 1975, 97,
1197-1204.

4134 J . Org. Chem., Vol. 64, No. 11, 1999
Vassilikogiannakis et al.
Ta ble 3. Syn /An ti Ster eoselectivity in th e En e Rea ction
of 2E-d 3 a n d 2Z-d 3 w ith MTAD in a Va r iety of Solven ts
Sch em e 5. Syn th esis of 2E-d 3 a n d 2Z-d 3
entry
substrate
solvent
syn/anti selectivity
Ta ble 2. Syn /An ti Ster eoselectivity in th e
P h otooxygen a tion of 2E-d 3 a n d 2Z-d 3 in a Va r iety of
Solven ts
1
2
3
4
5
6
7
8
9
2E-d 3
2Z-d 3
2E-d 3
2Z-d 3
2E-d 3
2Z-d 3
2E-d 3
2Z-d 3
2E-d 3
benzene-d6
benzene-d6
CDCl3
54/56
47/53
56/44
45/55
58/42
44/56
55/45
47/53
55/45
CDCl3
acetone-d1
acetone-d1
CD3CN
CD3CN
MeOH-d4
fast, without formation of any significant amounts of
byproducts or methoxy adducts33 in methanol. The defi-
nition of 2E-syn , 2E-a n ti, 2Z-syn, and 2Z-anti is the
same as that used in the description of the photooxygen-
ation reaction. The surprising result is that the syn/anti
stereoselectivity for each isomer is solvent independent
(Table 3), in contrast to the 1O2 reactions where the ratio
is highly solvent dependent. However, a small change
was found on going from 2E-d 3 (syn/anti ) 55/45) to 2Z-
d 3 (syn/anti ) 45/55).
In ter m olecu la r Kin etic Isotop e Effects in th e
Rea ction of MTAD w ith 1-OH a n d 2. To elucidate the
energy profile of triazolinedione addition to allylic alco-
hols, we measured the intermolecular kinetic isotope
effects in the reaction of MTAD with the chiral secondary
allylic alcohol 1-OH and the tertiary allylic alcohol 2. For
this purpose, we synthesized the deuterated allylic alco-
hols 1-OH-d 10 and 2-d 10 according to Scheme 6, through
entry substrate
solvent
sensitizera syn/anti selectivity
1
2
3
4
5
6
7
8
9
2E-d 3
2Z-d 3
2E-d 3
2E-d 3
2Z-d 3
2E-d 3
2E-d 3
2Z-d 3
2E-d 3
CCl4
CCl4
TPP
TPP
TPP
MB
TPP
MB
MB
RB
75/25
72/28
73/27
66/34
65/35
42/58
41/59
40/60
33/67
benzene
CHCl3
CHCl3
acetone
CH3CN
CH3CN
MeOH
MB
a
Abbreviations. TPP: tetraphenylporphine, MB: methylene
blue, RB: rose bengal.
1
the common precursor, mesityl oxide-d10
.
product ratio was measured by H NMR integration of
the appropriate peaks (see Experimental Section). The
results are summarized in Table 2. We define as syn the
adducts formed by allylic hydrogen abstraction which is
on the same side of the double bond as the hydroxyl. For
the case of 2E-d 3, the syn adduct is formed by hydrogen
abstraction, while for the case of 2Z-d 3 by deuterium
abstraction. In fact, adducts 2E-syn and 2Z-a n ti are
identical, and 2E-a n ti and 2Z-syn are also identical.
Triazolinedione additions were carried out by adding
solid MTAD directly into the NMR tube that contained
a solution of the allylic alcohols. The product ratio was
measured directly by 1H NMR integration of the ap-
propriate peaks (see Experimental Section). When neces-
sary the adducts were purified by flash column chroma-
tography. Reactions of 2E-d 3 and 2Z-d 3 with MTAD are
The kinetic competition of 1-OH versus 1-OH-d 10, and
for 2 versus 2-d 10 in chloroform, was monitored by GC
on a 5% methyl phenyl silicone capillary column, using
nonane as an internal standard. The deuterated alkenes
have sufficient separation from their perprotio analogues,
and their retention times are shorter. The results are
depicted in Scheme 6. The observed total isotope effects
are almost unity for the reaction of the secondary allylic
alcohol 1-OH, while for the tertiary allylic alcohol 2 a
small isotope effect was measured.
Discu ssion
(a ) P h otooxygen a tion of Secon d a r y Allylic Alco-
h ol 1-OH. We view the reaction as proceeding through
the intermediacy of a perepoxide or through an exciplex
(32) Preliminary results in the photooxygenation of 2E-d 3 have
been communicated earlier: Stratakis, M.; Orfanopoulos, M.; Foote,
C. S. Tetrahedron Lett. 1996, 37, 7159-7162.
(33) Smonou, I.; Khan, S.; Foote, C. S.; Elemes, Y.; Mavridis, I. M.;
Pantidou, A.; Orfanopoulos, M. J . Am. Chem. Soc. 1995, 117, 7081-
7087.

Stereochemistry in Ene Reactions
J . Org. Chem., Vol. 64, No. 11, 1999 4135
Sch em e 6. Syn th esis of Deu ter a ted Allylic
Alcoh ols 1-OH-d 10, 2-d 10 a n d In ter m olecu la r
Isotop e Effects in Th eir En e Rea ction s w ith MTAD
Sch em e 8. P r op osed Mech a n ism in th e
P h otooxygen a tion of Allylic Alcoh ol 2E-d 3
hydroxyl through hydrogen bonding reduces the favorable
oxygen-hydroxyl steering effect. Thus the threo/erythro
ratio decreases from 11/1 in chloroform to 3.3/1 in
acetonitrile. Also, for the anti-intermediate, the threo/
erythro ratio decreases from 4.3/1 in chloroform to 2.6/1
in acetonitrile. This could be attributed to the solvation
of hydroxyl group, which increases its bulkiness and
reduces the energy difference between TSanti-erythro and
Sch em e 7. P ossible Tr a n sition Sta tes in th e
P h otooxygen a tion of 1-OH
TSanti-threo
.
(b) P h otooxygen a tion of Allylic Alcoh ol 2. We
further examined the validity of the oxygen-hydroxyl
steering effect by performing the photooxygenation of the
tertiary allylic alcohol 2. Examination of the results in
Table 2 reveals that (i) the syn/anti selectivity does not
depend on the specific labeling of the methyl groups,
which indicates that the rate-determining step of the
reaction is the formation of the intermediate as has been
found for nonfunctionalized trisubstituted alkenes34 and
(ii) a remarkable difference in the syn/anti methyl
reactivity on changing solvent polarity.
In nonpolar solvents a favorable hydroxyl-oxygen
interaction during the formation of the perepoxide is very
important in stabilizing TSI where oxygen is syn to the
hydroxyl, as seen in the proposed mechanism of the
photooxygenation of 2E-d 3 (Scheme 8). On the other
hand in TSII, no hydroxyl-oxygen steering effect exists
and that transition state is higher in energy. The
interaction between oxygen and one allylic hydrogen
exists in both TSI and TSII. Thus, the only difference
between the two transition states is the stabilizing
oxygen-hydroxyl steering effect. The two perepoxides
collapse in a faster step through TSIII and TSIV, respec-
tively, to the ene products whose ratio reflects the relative
stabilities of TSI and TSII.
with the structural requirements of a perepoxide. As seen
in Scheme 3, in the photooxygenation of 1-OH-d 4 in
chloroform, the syn-intermediate is formed in 84% ratio
and results in a 11/1 threo to erythro diastereoselect-
ivity. On the other hand the anti-intermediate, formed
in 16% ratio, also affords threo diastereoselectivity (threo/
erythro ) 4.3/1), but to a minor extent compared to
the syn-intermediate. The diastereoselectivity in the
photooxygenation of 1-OH (threo/erythro ) 90/10) is
almost identical to that from the syn-intermediate,
since this intermediate predominates over the anti-
intermediate. For the syn-intermediate, it seems likely
that oxygen interacts with hydroxyl in a transition
state with the minimum 1,3-allylic strain and directs
the reaction diastereoselectivity, as proposed by Adam
and co-worker,16a,b who also considered the syn-interme-
diate to be predominant (Scheme 7). For the anti-
intermediate, we assume that the threo diastereoselec-
tivity could be attributed to the transition state TSanti-threo
in Scheme 7. In that transition state, oxygen approaches
the double bond from the face where the least steric
repulsions are developing. All other conformations, in-
cluding those leading to erythro adducts, seem to be less
favorable.
In polar solvents, hydroxyl coordinates with the solvent
instead of the oxygen. The effective size of -OH becomes
large enough for steric hindrance to increase significantly
the activation energy of TSI, thus leading to an almost
anti selectivity by favoring TSII. In methanol, for ex-
ample, the observed syn/anti ratio of 33/67 is approxi-
In the more polar acetonitrile, there is a decrease in
the threo/erythro ratio, from both syn and anti interme-
diates, which is more significant in the case of the syn-
intermediate. For the syn-intermediate, solvation of
(34) Stratakis, M.; Orfanopoulos, M.; Chen J . S.; Foote, C. S.
Tetrahedron Lett. 1996, 37, 4105-4108.

4136 J . Org. Chem., Vol. 64, No. 11, 1999
Vassilikogiannakis et al.
Sch em e 9. P ossible Tr a n sition Sta tes in th e
Rea ction of 1-OH w ith P TAD
mately the same as that observed for 2,4,4-trimethylpent-
2-ene (3), where a typical anti “cis effect” selectivity syn/
anti ) 25/75 was observed.5
Generally, the regioselectivity of singlet oxygen ene
reactions, as measured by the distribution of the ene
products, is almost solvent independent. Only in the
photooxygenation of R,â-unsaturated esters was a small
change in the distribution of the ene products found as
a function of solvent polarity, due to the different dipole
moments of the two transition states.35 The current
solvent dependence of ene stereoselectivity is the highest
ever reported.
(c) Rea ction of Tr ia zolin ed ion es w ith Ch ir a l
Secon d a r y Allylic Alcoh ol 1-OH. The negligible and
slight inverse kinetic isotope effect (kH/kD ) 0.98 ( 0.02)
observed in the intermolecular competition of 1-OH
versus 1-OH-d 10 clearly rules out the possibility that
allylic hydrogen abstraction occurs in a slow step, com-
pared to the preceding transition state of aziridinium
imide formation. If we consider formation of AI in the
rate-determining step, significant inverse R- and â-sec-
ondary isotope effects would be expected6,36 essentialy due
to the rehybridization of the olefinic carbons from sp2 to
sp3. It rather seems that in the addition of PTAD to
1-OH, the transition states of the two reaction steps are
close in energy, with formation of AI probably being
slightly higher. In that case, the inverse isotope effect in
the first step and the primary isotope effect of the second
step cancel.
cannot be ruled out. In TS4 a steering effect occurs
between the carbonyl functionality of PTAD and the
hydroxyl in a seven-membered ring transition state. A
similar stabilizing interaction between the carbonyl
group of PTAD and the silicon has been proposed to be
responsible for the observed cis stereochemistry in the
ene products derived from PTAD addition to allylsilanes.15b
However, the stability of transition state TS4 is ques-
tioned, since electronic repulsions between the carbonyl
group of MTAD and the oxygen of hydroxyl may also
develop. In polar solvents such as acetone, hydrogen
bonding between the solvent and the hydroxyl reduces
steering efficiency, thus leading to significant amounts
of the erythro adducts, as found the analogous 1O2
reaction.
We can explain the solvent dependent diastereoselec-
tivity observed in the reaction of 1-OH with PTAD,
considering the possible diastereomeric transition states
of AI formation. The contribution of each intermediate
(syn and anti) to the total observed diastereoselectivity
was measured in the reaction of 1-OH-d 4 with PTAD.
The threo/erythro selectivity measured from the syn-AI
when the reaction was carried out in nonpolar chloroform
was 55/5. We consider that, during the formation of the
syn-AI, the negatively charged nitrogen of PTAD inter-
acts with hydroxyl in a favorable six-membered ring
transition state where the 1,3-allylic strain between the
syn allylic methyl and the chiral carbon substituents is
minimized (Scheme 9). The transition state where hy-
drogen interacts sterically with the syn methyl (TS1) is
far more stable than the diastereomeric TS2, where two
methyl groups interact.
The threo selectivity observed from the anti-AI inter-
mediate (threo/erythro ) 32/8) can be rationalized in a
similar way to the explanation offered for the threo
selectivity found from the anti-perepoxide intermediate
in the photooxygenation of 1-OH-d 4. The most favorable
transition state leading to threo adduct is probably
TS3, in which PTAD attacks the double bond in a
conformation where the least steric repulsions are de-
veloping. An alternative threo-product transition state
(TS4) where OH is perpendicular to the olefinic plane
The high efficiency of hydroxyl in controlling the
diastereoselectivity of the ene reaction through the
steering effect with the electrophile can also be supported
by the low diastereoselectivity observed in the reaction
of 1-Et with PTAD (∼14% de). This result indicates that
steric reasons are less important for the high diastereo-
selectivity observed in the reaction of 1-OH with PTAD.
We could predict that steric factors could not lead to more
than 15% diastereomeric excess in the reaction of PTAD
with chiral allylic alcohol 1-OH, instead of the greater
than 80% observed in nonpolar solvents.
(d ) Rea ction of Tr ia zolin ed ion es w ith Ter tia r y
Allylic Alcoh ol 2. The significant results from the
reaction of MTAD with the tertiary allylic alcohols 2E-
d 3 and 2Z-d 3 (Table 3), are the following: (i) lack of
solvent effect on the syn/anti stereoselectivity; (ii) the
unexpected high reactivity of the syn methyl group
despite the remarkable steric hindrance, and (iii) a small
intermolecular kinetic isotope effect of kH/kD ) 1.15 (
0.02 as measured from the direct competition of 2 with
2-d 10.
The lack of solvent dependence in the syn/anti distri-
bution of the ene products is consonant with AI formation
occurring in a fast step comparing to the accompanying
step of hydrogen abstraction which is the rate-determin-
ing. If formation of AI occurred in the rate-determining
step, a significant change in the syn/anti allylic methyl
reactivity would be expected when changing the solvent,
as was observed in the corresponding reaction with
singlet oxygen. The normal kinetic intermolecular isotope
(35) Orfanopoulos, M.; Stratakis, M. Tetrahedron Lett. 1991, 49,
7321-7324.
(36) Elemes, Y.; Orfanopoulos, M. Unpublished results.

Stereochemistry in Ene Reactions
J . Org. Chem., Vol. 64, No. 11, 1999 4137
detector. Diethyl ether, diglyme, and THF were distilled
from Na/benzophenone. N,N,N′,N′-tetramethylethylenedi-
amine (TMEDA) and diisopropylamine were refluxed for 24 h
with LiAlH4 and then distilled and kept over 3 Å molecular
sieves.
effect of kH/kD ) 1.15 ( 0.02 on competition of 2 versus
2-d 10 is consistent with this assumption. Furthermore,
the small ratio change from the average syn/anti ) 55/
45 in 2E-d 3 to an average syn/anti ) 45/55 in 2Z-d 3 is
in accordance with the intermolecular competition ex-
periment (2 versus 2-d 10). An approximate value of kH/
kD ) 1.2 can be deduced, considering that in unlabeled
olefin, the syn and anti methyls are equally reactive (50/
50). If formation of AI occurred in the rate-determining
step, a significant inverse isotope secondary effect would
be expected. The fact that the isotope effect is normal,
but rather small, probably indicates that the hydrogen
abstraction is rate-determining, but its energy is close
to the preceding step of AI formation. Similarly, a small
value of kH/kD ) 1.25 measured in the reaction of PTAD
with trans-2-butene was explained in terms of a revers-
ible formation of the AI intermediate.2c The surprising
fact is that an extra methyl group at the allylic position
on going from 1-OH to 2 can alter the kinetic profile of
the reaction with triazolinediones.
4-Meth ylp en t-3-en -2-ol (1-OH). The compound was pre-
pared by MeMgI addition to 3-methyl-2-butenal at 0 °C, or by
reduction of mesityl oxide with LiAlH4, and was purified by
vaccum distillation. 1H NMR: 5.21 (d with allylic couplings,
J 1 ) 8.6 Hz, J 2 ) 1.1 Hz, J 3 ) 1.0 Hz, 1H), 4.56 (m, 1H), 1.71
(d, J ) 1.0 Hz, 3H), 1.69 (d, J ) 1.1 Hz, 3H), 1.40 (br s, 1H,
hydroxyl), 1.23 (d, J ) 6.3 Hz, 3H).
Ad d u cts fr om 1-OH a n d 4-P h en yl-1,2,4-tr ia zolin e-3,5-
d ion e. The ene products were obtained by adding solid PTAD
to a solution of 1-OH in several solvents and were purified by
flash column chromatography using ethyl acetate as eluent
(the diastereomers elute together). 1H NMR: 7.35-7.50 (m,
5H threo + 5H erythro), 5.12 (d, J ) 1.4 Hz, 1H erythro), 5.10
(s, 1H, erythro), 5.07 (s, 1H threo), 5.04 (s, 1H threo), 4.20-
4.37 (m, 2H threo + 2H erythro), 1.83 (s, 3H erythro), 1.75 (s,
3H threo), 1.24 (d, J ) 6.1 Hz, 3H threo + 3H erythro). MS,
m/z ) 275 (M+, 5).
(Z)-4-Meth ylp en t-3-en -2-on e-5,5,5-d 3. To a slurry con-
taining 1.32 g of NaH (60% in oil) in dry ether was slowly
added at 0 °C 3 g of acetylacetone (30 mmol). After stirring
for 20 min, 5.7 g of diethyl chlorophosphate (33 mmol) was
added. Stirring was continued for an additional 2 h, and then
the solution was quenched with saturated solution of NH4Cl
and extracted with ether to afford 7.3 g of enol phosphate (89%
The unexpectedly high syn methyl reactivity in the
reaction of 2 with triazolinediones (∼50%), could be
explained as follows. Although the electrophile probably
prefers to form the anti-intermediate due to steric
reasons, the small percentage of the syn-intermediate is
competitive with the anti-intermediate in the accompa-
nying rate-determining step of the reaction, because the
allylic hydrogen abstraction from the syn-AI results in
significant relief of the steric interactions, not only
between the bulky electrophile and the alkene, but also
between the cis substituents of the olefin (methyl and
dimethylhydroxyl). In other words, the ratio of the
formation of the syn-AI/anti-AI does not reflect the ratio
of the syn/anti products.
1
yield). H NMR: 5.44 (s, 1H), 4.20 (m, 4H), 2.26 (s, 3H), 2.15
(s, 3H), 1.33 (dt, J H-H ) 7.1 Hz, J H-P ) 1.1 Hz, 6H).
Subsequently, the enol phosphate was slowly added to a
solution containing 53.8 equiv (26.9 mmol) of (CD3)2CuLi at
-78 °C. The (CD3)2CuLi was prepared by addition of 53.8 mmol
of CD3Li (from CD3I and Li in ether) to 26.9 mmol of CuI at 0
°C. The solution was stirred at -78 °C for 4 h and then
quenched with a saturated solution of NH4Cl. The ether layer
was washed with 25% aqueous NH3 and then with brine. The
mesityl oxide-d3 was isolated after distillation in 62% yield
1
(1.70 g). The isomeric purity was 95%. H NMR: 6.07 (d, J )
Con clu sion s
1.2 Hz, 1H), 2.14 (s, 3H), 1.86 (d, J ) 1.2 Hz, 3H). MS, m/z )
101 (M+, 50).
As a mechanistic comparison we found that for both
electrophiles (1O2 and triazolinediones) behave via similar
manner in their ene reactions with secondary allylic
alcohols. The steering effect between hydroxyl and the
electrophiles is significant in the distribution of the
diastereomeric ene products. Polar solvents with hydro-
gen bonding ability minimize this interaction, and hy-
droxyl behaves sterically almost like a methyl group. For
the tertiary allylic alcohols, however, there is a discrep-
ancy in the stereoselectivity of the ene reactions, due to
the fact that the two electrophiles have different energetic
(Z)-4-Meth ylp en t-3-en -2-ol-2,5,5,5-d 4 (1-OH-d 4). The (Z)-
4-Methylpent-3-en-2-one-5,5,5-d3 (1.7 g, 16.8 mmol) was re-
duced with LiAlD4 (0.53 g, 12.6 mmol) in ether, to form the
allylic alcohol 1-OH-d 4 in 74% yield. The geometric purity of
1
the olefin was 94%. H NMR: 5.20 (br s, 1H), 1.70 (d, J ) 1.3
Hz, 3H), 1.21 (s, 3H). MS, m/z ) 104 (M+, 15).
1
P h otooxygen a tion of 1-OH-d 4. The H NMR data for the
products in the photooxidation of 1-OH have been reported
by Adam and co-workers.16b The product ratio of syn-threo/
syn-erythro can be found by integrating the allylic hydrogens
resonating at 1.73 and 1.80 ppm, respectively. The product
ratio of anti-threo/anti-erythro can be found by integrating the
olefinic region (5.07 ppm 1 H threo + 1H erythro, 5.09 ppm
1H threo, and at 5.12 ppm 1H erythro).
1
profiles. For O2, a solvent dependent steering effect was
observed, because formation of perepoxide occurs in the
rate-determining step, whereas for triazolinediones, the
absence of a steering effect is consonant with formation
of the aziridinium imide in a fast step. The proposed
change of the energy profile in the ene reactions of
triazolinediones on going from the secondary allylic
alcohol 1-OH to the tertiary 2 was supported by inter-
molecular kinetic isotope effects.
Ad d ition of P TAD to 1-OH-d 4. The product analysis was
accomplished by direct NMR integration of the appropriate
peaks in chloroform-d or acetone-d6. The product ratio of syn-
threo/syn-erythro can be found by integrating the allylic
hydrogens resonating at 1.75 and 1.83 ppm, respectively. The
product ratio of anti-threo/anti-erythro can be found by
integrating the olefinic region (5.04 ppm 1 H threo, 5.07 ppm
1H threo, 5.10 ppm 1H erythro, 5.12 ppm 1H erythro).
2,4-Dim eth ylh ex-2-en e (1-Et). This compound was pre-
pared by Wittig coupling of isopropylidenetriphenylphos-
ponium ylide with 2-methylbutyraldehyde. The ylide was
prepared by deprotonation of isopropyltriphenylphosphonium
bromide in dry diglyme using NaH as base, at 110 °C for 2 h.
The olefin was distilled out from the reaction mixture by a
slow stream of nitrogen and then purified by preparative GC
(SE-30, Tcolumn ) 60 °C). 1H NMR: 4.87 (d with allylic
Exp er im en ta l Section
Nuclear magnetic resonance spectra were obtained on 250,
360, and 500 MHz instruments. The spectra reported herein
were taken in CDCl3. Isomeric purities were determined by
NMR and by GC-MS on an HP-5 cross-linked capillary column
(5% phenyl methyl silicone) equipped with a 5971A MS

4138 J . Org. Chem., Vol. 64, No. 11, 1999
Vassilikogiannakis et al.
coupling, J 1 ) 8.0 Hz, J 2 ) 1.1 Hz, 1H), 2.20 (m, 1H), 1.68 (d,
J ) 1.1 Hz, 3H), 1.60 (d, J ) 1.1 Hz, 3H), 0.79-0.98 (m, 8H).
(Z)-2,4-Dim et h ylp en t -3-en -2-ol-5,5,5-d 3 (2Z-d 3). This
compound was produced in a similar way to 2E-d 3, by react-
ing the (Z)-ethyl 3-methylbuten-2-oate-4,4,4-d3 with 2.5
equiv of MeMgI in ether in 80% yield and 97% geometrical
purity. 1H NMR: 5.33 (d, J ) 1.3 Hz, 1H), 1.69 (d, J ) 1.3 Hz,
3H), 1.55 (br s, 1H, hydroxyl), 1.35 (s, 6H). 13C NMR: 17.5
(septet due to the D coupling), 26.7, 30.9, 70.1, 132.2,
132.9.
P h otooxygen a tion of 2E-d 3 a n d 2Z-d 3. Photooxidation
of 2E-d 3 in several solvents afforded two ene products, 2E-
syn and 2E-a n ti. Their ratio was measured by integration of
the appropriate peaks. In the photooxygenation of 2E-d 3 the
1H NMR is as follows: 5.15 (d, J ) 1.7 Hz, 1H of 2E-syn ),
5.07 (d, J ) 1.1 Hz, 1H of 2E-syn ), 4.27 (s, 1H of 2E-syn and
1H of 2E-a n ti), 1.86 (s, 3H of 2E-a n ti), 1.24 (s, 3H of 2E-syn
and 3H of 2E-a n ti), 1.22 (s, 3H of 2E-syn and 3H of 2E-a n ti).
In the photooxygenation of 2Z-d 3 the product analysis occurs
in a similar way.
Rea ction of 1-Et w ith P TAD. Two diastereomers were
formed in chloroform (56/44 ratio) and in acetone (58/42 ratio).
The stereochemistry of each isomer (which is threo and which
is erythro) was not assigned. 1H NMR: 7.37-7.55 (m, 5H threo
+ 5H erythro), 5.07 (br s, 2H threo + 2H erythro), 4.31 (d, J
) 10.8 Hz, 1H of the major diastereomer), 4.27 (d, J ) 11.0
Hz, 1H of the minor diastereomer), 2.04 (m, 1H threo + 1H
erythro), 1.81 (s, 3H threo + 3H erythro), 1.51 (m, 1H threo +
1H erythro), 1.01-1.24 (m, 1H threo + 1H erythro), 0.91 (m,
6H threo + 6H erythro).
(E)-Eth yl 3-Meth ylbu t-2-en oa te-4,4,4-d 3. To a flame-
dried flask containing 1.14 g (60 mmol) of CuI in 100 mL of
dry THF was added dropwise at -40 °C 55 mL of CD3MgI (1.0
M in ether, Aldrich). After 30 min, 22 mL of dry TMEDA was
added. The solution was cooled to -78 °C, and then 2.3 mL
(20 mmol) of ethyl tetrolate was syringed in. After 6 h the
reaction mixture was quenched with 8 mL of methanol and 4
mL of saturated solution of NH4Cl. GC analysis showed
approximately 1.5% of starting material present. The R,â-
unsaturated ester-d3 (1.8 g, 69% yield) was a mixture of E/Z
) 90/10. It is crucial to maintain the reaction temperature at
-78 °C; otherwise, the amount of the Z isomer increases
significantly. 1H NMR of the major isomer (E): 5.67 (d, J )
1.1 Hz, 1H), 4.14 (q, J ) 7.2 Hz, 2H), 2.16 (d, J ) 1.1 Hz, 3H),
1.27 (t, J ) 7.2 Hz, 3H).
Rea ction of 2E-d 3 a n d 2Z-d 3 w ith MTAD. Reaction of
2E-d 3 with MTAD in several solvents afforded two ene
products, 2E-syn and 2E-a n ti. Their ratio was measured by
integration of the appropriate peaks. In the reaction of MTAD
1
with 2E-d 3 the H NMR is as follows: 5.16 (d, J ) 1.4 Hz, 1H
of 2E-syn ), 5.07 (d, J ) 1.5 Hz, 1H of 2E-syn ), 4.36 (s, 1H of
2E-syn and 1H of 2E-a n ti), 3.09 (s, 3H of 2E-syn and 3H of
2E-a n ti), 1.90 (s, 3H of 2E-a n ti), 1.36 (s, 3H of 2E-syn and
3H of 2E-a n ti), 1.33 (s, 3H of 2E-syn and 3H of 2E-a n ti). In
the reaction of MTAD with 2Z-d 3 the product analysis occurs
in a similar way.
(E)-2,4-Dim eth ylp en t-3-en -2-ol-5,5,5-d 3 (2E-d 3). A solu-
tion of the above ester in dry ether was added dropwise at 0
°C to 2.5 equiv of MeMgI. After stirring at ambient temper-
ature for 5 h, the reaction was quenched with saturated
solution of NH4Cl and immediately transferred to a separat-
ing funnel, where 1-2 mL of pyridine were added, and the
organic layer was washed with saturated solution of NaHCO3
and brine. Removal of the ether left approximately a 1:1
mixture of 2E-d 3 and pyridine. Further purification was
achieved by vacuum distillation in the presence of solid K2-
CO3. In the absence of pyridine the labile37 tertiary allylic
alcohol dehydrates on standing. The geometrical purity was
90%. 1H NMR of the major isomer (E): 5.33 (d, J ) 1.2 Hz,
1H), 1.85 (d, J ) 1.2 Hz, 3H), 1.60 (br s, 1H, hydroxyl), 1.35
(s, 6H). 13C NMR: 17.8, 25.8 (septet due to the D coupling),
4-Meth ylp en t-3-en -2-on e-d 10 (m esityl oxid e-d 10). To a
dry flask containing 3 g (30 mmol) of diisopropylamine in 15
mL of dry THF was added at 0 °C 18.8 mL of n-BuLi (1.6 M in
hexane). The solution was cooled to -78 °C, and then 1.90 g
(30 mmol) of acetone-d6 were slowly syringed in. After 1 h,
the enolate reacted with 1.90 g (30 mmol) of acetone-d6.
Stirring was continued for 2 h, and then the solution was
warmed to room temperature and quenched with 1.5 mL of
saturated solution of NH4Cl (in D2O). The aldol adduct was
isolated by extraction with ether and was dehydrated neat,
on heating with a catalytic amount of p-toluenesulfonic acid.
The deuterated mesityl oxide was isolated by vacuum distil-
lation in 40% overall yield (1.3 g).
4-Met h ylp en t -3-en -2-ol-1,1,1,3,5,5,5,4′,4′,4′-d 10 (1-OH -
d
10). The mesityl oxide-d10 was reduced by LiAlH4 in ether to
30.4, 69.5, 131.9, 132.1. HRMS: calculated for C7D3H11
117.1233, found 117.1234.
O
afford 4-methylpent-3-en-2-ol-1,1,1,3,5,5,5,4′,4′,4′-d10 in 72%
yield. 1H NMR: 4.56 (s, 1H), 1.50 (br s, 1H, hydroxyl). The
deuterium content at the vinylic position was 88%, and at
the allylic positions 93%. The H/D scrambling occurred during
the acid-catalyzed dehydration of the aldol. MS, m/z)110 (M+,
5).
In ter m olecu la r Com p etition of 1-OH ver su s 1-OH-d 10
in th e Rea ction w ith MTAD. Perprotio olefin 1-OH and
deuterated olefin 1-OH-d 10 have sufficient separation on GC
to allow analysis, and the change in the ratio 1-OH /1-OH-
d 10 was easily monitored at several percentages of reaction
progress (20-30% conversion). Nonane was used as an internal
standard. For the estimation of the kinetic isotope effect the
following expression38 was used:
(Z)-Eth yl 3-Meth ylbu t-2-en oa te-4,4,4-d 3. To a dry flask
containing 60 mL dry THF and 1.28 g (32 mmol) of NaH (60%
in oil) were added dropwise 3.9 g (30 mmol) of ethyl acetoac-
etate at 25 °C. Immediate hydrogen evolution occurred. After
20 min, 5.5 g (32 mmol) of diethyl chlorophosphate was added,
and the resulting mixture was stirred for an additional 2 h.
The reaction mixture was quenched with aqueous NH4Cl,
extracted with ether, and washed with saturated solution of
1
NaHCO3, to afford the enol phosphate in 88% yield. H NMR:
5.26 (s with allylic coupling, J ) 1.0 Hz, 1H), 4.22 (m, 4H),
4.10 (q, J ) 7.1 Hz, 2H), 2.12 (d, J ) 1.0 Hz, 3H), 1.32 (dt, J 1
) 7.1 Hz, J 2 ) 1.1 Hz, 6H), 1.21 (t, J ) 7.1 Hz, 3H).
Subsequently, 7.5 g (26.5 mmol) of the above enol phosphate
was added dropwise at -78 °C to a solution of (CD3)2CuLi (26.5
mmol). The dimethyl-d6 cuprate was prepared by addition of
2 equiv of CD3Li to 1 equiv of CuI in dry ether at 0 °C. The
resulting mixture was stirred at low temperature for an
additional 4 h and then quenched with a saturated solution
of NH4Cl and extracted with ether. The organic layer was
washed with 25% aqueous NH3 and brine. Distillation afforded
1.86 g of pure ester (54% yield) in 97% Z geometrical purity.
1H NMR: 5.67 (d, J ) 1.2 Hz, 1H), 4.14 (q, J ) 7.2 Hz, 2H),
1.89 (d, J ) 1.2 Hz, 3H), 1.27 (t, J ) 7.2 Hz, 3H). MS, m/z )
131 (M+, 40).
kH log[1 - Hr/Ht]
)
kD
log[1 - Dr/Dt]
where Hr and Dr are the amounts of 1-OH and 1-OH-d 10 that
reacted, or by analogy, the amounts of adducts formed from
1-OH and 1-OH-d 10, and Ht and Dt are the initial amounts
of 1-OH and 1-OH-d 10, respectively.
2,4-Dim et h ylp en t -3-en -2-ol-1,1,1,3,5,5,5,4′,4′,4′-d 10 (2-
d 10). This compound was prepared by MeMgI to 4-methylpent-
3-en-2-one-d10 in ether. The isolation of the product was
(37) Vathke-Ernst, H.; Hoffman, H. M. R. Chem. Ber. 1981, 114,
1464-1475.
(38) Higgins, R.; Foote, C. S.; Cheng, H. ACS Chem. Ser. 1968, 77,
102-117.

Stereochemistry in Ene Reactions
J . Org. Chem., Vol. 64, No. 11, 1999 4139
accomplished following the experimental details in the prepa-
ration of 2E-d 3 or 2Z-d 3. H NMR: 1.50 (br s, 1H, hydroxyl),
Ack n ow led gm en t. This work was supported by
ΥΠΕΡ-1995 (Greece), NATO grant No CRG 931419 and
by NSF grant No CHE 94-23027.
1
1.35 (s, 3H). The H/D scrambling found in the synthesis of
2-d 10 is the same found in the synthesis of 1-OH-d 10. MS,
m/z ) 124 (M+, 10).
In ter m olecu la r Com p etition of 2 ver su s 2-d 10 in th e
Rea ction w ith MTAD. Identical procedure to that used in
the intermolecular competition of 1-OH versus 1-OH-d 10 were
followed.
Su p p or tin g In for m a tion Ava ila ble: Copies of 29 1H and
13C NMR spectra. This material is available free of charge via
the Internet at http://pubs.acs.org.
J O990258B