Oxa-ene reaction of enols of amides with 4-phenyl-1,2,4-triazoline-3,5- dione

Article abstract of DOI:10.1021/jo702083u

(Chemical Equation Presented) The reaction of 16 enols of amides with 4-phenyl-1,2,4-triazoline-1,3-dione gave open chain adducts rather than the [2 + 2] cycloadducts with a hemiaminal moiety, both in solid state and in solution. This assignment is based

Full text of DOI:10.1021/jo702083u

Oxa-ene Reaction of Enols of Amides with
4-Phenyl-1,2,4-triazoline-3,5-dione†
Ahmad Basheer and Zvi Rappoport*
Department of Organic Chemistry and the Lise Meitner MinerVa Center for Computational Quantum
Chemistry, The Hebrew UniVersity, Jerusalem 91904, Israel
zr@Vms.huji.ac.il
ReceiVed September 23, 2007
The reaction of 16 enols of amides with 4-phenyl-1,2,4-triazoline-1,3-dione gave open chain adducts
rather than the [2 + 2] cycloadducts with a hemiaminal moiety, both in solid state and in solution. This
1
assignment is based on X-ray crystallography, H and 13C NMR data, and IR spectra. The suggested
mechanism involves hydroxyl proton loss in a formal oxa-ene reaction. Mechanistic details and a possible
alternative are discussed.
Introduction
is with CH2N2,3 which reacts with different systems on either
the OH, the NHR, or the CdC groups. In the present work, we
extended this study to a possible reaction of the double bond
of the enol with the azo functionality in 4-phenyl-1,2,4-
triazoline-3,5-dione (PTAD) 3.4 This very reactive enophile5
was used as a reaction partner in [2 + 4]5,6 and [2 + 2]7
cycloadditions and in ene reactions.8 We desired to find out if
PTAD’s reaction with 1 would give the [2 + 2] cycloadducts
1,2-diazetidines 4, their isomeric open chain 1,2 adducts 5, or
both, or other species such as the enol imine 6.9
In recent years, we have prepared a large number of the rare
enols (1) of the well-known tautomeric amides (2), where Y
and Y′ are electron-withdrawing groups (EWGs). We studied
their properties, mostly H and 13C NMR, their solid state and
1
solution structure, their intramolecular hydrogen bonding, and
especially the 1 h 2 equilibria, defined by Kenol ) [1]/[2] as a
function of Y, Y′, R, the solvent, and the temperature.1,2
Results and Discussion
The reaction of 16 enols of amides 1 activated by different
combinations of two â-EWGs with PTAD 3 gave adducts (eq
Species 1 contain the OH, NHR, CdC, Y, and Y′ functional
groups and seem to be interesting synthones. However, except
for the very fast 1 h 2 reaction, the only study of their reaction
(3) Lei, Y. X.; Rappoport, Z. J. Org. Chem. 2002, 67, 6971.
(4) Radl, S. AdV. Heterocycl. Chem. 1997, 67, 119.
(5) Sauer, J.; Schroder, B. Chem. Ber. 1967, 100, 678.
* Corresponding author. Fax: 972-2-6585-345.
† Preliminary report: Basheer, A.; Rappoport, Z. The 69th Meeting of the
Israel Chemical Society, Tel Aviv, Israel, February 2-3, 2004, Abstract PB116.
(1) Preliminary report: Basheer, A.; Rappoport, Z. The 69th Meeting
of the Israel Chemical Society, Tel Aviv, Israel, February 2-3, 2004,
Abstract PB116.
(6) See, for example: (a) Burrage, M. E.; Cookson, R. C.; Gupte, S. S.;
Stevens, I. D. R. J. Chem. Soc., Perkin Trans. 2 1975, 1325. (b) Cookson,
R. C.; Gilani, S. S. H.; Stevens, I. D. R. J. Chem. Soc. C 1967, 1905. (c)
Jensen, F.; Foote, C. S. J. Am. Chem. Soc. 1987, 109, 6376.
(7) See, for example: (a) Pasto, D. P.; Chen, A. F.-T. Tetrahedron Lett.
1973, 713. (b) Koerner v Gustorf, E.; White, D. V.; Kim, B.; Hess, D.;
Leitich, J. J. Org. Chem. 1970, 35, 1155. (c) Cheng, C.-C.; Greene, F.;
Blount, J. F. J. Org. Chem. 1984, 49, 2917. (d) Nelsen, S. F.; Kapp, D. L.
J. Am. Chem. Soc. 1985, 107, 5548. (e) Moody, C. AdV. Heterocycl. Chem.
1982, 30, 1. (f) Adam, W.; Carballeira, N. J. Am. Chem. Soc. 1984, 106,
2874.
(2) (a) Mukhopadhyaya, J. K.; Sklenak, S.; Rappoport, Z. J. Am. Chem.
Soc. 2000, 122, 1325. (b) Mukhopadhyaya, J. K.; Sklenak, S.; Rappoport,
Z. J. Org. Chem. 2000, 65, 6856. (c) Lei, Y. X.; Cerioni, G.; Rappoport, Z.
J. Org. Chem. 2000, 65, 4028. (d) Lei, Y. X.; Cerioni, G.; Rappoport, Z.
J. Org. Chem. 2001, 66, 8379. (e) Lei, Y. X.; Casarini, D.; Cerioni, G.;
Rappoport, Z. J. Phys. Org. Chem. 2003, 16, 525. (f) Lei, Y. X.; Casarini,
D.; Cerioni, G.; Rappoport, Z. J. Org. Chem. 2003, 68, 947. (g) Basheer,
A.; Rappoport, Z. J. Org. Chem. 2004, 69, 1151. (h) Basheer, A.; Yamataka,
H.; Ammal, S.-C.; Rappoport, Z. J. Org. Chem. 2007, 72, 5297. (i) Song,
J.; Yamataka, H.; Rappoport, Z. J. Org. Chem. 2007, 72, 7605.
(8) For a recent review, see: Vougioukalakis, G.; Orfanopoulos, M.
Synlett 2005, 713.
(9) Elimination of water from 4 will give a substituted 3-imino-1,5-
diazabicyclo[3.2.0]heptane derivative.
10.1021/jo702083u CCC: $40.75 © 2008 American Chemical Society
Published on Web 12/12/2007
184
J. Org. Chem. 2008, 73, 184-190

Oxa-ene Reactions of Enols of Amides
1) having a 1:1 composition of the two reagents. The 1:2
adducts, similar to those observed in the addition of 3 to
diphenylketene,10 were not observed. The products can be the
[2 + 2] cycloadducts 4a-p, or the open chain isomeric
tautomeric adducts 5a-p, or 6a-p. The only previously
reported product was presented as 5o.11 The reactions of 0.25
M 3 with 5% molar excess of 1 in CHCl3 were rapid, mostly
finished in <1 min as judged by the disappearance of the red
color of 3, which was accompanied by a slower appearance of
a precipitate. To obtain approximate information on the relative
rates of several enols, these mixtures were diluted and the time
required for complete disappearance of the color of 3 was
recorded. For 1d, the color disappeared immediately even on
100-fold dilution and in 3 and 17 min on 1000- and 2000-fold
dilution, respectively. With 1h, the color disappeared in 1, 5,
9, and 150 min, respectively, on 10-, 50-, 100-, and 1000-fold
dilution. For the slowest compound, 1o, the color of the
undiluted solution disappeared in 45 min, whereas 2-, 5-, and
10-fold dilution caused its disappearance in 85, 285, and 7200
min, respectively. With the Meldrum’s acid derivative 1p, the
color disappeared in 15 min, 170 min, overnight, and 8 days,
respectively, with the original solution and its 2-, 5-, and 10-
fold diluted solutions.
On the basis of these data and that of Table S5, the
semiquantitative reactivity order of the enols is 1p < 1o < 1n
< 1g < (1i, 1j, and 1k) < 1h < 1d < other enols 1. The
corresponding Kenol (CDCl3) values are 1p (g502a), 1o (0.072a),
1n (0.412d), 1g (4.02h), 1i (4.02h), 1j (9.02h), 1k (8.12h), 1h
(13.32h), 1d (g502g), 1a (g502b), 1b (g502e), 1c (g502g), 1e
(g502g), 1f (g502g), 1l (g502b), and 1m (g502e). Hence, the
qualitative reaction rates depend mostly on the % enol in the
enol/amide mixture in CDCl3. The color disappeared im-
mediately with enols whose equilibrium percentage with the
amide 2 (e.g., 2e) in chloroform is high. For systems with a
lower Kenol value, its complete disappearance took from a few
minutes to 45 min at 0.25 M concentration. Exceptions are the
Meldrum’s acid enol 1p, which is fully enolic in CHCl3 but
reacted more slowly than most other enols, and the enol of
nitromalonamide, which is fully enolic in THF but did not react
at all with 3.
Under these reaction conditions, no reaction was observed
between 1b and cyclopentadiene or tetraphenylcyclopentadiene
and between 1f and tetraphenylcyclopentadienone or tetracya-
noethylene (TCNE). This is not surprising since 3 is ca. 102 to
104 more reactive than TCNE in Diels-Alder reactions with
1,4-diphenyl- and 2-chloro-1,4-butadienes.5
Structure Assignment. (a) In the Solid State. The solid state
structure of derivative i was determined by X-ray crystal-
lography to be that of the open chain adduct 5i. The ORTEP
drawing and the full data are given in the Supporting Informa-
tion. This structure (for atom numbering, see the following
designation) is unequivocally deduced from the C1-O1 and
C3-O2 bond lengths of 1.2081 (17) and 1.2137 (18) Å, which
indicate that they are CdO bonds. In both cases, no C-O bond
lengthening due to the amide resonance was observed. Likewise,
the C1-C2 and C2-C3 bonds are even longer (1.5669 (19)
and 1.5518 (19) Å, respectively) than a normal C-C bond,12
3
2
although C2-C3 is a Csp -Csp bond. A relevant comparison
is with the amide 2i, Me2NCOCH(CN)CONHPr-i, where the
C1-C2 and C2-C3 bond lengths are 1.545(3) and 1.532 (4)
Å, respectively, and the C1-O1 and C3-O2 bond lengths are
1.232 (3) and 1.220(3) Å, respectively. The remarkable similar-
ity in the bond lengths exclude structure 6i, and the N5-C1
distance of 2.473 Å excludes structure 4i.
The triazolidine N5-H bond length is normal at 0.84 (2) Å,
and the hydrogen is hydrogen bonded to the amide (formerly
enol) oxygen O1 with an H‚‚‚O1 hydrogen bond of 2.134 (19)
Å.
(10) Hall, J. H.; Krishnan, G. J. Org. Chem. 1984, 49, 2498.
(11) Hoffmann, R. W.; Schafer, W. Chem. Ber. 1972, 105, 2437.
(12) Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A.
G.; Taylor, R. J. Chem. Soc., Perkin Trans. 2 1987, S1.
J. Org. Chem, Vol. 73, No. 1, 2008 185

Basheer and Rappoport
FIGURE 1. 13C spectra of 5i. (a) In the solid state and (b) in CDCl3 solution.
The N5‚‚‚O1 nonbonded distance of 2.6763 (17) Å indicates
a moderately strong hydrogen bond, and the N5HO1 angle is
121.8 (16)°. The second N-H group is intermolecularly bonded
to the amido-carbonyl of another molecule of 5i with the
following parameters: N1-H 0.810 (19) Å, N1‚‚‚O2#1 2.037
(19) Å, and N1‚‚‚O2 2.8286 (17) Å and an NHO angle of 165.4
(17)°, where #1 denotes a symmetry transformation used to
generate equivalent atoms: -X + 1, -Y + 1, -Z.
The solid state 13C CPMAS NMR spectrum of 5i displayed
signals at 21.9, 26.3, 39.0, 41.4, 45.9, 71.0, 114.0, 116.5, 124.5,
125.9, 129.6, 131.7, 134.0, 154.1, and 159.8 ppm (Figure 1a).
The NMR spectra in CDCl3 displayed signals at 21.65, 21.82,
38.39, 38.49, 43.7, 71.1, 111.9, 125.7, 128.95, 129.22, 130.1,
153.4, 154.3, 157.2, and 158.9 ppm (Figure 1b). In general, the
spectra look much the same, except that the solid state spectrum
shows the two molecules observed in the X-ray diffraction and
is poorly resolved in the 155-160 ppm region as compared to
the solution spectrum.
(b) Structure in Solution. The 1H NMR spectra in solution
(Table S1 in the Supporting Information) were determined in
CDCl3, THF-d8, or DMSO-d6 and showed only the number of
signals expected for a single compound. When Y * Y′, both
CR and Câ are chiral, and 4 may display pairs of diastereomers,
whereas in compounds 5, only one set of signals is expected in
achiral medium. We conclude that either 4 does not exist in
186 J. Org. Chem., Vol. 73, No. 1, 2008

Oxa-ene Reactions of Enols of Amides
solution or if it is formed, its chiral CR center is lost by a rapid
ppm.16 We conclude that δ(C(OH)NH) is 75-88 ppm, mostly
ca. 85 ppm, and that δ(OH) and δ(NHR) are at the 2.5-5.2
ppm region. Since our compounds display only one 13C NMR
singlet at 69-79 ppm, assigned to Câ, and no unidentified H
NMR signals at 2.5-5.2 ppm, we exclude structure 4.
4 h 5 equilibration.
1
The H NMR spectra show Ph, N-Ar or N-Alk, and Y,Y′
1
signals and two additional protons: one for the NHR group,
and one is either the OH of 4 or the NH of 5. Of these, one
broad singlet is at the lowest field observed at room temperature.
The other is usually a singlet or a doublet for the NH-Pr-i
derivatives due to CH-NH coupling. Since when R is i-Pr or
t-Bu, the δ(NH) value is smaller than when R is Ar, and
assignment of this proton is clear. Its δ value of 7.65-11.58 is
much higher than that of hemiaminal C(OH)NHR protons.
Comparison with the amides 2 in the same solvent mostly show
a good correspondence with the higher field proton of the
adduct: adduct, δ(NH-4 or 5), δ(NH-2) ppm: 4c/5c, 8.14,
9.28;2g 4d/5d, 10.83, 10.30;2g 4e/5e, 6.31, 6.372g (CDCl3), and
8.83, 8.312g (DMSO-d6); 4f/5f, 6.22, 5.95;2g 4g/5g, 8.10, 9.14;2h
4h/5h, 8.77, 9.40;2h 4i/5i, 6.61, 6.682h (CDCl3) and 7.60, 7.312h
(THF-d8); 4j/5j 10.46, 10.38;2h 4k/5k 6.36, 6.88;2h 4o/5o 7.98,
9.26;2a and 4p/5p, 8.90, 10.992a. We conclude that all com-
pounds have structure 5.
In Table S3 in the Supporting Information, δ(13C) values for
the amides 2 and the corresponding adducts 5 (∆adduct-amide) in
the same solvent are compared. The major influence is on Câ,
which is attached to the triazolidinyl group. The differences in
parts per million are nearly constant, being 25.9 ( 0.9 for the
cyanomalonamides, 23.5 ( 1.4 for the cyanoester amides, and
18.9 ( 0.50 for the amido diesters. The latter lower values are
due to much higher δ values of 5, as compared to those for 2.
For all other groups, δ(5) are at a higher field than δ(2). They
are -0.75 to -2.17 for the CN and -0.63 -to -3.31 for the
amide CO with no apparent trend. In sum, the spectra of the
two species resemble each other reasonably, considering the
presence of the triazolidinyl group. Moreover, the strong effect
of the latter group on the δ(Câ) differences is not observed at
all for CR, suggesting that CR is not bonded to this group (i.e.,
1
the structure is 5 and not 4). NOESY, COSY, and H-13C
1H NOESY and COSY NMR spectra in CDCl3 and THF-d8
show a correlation between the i-Pr-H (δ 4.05 and 4.02) and
the N-H (δ 7.75 and 7.89) in 4b/5b/6b and 4i/5i/6i, respec-
tively. Hence, structure 6 is excluded. It is further excluded since
the NMR data for 4k/5k/6k show identical i-PrNHCO groups.
HSQC NMR spectra in THF-d8 indicate (by two N-H
couplings) that the acyclic 5i was obtained.
(c) IR Spectra. IR spectra of 5c, 5f, 5h, 5i, and 5l-o in
CHCl3 (Table S4 in the Supporting Information) display one
or two broad peaks at 3250-3357 cm-1, which are ascribed to
N-H absorptions,17 and most of the peaks appear at ca. 2880
or 2800 cm-1. CdO peaks for all compounds are at 1740-
1800 cm-1 (those for 3 are at 1780 and 1760 ppm)6a and 1680-
1730 cm-1. The weak CN absorptions observed at 2250 cm-1
for 5f and 5m are typical for CN systems substituted by strong
EWGs,18 but they are absent or very weak for most other
compounds. Only 5l, which is the only system with an
unsubstituted NH2 group, displays peaks at high wave numbers
of 3637 and 3553 cm-1, in addition to the other peaks, and no
compound has shown a peak >3360 cm-1, the range assigned
to O-H absorptions.17
Solvent Effect. Whereas the reaction of 1c with 3 in CHCl3
is complete in <1 min, at approximately the same concentra-
tions, the color disappeared in 2-3 min and after 17 min in
THF and in CH3CN, respectively, and was not completely
discharged in DMF-d7 even after 15 h. The percent enol of 1c
at equilibrium with 2c is 100, 89, 62, and 0 in CDCl3, THF-d8,
CD3CN, and DMF-d7, respectively. From the limited data, the
increase in solvent polarity reduces strongly the reaction rate
with 3. If the reaction intermediate is a polar aziridinium imide,8
we expect a rate increase for an ene reaction starting with two
neutral molecules. However, judging from literature data, this
is not the case. In the reaction of 3 with several alkenes, the
reactivity order in solvents CH2Cl2 > ClCH2CH2Cl > PhNO2
> C6H6 > EtOAc > THF with a 50-fold difference between
the extremes19 does not correlate with the solvent polarity. The
effect was tentatively ascribed to strong donor-acceptor interac-
The 13C NMR spectra in solution (Table S2 in the Supporting
Information) display the expected signals. The triazolidinyl
group behaves as a strong EWG, as deduced by the strong shift
of Câ to a lower field as compared to the amides 2. The major
difference expected between 4 and 5 (including 4i and 5i) is in
δ(CR), which is a carbonyl in 5i and an hemiaminal (C(OH)-
NHR) carbon in 4i. The adducts displayed three to five carbonyl
groups at 153.4-163.9 ppm. The two triazolidinedione CdO
groups are singlets <2 ppm apart at 153.6-156.1 ppm. In 3,
they are at 158 ppm.13 The CONRR′ carbonyls are at 155-161
ppm, and the ester carbonyls are at 158.6-165 ppm. The
differences are substituent- and solvent-dependent. More than
half of the amido carbonyls are doublets or multiplets with J
values of 2.5-4.0 ppm, due to coupling with the NH. The
RCONRR′ (R ) Alk, Ar) carbonyls resonate mostly at 160-
168 ppm (sometimes up to 171 ppm),14 in line with the δ values
in structures 5.
Hemiaminals are mostly unstable species that frequently
eliminate water to form an imine, but several of them are stable.
δ13C values for 14 R1R2(OH)NHR compounds (R1,R2 ) alkyl,
PhCH2 and R ) H, NH2, NMe2, N(c-(CH2)5) appear at 80.6-
86.9 ppm (75.4 ppm when R1 ) Me and R2 ) R ) H).15a The
values for δ(13C(RNH)(OH)(CF3)CO2Me) (R ) Ph, o-Tol, Ph2-
CH, PhCH2) with C(EWG)2, which are closer models for δ-
(13C)(4), are 84.5-87.3 ppm.15b We ascribe the unassigned 1H
NMR broad singlets at 4.31-4.66 and 4.66-2.46 ppm to the
OH and NH since they are affected by a temperature change:
δ(F3C13CH(OH)NRR′) is 84.5 ppm.15c For an hemiaminal Rh+
complex, δ(13C) is 84.6, δ(NH) is 5.16, and δ(OH) is 2.27
(16) El Mail, R.; Gerralda, M. A.; Hernandez, R.; Ibarlucea, L.; Pinilla,
E.; Rosario Torres, M. Organometallics 2000, 19, 5310.
(17) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric
Identification of Organic Compounds, 4th ed.; Wiley: New York, 1981;
pp 112-128.
(13) Zolfigol, M. A.; Ghorbani-Vaghei, R.; Mallakpour, S.; Chehardoli,
G.; Choghamarani, A. G.; Yazdi, A. H. Synthesis 2006, 1631.
(14) Kalinowski, H.-O.; Berger, S.; Braun, S. Carbon-13 NMR Spec-
troscopy; Wiley: New York, 1988; pp 209-214, 314.
(15) (a) Chudek, J. A.; Foster, R.; Young, D. J. Chem. Soc., Perkin Trans.
2 1985, 1285. (b) Dolensky, B.; Kvicala, J.; Paleta, O. J. Fluorine Chem.
2005, 126, 745. (c) Billard, T.; Langlois, B. R.; Blond, G. Eur. J. Org.
Chem. 2001, 1467.
(18) Juchnowski, I. N.; Binev, I. G. In The Chemistry of Functional
Groups, Supplement C: The Chemistry of Triple-Bonded Functional Groups;
Patai, S., Rappoport, Z., Eds.; Wiley: New York, 1983; Ch. 4, p 111.
(19) Ohashi, S.; Butler, G. B. J. Org. Chem. 1980, 45, 3472.
J. Org. Chem, Vol. 73, No. 1, 2008 187

Basheer and Rappoport
SCHEME 1. Suggested Mechanism for Formation of 5
tions between 3 and solvent. Greene et al.20 found the following
solvent effect on the ene reaction of 3 with tetramethylethyl-
ene: C6H6 > THF and CH2Cl2 > CH3CN, and from this and
literature data, they deduced a “lack of marked effect on solvent
polarity” and no correlation with ET. This was tentatively
ascribed to a diffuse charge distribution in the transition state
or to an unclarified solvent-solute interaction. A similar solvent
effect (CHCl3 > C6H6 > CH3CN) was found on the rates with
several CdC compounds.21 The solvent effect was ascribed to
the association of 3 with CHCl3, which decreases its LUMO
energy.
8. This can either undergo intramolecular charge recombination
to give the [2 + 2] adduct 4 or transfer a proton from the OH
group to the negatively charged nitrogen, either intramolecularly
or stepwise via 9, to give the 1,2 adduct 5 (Scheme 1, route a).
We note that an initially formed 4 can ring open in a
retroaldolization to give 5. The allowed photochemical variant
of the [2 + 2] cycloaddition, found by de Mayo26a and Eaton26c
gave in the reaction of cyclohexene with the enol of acetylac-
etone the bicyclic adduct, which was opened to give a CdO
group from the former OH. Similar ring openings were
recorded.26d A [2 + 2] adduct was considered as the intermediate
on the way to the open chain adduct in the [2 + 2] addition of
1,1,4,4-tetramethylbutadiene to 3,6c the [2 + 2] adduct of 3 with
adamantylideneadamantane was suggested to undergo the
reverse reaction,20 and the [2 + 2] adduct of 2-methylindene
Suggested Mechanism. Since [2 + 2] cycloadditions of
enophiles including 3 are well-documented,5 but thermal [π2s
+ 2s] reactions are symmetry forbidden,22 two step reactions
π
involving the intermediacy of aziridinium imides,20,23 radicals,24
or 1,4-dipoles25 were invoked. Similar polar two step processes
with a first step attack of the highly nucleophilic Câ of the
zwitterion 1 T 1′ on the enophile 3 will give a zwitterion 7 T
(24) Hall, J. H.; Bigard, W. E.; Farger, J. M.; Jones, M. L. J. Org. Chem.
1982, 47, 1459.
(25) (a) Turner, S. R.; Guibault, L. J.; Butler, G. B. J. Org. Chem. 1971,
36, 2838. (b) Wagener, B.; Butler, G. B. J. Org. Chem. 1973, 38, 3070. (c)
Hall, J. S.; Jones, M. L. J. Org. Chem. 1983, 48, 822. (d) Sepulveda-Arques,
J.; Simon, M. M. Tetrahedron Lett. 1985, 26, 6357.
(26) (a) De Mayo, P.; Takeshita, H.; Sattar, A. B. M. A. Proc. Chem.
Soc. 1962, 119. (b) de Mayo, P.; Takeshita, H. Can. J. Chem. 1963, 41,
440 and see: de Mayo, P. Acc. Chem. Res. 1971, 4, 7. (c) Eaton, P. Acc.
Chem. Res. 1968, 1, 50. (d) Carruthers, W. Cycloaddition Reactions in
Organic Chemistry; Pergamon: Oxford, 1990; p 353 and references therein.
(20) Cheng, C.-C.; Seymour, C. A.; Petti, M. A.; Greene, F. D.; Blount,
J. F. J. Org. Chem. 1984, 49, 2910.
(21) (a) Kinart, W. J.; Kinart, C. M. Phys. Chem. Liq. 2003, 14, 173.
(b) Kinart, W. J.; Kinart, C. M.; Tylak, I. Phys. Chem. Liq. 2002, 40, 405.
(22) Gilchrist, T. L.; Storr, R. C. Organic Reactions and Orbital
Symmetry, 2nd ed. 1979; Cambridge University Press: Cambridge, U.K.,
pp 173-212.
(23) Seymour, C. A.; Greene, F. D. J. Am. Chem. Soc. 1980, 102, 6385.
188 J. Org. Chem., Vol. 73, No. 1, 2008

Oxa-ene Reactions of Enols of Amides
gave the ene product on heating.27 The previous discussion on
the absence of NMR δ values of hemiaminals, or the absence
of a coalescence signal of 5 with the [2 + 2] adduct in our
system, exclude a significant concomitant concentrations of 4
and 5 in solution.
We view the formation of 5 as the oxa analogue of the
extensively studied ene reaction with 3, where an allylic
hydrogen is lost.8 In our variant, an enolic hydrogen is lost,
and an amide CdO bond is formed. The enol is suggested as
the alkene component (Scheme 1, route a) based on the faster
reaction of enols with a higher percentage of enols in the amide/
enol mixture (i.e., higher Kenol values) since at least half of the
reactants used are g98% enolic in CHCl3 or in the solid state.
The slower reaction of the Meldrum’s acid enol 1p is ascribed
to the loss of the very strong hydrogen bond, superimposed on
a possible steric retardation in the formation of 7.
However, due to the enols 1 h amides 2 equilibria in CHCl3,
it could be argued that 2 itself is the reagent, giving 5 via 2-
(Scheme 1, route b). The kinetic order will be the same in both
alternatives since the amide and the enol concentrations are
proportional, [2] ) [1]/Kenol. The methine carbon of 2 is acidic
due to its Y and Y′ and CONHR EWG substituents, and the
conjugate base 2- formed even in the absence of an added base
may be the nucleophile attacking 3, forming 9 directly and then
5. Alternatively, the proton formed by the ionization may first
protonate the nitrogen of 3, and 2- will then combine with 3‚
H+.28 Formation of 5o from 2o with 3 with no mechanistic
discussion was reported.11
Moreover, the slower reaction is that of 1o, whose equilibrium
percentage with 2a is only 7%.2a The lower reactivity of the
most enolic substrate 1p, by this route, is ascribed to the low
percentage of amide 2p and the relatively low nucleophilicity
of its highly stable anion, in addition to a contribution from
steric effects of the bulky anion of 1p.
ground state. Since 3 is also polar, and the transition state is
also zwitterionic, the lower rate in more polar solvents can be
accounted for by route a in Scheme 1 with more polar reactants
than the transition state. Moreover, the percent enol decreases
strongly in the more polar solvents,2 and this should reduce the
rate of route a (Scheme 1) in the more polar solvents. In contrast,
if the reaction proceed via route b in Scheme 1, both the
concentrations of the pre-equilibrium formed enolate ion and
the concentrations of the amide will be higher in a more polar
solvent, and the reaction rate will increase. This is in contrast
to the observation. Consequently, route a in Scheme 1 is more
consistent with the observed experimental data than route b.
The hydrogen transfer step can be either concerted with the
addition of 1 to 3 to form 7 or occur after it.29
Finally, is the aziridinium imide intermediate suggested in
the ene reactions of 38,20,23 relevant to the reaction of the enols
with 3? In contrast to most alkenes studied, which are not very
polar, the enols have definite positive and negative centers on
the formal alkene moiety. Whereas the interaction of the positive
nitrogen in the transition state with the δ-CYY′ carbon is
favorable, a simultaneous interaction with the positively polar-
ized δ+C(OH)NHR carbon is unfavorable. A reaction of the
electrophilic nitrogen of 3 with Câ of 1 to form 7 is favorable
(Scheme 1) with no need for interaction with CR.30
Conclusion
Reaction of enols of amides with PTAD gave adducts of
CYY′CONHR and H to the NdN bond in an oxa-ene reaction.
The suggested mechanism involves an initial nucleophilic attack
of the Y′YCδ- moiety of the enol to one PTAD nitrogen,
followed or coupled with proton transfer to the other nitrogen.
Experimental Section
Materials and General Methods. The enols were available from
our previous work.2 Compound 3 is commercially available. Melting
points are uncorrected. General methods are similar to those
reported.31
We attempted to probe this question. First, when the active
methylene compounds CH2YY′ (Y ) Y′ ) CO2Et, C4F9SO2
and Y ) CN, Y′ ) CO2Me, CONHR, and R ) i-Pr, t-Bu) were
added to 3 in CHCl3, the color did not disappear overnight.
Since the concentrations of the derived anions in the neutral
solvent may be too low to give a sufficiently rapid reaction,
we added Et3N to the reaction mixture of the slow 1o with 3.
The red color disappeared rapidly, but a control experiment
showed that the reaction of Et3N with 3 alone resulted in a rapid
disappearance of color. Second, when 0.5-0.05 mL of AcOH
was added to a mixture of 1c with 3 in CHCl3, the rapid color
disappearance was not slowed down, indicating that protonation
of the anion of 1c, if any, does not reduce its concentration to
affect sufficiently its reaction rate with 3. A control experiment
showed no apparent reaction of AcOH with 3.
The strong decrease of the reaction rate with the increased
solvent polarity is a better probe. As mentioned previously, the
ene reactions of 3 showed an irregular change of the rate with
solvent polarity, which was explained by a specific solvent-
solute interaction. Comparison of the polarities of the precursors
and the transition state of the reaction only suggested that the
latter is not highly polar, despite the polar intermediates that
follow it. Our system is more complicated since the enols are
highly polar with a significant zwitterionic contribution to the
Solid State NMR Spectroscopy. Solid state NMR spectroscopy
with MAS was carried out on a 300 MHz spectrometer operating
at 75.4 MHz for 13C and employing a triple-resonance probe
equipped with a 5 mm o.d. spinning module. The sample (78.6
mg) was packed into a high performance zirconium rotor and the
MAS frequency was maintained at 10.000 ( 2 Hz. Chemical shifts
were referenced to an external secondary standard ((2-13C)-Gly )
44 ppm). The contact time for 1H and 13C Hartman-Hahn matched
fields (50 kHz) was 3 ms. Refocused 13C spectra were acquired,
with delays set to one rotor period (100 µs) on each side of the
180° 13C pulse. High power dipolar coupling (DD) on the proton
channel (ca. 100 kHz) was used following the CP excitation.
Acquisition was of 6000 points with a dwell time of 25 µs. The
signal was averaged over 3072 transients with a repetition delay
of 3 s. The FID was Fourier transformed with an exponential
window function of 5 Hz.
General Procedure for Reaction of Enols of Amides with
4-Phenyl-1,2,4-triazoline-3,5-dione. The procedure used with
minor modifications regarding the reaction rate (deduced from the
disappearance of the red color of 3 and the appearance of the
precipitate) and the yields is demonstrated for 3-hydroxy-3-
isopropylamino-2-dimethylaminocarbonylacrylonitrile. When a so-
(29) For a stepwise mechanism in the ene reaction of 3, see: Elemes,
Y.; Foote, C. S. J. Am. Chem. Soc. 1992, 114, 6044.
(30) For aziridinium imides as “innocent bystanders” in the ene reaction,
see: Singleton, D. A.; Hang, C. J. Am. Chem. Soc. 1999, 121, 11885.
(31) Frey, J.; Rappoport, Z. J. Am. Chem. Soc. 1996, 118, 5169.
(27) Smonou, I.; Khan, S.; Foote, C. S.; Elemes, Y.; Mavridis, I. M.;
Pantidou, A.; Orfanopoulos, M. J. Am. Chem. Soc. 1995, 117, 7081.
(28) We note that if the reaction proceeds via 2-, 2- can also be formed
initially from 1.
J. Org. Chem, Vol. 73, No. 1, 2008 189

Basheer and Rappoport
lution of 1i (197 mg, 1 mmol) and 4-phenyl-1,2,4-triazoline-3,5-
dione (175 mg, 1 mmol) in chloroform (4 mL) was stirred for 1
min, the red color disappeared. The mixture was kept at room
temperature overnight, and the white precipitate that formed was
filtered and dried in air affording 320 mg (79%) of 5i, mp 182
-3 °C. Anal. calcd for C17H20N6O4: C, 54.84; H, 5.38; N, 22.58
Found: C, 54.72; H, 5.44; N, 22.77. Suitable crystals for X-ray
diffraction were obtained by dissolving crude 5i in THF and slow
evaporation of the solvent at room temperature.
Control Experiments. (a) No reaction was observed when
several active methylene compounds (CH2YY′, Y ) Y′ ) CO2-
CH2CCl3; Y ) CN, Y′ ) CO2Me; Y ) CN, Y′ ) CONHPr-i,
CONHBu-t (1.0 mmol)) were reacted with 3 (0.96 mmol) in CHCl3.
The red color persisted for >35 h, and the precursors were re-
isolated.
(b) To a solution of 3 (212 mg, 0.12 mmol) in CHCl3 (2 mL),
two drops of triethylamine (ca. 0.3 mmol) were added. The red
color disappeared immediately.
(c) When 1d (0.125 mmol) was reacted with 3 (0.12 mmol) in
CHCl3 (2 mL) in the presence of excess acetic acid (0.5 mL), the
red color disappeared immediately.
(d) To a solution of 3 (212 mg, 0.12 mmol) in CHCl3 (2 mL),
one to 10 drops of glacial acetic acid were added. The red color
persisted for >35 h.
1H NMR (CDCl3, 298 K, 400 MHz) δ: 1.17 (6H, d, J ) 6.5
Hz); 3.08 (3H, s); 3.33 (3H, s); 4.10 (1H, octet, J ) 6.8 Hz); 6.61
(d, J ) 7.7 Hz); 7.38-7.45 (m); 8.41 (s, br). 13C NMR (CDCl3,
298 K, 400 MHz) δ: 21.65 (q, J ) 1227.6 Hz); 21.82 (q, J )
126.5 Hz); 38.39 (q, J ) 143.3 Hz); 38.49 (q, J ) 143.8 Hz); 43.7
(d, J ) 140.3 Hz); 71.1 (s); 111.90 (s); 125.7 (d, J ) 164.4 Hz);
128.95 (dt, Jd ) 162.0 Hz, Jt ) 7.3 Hz); 129.22 (dd, J1 ) 163.1
Hz, J2 ) 5.6 Hz); 130.1 (m); 153.4(s); 154.3 (s); 157.2 (s); 158.9
1
Acknowledgment. We are indebted to the Israel Science
Foundation for support, to Prof. Asher Schmidt and Dr. Yael
Balazs for the solid state NMR spectra, to Dr. Shmuel Cohen
for the X-ray structure determination, to Dr. R. Hoffmann and
Prof. M. Orfanopoulos for discussions, and to Dr. Y. X. Lei
for a preliminary experiment.
(m, J ) 3.2 Hz, CdO). H NMR (THF-d8, 298 K, 400 MHz) δ:
1.15 (3H, d, J ) 6.4 Hz); 1.17 (3H, d, J ) 4.6 Hz); 2.99 (3H, s);
3.27 (3H, s); 4.05 (1H, octet, 6.8 Hz); 7.33 (1H, t, J ) 7.3 Hz);
7.42 (2H, t, J ) 7.7 Hz); 7.52 (2H, d, J ) 7.7 Hz); 7.60 (d, J )
8.1 Hz); 9.52 (s). 13C NMR (THF-d8, 298 K, 400 MHz)δ: 20.9 (q,
J ) 126.6 Hz); 36.91 (q, J ) 138.0 Hz); 37.39 (q, J ) 138.1 Hz);
42.9 (d, J ) 140.7 Hz); 70.7 (s); 112.7 (s); 125.3 (dt, Jd ) 163.8
Hz, Jt ) 6.2 Hz); 127.7 (dt, Jd ) 161.7 Hz, Jt ) 7.5 Hz); 128.4
(dd, J1 ) 162.3 Hz, J2 ) 8.1 Hz); 131.7 (s); 153.5 (s); 155.0 (s);
157.99 (d, J ) 3.2 Hz); 158.51 (m, J ) 3.5 Hz).
Supporting Information Available: CIF and ORTEP of 5i,
Tables S1 and S2 of 1H and 13C NMR of all adducts, Table S3 of
comparison of 13C NMR of amides 1 and adducts 5, Table S4 of
IR spectra of 5, and Table S5 of analytical data. This material is
available free of charge via the Internet at http://pubs.acs.org.
The full spectral data for 5a-p in several solvents are given in
Tables S1 and S2 of the Supporting Information. Melting points,
yields, and microanalyses are given in Table S5 of the Supporting
Information.
JO702083U
190 J. Org. Chem., Vol. 73, No. 1, 2008