Solvent-dependent competitive rearrangements of cyclic tertiary propargylamine N-oxides

Article abstract of DOI:10.1002/ejoc.200300461

In protic media, cyclic propargylamine N-oxides 1 undergo solvent-dependent competitive rearrangements leading to enamino aldehydes 5, acrylamides 3, and secondary amines 4. The ratios of the products are evaluated and the possible mechanism of the compet

Full text of DOI:10.1002/ejoc.200300461

FULL PAPER
Solvent-Dependent Competitive Rearrangements of Cyclic Tertiary
Propargylamine N-Oxides
[a]
[a]
[a]
[a]
´
´
´
´
´
´
Anna Szabo, Agnes Galambos-Farago, Zoltan Mucsi, Geza Timari,
[a]
[a]
´
´
Lelle Vasvari-Debreczy, and Istvan Hermecz*
Keywords: Ab initio calculations / Reaction mechanisms / Rearrangements / Solvent effects
In protic media, cyclic propargylamine N-oxides 1 undergo
solvent-dependent competitive rearrangements leading to
enamino aldehydes 5, acrylamides 3, and secondary amines
4. The ratios of the products are evaluated and the possible
mechanism of the competing reactions is interpreted by reac-
tion kinetics studies and ab initio calculations.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
rivatives[5] and benzylamine derivatives (e.g. pargylin).[6]
The O-allenylhydroxylamine product of the benzylamine
derivative is known to undergo a further sigmatropic [1,5]-
H shift rearrangement[6] (Scheme 1).
Certain propargyl-substituted tertiary amines are
irreversible monoamine-oxidase (MAO) inhibitors.[1] One
major metabolic degradation path of tertiary amines starts
with the oxidation of the tertiary nitrogen atom.[2] This can
be followed by enzymatic and chemical transformations.[3]
The chemistry of propargylamine N-oxides in protic media
N-Oxides having a hydrogen atom in a position that is β
to the nitrogen atom may undergo the characteristic Cope
elimination.[7]
We recently described a novel rearrangement of pargylin
is little known. It was our aim to study these transform- N-oxide (I: R1 ϭ Bn; R2 ϭ Me) in protic medium, leading
ations.
to a new enamino aldehyde with the appearance of the sec-
The Meisenheimer[2,3] sigmatropic rearrangement[4] of ondary amine and a new acrylamide-type product in the
tertiary propargylamine N-oxides I in aprotic media to give reaction mixture.[8] We suggested possible mechanisms for
O-allenylhydroxylamines has been described for cyclic de- these transformations: the enamino aldehyde major product
Scheme 1. Possible rearrangements of tertiary propargylamine N-oxides
[a]
Chinoin Pharmaceutical and Chemical Works Ltd.
is formed by an ionic mechanism through an isoxazolinium-
type intermediate state (route A), while the acrylamide by-
product results from the Meisenheimer rearrangement
´
1045 Budapest, To utca 1Ϫ5, Hungary
Fax: (internat.) ϩ 36-1-370-5597
E-mail: istvan.hermecz@sanofi-synthelabo.com
Eur. J. Org. Chem. 2004, 687Ϫ694
DOI: 10.1002/ejoc.200300461
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
687

I. Hermecz et al.
FULL PAPER
(route B) by further transformation of the primarily formed elimination, which is sterically hindered in these struc-
O-allenylhydroxylamine.
tures.[9] Further transformation of the O-allenylhydroxyla-
We have now extended these transformations to cyclic mine Meisenheimer product by a [1,5]-H shift is also ex-
propargylamine N-oxides where R1 and R2 form a 5- or 6- cluded in these structures because of the absence of suitable
membered ring, i.e. morpholine, piperidine, and pyrrolidine protons. Evidence from preparative experiments was sought
derivatives 1a, 1b, and 1c. The aim was to exclude the Cope for the route of formation of the secondary amine and acry-
Scheme 2
Scheme 3. Competing reaction routes of cyclic tertiary propargylamine N-oxides
Table 1. Distribution of the products obtained from N-oxide 1a and rate constants calculated for route A at 65 °C in alcohols with
different acities
Solvent
Anion/cation-solvating ability
Rate constant 104 k [sϪ1
]
Product [mol %], route
Acity
Basity
5a, A
3a,B1
4a,B2
7a,C
Others
MeOH
EtOH
nPrOH
iPrOH
tBuOH
0.75
0.66
0.63
0.59
0.45
0.50
0.45
0.44
0.44
0.50
0.78
1.14
1.47
2.16
3.31
33
47
42
57
69
11
14
14
11
8
29
17
15
13
11
6
5
8
14
8
21
17
21
5
4
Table 2. Distribution of the products from 1a in tBuOH at different temperatures and activation parameters calculated for route A
T
[°C]
Rate constant
∆H#
[kJ/mol]
∆S#
[J/molK]
Products [mol %], route
104 k [sϪ1
]
5a, A
3a, B1
4a, B2
7a, C
Others
50
65
79
0.60
3.31
15.0
98.62
Ϫ20.6
71
69
65
2
8
9
17
11
14
4
8
9
6
4
3
688
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 687Ϫ694

Solvent-Dependent Competitive Rearrangements of Cyclic Tertiary Propargylamine N-Oxides
FULL PAPER
lamide byproducts. The effects of the reaction conditions in agreement with our supposition that the rearrangement
on the rates of the competing rearrangements were studied. involves a cyclic (isoxazoline ring) transition state.
Reaction kinetics measurements were expected to yield in-
From ∆H# and ∆S# we have also calculated the free en-
formation on the mechanism of the rearrangements. Exper- ergy of activation (∆G#) for the reaction in tert-butyl al-
imental results were evaluated in the light of the ab initio cohol at 65 °C (105.6 kJ/mol). From that value and the re-
calculations we recently performed on these reactions.
action rates in various alcohols at 65 °C, the ∆G# values of
the reactions in various alcohols were calculated (see
Table 3).
Results and Discussion
Table 3. ∆G#B Ϫ ∆G#A and ∆G#AB values determined for reactions
performed in alcohols with different acities
Kinetics Studies
The rearrangements of cyclic propargylamine N-oxides
1aϪc were effected in alcohols; the reactions were moni-
tored by GC. The compositions of the mixtures were deter-
mined by using reference samples of the expected products
(all are known compounds) synthesized independently. Fi-
nal reaction mixtures contained the expected enamino alde-
hyde 5, acrylamide 3, and secondary amine 4 in all three
cases, together with the corresponding tertiary (parent) am-
ine 7 and byproducts. The O-allenylhydroxylamine 2 (the
Meisenheimer rearrangement product) appeared as an in-
termediate.
Solvent
Acity
∆G#AB
[kJ/mol]
∆G#B Ϫ ∆G#A
[kJ/mol]
ab initio
ab initio
MeOH
EtOH
nPrOH
iPrOH
tBuOH
0.75
0.66
0.63
0.59
0.45
109.7
108.6
107.9
106.8
105.6
107.1
Ϫ0.540
1.169
1.040
2.430
3.624
Ϫ14.400[a]
[a]
The correlation between the experimental and theoretically cal-
culated results is not sufficient and comes from the model contain-
ing only two molecules of MeOH.
A systematic study of the rearrangement was performed
on the morpholino model 1a.
To distinguish the rearrangement routes, we first pre-
pared the O-allenylhydroxylamine Meisenheimer product in
an aprotic medium (Et2O) and investigated its further trans-
Interpretation of the Results in the Light of Theoretical
Calculations
The results obtained from preparative experiments and
formations in a protic medium. In ethanol, the O-allenylhy- reaction kinetics measurements left several unanswered
droxylamine 2a gave exclusively the morpholine 4a and the questions. For example, the question of whether the two
acrylamide 3a. In response to base (triethylamine) addition, competitive reactions (routes A and B) proceed through dif-
the amount of the acrylamide increased (Scheme 2). The ferent transition states (TS) or pass through a common,
base is thought to abstract a proton from the oxaziridine rate-determining TS to a common intermediate from which
intermediate[8] (Scheme 1). These results led us to ascribe different products may arise in subsequent fast steps (see
the acrylamide 3 and secondary amine 4 to the Mei- Scheme 3) controlled by different experimental conditions.
senheimer route (B) for all three N-oxides 1aϪc.
To understand the energetic difference between the com-
Next, we investigated how the solvating abilities of the petitive routes A and B, theoretical methods, HF, MP2 and
solvents affected the proportions of the Meisenheimer route DFT (B3LYP) were applied.[12] The mechanism of the reac-
(B) and our novel rearrangement route (A). Reactions of 1a tions was studied on the morpholino model 3a, in vacuo Ϫ
were carried out in a series of alcohols with significantly mimicking the aprotic solvent with an isolated molecule Ϫ
different anion-solvating abilities (acity), but similar cation- and in the presence of two molecules of solvent (MeOH) Ϫ
solvating abilities (basity).[10] The formation of enamino al- required for proton transfer and simulating protic media.
dehyde 5 was followed [transformation of the starting com- Energy profiles based on these ab initio calculations are
pound 1 and the formation of 2 (route B) could not be presented in Schemes 4 and 5.
followed because of the instability of these compounds un-
In the aprotic medium model (Scheme 4), starting from
der GC conditions]. Rates were clearly of first order and the gauche conformation of 1a, the O1 and C5 atoms ap-
the rate constants were sensitive to the solvent (Table 1). proach each other and the forthcoming reaction passes
With decreasing anion-solvating abilities of the solvent, the through only one TS (10a: ∆G# ϭ ϩ67.4 kJ/mol, ∆S#
ϭ
participation of route A and the rate of the reaction in- Ϫ15.9 J/molK), which leads to only one final product, the
creased. O-allenylhydroxylamine derivative 2a (route BЈ) (Scheme 3).
For the same model (1a), we determined the rate constant We emphasise that under aprotic conditions the Mei-
at three different temperatures in tert-butyl alcohol. From senheimer rearrangement does not involve any intermedi-
the temperature dependence of the rate constants, we calcu- ate; it has a real [2,3] sigmatropic mechanism, in agreement
lated (according to Eyring[11]) the activation parameters, with the literature.
∆H# and ∆S# of the formation of enamino aldehyde 5a
(Table 2).
In the protic medium model, on the contrary, we have
found that a hydrogen bond stabilized the isoxazolinium-
The small negative entropy of activation indicates a con- type zwitterionic intermediate (see 8a in Scheme 5) for both
siderable decrease in randomness in the activated complex, transformations with the same TS (10a: ∆GA#B ϭ ϩ107.1 kJ/
Eur. J. Org. Chem. 2004, 687Ϫ694
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
689

I. Hermecz et al.
FULL PAPER
served between the H3 atom and the oxygen atom of the
second MeOH molecule.
For the stable intermediate 8a, two reaction routes are
open: without protonation it can follow route B, giving 2a
through TS 11a (calcd. ∆G# ϭ 2.2 kJ/mol); or it follows
route A, giving 5a. Route A can be realized by protonation
of the C4 atom of 8a by the first MeOH molecule, through
TS (12a) (calcd. ∆G# ϭ 16.6 kJ/mol). An unstable pro-
tonated intermediate 9a was recognized during the pro-
tonation reaction which, after proton abstraction, gives the
final enamino aldehyde product 5a. In a protic medium,
according to the ab initio calculations, the rate-determining
step of the reaction sequence is the cyclisation of the N-
oxide 1a to give the zwitterionic isoxazolinium intermediate
8a. Consequently, the activation parameters determined by
us in the reaction kinetics measurements for route A, apply
Scheme 4. Energy profile of route BЈ in aprotic medium, based on to the TS involved in the formation of the common inter-
ab initio calculations
mediate (8a in this model). The free energy of activation
(∆GA#B) determined for the methanolic reaction by kinetics
mol). In the stable intermediateϪsolvent complex 11a the measurements and obtained by ab initio calculations using
˚
first MeOH molecule forms a strong (1.80 A) hydrogen the model with two molecules of MeOH are in very good
bond with the nonbonding electron pair of the negatively agreement (see Table 3). In this suggested mechanism, the
charged C4 atom, and the second MeOH molecule is linked ratio of the products arising from routes A and B is finally
˚
to the first one. A very weak hydrogen bond (2.4 A) is ob- determined by the ratio of the fast parallel reactions of
Scheme 5. Energy profile of the competing reactions in protic medium based on ab initio calculations
690
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 687Ϫ694

Solvent-Dependent Competitive Rearrangements of Cyclic Tertiary Propargylamine N-Oxides
FULL PAPER
routes A and B, proceeding from this common intermedi- solvent than is ∆GB#, leading to the Meisenheimer product
ate 8a.
(route B). These results can be interpreted in terms that the
transition state of route A (12a) approaches a protonated
isoxazoline structure which no longer has an anionic
character; thus, a change in the anion-solvating ability of
the alcohol has less effect on the energy levels of TS 12a
than on the common zwitterionic intermediate 8a, contrary
the route B, where protonation does not occur.
The study of N-oxides 1b and 1c in ethanol and tert-
butyl alcohol revealed a similar effect of the solvent on the
product distribution as for 1a (Table 4).
Discussion of the Solvent Dependence of the Competing
Reactions
From the ratio of products arising from routes A and B
at a given temperature (Table 1) we determined the differ-
ence between the free energies of activation (∆GB# Ϫ ∆GA#)
of the competing routes in various alcohols at 65 °C, and
the free energy of activation (∆GA#B) of the rate-determining
step of these reactions (Table 3). Results indicate significant
solvent dependence not only in reaction rates, but also in
the product distribution.
Table 4. Product distribution in the rearrangements of 1aϪc
In Scheme 6, using a schematic energy level diagram, we
explain the above solvent dependencies, illustrating the en-
ergy levels of the starting and transition states in methanol
and in tert-butyl alcohol.
N-Oxide Solvent Temp.
[°C]
Products [mol %], route
5, A 3, B1 4, B2 7, C Others
1a
1b
1c
EtOH
tBuOH 65
EtOH 79
tBuOH 65
EtOH 79
tBuOH 65
79
48
69
35
61
44
74
13
8
25
14
17
4
14
11
13
14
25
16
5
8
4
9
3
3
20
4
23
2
11
3
In the rate-determining step (ring closure of the N-oxide
1a to TS 10a), the difference in the energy levels of the
starting state in methanol and in tert-butyl alcohol must be
larger than that in the transition state, where the hydrogen
bonds are weaker. This may explain the acceleration of the
reaction when changing the solvent from methanol to tert-
butyl alcohol.
In the product-distribution step, the value of ∆GB# Ϫ ∆GA#
The results in Tables 3 and 4 support experimentally the
increases with decreasing anion-solvating ability (acity) of structures estimated by ab initio calculations for the com-
the alcohol (see Table 3). The free energy of activation of mon intermediate 8 and for the transition states 10, 11, 12
8a to TS 12a (∆GA#), which leads to the enamino aldehyde involved in the rate-determining and product distribution
(route A), seems to be more dependent on the acity of the determining steps of the rearrangements.
Scheme 6. Schematic energy level diagram: solvent dependence of the rate-determining step and of the product distribution determining
steps of the rearrangements
Eur. J. Org. Chem. 2004, 687Ϫ694
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
691

I. Hermecz et al.
FULL PAPER
Effect of Bases
tion of the UV-active enamino aldehyde 5 and acrylamide
3 was followed by HPLC (Table 6).
The role of the base in these transformations was investi-
gated on the morpholino model 1a. We found that a non-
nucleophilic base shifted the ratio of the two transform-
ations towards the Meisenheimer route and, within this,
towards acrylamide formation (Table 5, Entries 2, 3). A
small amount of acid (e.g. 0.2 equiv. HCl) exerted the op-
posite effect (Entry 4).
Table 6. Enamino aldehyde 5 and acrylamide 3 formed from N-
oxides 1aϪc as a function of the pH
N-Oxide
pH ϭ 7
pH ϭ 11
5
3
5
3
Morpholine
Piperidine
Pyrrolidine
1a
1b
1c
59
51
48
0
0
0
15
12
23
8
8
4
Table 5. Effect of the base on the rearrangement of 1a (79 °C)
Entry Reaction medium
5a, A 3a, B1
Product [mol %], route
4a, B2 7a, C Others
The ratio of the enamino aldehyde decreased with in-
creasing pH, which can be ascribed to its hydrolysis at a
higher pH. The acrylamide was not detected in pH ϭ 7
buffer, whereas it did appear at pH ϭ 11. This is in agree-
ment with the results obtained in ethanol with triethylamine
(Table 5), where the base helped the formation of acrylam-
ide 3a.
Unlike the alcohol solutions, in aqueous buffered media
there are little differences between the results obtained for
the six-membered models 1a and 1b.
Beside rearrangement routes A and B, the loss of oxygen
(route C) also occurred, to various degrees, in the trans-
formation of the N-oxides, yielding the tertiary amine 7.
Route C is more pronounced in the presence of triethyl-
amine and in isopropyl alcohol, probably due to the oxidiz-
ability of these materials.
1
2
3
4
5
EtOH
48
25
12
63
1
13
24
42
2
14
16
21
14
3
5 20
12 23
14 11
EtOH ϩ0.2 equiv. NEt3
EtOH ϩ1 equiv. NEt3
EtOH ϩ0.2 equiv. HCl
EtOH ϩ0.2 equiv. NaOEt
3
0
18
90
6
The above results support our assumption: the negative
effect of bases (and the positive effect of acids) can be ex-
plained by the protonation step required to produce the
isoxazolinium cation intermediate. If this protonation is
slowed down, less isoxazolinium (and less enamino alde-
hyde) is formed; if it is promoted (by acid), the opposite is
the case.
Rearrangements depend on the negatively charged oxy-
gen atom of the N-oxide to attack the triple bond as a nu-
cleophile; thus, the N-oxide hydrochloride salts (1·HCl) re-
main intact in alcoholic solution.[8] Excess of a nucleophilic
base forces both rearrangements back drastically and brings
about other reactions (Entry 5) not discussed here. Among
others, product 6 was identified by GC-MS. The acrylamide
5 under similar conditions did not furnish 6.
Conclusion
The present results indicate that the recently revealed[8]
Results obtained for the six-membered ring models 1a rearrangement of propargylamine N-oxides I (R1 ϭ Bn;
and 1b (see Table 4) can also be explained by the effect of R2 ϭ Me), leading to enamino aldehyde (route A), is a sig-
bases: from 1b more acrylamide 3b is formed than from 1a, nificant transformation path, also for cyclic propargylamine
because of the more basic reaction medium, due to the N-oxides 1aϪc exclusively in protic media.[8] Thus, the
more basic character of the piperidine derivatives.
scope of this rearrangement can be extended to N-oxides I
We investigated the transformation of N-oxides 1aϪc in where R1 and R2 form a morpholine, piperidine or pyrroli-
aqueous buffered media at pH ϭ 7 and 11. Only the forma- dine ring (Scheme 1).
Scheme 7
692
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 687Ϫ694

Solvent-Dependent Competitive Rearrangements of Cyclic Tertiary Propargylamine N-Oxides
FULL PAPER
H), 4.04 (t, J ϭ 2.4 Hz, 1 H), 4.93 (d, J ϭ 2.4 Hz, 2 H), 13.0 (1 H)
ppm. C7H11NO2·HCl (177.63): calcd. C 47.33, H 6.81, N 7.89, Cl
19.96; found C 47.42, H 6.82, N 7.81, Cl 20.05.
The mechanism of the Meisenheimer rearrangement is
different in protic and aprotic medium. In an aprotic me-
dium, the [2,3] sigmatropic mechanism (route BЈ) is prob-
able, while in a protic medium, according to ab initio stud-
ies, both routes A and B pass through a common isoxazoli-
1b: White powder, 75% yield, m.p. 150Ϫ151 °C. IR (KBr): ν˜ ϭ
2124 cmϪ1 1H NMR ([D6]DMSO): δ ϭ 1.36Ϫ1.64 (m, 2 H),
.
nium-type intermediate 8 formed in the rate-determining 1.79Ϫ1.92 (m, 4 H), 3.66Ϫ3.83 (m, 4 H), 4.01 (t, J ϭ 2.4 Hz, 1 H),
4.81 (d, J ϭ 2.4 Hz, 2 H), 12.6 (1 H) ppm. C8H13NO·HCl (175.66):
calcd. C 54.70, H 8.03, N 7.97, Cl 20.18; found C 54.62, H 7.92,
N 7.96, Cl 20.25.
step. This common intermediate may transform, in the
product ratio determining steps, into the O-allenylhydroxyl-
amine Meisenheimer product 2 or the enamino aldehyde
product 5 (Scheme 7). In further transformations, the Mei-
senheimer product 2 partly degrades into the secondary am-
ine 4 and partly rearranges into the acrylamide 3.
Reaction kinetics measurements in alcohols with differ-
ent anion-dissolving abilities, in agreement with ab initio
studies, afforded information on the structure of the tran-
sition state involved in the rate-determining step and on
those involved in the product distribution determining steps
of the competing routes, supporting our assumption on the
mechanism of these reactions.
1c: White powder, 72% yield, m.p. 98Ϫ99 °C. IR (KBr): ν˜ ϭ 2129
cmϪ1. 1H NMR ([D6]DMSO): δ ϭ 2.12 (m, 4 H), 3.81Ϫ3.96 (m, 4
H), 3.76 (t, J ϭ 2.8 Hz, 1 H), 4.89 (d, J ϭ 2.8 Hz, 2 H), 12.7 (1 H)
ppm. C7H11NO·HCl (161.63) calcd. C 52.02, H 7.48, N 8.67, Cl
21.93; found C 49.95, H 7.59, N, 8.60, Cl 22.06.
Transformation of the O-Allenylhydroxylamine 2a in Alcoholic Me-
dium: The O-allenylhydroxylamine 2a (2 mmol) was refluxed for
1 h in ethanol (10 mL) containing NEt3 (2 mmol). The composition
of the reaction mixture was determined by GC.
General Procedure for the Rearrangement of the N-Oxides 1a؊c in
Alcoholic Media: The hydrochloride salt of the N-oxide 1aϪc
(1 mmol) was dissolved/suspended in the appropriate alcohol
(10 mL). The N-oxide base was liberated by the addition of 1 mmol
of NaOEt. The reaction mixture was stirred at the given tempera-
ture until the N-oxide was completely transformed (as followed by
TLC). Experiments investigating the effect of the base (Table 5):
Base 1a (1 mmol) was refluxed in ethanol (10 mL) in the presence
of the appropriate amount of NEt3 or NaOEt, except for Entry 4
of Table 5, where 1 mmol of 1a·HCl was made to react in the pres-
ence of 0.8 mmol of NaOEt. The compositions of the reaction mix-
tures were determined by GC (Scheme 8, Table 7). The amounts of
the acrylamide 3 and secondary amine 4 have been corrected in
each case with the amount of the Michael adduct 13 formed from
them; see Tables 1Ϫ5.
Experimental Section
General: Compounds 1a and 2a were prepared according to litera-
ture procedures without any significant modification.[5a] Com-
pounds 7aϪb[13] and 7c[14] were obtained by propargylation of
4aϪc, compounds 5a,[15] 5b,[16] 5c[17] by Michael addition of 4aϪc
to propargyl aldehyde, compounds 3a,[18] 3b,[19] 3c[20] by acylation
of 4aϪc by acryloyl chloride. All other reagents and solvents were
used as obtained from commercial sources without further purifi-
cation. Melting points were determined with a Büchi 535 apparatus
and are uncorrected. IR spectra were recorded with a Bruker IFS-
28 instrument. Elemental analyses were performed with a Carlo
Erba Mod 1106 Analyser. NMR spectra were recorded with a
Bruker DRX-400 spectrometer operating at 400 MHz; chemical
shifts (δ) are given in ppm and were measured with reference to
residual solvent peaks. Gas-phase chromatograms of the rearrange-
ment reaction mixtures were recorded with an HP-6890 chromato-
graph equipped with a flame ionization detector under the follow-
ing conditions: RTX-1301 capillary column, helium as the vector
gas (initial flow: 0.6 mL/min), temperature of injector 250 °C and
detector 280 °C, temperature of the oven: isothermal 100 °C for
15 min, then 100Ϫ250 °C (10 °C/min), then 250 °C for 30 min.
HPLC analyses were performed with a Waters Alliance 2690 chro-
matograph equipped with a 996 PDA detector, column: µBonda-
Pak C18 10µm, mobile phase: 100 m (NH4)3-phosphate/aceto-
nitrile (8:2) pH ϭ 7.5, flow rate: 1.0 mL/min.
Scheme 8
Crossover Procedures: The acrylamide 3a (1 mmol) was dissolved
in ethanol (10 mL) and refluxed for 3 h in the presence of NaOEt
(1 mmol). No sign of the formation of 6 was detected by GC, indi-
cating that 6 is not formed from 3a by Michael addition. The acryl-
amide 3a (1 mmol), dissolved in ethanol (10 mL), was refluxed for
3 h in the presence of morpholine (1 mmol). The formation of 13
was confirmed by GC. Conversion by GC: 64%.
N-Oxide Hydrochloride Salts 1a؊c: To a stirred solution of 7aϪc
(20 mmol) in chloroform (10 mL) 3-chloroperbenzoic acid (7.6 g,
55 mmol) in chloroform (50 mL) was added dropwise at 0Ϫ5 °C
over a period of 30 min. After 2 h of stirring below 5 °C, a satu-
rated hydrogen chloride solution in diethyl ether (10 mL) was ad-
ded dropwise and the reaction mixture was stirred for 1 h. The
precipitated 3-chlorobenzoic acid was filtered off, the filtrate was
concentrated to dryness in vacuo at 30 °C. The oily residue was
triturated with diethyl ether (50 mL ϩ 2 ϫ 30 mL). The crystalline
product was filtered off, washed with diethyl ether, and recrystal-
lised from ethanol.
General Procedure for the Rearrangements of the N-Oxides 1a؊c in
Aqueous Buffered Media: The hydrochloride salt of the N-oxide
1aϪc (1 mmol) was dissolved in 1  NaOH solution (1 mL), then
it was diluted to 20 mL with the appropriate buffer. The reaction
mixture was stirred at 65 °C until the N-oxide had disappeared.
The amounts of the enamino aldehyde 5aϪc and acrylamide 3aϪc
1a: White powder, 86% yield, m.p. 138Ϫ139 °C. IR (KBr): ν˜ ϭ products were determined by HPLC from the aqueous phase. The
1
2128 cmϪ1. H NMR ([D6]DMSO): δ ϭ 3.82 (m, 4 H), 3.95 (m, 4 amount of the acrylamide was corrected with the amount of the
Eur. J. Org. Chem. 2004, 687Ϫ694
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
693

I. Hermecz et al.
FULL PAPER
Table 7. GC retention times of the products
N-Oxide
Product/Rt [min]
1
Products with reference standards
Products identified by GC-MS
4
7
3
5
2
6
13
a
b
c
12.5
10.1
7.9
22.0
20.6
16.9
29.2
29.3
28.8
35.6
36.3
35.0
26.6
26.2
24.7
29.5
28.8
27.5
46.7
44.3
40.0
[4]
R. A. W. Johnstone, Mech. Mol. Migr. 1969, 2, 249Ϫ266.
[5a] A. H. Khuthier, M. A. Al-Iraqi, J. Chem. Soc., Chem. Com-
Michael adduct 13aϪc as measured by GC from the chloroform
extract of the cooled reaction mixture; see Table 6.
[5]
mun. 1979, 9Ϫ10. [5b] J. C. Craig, N. N. Ekwuribe, L. D. Gruen-
[5c]
Reaction Kinetics Procedures: The hydrochloride salt of the N-oxide
of 1a (1 mmol) was dissolved/suspended in the solvent (10 mL) and
the free N-oxide base was liberated in situ by 1 mmol of NaOEt.
The reaction mixture was stirred at the given temperature (main-
tained constant within Ϯ0.2 °C). The kinetics of the N-oxide re-
arrangement leading to the enamine aldehyde product 5a was stud-
ied by monitoring the formation of 5a, injecting the cooled samples
from the reaction mixture into the GC. In all the cases, the value
of Aϱ was determined experimentally for each run by leaving the
reaction mixture at the specified temperature until there was no
further change in the concentration of the enamino aldehyde prod-
uct. All the kinetics runs were carried out to 90% completion. Rate
constants were calculated from the slope of ln(Aϱ Ϫ At) vs. time.
The error in k is less than 3% for all examined alcohols. The en-
thalpy of activation and the entropy of activation were calculated
according to Eyring from the linear regression of log (k/T) (ob-
tained from experiences in tert-butyl alcohol) vs. 1/T by the least-
squares method.
deke, Tetrahedron Lett. 1979, 4025Ϫ4028.
A. H. Khuthier,
[5d]
M. A. Sheat, J. Prakt. Chem. 1989, 331, 187Ϫ194.
A. H.
Khuthier, M. A. Sheat, J. M. A. Al-Rawi, Z. S. Salih, J. Chem.
Res. (S) 1987, 184Ϫ185.
[6]
G. Hallström, B. Lindeke, A. H. Khuthier, M. A. Al-Iraqi,
Tetrahedron Lett. 1980, 667Ϫ670.
[7] [7a]
A. C. Cope, T. T. Foster, P. H. Towle, J. Am. Chem. Soc.
[7b]
1949, 71, 3929Ϫ3935.
C. H. DePuy, R. W. King, Chem.
Rev. 1960, 60, 448Ϫ451.
[8]
[9]
´
A. Szabo, I. Hermecz, J. Org. Chem. 2001, 66, 7219Ϫ7222.
A. C. Cope, N. A. LeBel, J. Am. Chem. Soc. 1960, 82,
4656Ϫ4662.
[10]
C. G. Swain, M. S. Swain, A. L. Powel, S. Alunni, J. Am. Chem.
Soc. 1983, 105, 502Ϫ513.
[11] [11a]
[11b]
H. Eyring, J. Chem. Phys. 1935, 3, 107.
Chem. Rev. 1960, 60, 448Ϫ451.
H. Eyring,
[12]
[13]
´
Z. Mucsi, A. Szabo, I. Hermecz, J. Mol. Struct. Theochem.,
accepted for publication.
H. Tsou, N. Mamuya, B. D. Johnson, M. F. Reich, B. C.
Gruber, F. Ye, R. Nilakantan, R. Shen, C. Discafani, R. De-
Blanc, R. Davis, F. E. Koehn, L. M. Greenberger, Y. Wang, A.
Wissner, J. Med. Chem. 2001, 44, 2719Ϫ2734.
J. H. Biel, F. DiPierro, J. Am. Chem. Soc. 1958, 80, 4609Ϫ4611.
H. Quast, T. Hergenröther, Liebigs Ann. Chem. 1992, 581Ϫ590.
F. Wille, L. Saffer, Justus Liebigs Ann. Chem. 1950, 568, 34Ϫ46.
D. Berger, P. Wilhelm, M. Neuenschwander, Helv. Chim. Acta
1999, 82, 326Ϫ337.
Acknowledgments
We thank Dr. David Durham for his assistance towards the linguis-
tic improvement of this paper.
[14]
[15]
[16]
[17]
[1] [1a]
R. H. Abeles, A. L. Maycock, Acc. Chem. Res. 1976, 9,
[18]
[19]
[1b]
´
J. Gaal, I. Hermecz, in: Inhibitors of Monoamine
J. Taillades, L. Garrel, P. H. Lagriffoul, A. Commeyras, Bull.
Soc. Chim. Fr. 1992, 191Ϫ198.
313Ϫ319.
Oxidase B (Ed.: I. Szelenyi); Birkhäuser, Basel, 1999, pp.
M. K. Scott, E. W. Baxter, D. J. Bennett, R. E. Boyd, P. S.
Blum, E. E. Codd, M. J. Kukla, E. Malloy, B. E. Maryanoff,
C. A. Maryanoff, M. E. Ortegon, C. R. Rasmussen, A. B. Re-
itz, M. J. Renzi, C. F. Schwender, R. P. Shank, R. G. Sherill, J.
L. Vaught, F. J. Villani, N. Yim, J. Med. Chem. 1995, 38,
4198Ϫ4210.
75Ϫ108.
[2] [2a]
S. B. Park, P. Jacob, N. L. Benovicz, Chem. Res. Toxicol.
1993, 6, 880Ϫ888. [2b] A. M. Weli, B. Lindeke, Biochem. Pharm-
acol. 1985, 34, 1993Ϫ1998.
[2c]
M. Katagi, M. Tatsuno, A.
Miki, M. Nishikawa, K. Nakajima, H. Tsuchihashi, J. Chroma-
togr. B 2001, 759, 125Ϫ133.
G. Hallström, B. Lindeke, A. H. Khuthier, M. A. Al-Iraqi,
Chem. Biol. Interact. 1981, 34, 185Ϫ200.
[3]
[20]
A. Pouilhes, S. E. Thomas, Tetrahedron Lett. 1989, 2285Ϫ2288.
Received July 24, 2003
694
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 687Ϫ694